OverUnity Research

Being discussed here

http://www.energeticforum.com/renewable-energy/20199-energy-conference-graham-gunderson.html
All comments welcome

there was some mention of a Mcfree MO ...but all is speculative ATM ,


respectfully

Chet K

Graham is a recent member of this forum posting as energyiseverywhere.

Smudge

Graham seems to make an amazing impression on the people he meets , and he is Open sourcing his work too !!
Our world needs more men Like this !

Quote from: Chet K on 2016.07.12, 10:10:22
Graham seems to make an amazing impression on the people he meets , and he is Open sourcing his work too !!
Our world needs more men Like this !

Well i went and read the thread at EF,as i to think Graham is a decent guy.

Everything was looking great,and some very high COP+ numbers were being thrown around.
I was thinking that this is not that hard of a build,but i was waiting for the hiccup--there's always a hiccup--that part that is hard to find or no longer exists,that is needed for the device to work--and there it was,right on que

Quote: The big problem with replication is going to be getting the high permittivity "C" cores. Graham believes that they are no longer available commercially, however if money were available (say $5K) a custom lot could be contracted for. So if you have any of they "fossils" laying around you had better hang on to them.

Oh well,maybe Graham will share a bit here on this forum,and give us some means of replicating his transformer.


Brad

Yes the core does need a much closer look, it could hold a very big clue.

a clue which may assist in other projects too.

perhaps casting a C core is one possibility for experiments ,T1000 has mentioned home brew cores a Key in some of his
investigations.

Quote from: TinMan on 2016.07.12, 14:06:26

Oh well,maybe Graham will share a bit here on this forum,and give us some means of replicating his transformer.

Brad

Haven't seen him sign in so methinks the daily rubbish that is appearing on this forum could be putting him off.  I think someone should message him to get his interest.

Smudge

Quote from: Smudge on 2016.07.12, 15:32:06
Haven't seen him sign in so methinks the daily rubbish that is appearing on this forum could be putting him off.

How do you define rubbish ?

As this is an overunity research forum I define rubbish as anything that is clearly not overunity research.

Smudge

Your perspective of strict adherence to scientific practices is a narrow view and fails to take into account that this forum, is a forum of apes choosing to communicate, and apes can be emotional creatures at times.

The moderators here defined overunity within the Power and Creed mission statement which I commented on a while back:

http://www.overunityresearch.com/index.php?topic=3037.msg49180#msg49180

Quote from: evolvingape on 2015.05.31, 00:33:00
COEFFICIENT OF PERFORMANCE (COP)
Again we must examine this parameter in the context of open and closed
systems.

In a closed system, the COP will be equal to the efficiency in a sense, but is
expressed as a ratio as follows: 1:1 (η=100%), 0.8:1 (η=80%), etc. So one
should conclude from the discussion so far, that the COP in a closed system will
never be higher than 1:1.

In an open system, the COP could be anywhere from 0.1:1 to 106:1. It all
depends on the efficiency of the device (with low COP's), and how much energy
or power is freely added to the system from the "outside" with a given input
power.

COP in open systems is computed by taking the ratio between the freely added
"outside" power (POopen), PLUS the output power (if any) supplied by the closed
system (POclosed), to the user-supplied input power (PI).
In equation form:

COP = (POopen + POclosed)/PI   OR  = POtotal/PI

Do we care if we are measuring collected open-system power PLUS closed system
power on the output? No. What we care about is obtaining more total
power on the output of the DUT, than we are supplying for device operation.

OVERUNITY
As already discussed, overunity is not possible in closed systems, and therefore
can only exist in open systems. Overunity then is achieved any time a device or
system exhibits a COP>1.

I chose to study energy, and the history of war as it pertains to geopolitics.

Both are important to me, and I chose to intervene in geopolitics while still undecided whether to give you hhop gen 3 or not..

..ah.. the good old days..  >:-)

Specialty cores and highly defined switches are all rubbish. The cores hold energy in a subdued mode and the switches are too slim in their transmission windows. This won't control or coerce the energy to do anything. Everybody has the better answer, no? Yeah right. None of this executes what Telsa was doing and that was metal cores for coupling, air cores for resonance and spark gaps for broad band transmission. No real technology, just straight forward answers. Maybe in the future cores will be glass and the flux will be photons but as of this writing only the cromagnon,brute force works when dealing with magnetic flux. Why? Because flux is a property of a very large medium. Any new technological attempts at this point is like hunting T-rex's with a fly swatter. O0

Quote from: Smudge on 2016.07.12, 16:21:42
As this is an overunity research forum I define rubbish as anything that is clearly not overunity research.

Smudge

Well that's me out the door......

The adage " all work and no play " springs to mind.

Cheers Grum.

Quote from: Grumage on 2016.07.12, 17:51:35
Well that's me out the door......

The adage " all work and no play " springs to mind.

Most of us have one foot out the door at all times, survival mechanism.

This forum certainly allows for differing points of view and self moderation.

Real new tech stands unrefuted for eternity and becomes law..  :-*

images from Reiyuki at the conference

http://imgur.com/a/oYyM7

Quote from: Smudge on 2016.07.12, 08:19:32
Graham is a recent member of this forum posting as energyiseverywhere.

Excellent!
I wonder if he'll remember me from the Vortex-L list

Anyway, what is the claimed O/I energy ratio? 

I read something about 1.03 but that is is not a "huge" ratio, especially when the inherent difficulties of measuring the energy content of high crest factor I*V waveforms, is considered.

Reiyuki is being added to the forum , he was at Graham's demonstration and took many notes.

Quote from: verpies on 2016.07.13, 15:27:19
Anyway, what is the claimed O/I energy ratio? 

I read something about 1.03 but that is is not a "huge" ratio, especially when the inherent difficulties of measuring the energy content of high crest factor I*V waveforms, is considered.

TK has said many times that all that is needed for an electrical self runner is a COP of 1.3

1.03 is way to close to the margin of error for it to be accepted on measurement alone, so they would have to loop it anyway for credibility.

Here we have an "unofficial  " block  Schem form Reiyuki

also some perhaps related info from the energetic thread

https://electropub.wordpress.com/2012/05/21/non-linear-load-using-mosfets/

energetic thread
http://www.energeticforum.com/renewable-energy/20199-energy-conference-graham-gunderson.html

Quote from: verpies on 2016.07.13, 15:27:19
Anyway, what is the claimed O/I energy ratio? 

I read something about 1.03 but that is is not a "huge" ratio, especially when the inherent difficulties of measuring the energy content of high crest factor I*V waveforms, is considered.

According to one guy on energeticforum who knows Graham and helped him set up his equipment.

Quote
Graham was making his measurements using two different approaches:

1) Two Power Analyzers one on the input and one on the output

2) A Tektronix (2000) 4 channel scope with advanced math functions and probes

Both sets of instruments agreed on the values being measured. From the warm up moment the unit started at 1.53 watts in and 9.43 watts out or a initial COP of 6.16. The load was a 12V automotive lamp that operated during the entire presentation. As time went on the COP improved an hour later the input wattage was down to zero (with four places of accuracy) with the output still at 9.5. That is a COP of infinity, but lets say around 50. I was in the front row to observe this. Graham reported that he has observed the input value go negative like -.25 watts in.

What do you make of that?

Smudge

Quote from: Chet K on 2016.07.14, 13:38:50
Here we have an "unofficial  " block  Schem form Reiyuki

also some perhaps related info from the energetic thread

https://electropub.wordpress.com/2012/05/21/non-linear-load-using-mosfets/

energetic thread
http://www.energeticforum.com/renewable-energy/20199-energy-conference-graham-gunderson.html

Yep.   I'm also here now too, if anyone has some questions or needs more pics/notes.  Just post/pm. :)

The best question I think to ask for any claimed COP>1 device, is what does it do differently from the thousands of other attempts that failed, and what similarities does it share with others that also claim to succeed.

The most unique thing I see in his system is the 'interrupted synchonous rectifier' on the output.
Functionally, the entire circuit is very similar to Kapandze/Kacher where HV impulses are applied at precise part of a push/pull resonant tank circuit.  But practically, I think only Magpwr on ou.com was the only other one ever mentioning a 'nanosecond interrupter' on the LOAD portion of a circuit.  It's a brilliant idea, and could potentially saves us a lot of complexity in other areas.

The other is the 'push/pull/sleep' timing cycle on the resonant input side.  I suppose any PWM driver might give us similar results as well, but it means we have delays that give room for a compression-expansion cycle.

-Rei

Hello Reiyuki.

Welcome to OUR.

I have to agree with your sentiments regarding not tying down the output in a physical way.   ;)

Your reference to Magpwr just made me notice that he has been absent from OU.com for quite some time now...... Hmmm......

Cheers Grum

Quote from: Reiyuki on 2016.07.14, 17:24:29
interrupted synchonous rectifier' on the output

Quote from: Reiyuki on 2016.07.14, 17:24:29
'nanosecond interrupter' on the LOAD portion of a circuit

Are you describing chopping of the output power wave creating a localised resonant feedback loop ?

Quote from: evolvingape on 2016.07.14, 18:35:25
Are you describing chopping of the output power wave creating a localised resonant feedback loop ?

Exactly.  When current's flowing through a coil, any time you open that circuit you get a big voltage spike.

In this circuit, it is the equivalent of a bridge rectifier running into a large capacitor bank, 99% of the time.  At a very brief instant, the circuit is opened, causing a BIG voltage spike and doing *something* within and/or between the transformer cores.  Regauging, changing permeability, flux compression, BEMF feedback, we're not sure what the actual mechanism is, only what rough parameters the effects show up under.

Quote from: verpies on 2016.07.13, 15:27:19
Anyway, what is the claimed O/I energy ratio? 

Output variable between 5-10 watts total on his setup.  The input is a resonant tank and depending on timing, it can go to zero or slightly negative.  So, COP goes up past undefined, but the sum output to both windings are limited.

Quote from: Reiyuki on 2016.07.14, 19:32:47
Exactly.  When current's flowing through a coil, any time you open that circuit you get a big voltage spike.

In this circuit, it is the equivalent of a bridge rectifier running into a large capacitor bank, 99% of the time.  At a very brief instant, the circuit is opened, causing a BIG voltage spike and doing *something* within and/or between the transformer cores.  Regauging, changing permeability, flux compression, BEMF feedback, we're not sure what the actual mechanism is, only what rough parameters the effects show up under.



Output variable between 5-10 watts total on his setup.  The input is a resonant tank and depending on timing, it can go to zero or slightly negative.  So, COP can go up past undefined, but the sum output to both windings are limited.

Quote from: Reiyuki on 2016.07.14, 19:32:47
Exactly.  When current's flowing through a coil, any time you open that circuit you get a big voltage spike.

In this circuit, it is the equivalent of a bridge rectifier running into a large capacitor bank, 99% of the time.  At a very brief instant, the circuit is opened, causing a BIG voltage spike and doing *something* within and/or between the transformer cores.  Regauging, changing permeability, flux compression, BEMF feedback, we're not sure what the actual mechanism is, only what rough parameters the effects show up under.

My models for hhop electromagnetic are immature but they support this line of inquiry.

Quote from: Reiyuki on 2016.07.14, 19:32:47
When current's flowing through a coil, any time you open that circuit you get a big voltage spike.

Is this statement correct ?

Seems to me that in order for current to 'flow' the circuit must be open to begin with.

A static potential could surge if suddenly released from the tank, and conversely it could surge if suddenly prevented from flowing by closing the circuit.

I don't see how you can have a voltage (electrostatic pressure) spike by opening a circuit that is already open..

Unless it is being chopped very rapidly on the output side of the circuit maintaining some back pressure resistance to flow ?

Rei

Welcome to the forum and thank you for taking the time to attend the conference and also for your diligent gathering of information.

From the photos, I could not tell if the ferrite gaps were adjustable in any manner. I was guessing that the nylon rods and plexiglass holders could possibly allow for some gap or position adjustment.

Any comments on that?

Kind Regards
ION

Quote from: ION on 2016.07.15, 00:36:02
Rei

Welcome to the forum and thank you for taking the time to attend the conference and also for your diligent gathering of information.

From the photos, I could not tell if the ferrite gaps were adjustable in any manner. I was guessing that the nylon rods and plexiglass holders could possibly allow for some gap or position adjustment.

Any comments on that?

Kind Regards
ION

Between the two ferrites, I believe it was 2ea 5mil mylar sheets, so 10mil spacing.  The ceramic magnets were also gapped in a couple places, and the variable nylon mount screws suggests he thought there was some 'space tuning' involed.

Also reminding that the top(input) was high permeability and the lower (output) was low permeability.      He said the strip of magnets on the top was to presaturate the top ferrite 'at the knee' of the BH curve.

I'm a little confused; where does the COP of 1.03 come from when there are claims of Pin=0W and Pout=9W?

Quote from: poynt99 on 2016.07.15, 12:15:49
I'm a little confused; where does the COP of 1.03 come from when there are claims of Pin=0W and Pout=9W?

That comes from Reiyuki's first post on energetic forum.  A later post from Spokane1 stated that Graham had demonstrated at the previous year's conference a different transformer system that showed the small COP.  He knows Graham, lives in the same town (Spokane) and helped Graham set up his new design at this latest conference, and it is he who said the Pin=0W and Pout=9W.  Then a later post from Reiyuki states output range 5-10 watts with input tuned allowing COP to be undefined or negative.

To help me get to grips with the magnetic system I have sketched the layout see image below.  I think Reiyuki may be wrong with his interpretation of the direction of the sideways magnets.  It makes more sense if they create a crossed field and that is what I show in my image.  My thoughts are then associated with the sideways static magnetization vector being driven to precess as shown in the second image.  Because the top arm of the core is biased at its saturation knee, the AC drive on the primary will result in unidirectional field pulses driven round the rest of the core.  Those pulses drive the precession.  Maybe there is some lag in the response to those pulses, and Graham's circuitry is tuned to capture the resulting delayed magnetization pulses arriving in the bottom low perm core.  Just food for thought.

Smudge 

I have just noticed that the energetic forum thread started on the day of last year's conference (where Graham presented his low COP transformer) then jumps to this year's conference that took place just recently (where Graham presented his new infinity COP device).  Perhaps that explains why we have different COP's mentioned.  It is that new design in which we are interested.

Smudge

It would be nice to see a confirmed block diagram of his setup and the measurement points, along with wave forms etc. Does anyone know if they exist?

What frequency is the device operating at?

Quote from: poynt99
It would be nice to see a confirmed block diagram of his setup and the measurement points, along with wave forms etc. Does anyone know if they exist?

What frequency is the device operating at?

Supposedly, the effect would manifest around 20kcps-50kcps.  At presentation, I believe it was tuned around 50kcps.

Quote from: poynt99 on 2016.07.15, 15:42:53
It would be nice to see a confirmed block diagram of his setup and the measurement points, along with wave forms etc. Does anyone know if they exist?

What frequency is the device operating at?

The only waveforms I've seen were in a link posted earlier in this thread, copy attached.  From the cursor positions it looks like 20 uS covers one sine wave input cycle plus one half cycle of dead time.  That's a rep rate of 50KHz but the sine wave is 75KHz.

Smudge

Having used saturating cores to get fast pulses in radar modulators, and being aware of saturation as a pulse sharpener, it strikes me that he could be using saturation to get a fast magnetic impulse.  That would fit with his "magnetic implosion" for describing the transformer.  Then that fast impulse applied to the cross-field area could "kick" it into precessional oscillation.  Just another thought.

Smudge

Just looked at some images and Reiyuki is correct with his interpretation of the sideways magnets.  There is not a crossed field since the field both sides points into the core.  I would guess that the core between those side magnets is saturated near the surface but not at the center.  There could be interesting things happening there with propagation of magnetic impulses through this central channel.

Smudge 

Thanks Reiyuki and smudge.

50kHz should make power measurements relatively easy.  :)

FYI,

I added all the pics (available so far) to the first post.

Quote from: Reiyuki on 2016.07.15, 16:36:08
Supposedly, the effect would manifest around 20kcps-50kcps.  At presentation, I believe it was tuned around 50kcps.

During the presentation, do you recall if Graham was taking the input power measurements at the DC source or were they taken farther downstream in the input switching circuitry?

pm

Quote from: partzman on 2016.07.16, 23:28:34
During the presentation, do you recall if Graham was taking the input power measurements at the DC source or were they taken farther downstream in the input switching circuitry?

pm

Poynt had the same questions :).   Input was measured just before the primaries, output was measured just after the secondary, on the output capacitor bank and the load.
Two methods were used, one was with a 4 channel scope, differential volt probes, precision hall current sensors, and the scope's internal math.  The other was a pair of precision load meters, and during the presentation they were within a fraction of a watt from each other.

Quote from: Reiyuki on 2016.07.16, 23:56:43
Poynt had the same questions :).   Input was measured just before the primaries, output was measured just after the secondary, on the output capacitor bank and the load.
Two methods were used, one was with a 4 channel scope, differential volt probes, precision hall current sensors, and the scope's internal math.  The other was a pair of precision load meters, and during the presentation they were within a fraction of a watt from each other.

OK thanks Reiyuki!

Another question that stems from my somewhat color blindness, I've attached a waveform copy with what I think are the correct ID's for each trace but I'm not sure. Would you confirm or correct this please?

Thanks again,
pm

Quote from: partzman on 2016.07.17, 15:00:38
Another question that stems from my somewhat color blindness, I've attached a waveform copy with what I think are the correct ID's for each trace but I'm not sure. Would you confirm or correct this please?

That's correct.

Here is my analysis of the waveforms supplied by Reiyuki.  Enjoy!

Smudge

Quote from: Smudge on 2016.07.18, 14:13:01
Here is my analysis of the waveforms supplied by Reiyuki.  Enjoy!

Smudge

Thanks for the great analysis! The input H bridge would make it easy to freeze the primary current.

The output current during the no load connection is definitely strange and at first glance appears to possibly be a resonant current but where could this be measured or sensed? Or is this waveform possibly output voltage?

pm

Yes I wondered whether it was output voltage, that would make more sense.  Interestingly, because during that load period the core is saturated up top somewhere, the output could be decoupled from the input.

Smudge

I'm firing up a Google drive folder so we can better coordinate notes/analysis/datasheets/schematics/etc.

https://drive.google.com/drive/folders/0B0_1dJJ_ezEmUDlsNDZUYWlrbE0


Anyone just PM or email me if they want write access to contribute/organize.

Quote from: Smudge on 2016.07.18, 18:59:50
Yes I wondered whether it was output voltage, that would make more sense.  Interestingly, because during that load period the core is saturated up top somewhere, the output could be decoupled from the input.

Smudge

That is an interesting thought!  Also assuming that the upper core half is saturated during the apparent current clamped time, what is causing the current to increase during that same time period? This would seem to indicate that energy is supplied to the primary from a supply source or ....? If the output action is the cause of the increase, then theoretically once started, there would be no outside power needed on the primary!?!

I think the core assembly itself is possible to simulate using gyrator-capacitance modeling where capacitors are used to model the flux paths. This would allow the modeling of cores with different ur's, gaps, leakage inductances, and saturation characteristics with non-linear capacitances. What isn't clear is how one would model the PM's that face the core!

pm

Looking at Reiyuki's Unofficial (probably incorrect) diagram (his words) I see that he refers to the output being interrupted for between 50nS and 1uS.  That throws a different light on the output current waveform, see image below.  I now have the view that the load is connected for most of the time, with only a small interrupt.  It is interesting that the negative portions of the load current are roughly following the negative portions of the input current.  During the interrupt period the load current drops, then there is a recovery period during which the input voltage is at zero.  Perhaps the zero level for that current waveform is near its lower level, making most of the waveform positive????.  In any case the combination of load interrupt, input voltage dwell and core saturation is causing the input current to be reactive and therefore not draw power.  Can't see why this should be but the brain cells are still working on this.

Smudge

Reiyuki,

Thanks for posting the info on the google drive and I have a question regarding the raw photo "notes4.jpg".  I assume this is taken from your notes while questioning Graham on the synchronous rectification and it raises a question for me.  As shown, the Cree mosfet substrate diode would be the conducting element for current flow to the output with the gate drive being off.  I am trying to tie in how this might relate to Smudge's recent post on the near continuous output current with the short interruption period so do you think this is a relatively accurate generalized output circuit description?

pm

Quote from: Chet K on 2016.07.12, 10:10:22
Graham seems to make an amazing impression on the people he meets , and he is Open sourcing his work too !!
Our world needs more men Like this !

Quote from: TinMan on 2016.07.12, 14:06:26


Oh well,maybe Graham will share a bit here on this forum,and give us some means of replicating his transformer.


Brad

Good day all:
For those who do not read at EnergicForum I post this from today:

Quote from: Spokane1;290549Fellow Researchers:

I have had the pleasure to be in contact with Graham Gunderson and if I recall his words correctly he wishes to pass on these comments.

1. He encourages researchers to explore and reproduce his work to date, however he plans several improvement and modifications to the device that was presented at the conference. So, it will be a moving target. To him there was no point in drafting schematics when he plans to change every thing. (The working technology is less than a month old)

2. He will eventually post on his web site some written material when he gets around to it.

3. He doesn't want to answer a lot of technical questions or teach people advanced electronics (and what he is doing is advanced in some areas). This is a continuing work in progress and not a finished product. He would rather be undisturbed for a few months.

4. He doesn't plan to do another demonstration until he can "Close the Loop"

5. He doesn't know why this system works since its performance was way above any of his expectations. He was hoping for a COP of 1.05. He does have some interesting measurement observations which I shall share later.

6. He would like to move into small manufacturing by selling his circuit boards and other subsystems for general OU work, along with white papers and plans. Maybe some assembled modules. He would also like to advance his technical interests in music.

7. For those of you with the Big Bucks (Probably nobody here) the technology is for sale in this infant form for someone who has the means to take a big gamble.

Sticker price $2 million US Lump sum (this is so he can end up with $1 million after taxes)

This includes 90% of the rights to the provisional patent, copies of all lab notes and written material, data sheets, catalog cuts, etc. The operating prototype as it was shown at the convention w/ possible improvements. And 40 hours of training.  This quote is good for 30 days  (until Aug. 19, 2016)

Offer does not include the instrumentation.

The price will increase as successful replications, improvements, performance enhancements, university demonstrations, and inflation dictates.


This could turn in to one great adventure.

Mark McKay, PE

take care, peace
lost_bro

Quote6. He would like to move into small manufacturing by selling his circuit boards and other subsystems for general OU work, along with white papers and plans. Maybe some assembled modules.

Worthy goals - strongly supported!  That would lead to rapid dissemination of the new information, which is great IMO.

This approach - along with one or several good conferences to make some big money quite fast for Graham (nothing wrong with that!) would be better, IMHO, than the patent route which we discussed before.

I would be glad to help in the kit and conference approach - strongly think this would help Graham with funds quickly.

It will be interesting to explore the options here and see if things can move in a direction which will give Graham what he wants and the world what it needs !

O0

Quote from: Chet K on 2016.07.20, 01:46:58
It will be interesting to explore the options here and see if things can move in a direction which will give Graham what he wants and the world what it needs !

O0

Well said, Chet.  You have a good heart.

PhysicsProf
This house is filled with Plenty of good hearts  O0

And it seems we have some tentative financial commitments,

its been a while since we went down this road for an inventor [raising funds thru Angel donors]

Open source Of course !

In the OU forum it is said that the white dots on the magnets do not represent the polarity but something to do with the way they are put into a mould, (sorry mold :-[).  The magnet field runs around the top loop of magnets, so there is indeed a crossed field in the main core.  Maybe precession does play its part.

Smudge

Just an Update
Member Spokane1 will be joining the discussion here too.  [Peter will be adding him when he gets the time.

This might explain some things about Graham's transformer, see attached pdf.

Smudge

Gentlemen,

I was over at Grahams shop last Monday (after I had brought over some Panda Express Chow). I was dropping off a roll of Litz wire that I had so that Graham could try some ideas with his transformer to reduce the number of switching elements in the H-Bridge from four down to two. I believe he wants to go to a center tapped arrangement and make something like a Dr. Tesla's 1892 oscillator with two pancake coils with a space between them.

At the time I was there he had his conversion transformer taken apart with only one set of PM's attached. The attached photo shows how the PM's are supporting themselves with out the clamping hardware or any spacers. This should establish how the polarity has been arranged.

Last Saturday was not a good day for him since his machine failed to improve with any of his new tuning ideas. He blew a transistor in his H-Bridge and was wondering why that had happened since those FETS are rated at 1000 PIV while his measured peak voltage is only 800V.

What he did find out (that didn't make the day a complete loss) was that if he were to input energy into the output side; than the regular input signal would be reconstructed at the input side with no applied power on that end. He thought that was really something since that wave is discontinuous.

How many photos can we post on the kind of forum? Or maybe how much memory can we use?  I took 40 additional pictures of the machine while there.

There are a lot of details in this construction. You have to be about 2" away from the device with the camera to begin to see them. Without Graham's guidance I would have never seen the surface mount components from the present photos and I think I know what I'm looking at.

The tuning of this device is fragile. Graham said that during development he could hold a few Barium Magnets in his hand about 24" away, twist his hand 90 degrees and the intensity of his load lamp would vary by 40%. (my guess)

Those 18 PM's on the top of the device account for about a 20%(my guess) increase in performance - that is why they are there.  It is really strange to me to know that such small external magnetic field changes could have such a profound impact on a closed core device.

Also, Gram said that by adding a 100 pf Silver-Mica capacitor between the Source and Drain of one of the Back End FETS would completely destroy the OU effect.

Graham also showed me some two variations of some simple magnetic probes he fabricated. They were composed of a small inductor, a diode and an LED. He developed these to discover magnetic leaks. What he told me was that he found several leaks where there shouldn't be any. The main one he described and showed me (the machine was not running) was at the very bottom center (of one side) of the main core. He said that his little tester was just " going wild" when by rights all of the flux at that location should have been going through the ferrite.

The impression I get from Graham is that he wants to play with his new child. At the moment that is far more rewarding than drawing up schematics, writing reports, and answering dumb questions (which I ask a lot of). Graham presently works full time as a repair technician for a local aviation electronics company on the East side of town. Therefore time is at a premium.  If I were him I would have the same attitude. He has two sons and is planning to get married, so his plate is already fully.

I have no doubt that Graham is onto something big - but right now very fragile. I have known Graham since 2006 and his technical knowledge makes me feel like a first grader talking to a Ph.D. I can barley form what I think are meaningful questions and he takes the time to explain some complex concepts that even a first grader can understand. I have the highest regard for his candor and morel integrity. (Besides he likes Chicken Wings and Pizza)

Graham said that the lower ferrite core came from FoxCube and if I recall the part number correctly it was an HC90 (or something close to that). It is 100mm on the long side and looks to be 4" wide with 1" arms (or legs)

Graham does want to keep the part number of the other core (upper) proprietary at this time. He wants to keep something for himself (or the future owners). I would have kept a lot more than that. I would guess that Graham wants others to pay their dues before he hands over a possible age changing technology to the masses. I'm sure that who ever makes the purchase will discover what that part number is and have to contract with a core shop to make a custom batch. Graham thinks this could be done for about $2,500. I think that is a low figure.

Has anybody had any luck sourcing a set of cores? I doubt it. Never the less I'm planning to move ahead with a 5" steel laminated UI core. I know this is not the same that Graham used, but it is what E.V. Gray's team used to achieve a COP of 282 at the 7.5 kW level. So there is magic in that iron somewhere. The Gray approach was a pulsed method that used a moving spark gap to generate the 46 kHz RF. Graham's system is continuous with the PM's providing the saturation field and the RF being continuously injected.

The bid difference between the Gray system and the Gunderson system is the in the Gray system the OU was harvested as Force - Huge amounts of force in fact more force than what can be accounted for from classical magnetic repulsion.

Mark McKay, PE

After much scouring, the closest cores I could find were these.  May not be the right size, but they're at least in the ballpark.

http://www.surplussales.com/inductors/FerPotC/FerPotC-5.html


It might be possible to lower the permeability artificially, possibly by heating the ferrite to a high temperature for a period of time.  Anyone have ideas on that?

Mark
Welcome to the forum

Spokane1
Quote
How many photos can we post on the kind of forum? Or maybe how much memory can we use?  I took 40 additional pictures of the machine while there.

There are a lot of details in this construction. You have to be about 2" away from the device with the camera to begin to see them. Without Graham's guidance I would have never seen the surface mount components from the present photos and I think I know what I'm looking at.

End Quote

member Reiyuki has posted a File share link for Pics here * Note I see he posted whilst I was chiseling on the keyboard.

https://drive.google.com/drive/folders/0B0_1dJJ_ezEmUDlsNDZUYWlrbE0

* However I will get an answer to that question and see if we can get those posted .


  Graham truly is an amazing fellow for sharing all he has and we have nothing but gratitude over here for the privilege to investigate this,
and will hopefully be viewed as an asset to him on this newfound path with "his" new baby Device .

truly a privilege indeed !!
* To Note
Myself and others are actively investigating the Financial options so this can be completely open sourced and Graham can have his needs met
and the world can too .

respectfully

Chet K

I hope Gunderson has the ability to read Smudge's latest thoughts on the operation of his (Gunderson's) "transformer" as it contains a viable explanation for the source of excess energy, and may aid Graham in fine tuning his device.

Smudge's PDF's on this Topic have been placed at a few other forums Too.[with his consent]

And I believe they can be read here by passers by .

I know Mark M [Spokane1] has Noted Cyril's contributions here.

http://www.energeticforum.com/renewable-energy/20199-energy-conference-graham-gunderson-2.html [post number 33]

Chet K

Also to Note
A new builder and experimenter will be joining here shortly
Ben K4zep [Always Luved this mans work]

may take a while as I'm on the road today.

respectfully

Chet K

If what is claimed is true, and can be repeated, and measured by independent qualified engineers  , getting the money is not a problem i think.

I think we, peoples who take their times to research in this subject, and love it, will be able to collect that amount of money in no time, if this is verified, and work as claimed.

Quote from: ION on 2016.07.21, 13:28:29
I hope Gunderson has the ability to read Smudge's latest thoughts on the operation of his (Gunderson's) "transformer" as it contains a viable explanation for the source of excess energy, and may aid Graham in fine tuning his device.

Dear Ion,

I'm afraid that Graham has sworn off theories for awhile. I attempted to explain to him my concept of the McFreey idea and he didn't want to here it, I don't blame him. Why try to wrap your mind around a new theory when you are already observing more than one phenomena not addressed by classical theory. Theory has its place and has served us well, but in this case he needs to collect more observations and measurements before someone can compose something that will lead to some better engineering judgment.

But that doesn't mean people should give up on attempting to piece together a theory of the moment that helps visualize what might be going on with the limited information available.

Spokane1

The Graham Effect may be peculiar to the high perm ferrite,  Here are some more thoughts on the subject.

Smudge

That's very true !  We can have a nice theory which doesn't explains facts or a simple theory matched to the experiment results. That was mentioned by Tesla.

Quote from: Chef on 2016.07.21, 17:24:32
If what is claimed is true, and can be repeated, and measured by independent qualified engineers  , getting the money is not a problem i think.

Dear Chef,

Unless someone has cash money sitting on the kitchen table (or a bank draft) Graham is not currently interested in having his present setup evaluated by "independent qualified engineers" for the sake of verification alone. A serious customer will get the required verification (with their own equipment even) otherwise Graham doesn't want some 3rd party to tell him that he is suffering from "instrument error". He wants to discover that for himself - if that is the case. He wants to close the loop and completely remove all these potential instrument issues.

Money might be hard to come by for such an infant and fragile process. Heck, the inventor of the copy machine was rejected by several companies even with a crude working prototype in hand.

Spokane1

Quote from: Spokane 1 on 2016.07.21, 18:48:50
Dear Chef,

Unless someone has cash money sitting on the kitchen table (or a bank draft) Graham is not currently interested in having his present setup evaluated by "independent qualified engineers" for the sake of verification alone. A serious customer will get the required verification (with their own equipment even) otherwise Graham doesn't want some 3rd party to tell him that he is suffering from "instrument error". He wants to discover that for himself - if that is the case. He wants to close the loop and completely remove all these potential instrument issues.

Money might be hard to come by for such an infant and fragile process. Heck, the inventor of the copy machine was rejected by several companies even with a crude working prototype in hand.

Spokane1

I don't see how verifying the claim by some independent engineer payed by us, a group!,  would be different, as one, payed by somebody else who want to buy it alone.  I don't understand also his intention on this sell out fast for 2 million bucks, but it's his project, so he can do what he want.

In any way, I will follow the story, quite excited to see, how it will unfold.

Quote from: Chef on 2016.07.21, 19:27:32
I don't see how verifying the claim by some independent engineer payed by us, a group!,  would be different, as one, payed by somebody else who want to buy it alone.  I don't understand also his intention on this sell out fast for 2 million bucks, but it's his project, so he can do what he want.

In any way, I will follow the story, quite excited to see, how it will unfold.

Funny that for a while, Graham WAS an independent engineer sent to verify other people's systems. ;D
We can hire him to test it!/sarcasm

Quote from: Reiyuki on 2016.07.21, 00:04:19
After much scouring, the closest cores I could find were these.  May not be the right size, but they're at least in the ballpark.

http://www.surplussales.com/inductors/FerPotC/FerPotC-5.html


It might be possible to lower the permeability artificially, possibly by heating the ferrite to a high temperature for a period of time.  Anyone have ideas on that?

Dear Reiyuki,

Great Work in chasing those down.

Those certainly look like the low permittivity model in question.  I suppose you already bought them all.  Now to find the high u model.

Also, if a builder wants to be true to the original machine they will have to replicate Graham's custom Swinging choke that drives the H-Bridge. That device takes four (4) of these "U" cores and two smaller "I" cores that are about 0.75" square.

Spokane1

Quote from: Chet K on 2016.07.12, 10:10:22
Graham seems to make an amazing impression on the people he meets , and he is Open sourcing his work too !!
Our world needs more men Like this !

Open sourcing his work? For a mere 2 million dollars? 

Quote from: Smudge on 2016.07.14, 14:29:58
According to one guy on energeticforum who knows Graham and helped him set up his equipment.

QuoteAs time went on the COP improved an hour later the input wattage was down to zero (with four places of accuracy) with the output still at 9.5. That is a COP of infinity, but lets say around 50.

What do you make of that?

Smudge

If the input power was 0.000 Watts, why did it still have to be connected to the power supply in order to keep running?

What do I make of that? How about "measurement error"? Is anyone willing to consider that the measurements are in error?


Yes, show me any device with electrical inputs and outputs that _truly_ produces a COP of at least 1.3 and I will show you how to self-loop it so that it does NOT need to be connected to any external power supply, once started.  If it uses "0.000 Watts input" that makes it even easier. I've stated this many times and I state it again.

Oddly enough ... not a single claimant has ever contacted me to see if I can really do it. Why do you suppose that is?

Welcome to the forum, Mark! 
You noted that Graham had taken his device apart, at least partly, and noted (and thanks for your observations!):

"The tuning of this device is fragile. Graham said that during development he could hold a few Barium Magnets in his hand about 24" away, twist his hand 90 degrees and the intensity of his load lamp would vary by 40%. (my guess)

Those 18 PM's on the top of the device account for about a 20%(my guess) increase in performance - that is why they are there.  It is really strange to me to know that such small external magnetic field changes could have such a profound impact on a closed core device.

Also, Gram said that by adding a 100 pf Silver-Mica capacitor between the Source and Drain of one of the Back End FETS would completely destroy the OU effect.

Graham also showed me some two variations of some simple magnetic probes he fabricated. They were composed of a small inductor, a diode and an LED. He developed these to discover magnetic leaks. What he told me was that he found several leaks where there shouldn't be any. The main one he described and showed me (the machine was not running) was at the very bottom center (of one side) of the main core. He said that his little tester was just " going wild" when by rights all of the flux at that location should have been going through the ferrite.

The impression I get from Graham is that he wants to play with his new child."

I can understand that!
No need to respond to skeptics yet (IMHO).

One caution:  when a device exhibits OU, it seems the temptation is to play with it, take it apart, try changing this and that -- until the device has lost its "magic".  I would recommend that Graham take measurements of the device basically AS IS - and never take it apart - until he builds a second one to play with and dissect.

A working OU device that can be measured and observed is worth more than thousands of theories!  worth much more than trying to please the skeptics or answer every question being raised. 

--Prof Jones

  Mark also noted,

QuoteHas anybody had any luck sourcing a set of cores? I doubt it. Never the less I'm planning to move ahead with a 5" steel laminated UI core. I know this is not the same that Graham used, but it is what E.V. Gray's team used to achieve a COP of 282 at the 7.5 kW level. So there is magic in that iron somewhere.

The Gray approach was a pulsed method that used a moving spark gap to generate the 46 kHz RF. Graham's system is continuous with the PM's providing the saturation field and the RF being continuously injected.

The bid difference between the Gray system and the Gunderson system is the in the Gray system the OU was harvested as Force - Huge amounts of force in fact more force than what can be accounted for from classical magnetic repulsion.

Mark McKay, PE


------------------------

Two things -
1.  My guess at this point is that the "magic" in Graham's device is not something in that particular core he used, that is, I would encourage replication efforts even if that particular core is not readily available.

2.  What is this device by E.V. Gray?  sounds interesting.

Someone mentioned that not only has Gunderson measured 0W input, but at times the input power went negative. What immediately pops into my mind is that if he was measuring the input power properly to begin with with all probe polarities correct, the input power measurement should be negative anyway.

Would someone in-the-know be so kind as to draw out a simple block diagram of the input source, device, and output load, along with the measurement probe positions? If no one knows, perhaps Graham himself would be open to providing such a diagram (it wouldn't give away any secrets so why not?)?

Quote from: TinselKoala on 2016.07.22, 09:25:55
What do you make of that?

Smudge


If the input power was 0.000 Watts, why did it still have to be connected to the power supply in order to keep running?

What do I make of that? How about "measurement error"? Is anyone willing to consider that the measurements are in error?


Yes, show me any device with electrical inputs and outputs that _truly_ produces a COP of at least 1.3 and I will show you how to self-loop it so that it does NOT need to be connected to any external power supply, once started.  If it uses "0.000 Watts input" that makes it even easier. I've stated this many times and I state it again.

Oddly enough ... not a single claimant has ever contacted me to see if I can really do it. Why do you suppose that is?

TK,

I think the measurements are more than likely correct, but the issue IMO is where the measurements are taken. Without any schematic to reference it is pure speculation on my part but from the scope traces and comments made, we see a device with reactive input and resistive output and it is at these points that the measurements were taken. 

So, if I am correct in my overall view of this device, what Graham has disclosed is only part of the story as the generating means that produces the input sine waveform must be able to return the so-called negative energy back to the DC source that produces the sine waveform. This is required for the overall device to be OU and for any attempted looping to be successful. At the operating frequency range of 20kHz-50kHz this is not an impossibility but I don't believe the demonstration device included such circuitry. However, this is again speculation on my part.

The apparent sensitivity of the device is also a problem for looping IMO as even though it operates in relatively medium frequencies, it seemingly has certain characteristics of a device operating at much higher frequencies.

pm 

Just an update to All and Spokane
Just hung up with Stefan H from OU.com
He is trying to sort out a proper and respectful path
At his forum for this with the Mentioned above moderator??

I am traveling around these next few days ,and staying on top of this will Be a challenge.

To mention
Stefan considers Graham an old friend

Once the noise can be quieted and the discussion taken more respectfully over there
we should be able to get a lot more accomplished in a lot less bandwidth at all the groups.

Respectfully
Chet K

Quote from: PhysicsProf on 2016.07.22, 12:36:03
  Mark also noted,
Two things -
1.  My guess at this point is that the "magic" in Graham's device is not something in that particular core he used, that is, I would encourage replication efforts even if that particular core is not readily available.

2.  What is this device by E.V. Gray?  sounds interesting.

Dear PhysicsProf,

Drop me an email to mmckay@simplexgrinnell.com and I shall send you more than you want to read on the subject of E.V. Gray.

Spokane1

Interesting is that some time ago I planned to build similiar device but working at 50Hz. It would be much easier if you follow my advice and try with a normal iron core transformers. It is a saturable converter kind  O0

Quote from: Chet K on 2016.07.22, 15:17:59
Just an update to All and Spokane
Just hung up with Stefan H from OU.com
He is trying to sort out a proper and respectful path
At his forum for this with the Mentioned above moderator??

I am traveling around these next few days ,and staying on top of this will Be a challenge.

To mention
Stefan considers Graham an old friend

Once the noise can be quieted and the discussion taken more respectfully over there
we should be able to get a lot more accomplished in a lot less bandwidth at all the groups.

Respectfully
Chet K

Your premise to proceed is fundamentally flawed.

If you require a public audience for inspiration from an anonymous it is a valid path. If you wish to disclose a viable theory or working prototype it is a viable path.

If you need to further develop R&D a private thread on this forum offers that facility.

Anything else is grandstanding for a reason.

I predict this is the path that will be chosen due to 'wriggle room': "If you wish to disclose a viable theory or working prototype it is a viable path."

;D

Mark - done.  Thanks again.  Pls let me know if there's anything I can do to help. 
Good work out there in Spokane!

Question - is there a VIDEO of Graham's presentation at the conference?
If so - when available?

I suggest that someone SELL this video for Graham - making sure that he gets the proceeds.  I'd buy it!
Maybe a short trailer on Youtube to generate interest.

Also, are you guys thinking about crowd-funding?  I've done 3 or 4 projects funded for people, using YouCaring.

Quote from: poynt99 on 2016.07.22, 12:51:52
Someone mentioned that not only has Gunderson measured 0W input, but at times the input power went negative. What immediately pops into my mind is that if he was measuring the input power properly to begin with with all probe polarities correct, the input power measurement should be negative anyway.

Would someone in-the-know be so kind as to draw out a simple block diagram of the input source, device, and output load, along with the measurement probe positions? If no one knows, perhaps Graham himself would be open to providing such a diagram (it wouldn't give away any secrets so why not?)?

Perhaps this will help

Spokane1

Quote from: PhysicsProf on 2016.07.22, 19:53:47
Mark - done.  Thanks again.  Pls let me know if there's anything I can do to help. 
Good work out there in Spokane!

Question - is there a VIDEO of Graham's presentation at the conference?
If so - when available?

I suggest that someone SELL this video for Graham - making sure that he gets the proceeds.  I'd buy it!
Maybe a short trailer on Youtube to generate interest.

Also, are you guys thinking about crowd-funding?  I've done 3 or 4 projects funded for people, using YouCaring.

Dear PhysicProf

The whole purpose of the EE Conference is to produce and then sell these lectures in a DVD download for $20 - $40. Aaron Murakami is the publisher and puts out one of these every couple of weeks. He has a vote as to which lecture should be available first. I think is was a toss up between Graham and Eric Dollard - last time I checked.

These lectures are heavily advertised at:

http://www.energeticforum.com/renewable-energy/20199-energy-conference-graham-gunderson-2.html

so far I haven't seen Grahams's July 9  presentation come out.

The Lecturer gets 10% of the gross every month. There is not a whole lot of money to be made from these sales.


Graham hasn't mentioned crowd funding and I kind of doubt he will consider it for several months. He has spent 15 years looking for this phenomena and he is going to play with it. He is pretty well set up with the equipment and supplies that Mark Gobles gave him several years ago. However if he needs something outside of his present inventory like Litz wire or more of those Cree SiC FETS, then he is going to have to save for awhile. I have a huge inventory of 40's to 60's electronics, but there is little there that he can use.

Graham hasn't asked for any money (that I know of). But if you sent him a dozen of those expensive CREE SiC 1.2 kV 90 AMP Power Switching FETS I'm sure he would smile from ear to ear and you would have his gratitude. Perhaps you might know how he could get an institutional discount from DigiKey or Mouser. Other than Litz wire and  some 6" x 8" x 1" High Quality Barium Ferrite PM's (for his Sweet VTA work) I'm not aware of anything specific he is short of right now. Next week could be a different story.

Spokane1

Very nice Spokane1. Obviously you've made a few diagrams in your time.  ;)

One thing that sticks out immediately, is the logic controller feed to the output driver chips (the stage before the output storage caps). I would strongly urge Graham to measure the power being delivered by that feed to make sure it isn't adding significant power to the output.

The other alarm bell that is going off for me is the fact that that same output driver stage has 3 power supplies in it. They too could indirectly be contributing power to the load.

There could indeed be significant power being sourced by the logic controller and the power supplies I mentioned above, but determining if a significant portion of that power is getting to the output could be tricky.

If they are sourcing only milliwatts of power, then obviously they are not an issue. However, if the measured power is significant relative to the output power, then further investigation should ensue.

It would be nice to see detailed input power traces on the scope, and how the power was being computed (no. of cycles, MEAN, etc.). The output power measurement could be getting skewed by what I mentioned above, but the input power measurement seems more interesting since it apparently can go to 0W. I've seen that before and it was attributed to parasitic wiring inductance that was not compensated for. Once corrected, all was back to normal.

Regarding the input power measurement and a potential source of error...

Really? You cannot possibly be serious. I can't believe that people are actually falling for this. Look at all the power supplies that ARE NOT BEING MEASURED but which are providing power to the apparatus.

The proper locations for measuring the INPUT POWER are shown in my markup of Spokane1's block diagram.

Power supplied to _all components_ that are required for the performance of the entire apparatus _must_ be considered as comprising the input power!


Sheesh. No wonder nobody will answer my question: If it draws 0.000 W INPUT, why can't it be disconnected from the power supply and keep running? Now we know the real answer. The "0.000 W" and the other "input power" measurements are not the _actual_ input power to the apparatus! Even if they are correct, which they almost certainly are not, they ignore the power supplied to other components without which the apparatus would not work at all.

Quote from: TinselKoala on 2016.07.23, 04:56:44
Really? You cannot possibly be serious. I can't believe that people are actually falling for this. Look at all the power supplies that ARE NOT BEING MEASURED but which are providing power to the apparatus.

The proper locations for measuring the INPUT POWER are shown in my markup of Spokane1's block diagram.

Power supplied to _all components_ that are required for the performance of the entire apparatus _must_ be considered as comprising the input power!


Sheesh. No wonder nobody will answer my question: If it draws 0.000 W INPUT, why can't it be disconnected from the power supply and keep running? Now we know the real answer. The "0.000 W" and the other "input power" measurements are not the _actual_ input power to the apparatus! Even if they are correct, which they almost certainly are not, they ignore the power supplied to other components without which the apparatus would not work at all.

Good Morning Mark, TK, P99 and all.

It appears that a lot of serious experimenters are here!!! This is very good!

The excellent questions about the power measurements are all point on. Perhaps
at this early conceptual stage with the investigation of this device it would be wise to to note
that Graham was showing the concept of a transformer that in itself appears to have
excess power producing capacity. He himself was not satisfied with the device and
is continuing to work on it as shown in pictures.

Let us assume that the measurements shown of and around
the transformer itself do show excess output internally. That is a good first step. 
As power supply losses appear to not have been included in
the calculations, and as all measurements can be questioned, perhaps 
closing the loop is the only way to show that this type of transformer
is a self contained power producing device.  That is, not only powering itself but
providing excess power as a gift to the world.

It would appear that Graham has gone to great lengths in his output power supply to work at a high
as efficiency as possible by the switching circuitry using FETs as rectifiers, snubbers, very high speed isolaters, etc. as can
be gleaned from the pictures.  That was probably his first steps when looking at the device overall and down the road. 
Unfortunately we do not have schematics of this device! As an experimental builder, this drives me up the wall.
But, here, we have been shown a dream, with excellent power measurements up to a point.  I think everyone here
can take these ideas and run with it and see where we go. All of us has had a "EURIKA" moment (I know I have) that upon
further measurements turned the "Eurika" into crow we had to eat.  But we keep flailing away! 

On Overunity.Com page 4, comment 53, gotoluc has shown a waveform very similar to Grahams with description as to how he produced it.
Perhaps he can give us more insight into his circuitry etc.  Anything at this point would be helpful.

Respectfully to all,
Ben K4ZEP


 

Well, if Gunderson wants to show that the transformer itself is OU, then perhaps the _output_ measurement should be made right at the actual output of the transformer itself, just as the "input" power is being measured right at the input to the transformer. And special care should be taken to account for all phase differences and shifts so that the actual _real_ power is reported, not some mishmash of skewed probe timings, reactive and apparent power.

The thread on OU dot com is bringing up some other critical issues. If I am reading correctly, now we are told that the Hall Effect current probe used on the "input" is positioned very close to the transformer structure itself. This means that it is exposed to external magnetic fields, both DC from the permanent magnets and AC from the oscillating current. This will screw up the accuracy of the Hall sensor. These probes need to be operated so that the _only_ magnetic field that they can possibly sense is that induced by the current in the wire they are actually trying to measure.

They also have a significant delay in response, which can be enough to cause incorrect results in power calculations due to a phase shift. And this problem is worse at high frequencies. High-end digital scopes have a "probe skew" or de-skew function that compensates for the time delay in active probes (diff. voltage probes and Hall effect current probes both suffer from this issue). But the user has to actually use this function properly! This is another reason why Hall effect current probe readings should be compared to readings taken at the same time by another method, such as V_drop across a current-viewing resistor. I have demonstrated in a video how improper use, or rather non-use, of probe deskewing can produce "overunity" measurements from a pulse motor setup.

Also, the frequency response of the probe has to be considered, as fast risetime spikes may not be accurately picked up by the probe, whereas a passive voltage probe across a CVR can pick them up. Spikes can contain a lot of power, or a little power. In either case it is important not to inadvertently filter them out of your power measurements by not understanding the limitations and proper usage of active probes.  I have also demonstrated in a video how a small amount of inductance (just 3 inches of wire) can make a large difference in measurements made using current probes on a signal that is naturally spiky.


Gotoluc has described his coil-capacitor setup used to produce his waveforms on the other forum. It is very simple, just a tank circuit made of an ordinary 10 mH ferrite-cored coil in parallel with an appropriate capacitor, tuned so as to give a single cycle response to the 33 percent duty-cycle of the stimulating signal from his function generator.

Is his setup overunity? Of course we know that "real OU" doesn't show up on scope waveforms. Does it?   >:-) 

Certainly, self-looping would be a powerful demonstration of an OU effect, much better than any instrumental measurements. But with all those power supplies in Gunderson's circuit, and just a car taillight bulb's worth of power on the output ... it ain't gonna happen.

Quote from: TinselKoala on 2016.07.23, 12:55:37
Well, if Gunderson wants to show that the transformer itself is OU, then perhaps the _output_ measurement should be made right at the actual output of the transformer itself, just as the "input" power is being measured right at the input to the transformer. And special care should be taken to account for all phase differences and shifts so that the actual _real_ power is reported, not some mishmash of skewed probe timings, reactive and apparent power.

Agreed.

And the wiring should be cleaned up so that it is just a short run between the h-bridge and transformer.

Poynt, partzman and TK have diligently explained the nuances of careful measurement techniques based on their many years of bench time performing and debugging those kinds of tests.

I understand that there are many  "faithers" that have a deep disdain for people who like to use good measurement techniques. Then, they must remember that Graham's claims are solely based on measurement with a high end o'scope and a Clark Hess

As such his measurements are only as good as the limits of the equipment and the skill in using such, as has been pointed out.

So, faith based beliefs regarding the validity of the OU claims need to be "brought down to earth" with at least two other alternative measuring techniques, in parallel with the use of the high end o'scope and it's math calculations and the Clark Hess, as a double and triple check. This would then be good science.

Unfortunately. since Graham is holding some of the "secret sauce" cards to his chest it may not be possible to accurately replicate his claims. That he has disassembled the device only adds to the confusion.

Ben and Mark, welcome to OUResearch forum.

Dear All.

With regard to the output after the rectification and smoothing capacitors surely this could be a simple VxI measurement? The load is virtually resistive being an incandescent lightbulb.

I have always favoured ION's solution though, the load being a suitable high wattage dissipation resistance with a thermocouple attached, no " funny " waveforms can affect that arrangement. I have been toying with an idea of attaching the tip of one of those $/£ store digital thermometers to a partially drilled Aluminium cased 25 W resistor. Calibration would be fairly simple just VxI.

Am I being too simple or just pure simple?  :)

Cheers Grum.

Quote from: ION on 2016.07.23, 13:24:59
Poynt, partzman and TK have diligently explained the nuances of careful measurement techniques based on their many years of bench time performing and debugging those kinds of tests.

I understand that there are many  "faithers" that have a deep disdain for people who like to use good measurement techniques. Then, they must remember that Graham's claims are solely based on measurement with a high end o'scope and a Clark Hess

As such his measurements are only as good as the limits of the equipment and the skill in using such, as has been pointed out.

So, faith based beliefs regarding the validity of the OU claims need to be "brought down to earth" with at least two other alternative measuring techniques, in parallel with the use of the high end o'scope and it's math calculations and the Clark Hess, as a double and triple check. This would then be good science.

Unfortunately. since Graham is holding some of the "secret sauce" cards to his chest it may not be possible to accurately replicate his claims. That he has disassembled the device only adds to the confusion.

Ben and Mark, welcome to OUResearch forum.

Hi Ion and all,

Good to be here too! I had backed off OU research for quite a few months but this new development has stoked the old fires one more time!  Now have to stay
loose and try not to get totally obsessed with it and burn out again.

You can instrument, measure till your red in the face.  It all boils down to:  Will the device power itself and will it also output excess power.
Actually, I would be jumping up and down if it just self powered!  All the measurement tools are just that, tools!  They can guide or point in a direction or give
some indication if what you are doing is better or worse than your last experiment, but again, their just tools.  I'm glad I am doing this for fun and the pure enjoyment of doing it. 
Grahams device might not be the final answer but it is a step in the right direction me thinks.  More power to him and his having the guts to show it to the world.
Lets encourage him as every experimenter needs encouragement.  It's not how many times you find something that does work, its how many times something you found
(a theory) that doesn't work but you keep plugging away!

Or, how many simulations have you see for "Besslers Wheel" that work?  Then how many Bessler wheels have you seen working. ;)  Sorry, Off Topic!

Respectfully
Ben K4ZEP

Quote from: Grumage on 2016.07.23, 14:36:38
Dear All.

With regard to the output after the rectification and smoothing capacitors surely this could be a simple VxI measurement? The load is virtually resistive being an incandescent lightbulb.

I have always favoured ION's solution though, the load being a suitable high wattage dissipation resistance with a thermocouple attached, no " funny " waveforms can affect that arrangement. I have been toying with an idea of attaching the tip of one of those $/£ store digital thermometers to a partially drilled Aluminium cased 25 W resistor. Calibration would be fairly simple just VxI.

Am I being too simple or just pure simple?  :)

Cheers Grum.

Hi Grum,

For a quick and moderately accurate measurement, it would work like a champ, but response would be sluggish! 

Respectfully
Ben K4ZEP

Quote from: ION on 2016.07.23, 13:24:59
Poynt, partzman and TK have diligently explained the nuances of careful measurement techniques based on their many years of bench time performing and debugging those kinds of tests.

I understand that there are many  "faithers" that have a deep disdain for people who like to use good measurement techniques. Then, they must remember that Graham's claims are solely based on measurement with a high end o'scope and a Clark Hess

As such his measurements are only as good as the limits of the equipment and the skill in using such, as has been pointed out.

So, faith based beliefs regarding the validity of the OU claims need to be "brought down to earth" with at least two other alternative measuring techniques, in parallel with the use of the high end o'scope and it's math calculations and the Clark Hess, as a double and triple check. This would then be good science.

Unfortunately. since Graham is holding some of the "secret sauce" cards to his chest it may not be possible to accurately replicate his claims. That he has disassembled the device only adds to the confusion.

Ben and Mark, welcome to OUResearch forum.

Whaat?  RULE NUMBER ONE for OU inventors is this:

If you have built a device that produces OU, Never, Never, NEVER take it apart !!!

If you absolutely must take something apart, build another identical one and take _that one_ apart !!!!

:o

Quote from: TinselKoala on 2016.07.23, 18:31:23
Whaat?  RULE NUMBER ONE for OU inventors is this:

If you have built a device that produces OU, Never, Never, NEVER take it apart !!!

If you absolutely must take something apart, build another identical one and take _that one_ apart !!!!

:o

OH YES, a thousand times NEVER NEVER~~~~~~~  Once the goose starts laying golden eggs..........don't kill the goose to see how she works!!!!!

Respectfully
Ben K4ZEP

Dear All.

I have just re read the last 3 pages of OU.com I cannot find any reference to disassembly, did I miss it ?   ???

Cheers Grum.

That's the word on the street, mighta been mentioned somewhere .

but i do believe its accurate.

Rule number one..never take it apart
Rule number two ,"see rule number one"

Chet

Quote from: Grumage on 2016.07.23, 19:14:06
Dear All.

I have just re read the last 3 pages of OU.com I cannot find any reference to disassembly, did I miss it ?   ???

Cheers Grum.

It was on this forum, in post #57:

http://www.overunityresearch.com/index.php?topic=3319.msg56251#msg56251

At the time I was there he had his conversion transformer taken apart with only one set of PM's attached................................


Itsu

Here are the release dates just posted by Aaron for the videos from the conference:


Release schedule so far:

7/25/2016 PAUL BABCOCK – BABCOCK DC MOTOR DISCLOSURE
7/27/2016 AARON MURAKAMI - COLD ELECTRICITY CIRCUIT
7/29/2016 PANEL DISCUSSION
8/1/2016 JIM MURRAY – FUNDAMENTALS OF THE TRANSFORMING GENERATOR
8/3/2016 GRAHAM GUNDERSON – MAGNETIC IMPLOSION TRANSFORMER
8/5/2016 MICHAEL WATERS – OPEN SYSTEM PHYSICS AND THERMODYNAMICS
8/9/2016 PETER LINDEMANN - BEDINI SG: BEYOND THE ADVANCED HANDBOOK
8/11/2016 KEN ROCHON - USING SOCIAL MEDIA TO SPREAD THE BREAKTHROUGH ENERGY MESSAGE
8/15/2016 ERIC DOLLARD – ELECTRICAL TRANSMISSION IN MULTIPLE COORDINATE SYSTEMS
8/18/2016 MORAY B. KING – FROM NANOBUBBLES TO BALL LIGHTNING: AN OVERVIEW OF NEW ENERGY POSSIBILITIES
8/23/2016 PROFESSOR ROBERT M. HARALICK – PARAMAGNETIC AND DIAMAGNETIC FIELD EXPERIMENTS
8/25/2016 AARON MURAKAMI – PULSED PLASMA MOTOR – KEY TO THE GRAY MOTOR?
8/27/2016 AARON MURAKAMI - SCOTCH YOKE BOB TEAL MOTOR

So it looks like Aug 3rd you will be able to purchase the video.

Quote from: Itsu on 2016.07.23, 21:04:50
It was on this forum, in post #57:

http://www.overunityresearch.com/index.php?topic=3319.msg56251#msg56251

At the time I was there he had his conversion transformer taken apart with only one set of PM's attached................................


Itsu

Thanks Itsu.

To my mind that meant Graham had only removed the BaFe magnets only. Let's hope that was the case?

Cheers Grum.

Quote from: Grumage on 2016.07.23, 23:23:57
Thanks Itsu.

To my mind that meant Graham had only removed the BaFe magnets only. Let's hope that was the case?

Cheers Grum.

I think he means that he removed the 'biasing magnets' that were sitting above the transformer

Well it seems like another claim that reinforces the reputations of those involved as the sellers of books and video's which they promise lead to OU when they really do not.
It is my contention that they know what is going on and have engineered the setup to show the same or similar type of OU that the Inductive resistor showed. Sneaking power through the switches/ic's maybe and big claims made with mediocre measurement.

Hows this for a novel approach ? Measure the input power as all the power consumed by the entire setup, in other words measure all the input power. duh all the power consumed needs to be considered.

This seems like a common and well worn trick. Any serious experimenter should be aware of this issue.

Is the input AC ? if so why ?

Just use a DC power supply and measure the input at the output of that.

I honestly am a bit confused as to why a setup promoted by those people is being discussed as legit by anyone.

Has the Ainsley affair been forgotten by some people. Gee wizz.

..

Quote from: Spokane1Also, Gram said that by adding a 100 pf Silver-Mica capacitor between the Source and Drain of one of the Back End FETS would completely destroy the OU effect.

By "Back End FETS" are you referring to the H-bridge FETs or the ones in the "Synchronous Diode" ?

Friends, there is a major clue contained here in this statement of "Gram's".

Quote from: Farmhand on 2016.07.24, 01:20:05
Well it seems like another claim that reinforces the reputations of those involved as the sellers of books and video's which they promise lead to OU when they really do not.
It is my contention that they know what is going on and have engineered the setup to show the same or similar type of OU that the Inductive resistor showed. Sneaking power through the switches/ic's maybe and big claims made with mediocre measurement.

Hows this for a novel approach ? Measure the input power as all the power consumed by the entire setup, in other words measure all the input power. duh all the power consumed needs to be considered.

This seems like a common and well worn trick. Any serious experimenter should be aware of this issue.

Is the input AC ? if so why ?

Just use a DC power supply and measure the input at the output of that.

I honestly am a bit confused as to why a setup promoted by those people is being discussed as legit by anyone.

Has the Ainsley affair been forgotten by some people. Gee wizz.

..

Mainly, I agree with you, especially about measuring all the input power to whatever parts are necessary for the device to function.  But we might as well give the claimants the "benefit of the doubt" at the moment. I think there are enough inconsistencies in the information we have to cause us seriously to question the validity of the claims, and it is here we should focus, not on the past reputations or behaviours of the "principals" involved. Yet.

But we also have the information that Gunderson will not be cooperating with us in delving into the matter. No more demonstrations, no "advanced" information, nothing. I think someone who was really doing "open source" research would be more than happy to address issues such as providing a full working schematic, doing proper placement of Hall probe so as not to be influenced by external magnetic fields, proper probe timing deskewing, proper measurement of all input power, addressing questions like "0.000 Watts input" but can't be disconnected from the _real_ input power source, etc. etc.

I'd like to know a lot more "advanced" technical details, but we have been told that we aren't going to get them. For example, how does a mosfet H-bridge manage to put out such a nice sinusoidal voltage waveshape when it is being driven by a rectangular pulse train? Of course all I have to go on is Spokane1's block diagram, Reiuki's photo of the "pwm pulses" sketch and one very poorly presented scopeshot from Gunderson's demonstration. (One great thing about DSOs is that they have the ability to produce really nice screenshots that have almost all information of interest displayed on them in very clear graphics. So why do we need to have an oblique, poorly focussed, color-shifted photograph of the scope, instead of a proper screenshot, to work with? Why, it is almost as though the presenter doesn't really want us to see things clearly....)

Interesting comments all.

So the show goes on.

The financial rewards are sufficient to keep some of the
purveyors in business while living a good life.

The personnel who make up the Deep Space Program
(U.S. Navy) and numerous other Black Projects know
well the true source of Over-unity or Free Energy.
They also are prepared to "pay the price" and meet other
imposed conditions in order to access the energy for their
less than noble (i.e. making war) purposes.  They have
discovered that the Other Worldly energy they so crave to
implement their Star Wars plans does not come free at all
but with burdensome "strings attached."

As with all things in today's World;  nothing is as it seems.
We are being played.

The Players will make their appearance upon Planet Earth in
the not too distant future.

Daniel Pomerleau is one who has been given limited access
to this kind of energy for purposes of demonstration. (http://sustainable-lifestyle.empowernetwork.com/blog/daniel-pomerleau-free-energy-wizard)  It is
real.

@Smudge:

In your pdf analysis of the waveforms in that scopeshot, you do a rough estimation of the input power by eyeballing and multiplying the input Voltage and input Current. Then you state,

QuoteIt is seen by the symmetry of the power waveform that the average input power is near zero, which is what was demonstrated.

What does this tell us about the nature of the power measured at this point? Is this consistent with the output of an H-bridge?

The "H Bridge" is quite a switching circuit which, when properly driven
with appropriate pulse frequency and width, can produce an output either
as a Modified Sine Wave or a Pure Sine Wave.

Is Graham driving his special H Bridge in such a manner to achieve the
latter output; the Pure Sine Wave?

.99 says that it is legitimate to measure the "input power" at the input to the transformer. As Smudge has shown, the _ average power_ at this point is indeed very close to zero. This is because the voltage and current waveforms are 90 degrees out of phase! We have a name for this kind of power.

So, if I apply this same reasoning to several of my own devices, such as the Micro QEG or my wireless power transmission systems, and count as "input power" only the voltage and current waveforms measured at the input to the _coil_ which forms the primary of an air-core transformer, I will also get an "average power" of quite near zero at that point. And it is clear that I can extract useful real power from the secondary of that transformer, to light up a light bulb, charge capacitors to high voltage, or even power a motor. Hence, by the same reasoning that Gunderson apparently is using, I am also producing an overunity COP, and a large, consistent one at that. And from a much simpler, cheaper circuit that anyone can build, that doesn't use any unobtanium parts and for which I have published complete details.


I can haz 2 million cheezburgers now?

Quote from: muDped on 2016.07.24, 05:18:59
The "H Bridge" is quite a switching circuit which, when properly driven
with appropriate pulse frequency and width, can produce an output either
as a Modified Sine Wave or a Pure Sine Wave.

Is Graham driving his special H Bridge in such a manner to achieve the
latter output; the Pure Sine Wave?

The key being "appropriately driven". Radically different PWM signals must be applied to each side of the H-bridge. This is quite a complicated means of producing a sine wave output. And in Gunderson's implementation yet another PWM signal must be applied to get the 33 percent dead time between each full cycle of the sine wave output.  The H-bridge's output is a critical part of the functioning of the device and its input power should be counted as part of the input for the overall device. But if not... then I can claim "OU" too for my devices, since as I have explained, the input to the primary of the air-core transformer in my devices also can be measured as "zero average power".

My devices use a much simpler Royer oscillator feeding a tuned tank circuit to achieve the pure sine wave output at a given frequency. Producing the gated output would involve a slight amount of further complexity, easily enough accomplished with a 555 circuit and an additional mosfet.  However this is not necessary for my devices, because the pure, continuous sine wave is very effective in producing the "zero input power" measurement at the primary coil input, which can be effectively transformed into real power at the secondary's output.

http://www.youtube.com/watch?v=xyVZWkYAvkk

When I was making these videos I had no idea that I only needed to count the average input power to the primary coil as "input", so I showed the actual DC power input to the oscillator as my input!

I'm in the exact same requirement for other device : a pure sine wave output with quite high voltage and a variable frequency.
Can it be made with a op-amp based sinewave generator and a power amplifier on transistors ? An inverter charging HV capacitor and a transistor amplifier driven by sinewave signal connected to that capacitor generating power sine output ?
Is that possible ?

Given this scenario, can we get real power at the load?

Quote from: TinselKoala on 2016.07.24, 05:39:11
.99 says that it is legitimate to measure the "input power" at the input to the transformer. As Smudge has shown, the _ average power_ at this point is indeed very close to zero. This is because the voltage and current waveforms are 90 degrees out of phase! We have a name for this kind of power.

So, if I apply this same reasoning to several of my own devices, such as the Micro QEG or my wireless power transmission systems, and count as "input power" only the voltage and current waveforms measured at the input to the _coil_ which forms the primary of an air-core transformer, I will also get an "average power" of quite near zero at that point. And it is clear that I can extract useful real power from the secondary of that transformer, to light up a light bulb, charge capacitors to high voltage, or even power a motor. Hence, by the same reasoning that Gunderson apparently is using, I am also producing an overunity COP, and a large, consistent one at that. And from a much simpler, cheaper circuit that anyone can build, that doesn't use any unobtanium parts and for which I have published complete details.


I can haz 2 million cheezburgers now?

Yes you may, but first let's reason thru this together for a moment.  As I posted earlier, Graham's device is exactly as you say and that is, it utilizes a reactive input that produces a real resistive power output thru a circuit that we will assume is unique in function with our limited information. Forget for a moment how or what develops the sine wave input (see below) and focus on the fact that the input power averages near zero or below in reference to the sine wave and not in reference to any sine generating means. This is the same as your Micro QEG. So my point is, in this context using these "references", Graham's device as well as your own Micro QEG produce OU as indicated by the measurements taken.

Maybe at this point we should examine what possibly could be done to utilize this situation! What type of circuit design would allow us to generate a sine wave from a DC source and take advantage of the reactive input power? Impossible one might say, but actually it is done at 50/60Hz in high quality DC to pure sine wave AC inverters that handle both inductive and capacitive loads. If one investigates these inverter types, the efficiency of returning the negative energy to the DC supply (this is actually what they do) is around 50% or so. The typical reactive to resistive power devices I've designed exhibit a high reactive power (VARs) to resistive power ratio in the order of 5-15:1 which would be problematic with such a low efficiency return to the DC supply even at low frequencies.

So we have defined the problem. Is it solvable?

BTW, the switching H bridge produces the sine wave with the parallel resonant network by virtue of the "swinging choke" connected to the DC supply that provides a near constant current to the zero voltage controlled switches. This is basically a variation of the common current fed parallel resonant inverter design that was used for years in the early solid state ballasts for fluorescent lighting using two switching devices and was extremely efficient and simple. 

IMO because of the input choke, Graham's device as demonstrated would not produce OU when measuring even just the H bridge inputs at the DC source and would not self-power with the output connected to the input. I would love to be proved wrong!

pm   

Quote from: poynt99 on 2016.07.24, 13:19:54
Given this scenario, can we get real power at the load?

Theory and practice tells us that we cannot get real power at the load if the voltage and current are exactly 90 degrees out at the input.

If the input power factor is "near zero", then the difference will be sufficient to produce some real power at the load, and this small difference fools many into believing they have achieved OU, however it all comes out in the wash when a careful tally is made using very careful measurement.

There is, however,  the possibility (in e.g. an OU device) that the zero power factor in a pure (ideal) reactive circuit can act as a stimulus to some other heretofore unknown process that releases real power to a load.

High circulating amps and high voltage of e.g. a very high Q tank circuit may act as such a stimulus to the right materials under the right conditions to release real power to the load. The important thing is not to kill the Q by trying to extract power directly.

So, to me the answer is no and then yes.


edit: sorry partzman, I was typing this while you had posted and did not see your post.

Quote from: ION on 2016.07.24, 14:01:11
Theory and practice tells us that we cannot get real power at the load if the voltage and current are exactly 90 degrees out at the input.

Agreed.

And as TK's devices are not OU, there are two possible reasons why real power is delivered to his loads;

1) The phase relationship isn't exactly 90º.
2) The measurement isn't correct (i.e. there is phase skew).

I suspect it is a combination of the two, and I also suspect this is what Gunderson will find if he digs deep enough.

Quote from: poynt99 on 2016.07.24, 14:19:43
Agreed.

And as TK's devices are not OU, there are two possible reasons why real power is delivered to his loads;

1) The phase relationship isn't exactly 90º.
2) The measurement isn't correct (i.e. there is phase skew).

I suspect it is a combination of the two, and I also suspect this is what Gunderson will find if he digs deep enough.

Also agreed, and more times than we can count it is 1) or 2), however we must be willing to also consider the third possibility i.e  in an supposed OU device that the zero power factor in a pure (ideal) or near pure reactive circuit can act as a stimulus to some other heretofore unknown process that releases real power to a load.

In the Sweet VTA, the SM devices and others, we often see input and output coils at right angles to each other, such that an output power cannot be directly related to the input by normal Faraday induction.

Smudge has a few papers that offer some promise of possible explanations for OU effects based on stimulating the right materials in the correct way. These papers also show coils, magnets and ferrite materials or a liquid being stimulated at right angles. Now whether the stimulation can act as a release agent is TBD. NMR and NAR seem to be viable areas to explore.

It might be noted that in a very high Q tank circuit power will be lost either through imperfect materials (dielecric or resistive) or radiation loss. The imperfect materials we will have to live with, however there are non-radiative structures that might be interesting to look at. Due to their low radiative loss, they my require very little real power input, yet produce many thousands of amperes of circulating currents at near zero power factor. Such fields would almost have to be noticed by e.g. nearby protons or magnetic dipoles.

Quote from: TinselKoala on 2016.07.24, 05:00:46
@Smudge:

In your pdf analysis of the waveforms in that scopeshot, you do a rough estimation of the input power by eyeballing and multiplying the input Voltage and input Current. Then you state,

What does this tell us about the nature of the power measured at this point? Is this consistent with the output of an H-bridge?

Don't know enough about the inner workings of the H bridge to answer that.  I was just trying to make sense of what little information we had been given so that I could take things further in my mind.  Posted it to help others understand things better.  It could all be wrong because I didn't know where the zero current line was on the scope shot, I assumed it was in the center of the screen.

Can I say that I have known Graham for many years, although we have never met in person.  First came across him on Dave Squire's OU Builders forum, then we both worked for MPI, followed by Chava Energy.  Have been involved with conference calls with him.  I have great respect for his abilities and know he fully understands the limitations of the measurements he takes.  I think he truly believes he has hit on something good, after many years of achieving only marginal overunity COP's.  Perhaps it was a bit naive for him to go public, since it was bound to create the enormous amount of interest, criticism, demand for information and general harassment that it clearly has.  I think I would be overwhelmed and want to disappear into the woodwork.

As regards TK's observations on measurements, if you use a scope to take measurements do you include the AC input power to the scope on the basis that some of that power might inadvertently get directed into your power measurement?  I think not.  Clearly the whole thing on display by Graham is not OU, but I think it quite legitimate to attempt to measure power flows somewhere inside the device then draw conclusions.  And knowing Graham I think he will have taken into account things like measurement accuracy, phase or time delays in the probes, yes even the effect of external magnetic fields on the hall probes etc.. 

As regards his $2M offer he knows no one will take him up.  Muammer Yildiz demonstrated his magnetic motor numerous times, and made the same offer (I witnessed it at Delft University).  Without some credible route to get from a demonstration of a device that no-one know how it works, that has been got to work by a lot of tinkering with components, to something that will produce a profit for the investor, then just forget it.

For my part I am prepared to believe that his device could be doing something remarkable, and I would like to know why that could be.  So I will continue my deliberations and post my thoughts here.  If Graham ultimately finds he has been misled then I am sure we will know soon enough.

Smudge     

QuoteHigh circulating amps and high voltage of e.g. a very high Q tank circuit may act as such a stimulus to the right materials under the right conditions to release real power to the load. The important thing is not to kill the Q by trying to extract power directly.

There has to be a reception device of equal q before the load. Tesla showed this in his coupling patents.

Quote from: Smudge on 2016.07.24, 16:05:30
Don't know enough about the inner workings of the H bridge to answer that.  I was just trying to make sense of what little information we had been given so that I could take things further in my mind.  Posted it to help others understand things better.  It could all be wrong because I didn't know where the zero current line was on the scope shot, I assumed it was in the center of the screen.

That's not the only or even the primary reason it could be all wrong. But on the photo of the scopeshot you can see (barely) the positions of the baseline indicators on the left side of the trace display area.

Quote

Can I say that I have known Graham for many years, although we have never met in person.  First came across him on Dave Squire's OU Builders forum, then we both worked for MPI, followed by Chava Energy. 

I am truly sorry and even shocked to read that. I'm sure our old friend MarkE would be rolling (or perhaps even LOLing) in his grave over that disclosure.  I've read much of what you have written on this forum and I'm impressed by your level of knowledge and your willingness to engage, but I have to admit that knowing you worked for MPI and Chava has resulted in a certain hike in my level of skepticism.

Quote

Have been involved with conference calls with him.  I have great respect for his abilities and know he fully understands the limitations of the measurements he takes.  I think he truly believes he has hit on something good, after many years of achieving only marginal overunity COP's.  Perhaps it was a bit naive for him to go public, since it was bound to create the enormous amount of interest, criticism, demand for information and general harassment that it clearly has.  I think I would be overwhelmed and want to disappear into the woodwork.

Yes, perhaps that is the wisest choice for him at this point. Achieving "only marginal overunity COPs"? It never happened, not even marginally. And it certainly isn't happening with this apparatus.

Quote


As regards TK's observations on measurements, if you use a scope to take measurements do you include the AC input power to the scope on the basis that some of that power might inadvertently get directed into your power measurement?  I think not. 

Well, you certainly must take into account the effect of the scope and other instruments on the claims of OU, although not in the silly way you suggest. After all, the _only_ place that "Overunity" or excess COP shows up in any of these devices (here I include Gunderson's other work, the patents assigned to MPI and later Chava, etc.) is in the instrumental measurements. But you cannot run any useful loads using the pretty squiggly lines on an oscilloscope or the numbers spit out by a mis-used power analyzer. Too bad.

Quote

Clearly the whole thing on display by Graham is not OU, but I think it quite legitimate to attempt to measure power flows somewhere inside the device then draw conclusions. 

Then it appears to me that you are forced to conclude that my devices are also OU. Are you ready for that? Somehow I don't think so.

Quote

And knowing Graham I think he will have taken into account things like measurement accuracy, phase or time delays in the probes, yes even the effect of external magnetic fields on the hall probes etc.. 

I am prepared to believe that Graham knows something about what he is doing, and this means to me that he is perilously close to running a deliberate deception in order to stir up interest and obtain funding. The more charitable interpretation is that he is indeed fooling himself. Experimenter bias can be very difficult to overcome, especially when there is money and fame involved. The association with Mark Goldes, MPI, Chava, etc. runs more than one big red flag up the pole. However, until we hear credible information about, for example, the probe skew correction, the probe position wrt both stray inductances and external magnetic fields, we are just guessing. Tell me, did _you_ know about or consider the effect of probe skew, and the ability of the scope to correct for it, before I mentioned it?

http://www.youtube.com/watch?v=IgvFHejoQEk

http://www.youtube.com/watch?v=KWDfrzBIxoQ

Quote

As regards his $2M offer he knows no one will take him up.  Muammer Yildiz demonstrated his magnetic motor numerous times, and made the same offer (I witnessed it at Delft University).  Without some credible route to get from a demonstration of a device that no-one know how it works, that has been got to work by a lot of tinkering with components, to something that will produce a profit for the investor, then just forget it.

Through my own network of associations I could provide just such a credible route -- if only the major issues could be overcome first. Demonstrate true, reproducible OU to me, so that I can demonstrate it to my contacts, and the veritable Gates of Heaven would open unto the inventor, and maybe even a tiny bit of that largesse would wind up falling on me.  But this present claim, coming from associates of MPI and Chava ... well, let's just say there are some high hurdles to overcome before credibility can be attained. The first one would be to account properly for all the power input to all parts of the system that are necessary for it to light up that little bulb.

Demonstrate true OU from a device with electrical inputs and outputs, with COP greater than 1.3 to ME and I will show you how to self-loop the system so as not to need any external power supply for continued operation once it is started. That's my big claim, which nobody has challenged yet-- because nobody actually has such a device.

And please don't tell me that you actually believe Yildiz's claims are true. A great deal of analysis occurred after SA's visit to one of Yildiz's demonstrations that absolutely demolished any hint of credibility there.

Quote

For my part I am prepared to believe that his device could be doing something remarkable, and I would like to know why that could be.  So I will continue my deliberations and post my thoughts here.  If Graham ultimately finds he has been misled then I am sure we will know soon enough.

Smudge     

The "remarkable" thing that this device is doing is that it is producing measurements that are fooling instrumentation which is not being used or interpreted properly. Clean up the act, measure properly, perform true experiments with correct controls, invite independent and competent outside investigators to work with the actual apparatus that produces the "OU" measurements in Gunderson's hands... then you might have something to write home about. "IF" Gunderson is being misled, it is he himself doing the misleading, with great help from a few other people. He is doing nothing to clear up the issues and has even stated outright, apparently, that he will not be doing so in the future.


All of that being said... sure, go ahead, all you experimenters, try to replicate Gunderson's device. It will at least keep you busy and off the streets! Maybe we will all learn something about metrology from your efforts.

Quote from: poynt99 on 2016.07.24, 14:19:43
Agreed.

And as TK's devices are not OU, there are two possible reasons why real power is delivered to his loads;

1) The phase relationship isn't exactly 90º.
2) The measurement isn't correct (i.e. there is phase skew).

I suspect it is a combination of the two, and I also suspect this is what Gunderson will find if he digs deep enough.

Bingo.

Here's an example scopeshot of the "input power" to the primary coil of my microQEG. Is the instrument reliable and accurate enough? (There is a slight, non-adjustable DC offset on the scope's CH1 which affects both phase and amplitude measurements of that signal). Is my measurement technique good enough? (I have passive scope probes attached quite near the oscillator "coil power supply" output, I'm using a noninductive current sense resistor, etc.etc.) Are the probes properly compensated? Is it legit to think that equal length ground connections will equate to equal phase delays for both probes? Is the scope's "average power" computation accurate? Have I programmed the scope correctly? The output of the air-core loose-coupled secondary is driving a small light bulb (and keeping two capacitors charged) at around 800 mW (7.2 VDC @ 110 mA). Have I measured this output power properly? Does my measurement concur with other techniques of measurement of the output power, like bulb illumination? Am I simply lying about the whole thing?  If I were to present this device at an "Overunity Conference" I'm quite sure I could stir people up and get them to believe it was indeed exhibiting OU performance. And who would speak out to challenge that claim? Why, I'd have to set up a new alias and debunk my own device!   ^-^

Quote from: poynt99 on 2016.07.24, 13:19:54
Given this scenario, can we get real power at the load?

A little late in responding but YES!

pm

After taking a close look at G's scope shot, it appears that the input power approaches ~500W reactive. This is a really high reactive to resistive power ratio which would make any looping attempts next to impossible IMO.

pm

Quote from: partzman on 2016.07.25, 13:13:56
A little late in responding but YES!

pm

Really?

How is that possible if the source is supplying an average power (real) of 0W?

Quote from: poynt99 on 2016.07.25, 13:53:09
Really?

How is that possible if the source is supplying an average power (real) of 0W?

I would also like to know how it is possible or where does the real power come from?

If there are 500 VA at near zero power factor, you don't have to be off by much to get 5 to 10 Watts of real power hidden by some small non-zero portion of the power factor, enough to light a small lamp.

BTW, does anyone know the type of lamp and how bright or dimly it was lit? This would allow an approximate back calculation of the actual power factor.

Quote from: poynt99 on 2016.07.24, 14:19:43
Agreed.

And as TK's devices are not OU, there are two possible reasons why real power is delivered to his loads;

1) The phase relationship isn't exactly 90º.
2) The measurement isn't correct (i.e. there is phase skew).

I suspect it is a combination of the two, and I also suspect this is what Gunderson will find if he digs deep enough.

OK,i must have missed something some where here,as i see nothing special about being able to produce usable power on the output,when the input phases are 90* out.
Using my hybrid toroid transformer,i can sweep from 85* through to 95* with regard to the voltage/current phase relationship,and can produce usable power on the secondary side of the transformer.

In the scope shot below,i have an AC 24 volt LED (LED with built in rectifier)on the output of the transformer,and it will remain brightly lit throughout the 85 to 95* phase swing. The scope shot below is very close to(if not ,exact) showing a 90* phase shift between voltage and current,and the LED is lit brightly--so what am i missing here?


Brad

Brad

Where are the probes placed for this measurement? If it is input to output phase shift, this is something different than what Poynt is asking.

Capacitance coupling between windings can relate to transfer of power.

So, are you measuring power  (current and voltage phase difference at the input to your transformer) or are you measuring input / output phase difference?

Quote from: poynt99 on 2016.07.25, 13:53:09
Really?

How is that possible if the source is supplying an average power (real) of 0W?

Well, I'm sure you and others have seen my MEI experiments that were recently posted on my bench but I will attach a sample here with a schematic for others to view.

The example below is extreme in that the current lead is >90'. This is made possible by the physical coil structure with it's distributed capacitance and electrical connections that permit the magneto electric induction to take place. What you see is the result. These measurements were taken with extreme care and are relative to the individual points measured and do indicate that resistive power is able to be produced with reactive input and with gain.

With an output power of 33.2 watts rms, the VAR to watts average ratio ~11:1 for this example.

So, is this device OU when considering the power required to supply the class E amplifier? No.

pm

Edit: The DUT includes T1, RL and Rs. 

What does, or would the sim say?

Quote from: poynt99 on 2016.07.25, 22:06:04
What does, or would the sim say?

Dear poynt99,

Graham (at the conference) said that he used LT Spice to simulate his machine with the weird wave forms and found that the simulator actually displayed a positive output energy in excess of the input energy. Neither he nor I have an idea as to how a classical calculator can come up with a non-classical phenomena. I do know that LT Spice can process non-world components like negative resistance and negative transformer coupling factors and who knows what else..

I'm sure that Graham is not going to be going back to the simulator for a while.

Spokane1

Spokane 1,

I am fairly confident that if the simulator is indicating OU, then either there is leakage from other components (as we've has postulated about GG's device), or the measurements were not done correctly.

Partzman had a sim circuit that was indicating OU, which later turned out to be a measurement error (once I helped him with the proper way to do the power measurement in SPICE).

Chances are that GG's sim was not being measured properly. If you could obtain the sim files, we can perhaps sort out the error.

Quote from: ION on 2016.07.25, 15:23:46
Brad

Where are the probes placed for this measurement? If it is input to output phase shift, this is something different than what Poynt is asking.

Capacitance coupling between windings can relate to transfer of power.

So, are you measuring power  (current and voltage phase difference at the input to your transformer) or are you measuring input / output phase difference?

It can be either way ION,as the voltage trace for the input is in phase with the output,but the current trace on the input is 90* out of phase with both voltage input,and output across the load-->or 370* out of phase with the load,depending how you wish to place your scope across the AC output-of course :)

So 1-the input voltage and current are 90* out of phase.
2-the output voltage and current are 90* out of phase with the input current
3-the output voltage and current are in phase with the input voltage.
4-the input current is 90* out of phase with the input voltage,and output voltage and current.


Brad

Quote from: poynt99 on 2016.07.25, 23:04:08
SPICE).

If you could obtain the sim files, we can perhaps sort out the error.

Dear poynt99

Thank you for the review. I doubt we will be getting those sim file for many moons if ever. By then we will be much more involved in more serious matters or this will be another dead horse.

Spokane1

Let the Junk Box Replication Experiments Begin!

First Step is to Decode the logic board. This is where the timing signals are derived and control both the H-Bridge and the Backend Diode. I think I have most of it figured out - now to draft it and get my head around how it works. I'm sure there are several other approaches that could be used.

Quote from: TinMan on 2016.07.25, 23:44:01
It can be either way ION,as the voltage trace for the input is in phase with the output,but the current trace on the input is 90* out of phase with both voltage input,and output across the load-->or 370* out of phase with the load,depending how you wish to place your scope across the AC output-of course :)

So 1-the input voltage and current are 90* out of phase.
2-the output voltage and current are 90* out of phase with the input current
3-the output voltage and current are in phase with the input voltage.
4-the input current is 90* out of phase with the input voltage,and output voltage and current.


Brad

I can only guess that there may be a sneak path through capacitance coupling between the windings and back to the scope ground on the secondary that is lighting the LED. If there is no ground path on the secondary then this does not apply, but since you are giving output phase shift with respect to input, I assume you have a scope ground also on the secondary.

It is very difficult to eyeball an exact 90* shift when many cycles are displayed on the scope, much better resolution if only one cycle is shown, but still difficult to be exact.

Since you are varying the phase shift, this may also not apply.

Have you attempted to measure input and output power (with a non-inductive resistive load) to see if you have a COP> 1 or tried to loop the transformer for sustained stand alone operation? Maybe you have done all this on another thread.

Perhaps continue to pursue this if you think you have something.

Quote from: ION on 2016.07.26, 01:46:21
I can only guess that there may be a sneak path through capacitance coupling between the windings and back to the scope ground on the secondary that is lighting the LED. If there is no ground path on the secondary then this does not apply, but since you are giving output phase shift with respect to input, I assume you have a scope ground also on the secondary.

It is very difficult to eyeball an exact 90* shift when many cycles are displayed on the scope, much better resolution if only one cycle is shown, but still difficult to be exact.

Since you are varying the phase shift, this may also not apply.

Have you attempted to measure input and output power (with a non-inductive resistive load) to see if you have a COP> 1 or tried to loop the transformer for sustained stand alone operation? Maybe you have done all this on another thread.

Perhaps continue to pursue this if you think you have something.

I did have a thread on this transformer,and if i recall correctly,Smudge was analysing  the results from some test we carried out,which did seem to show a negative resistance result.

I would need a 4 channel scope with isolated grounds to confirm this,but at the moment all my spare cash (+ some),is being spent on the issue mentioned in my last email to you,and i think you would agree that that is priority  ATM..

What i dont understand is why it would seem  so odd that real power cannot be produced from the output of the transformer,when the voltage and current are 90* out of phase on the input side?.
After all,it is current flow through the primary that produces the magnetic field-not the voltage across it,and this we can see with regard to super conductors. The only problem with this in my toroid transformer,is the current and voltage on the output is in phase with the voltage on the input-not the current,so not sure what is going on there?.

Perhaps this phase relationship thing only complies with your standard type transformer?.


Brad

Quote from: TinMan on 2016.07.26, 04:23:31
After all,it is current flow through the primary that produces the magnetic field-not the voltage across it

Brad,

It's a bit more complicated than that.  The magnetic field in a transformer is produced from both the primary current and the secondary current.  These two currents almost cancel each other out, so it is the difference between primary and secondary ampere-turns that actually creates the field, it is not just the primary current.  In fact with the transformer under load most of the primary current does not produce any field because of that cancellation, it is only the small so-called magnetizing current that creates the field.  And you can't divorce the voltage from the field since it is related to it.  If you do the math you find that for transformers connected to low impedance voltage sources, as is normally the case in power transformers, it is the voltage and frequency that determines the magnetizing current hence the magnetic field.  The much larger primary load current, that is opposed by the secondary load current, does not produce any field.  But it does present a mmf to the magnetic circuit, in the same way that the secondary current produces an opposing mmf.  It is that balance of opposing mmf's that is crucial to the way the transformer works, but unfortunately this isn't taught in electrical engineering courses.  They teach a transformer equivalent circuit that has at its heart a "perfect" transformer that somehow, magically, without the use of magnetic fields, transforms voltage according to the turns ratio and transforms current according to the reciprocal of the turns ratio.  Then the "imperfections" of a practical transformer, that includes the need for a magnetic field, are accounted for by adding components around the outside of that magical component.  Unfortunately this approach completely hides the inner workings of the magnetic domain.  So when you get unusual transformer designs, be it like your buried secondary or like Cobb's double winding, few people know how to properly analyze them. 

Smudge

Quote from: ION on 2016.07.25, 14:18:08
I would also like to know how it is possible or where does the real power come from?

Sorry for the delay ION but just got back to my computer.

In my MEI designs there is no outside source of energy as best I can determine at this time. The transformer construction and electrical properties allow real power to be developed from the reactive input.

Quote
If there are 500 VA at near zero power factor, you don't have to be off by much to get 5 to 10 Watts of real power hidden by some small non-zero portion of the power factor, enough to light a small lamp.

I agree totally. One design objective is to obtain as low a ratio of VAR/watts as possible.

Quote
BTW, does anyone know the type of lamp and how bright or dimly it was lit? This would allow an approximate back calculation of the actual power factor.

No idea here.

As a side note, I see from a post by Spokane on OUdotcom that Graham has decided not to open source his new technology.

pm

Quote from: Spokane 1 on 2016.07.26, 01:28:14
Let the Junk Box Replication Experiments Begin!

Hi All!

Just a few of the things to understand and consider!  If you GOOGLE "H Bridge motor drivers"......thousands of articles. 

https://www.embeddedrelated.com/showarticle/424.php
http://www.modularcircuits.com/blog/articles/h-bridge-secrets/h-bridges-the-basics/

It gets complicated fast!  With the voltage ratings and current ratings of those H components.  I'm real curious which is blowing, the FET's or the Diodes in them!  Who knows.

We have to "forward, reverse and float inductor off and or maybe short to ground!  Too bad GG has decided to not open source!  Probably he was made an offer that he can't refuse! :-X

Ben K4ZEP

Quote from: ION on 2016.07.25, 14:18:08
I would also like to know how it is possible or where does the real power come from?

If there are 500 VA at near zero power factor, you don't have to be off by much to get 5 to 10 Watts of real power hidden by some small non-zero portion of the power factor, enough to light a small lamp.

BTW, does anyone know the type of lamp and how bright or dimly it was lit? This would allow an approximate back calculation of the actual power factor.

The bulb is an automotive brake light bulb, I believe, maybe a #1156.

From one of the images it looks like it is pretty bright. The one I have just tested on my bench draws almost exactly 2 amps when supplied with 12 VDC and it's pretty damn bright and white at that 24 watt power.  At 6 VDC it draws 1.4 amps (8.4 Watts) and is pretty bright but yellowish.

See the images below. There are some things to note in the face-on image showing the two Clarke-Hess power analyzers.

First, the bulb does not seem to be glowing in that image. Yet the "output" C-H is showing a reading. Second, the "input" C-H has its Current input sense bypassed. This would normally be in series between the DUT and its power supply. Instead of the series connection through the instrument the leads are simply shorted with a red cliplead. So naturally the instrument reads Voltage but no current, and therefore should naturally read 0 on the Watts (leftmost set of numbers), but is showing a slight Current indication. Next, both the voltage and current sense leads are fed through toroids, probably to suppress spikes.


The final image is from my cellphone camera showing the brightness of the #1156 at just under 13 Watts DC. I took this with the cellphone camera instead of the DSLR because it's comparable to the "end_view.jpg" image from the Gunderson demo.

Quote from: K4zep on 2016.07.26, 13:03:10
Hi All!

Just a few of the things to understand and consider!  If you GOOGLE "H Bridge motor drivers"......thousands of articles. 

https://www.embeddedrelated.com/showarticle/424.php
http://www.modularcircuits.com/blog/articles/h-bridge-secrets/h-bridges-the-basics/

It gets complicated fast!  With the voltage ratings and current ratings of those H components.  I'm real curious which is blowing, the FET's or the Diodes in them!  Who knows.

We have to "forward, reverse and float inductor off and or maybe short to ground!  Too bad GG has decided to not open source!  Probably he was made an offer that he can't refuse! :-X

Ben K4ZEP

You might find more relevant information if you look for "H-bridge Solid State Tesla Coil Drivers"  where you will also find thousands of articles.

I'm sure Gunderson's decision not to open-source is based on a desire not to be proven wrong by a bunch of amateurs who think they know how to measure stuff.

An offer he can't refuse? Not likely, but.... one had better be careful with those kinds of offers.... if you don't deliver what you've claimed you might end up sleeping with the fishes.

As to the exact failure mode of the mosfets... sometimes it can be hard to tell. Even if you _can_ find all the pieces.

Quote from: Smudge on 2016.07.26, 10:25:35
Brad,

It's a bit more complicated than that.  The magnetic field in a transformer is produced from both the primary current and the secondary current.  These two currents almost cancel each other out, so it is the difference between primary and secondary ampere-turns that actually creates the field, it is not just the primary current.  In fact with the transformer under load most of the primary current does not produce any field because of that cancellation, it is only the small so-called magnetizing current that creates the field.  And you can't divorce the voltage from the field since it is related to it.  If you do the math you find that for transformers connected to low impedance voltage sources, as is normally the case in power transformers, it is the voltage and frequency that determines the magnetizing current hence the magnetic field.  The much larger primary load current, that is opposed by the secondary load current, does not produce any field.  But it does present a mmf to the magnetic circuit, in the same way that the secondary current produces an opposing mmf.  It is that balance of opposing mmf's that is crucial to the way the transformer works, but unfortunately this isn't taught in electrical engineering courses.  They teach a transformer equivalent circuit that has at its heart a "perfect" transformer that somehow, magically, without the use of magnetic fields, transforms voltage according to the turns ratio and transforms current according to the reciprocal of the turns ratio.  Then the "imperfections" of a practical transformer, that includes the need for a magnetic field, are accounted for by adding components around the outside of that magical component.  Unfortunately this approach completely hides the inner workings of the magnetic domain.  So when you get unusual transformer designs, be it like your buried secondary or like Cobb's double winding, few people know how to properly analyze them. 

Smudge

Most useful insight.

Spokane1

Quote from: TinselKoala on 2016.07.26, 14:18:50
The bulb is an automotive brake light bulb, I believe, maybe a #1156.

From one of the images it looks like it is pretty bright. The one I have just tested on my bench draws almost exactly 2 amps when supplied with 12 VDC and it's pretty damn bright and white at that 24 watt power.  At 6 VDC it draws 1.4 amps (8.4 Watts) and is pretty bright but yellowish.

Dear TK,

Your assessment of the load lamp seems pretty accurate. The second photo you posted is the one I took at the convention. Graham had just shut the machine off about 10 seconds prior, so we are not going to get any useful lamp brightness information from that one.

Spokane1

In regards to Heisenberg's quote:
And you will find the enemy too or he will find you.
Most assuredly if one peeps, the fox will find the henhouse.
The technology is not allowed here, at all.

Any argument is easily beaten down by the history of prior prominents. They were all intelligent and esteemed at one point.
Tesla, Schauberger, Keely, E.V. Grey, Stanley Meyers, Moray. The list goes one.
I wonder who the next prairie dog will be?

Quote from: giantkiller on 2016.07.26, 16:12:01
In regards to Heisenberg's quote:
And you will find the enemy too or he will find you.
Most assuredly if one peeps, the fox will find the henhouse.
The technology is not allowed here, at all.

Any argument is easily beaten down by the history of prior prominents. They were all intelligent and esteemed at one point.
Tesla, Schauberger, Keely, E.V. Grey, Stanley Meyers, Moray. The list goes one.
I wonder who the next prairie dog will be?

Tesla would roll in his grave if he saw your list. The inventions of Nikola Tesla, or derivations of them, are all around you, from the radios that connect you to the internet, the logic gates that run your computer, the AC motors that run your appliances, and many more. Those others.... frauds, fakers, deluded old men, none of whom has contributed a single _real_ thing in all their lifetimes of tomfoolery.

Quote from: Spokane 1 on 2016.07.26, 14:46:07
Dear TK,

Your assessment of the load lamp seems pretty accurate. The second photo you posted is the one I took at the convention. Graham had just shut the machine off about 10 seconds prior, so we are not going to get any useful lamp brightness information from that one.

Spokane1

That makes sense. The bulb is still connected to the capacitor bank and has drained it down to about 1 volt according to the C-H analyzer and looks like it is drawing about a quarter of an amp. My own test #1156 draws about 0.6 A at 1 volt, so maybe Gunderson's bulb is a different type, but in the same package type. Perhaps it is a #1157 and he is using the higher-resistance filament in that dual-filament bulb. I'll have to dig around to see if I can find a  #1157 in my bulb stash.

Now.... what about the "input" C-H setup? Did Gunderson conduct the demonstration with it wired the way it is shown in the photo? I mean, with the Current side not connected and bypassed with the red cliplead? If that's how he got his "0.000 W" readings..... well, just go ahead and insert a ROFL here.  Just kidding.... he would not have done that. Surely not.... still, the setup does look rather suspicious.

Quote from: TinselKoala on 2016.07.26, 21:59:36
That makes sense. The bulb is still connected to the capacitor bank and has drained it down to about 1 volt according to the C-H analyzer and looks like it is drawing about a quarter of an amp. My own test #1156 draws about 0.6 A at 1 volt, so maybe Gunderson's bulb is a different type, but in the same package type. Perhaps it is a #1157 and he is using the higher-resistance filament in that dual-filament bulb. I'll have to dig around to see if I can find a  #1157 in my bulb stash.

Now.... what about the "input" C-H setup? Did Gunderson conduct the demonstration with it wired the way it is shown in the photo? I mean, with the Current side not connected and bypassed with the red cliplead? If that's how he got his "0.000 W" readings..... well, just go ahead and insert a ROFL here.  Just kidding.... he would not have done that. Surely not.... still, the setup does look rather suspicious.

Dear TK,

I have wondered about that as well. I didn't notice it at the time but there is no doubt that there is no current inputs to the Wideband Power Analyzer for the input to the H Bridge. My photo was taken just after the machine was turned off so this is how it operated during the entire presentation.

That is going to be the first question I ask Graham the next time we go out for Pizza and Chicken Wings.

I know this one presumed omission might be a game stopper for anyone reviewing this material and that would be a legitimate call.

I am going to take the path that assumes that Graham knows what he was doing and had some technical work around to account for the missing connections and why his approach provided more accurate data. If there was no current data then the instrument would have been unable to calculate a 1.53 watt input value when the demonstration first started. I would say that the instrument only registered the 0.000 reading during the last 5 minutes of the show.

This is a good catch. Did any one else besides yourself notice it?

Spokane1

Doesn't the "input" C-H monitor the _output_ of the H-bridge going to the input of the transformer?

I have no idea if anyone else noticed it. I'm fairly familiar with the use of the Clarke-Hess 2330 myself, though, so I know what to look for on the instrument. I str that .99 poynted out the toroid filters on the patch wiring. He probably noticed the Current bypass too.

I'm not sure what effect this filtering would have on the overall measurement problem. Since the Current patch lead is making a series connection with the C-H, going from the H-bridge to the transformer, the filtering would keep spikes out of both the C-H and the transformer. At first I thought maybe the spikes could be sneaking energy into the transformer that was deliberately being filtered out of the C-H. But since the circuit is a series one for the Current, that's probably not the case.

If, that is, the C-H is being patched in properly to begin with, which I can't tell from the photo.

Performed the whole demonstration with the "input" C-H current input bypassed/disconnected? That's hardly kosher IMHO.  Under certain conditions the C-H display will hold the last value sensed, which might explain the 1.53 W starting reading you saw.

Here's the brochure with specifications for the C-H 2330.
http://www.clarke-hess.com/2330.html

I recently saw one listed, manual included, for 300 dollars US on Ebay. I have a friend about 100 miles away who has one I can borrow but it costs me about 30 dollars to make the round trip to pick it up, and the same to return it.

It's pretty hard to tell from the low-resolution photo I have to work with but here are my best "guesses" as to the settings on the two C-H analyzers.

"input" (RH) instrument settings, from R to L:
Voltage Range 200V, showing 131.1 V
RMS/PEAK in RMS mode, Auto range OFF
Current (disconnected) Range 50 mA, showing 00.01 mA
Function: PWR, W range, 00.000 W indicated

"output" (LH) instrument settings, from R to L:
Voltage Range 2V, showing 0.961V
RMS/PEAK in RMS mode, Auto range ON
Current Range 500 mA, showing 254.3 mA
Function: Px10, mW range, showing 243.18 mW

Quote from: TinselKoala on 2016.07.26, 14:30:45
You might find more relevant information if you look for "H-bridge Solid State Tesla Coil Drivers"  where you will also find thousands of articles.

I'm sure Gunderson's decision not to open-source is based on a desire not to be proven wrong by a bunch of amateurs who think they know how to measure stuff.

An offer he can't refuse? Not likely, but.... one had better be careful with those kinds of offers.... if you don't deliver what you've claimed you might end up sleeping with the fishes.

As to the exact failure mode of the mosfets... sometimes it can be hard to tell. Even if you _can_ find all the pieces.

Morning TK,

I see you have let the smoke out of a few devices yourself.  Soon after I do that wife says "OK, what you burning up now!"
Your direction to search "H-bridge Solid State Tesla Coil Drivers" is very good, .  Looking at some of
the schematics there from experienced coil builders looks very much like what I think he has done on the H driver board.
One of many sites, http://www.stevehv.4hv.org/SSTCindex.htm,  really shows what we can expect to have to do in the future!

Ben K4ZEP

Quote from: Spokane 1 on 2016.07.26, 23:01:56
This is a good catch. Did any one else besides yourself notice it?

Spokane1

Apparently (as is often the case), no one reads my posts. (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488798/#msg488798) (or perhaps a very short memory is to blame) C.C ;)

I'm the one that first pointed out that the current side of the analyzer was bypassed (see post in link above).

Then just 10 posts later, TK himself quoted my post (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488811/#msg488811) where he first mentions the bypass, so indeed he knows who else (first) noticed it. ;)

This seems to happen a lot in general.

Quote from: poynt99 on 2016.07.27, 13:04:04
Apparently (as is often the case), no one reads my posts. (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488798/#msg488798) (or perhaps a very short memory is to blame) C.C ;)

I'm the one that first pointed out that the current side of the analyzer was bypassed (see post in link above).

Then just 10 posts later, TK himself quoted my post (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488811/#msg488811) where he first mentions the bypass, so indeed he knows who else (first) noticed it. ;)

This seems to happen a lot in general.

Dear poynt99,

Then you can claim the Wendy's Country Fried Chicken Burger with Bacon and a small Frosty reward for being most observant, but you are going to have to come to Spokane to collect your prize.

I hear you when you grouse about people not reading the past posts. I'm guilty of that myself.

Do you have any idea as to how Graham could have made an attempt at an energy input reading with the current connections on the C-H disconnected?

Attached is a revised one line diagram showing some of the instrument connections.

Quote from: poynt99 on 2016.07.27, 13:04:04
Apparently (as is often the case), no one reads my posts. (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488798/#msg488798) (or perhaps a very short memory is to blame) C.C ;)

I'm the one that first pointed out that the current side of the analyzer was bypassed (see post in link above).

Ah...  that's the post where you quoted _my_ post where I said,

Quote from: TKFurthermore--- the Clarke-Hess power analyzers with which I am familiar require several connections to a circuit in order to monitor it properly. To monitor Voltage the instrument is connected like any DMM, across the supply. To monitor Current the instrument is patched _in series_ with the supply and circuit being monitored. Obviously to compute a Power reading both connections must be made.

Quote
Then just 10 posts later, TK himself quoted my post (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488811/#msg488811) where he first mentions the bypass, so indeed he knows who else (first) noticed it. ;)

This seems to happen a lot in general.

:D     ^-^

Quote from: TinselKoala on 2016.07.27, 14:41:12
Ah...  that's the post where you quoted _my_ post where I said,
:D     ^-^

Yeah, obviously.

So what?

Quote from: Spokane 1 on 2016.07.27, 14:07:43
Dear poynt99,

Then you can claim the Wendy's Country Fried Chicken Burger with Bacon and a small Frosty reward for being most observant, but you are going to have to come to Spokane to collect your prize.

I hear you when you grouse about people not reading the past posts. I'm guilty of that myself.

Do you have any idea as to how Graham could have made an attempt at an energy input reading with the current connections on the C-H disconnected?

Attached is a revised one line diagram showing some of the instrument connections.

The "input" C-H is connected as a voltmeter, presumably correctly across the supply. The Current patch wiring is not connected to the instrument and is connected to a Red Cliplead which goes out of the camera frame to the right. If this Red Cliplead is in fact two Clipleads, which connect over to some current-monitoring instrument like a DMM, then Gunderson could have used the V reading from the C-H and the I reading from this off-screen DMM to compute a hand-waving sort of power measurement.

But I'm "guessing" that there is only one Red Cliplead that simply makes the short-circuit connection between the two banana plug ends of the Current patch wiring so as to complete the circuit to the DUT. Which means of course that the C-H isn't giving a valid power reading at all, and no other current reading instrument is in evidence. Only the people who were present at the demonstration can tell us more than this.

Was this lack of a current connection to the C-H mentioned during the demonstration? Did Gunderson ever refer to the disconnected instrument's invalid power reading during the demonstration? It would be very easy for people watching, who might not be familiar with the use of the C-H or power measurements in general, to "assume" that the Power reading shown in the C-H's Function window is really indicating the input power to the DUT.

Quote from: Spokane 1 on 2016.07.27, 14:07:43
Do you have any idea as to how Graham could have made an attempt at an energy input reading with the current connections on the C-H disconnected?

Quite simply, he would not have been able to obtain an input power measurement with the current path bypassed.

It's only a guess, but perhaps he found that the CH analyzer was causing some problems with the scope measurement when done simultaneously, or the current measurement itself was creating an issue. Not sure how though, since the only difference would be running through a bit more wiring, and a non-inductive CSR.

But again, the fact the H-bridge output is running through that much wiring (looped through a toroid) is troubling itself.

Quote from: poynt99 on 2016.07.27, 13:04:04
Apparently (as is often the case), no one reads my posts. (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488798/#msg488798) (or perhaps a very short memory is to blame) C.C ;)


This seems to happen a lot in general.

I am relatively new to this forum but I have noticed that when I click on a link to a thread that has something new I get taken to the last post.  I have mistakenly several times only read that post thinking it was the latest.  Then I have seen a quote or reference to something I didn't see and when looking back at the thread I have found the post or posts that I had missed.  On most other forums I belong to when you click on "new" or "unread" they will take you to the first new post in that thread that you haven't read.  Am I missing something in the way I open threads when looking for new postings?  Or does this forum not have a way to take you to the first post you haven't read?  Thanks for any insight into what is going on.

We'll check it out.

Well
I was confused [ ...and then I wasn't
yes Poynt had mentioned that early on

Quote from: CITFTA on 2016.07.27, 15:17:52
I am relatively new to this forum but I have noticed that when I click on a link to a thread that has something new I get taken to the last post.  I have mistakenly several times only read that post thinking it was the latest.  Then I have seen a quote or reference to something I didn't see and when looking back at the thread I have found the post or posts that I had missed.  On most other forums I belong to when you click on "new" or "unread" they will take you to the first new post in that thread that you haven't read.  Am I missing something in the way I open threads when looking for new postings?  Or does this forum not have a way to take you to the first post you haven't read?  Thanks for any insight into what is going on.

I see an Unread Posts button on the Home Page menu bar.

Smudge

Quote from: poynt99 on 2016.07.27, 15:11:41
Quite simply, he would not have been able to obtain an input power measurement with the current path bypassed.

It's only a guess, but perhaps he found that the CH analyzer was causing some problems with the scope measurement when done simultaneously, or the current measurement itself was creating an issue. Not sure how though, since the only difference would be running through a bit more wiring, and a non-inductive CSR.

But again, the fact the H-bridge output is running through that much wiring (looped through a toroid) is troubling itself.

Dear poynt99

Come to think of it Graham did mention that he was having problems with the C-H Instrument interfering with the circuit, we will have to review his presentation to see what he actually said. But, it was something about the sampling rate and the output of the logic system conflicting. He had to make a slight adjustment to the logic system to change the frequency so that the two systems wouldn't interact.  For all I know he might have pulled the current connections off in order to continue with the presentation. If that is the case then I have been reporting on a bogus observation, which means that his machine is only running at a COP of 6 and not 50.

I shall pay a lot better attention when the DVD comes out.

Graham did spend a fair amount of time explaining how the instruments did their job and how they were technically different in their measurement approach.

Spokane1

Quote from: Smudge on 2016.07.27, 16:26:15
I see an Unread Posts button on the Home Page menu bar.

Smudge

Thanks!  I just did not see that before.  I tried it and it works great!

Quote from: poynt99 on 2016.07.27, 15:11:41

But again, the fact the H-bridge output is running through that much wiring (looped through a toroid) is troubling itself.

Dear poynt99,

Does this photo with a view from the top help with understanding the connections any better? This is one of Reiyuki's posted photos.

Quote from: Spokane 1 on 2016.07.27, 16:52:40
Dear poynt99

Come to think of it Graham did mention that he was having problems with the C-H Instrument interfering with the circuit, we will have to review his presentation to see what he actually said. But, it was something about the sampling rate and the output of the logic system conflicting. He had to make a slight adjustment to the logic system to change the frequency so that the two systems wouldn't interact.  For all I know he might have pulled the current connections off in order to continue with the presentation. If that is the case then I have been reporting on a bogus observation, which means that his machine is only running at a COP of 6 and not 50.

  It was something along the lines of:  'Around 30khz, the pulses would hit a harmonic of the load meters internal clock and false-trigger them'.  This would probably cause things like resets, insane signals, and phantom button presses.

Quote from: Reiyuki on 2016.07.27, 18:35:36
  It was something along the lines of:  'Around 30khz, the pulses would hit a harmonic of the load meters internal clock and false-trigger them'.  This would probably cause things like resets, insane signals, and phantom button presses.

Or simply readings he didn't like very much.  (I have occasionally noticed some small effect on circuit performance myself from inserting a C-H 2330 inline, so I'm not discounting the possibility altogether.)

Besides, the scope shot shows no signal "around 30 kHz".  The main repetition frequency shown on the scope is almost exactly 50 kHz and the sinusoid frequency is about 73-75 kHz.

Quote from: Spokane 1 on 2016.07.27, 16:52:40
Dear poynt99

Come to think of it Graham did mention that he was having problems with the C-H Instrument interfering with the circuit, we will have to review his presentation to see what he actually said. But, it was something about the sampling rate and the output of the logic system conflicting. He had to make a slight adjustment to the logic system to change the frequency so that the two systems wouldn't interact.  For all I know he might have pulled the current connections off in order to continue with the presentation. If that is the case then I have been reporting on a bogus observation, which means that his machine is only running at a COP of 6 and not 50.

I shall pay a lot better attention when the DVD comes out.

Graham did spend a fair amount of time explaining how the instruments did their job and how they were technically different in their measurement approach.

Spokane1

If he adjusted the frequency so the systems didn't interact, why didn't he plug the C-H current monitor back in then?

So where did the data indicating COP 6 come from then?

Quote from: TinselKoala on 2016.07.27, 18:51:08
Or simply readings he didn't like very much.  (I have occasionally noticed some small effect on circuit performance myself from inserting a C-H 2330 inline, so I'm not discounting the possibility altogether.)

He mentioned it more like sporatic glitches (like the meter automatically changing modes) than reliably giving false readings.  Of course you're right; one does not automatically rule out the other ;)

Quote
So where did the data indicating COP 6 come from then?

  It was on the initial presentation, the ~1.5w in ~8-9w out, with both the scope math and load meters quote: 'in agreement'.

  A problem is that he moved the probes around a few times during the presentation to show different features (zooming into the ~1kv interrupt spike and that kind of thing).  A lot of pictures are from later on that may not have had the probes hooked up, or hooked up in a different position.

Quote from: Spokane 1 on 2016.07.27, 16:52:40
Graham did spend a fair amount of time explaining how the instruments did their job and how they were technically different in their measurement approach.

Spokane1

That part will be very interesting to review. Looking forward to it.

Quote from: CITFTA on 2016.07.27, 17:19:07
Thanks!  I just did not see that before.  I tried it and it works great!

wow.. who knew.. i didn't click on it either that i recall.. just like's Mike's radio station.. what does this mean ? is it a common trait amongst serious people to be blinkered.. how does this affect us ? how does your observation of the situation change your behaviour ? phew.. take a breath!  >:-)

Chips on the logic board:

TS556CN dual timer
MM74HC123AN dual monostable multivibrators ( four chips )
74AC74P dual D-type flip-flop
74AC14PC hex Schmitt trigger inverter
TS372 dual comparator

And the TO-220 devices at the top right of the board are presumably voltage regulators, like perhaps LM7805 (non heatsinked). I still can't figure out that part of the circuit. It's possible that the heatsinked device isn't actually in the circuit, since the only "output" I can see is directly from the 7805 (?) to the power rails of the breadboard. Or the heatsinked device could be another 78xx regulator stepping down voltage from a higher voltage DC rail to act as input to the 7805. I just can't tell how they are wired from the photo, except that they are getting input power from the topmost breadboard rail and the only output is from the one device to the other lower voltage power rails. And of course a monitoring LED+resistor to show that the lower voltage rail is powered.

I've filled in the chip part numbers on Spokane1's annotated photo of the breadboard.

Nobody has made a post in the Gunderson thread on Energetic for several days. I'm beginning to think that they don't really believe in Gunderson over there !

GG certainly seems old school. Most or all of that logic circuitry could be done with a single $5 microcontroller.

But good for him for designing it with discretes. I like it. :)

Quote from: poynt99 on 2016.07.28, 12:20:54
GG certainly seems old school. Most or all of that logic circuitry could be done with a single $5 microcontroller.

But good for him for designing it with discretes. I like it. :)

I " cut my teeth " with the 74 series TTL. Greatest sense of achievement was a working 4 way traffic light scheme.

Those were the days.

Dear Builders and Measurement Specialists,

Here is my first pass at wiring diagrams for the Graham Gunderson Logic controller. There are two versions. One is the way things appeared to be wired during the presentation. The other is the way the system was connected a week after the convention in Graham's shop. In that version he has disabled some of the chips and appears to have attached one or two pulse generators to possibly sweep input signals.

There is a fair amount of detail in these drawings. I have included pdf versions so you can study the small text. The resistors came in pretty good, but I couldn't see what the values were for the capacitors or the trim pots.

The next step is to develop a logic diagram so that we can figure out what is happening.

Spokane1

OK, so I posted this on O U dot com so I'll go ahead and post it here too:

@Spokane1:
That's nice work! I know how tedious that kind of thing is, especially when working from photographs that may not be too clear or from good angles.

But.... It seems from your schematics that the entire bit of circuitry that includes the components associated with U4, U2 and U1 is a Red Herring that does not connect to the rest of the circuit at all, other than being powered by the Vcc supply. This is the case in both the Experimental and the Presentation schematics. I'm sure you noticed this when you were drawing them up. This simplifies the problem greatly.

So here's what I think it's doing, based on your schematics.  We appear to have the 556 timer (U3) providing a clock signal that goes unchanged to both the H-Bridge and the Synchronous Diode. And we also have in the Presentation circuit, two monostable multivibrators U5 and U7. U7 appears to be an oscillator that provides pulses to the U8 hex inverter, which go through two of U8's gates then fed to the H-bridge. But this same signal through only one gate of the U8 is then fed back to the other monostable multivibrator U5 to trigger it, I believe, and it provides a pulse that goes to the Synchronous Diode. The U8 hex inverter also passes the input signal from U7 through three gates which then goes to the H-bridge. So the hex inverter's two outputs to the H-bridge are cleaned up, squared-off and oppositely phased versions of the pulse signal from U7. The U6 is not used in the Presentation circuit. There doesn't appear to be anything that synchronizes the U3 556 clock with the U7-U5 clock system. The U3 clock is fixed frequency and pulse width but the U7 and U5 multivibrators have trimpots for adjustments of either pulse width or frequency or both.

In the Experimental version, the U7 oscillator isn't used and an external FG (or two) is presumed to provide the clock signal for the U8 hex inverter and also the feedback signal for triggering the U5 monostable multivibrator.


Did you notice that my "mockup" of the U4 circuit (the TS372 dual comparator) oscillates, when it oscillates,  at exactly 60 Hz? This is undoubtedly due to those essentially floating reference pins 2 and 5, making the thing so sensitive and unstable that it picks up EMI from the house wiring, power supplies, etc. and oscillates in step with it, when the voltage supplied to the 1k resistor is just right. That is, if I built it right to begin with and if my TL082 is behaving as the TS372 would. Of course this issue is moot because this section of the circuitry isn't actually connected to anything in both the Experimental and the Presentation schematics.

---------

It would be great if someone else with TTL experience could correct, add to or subtract from my guesstimation.  Unfortunately I don't have any 74HC123 chips in my stash but at least I can try to breadboard the 556 clock portion later on today FWIW.

Hmmm... OK, I've breadboarded the 556 timer. The only flavor of 556 that I had in my box is a TI brand NE556N, and with Vcc = 5 V, and using the timing components listed on Spokane1's schematic, it oscillates at about 292 kHz.  Going up to Vcc = 9V it goes up to about 313 kHz.

This doesn't seem to correspond to any of the frequencies in the Gunderson scopeshot though.

Changing the capacitor on Pin 8 to 10 nF instead of 1 nF brings the frequency down to about 72 kHz, and shortens the duty cycle to about 57 percent, which makes a bit more sense. Maybe.

To All,

Here is a possible simplified version per TK's initial review.

Graham said that only 25% of the circuit was being used. Apparently that is a correct statement.

Spokane1

Dear TK,

Perhaps this is a good time to review Reiyuki's timing diagram.

The attached photo is from his note book. These are timing notes that Graham wrote into his book on Sunday July 10, 2016 during lunch.

Gotoluc's outstanding  work is done by using an oscillator and adjusting the frequency to get the desired input response.  Graham's system is a little more complex than that. You can see how two timing pulses are processed through the logic to get that backend control pulse coordinated. This is what I want to figure out at this point.

Dose the logic we have so far appear to function in this way? I haven't dug into it yet.

Spokane1

Quote from: Spokane 1 on 2016.07.28, 17:24:50
Dear TK,

Perhaps this is a good time to review Reiyuki's timing diagram.

The attached photo is from his note book. These are timing notes that Graham wrote into his book on Sunday July 10, 2016 during lunch.

Gotoluc's outstanding  work is done by using an oscillator and adjusting the frequency to get the desired input response.  Graham's system is a little more complex than that. You can see how two timing pulses are processed through the logic to get that backend control pulse coordinated. This is what I want to figure out at this point.

Dose the logic we have so far appear to function in this way? I haven't dug into it yet.

Spokane1

Yes, I think so. The line going from U8 to U5 triggers one pulse generator in U5 which triggers the other one, and this is synchronized with the signal to the H-bridge from U8 which is "processing" the pulse signal from U7. What I still don't understand is how this system is synched to the 556 clock system which is also sent to the H-bridge.

It turns out that my local supplier allegedly has some 74HC123 (or maybe HCT versions) in stock so I'm off to try to pick up 3 of them to play around with. We shall see what we shall see.

Quote from: TinselKoala on 2016.07.28, 17:59:24
It turns out that my local supplier allegedly has some 74HC123 (or maybe HCT versions) in stock so I'm off to try to pick up 3 of them to play around with. We shall see what we shall see.

Dear TK,

AT least you live some place where you have a supplier. such is not the case in Spokane with a population of about 400,000. I will have to put in an order to DigiKey or Mouser and then wait a week.

Keep in mind that all my capacitor values on those drawings are WAG's.

Again thanks for all the footwork you are doing on this.

Spokane1

OK... after a rather frustrating all-night session I have finally gotten some results from the U5-U7-U8 circuit.

The 74HC123s do not oscillate on their own. Apparently there must be an input from an external clock for the system to work. I supplied a clock signal from one of my function generators to Pin 10 of U7. I could not get it to work with a signal from the 556 circuit that I tested earlier. The result is shown in the screenshot below.

The CH1 Yellow trace is the Pin 13 output of U5 which goes to the Synchronous Diode. The CH2 Light Blue and CH3 Purple (Cyan and Magenta?) traces are the complimentary outputs of the 7414 U8 which go to the H-bridge. The CH4 Dark Blue trace is the input from my FG to Pin 10 of U7.

The two pots on U5 control the Yellow trace pulse width and pulse delay. One of the pots on U7 controls the pulse width of the two complimentary outputs of U8 to the H-bridge. I couldn't see any effect of the other pot on U7 but that may be due to where I'm injecting the clock signal. Too burned out at the moment to track that one down.

I'm not completely sure if the U7 chip is working on both sides. I'll try replacing it with a fresh chip to see if anything changes.

There may be a mistake on Spokane1's schematic because 1) there apparently needs to be a clock input to the 74hc123 system for it to work at all,  and 2) otherwise there is no synchronization between the 556 signal to the H-bridge and the signals from the 74HC123 system to the H-bridge and the synchronous diode.

Also the data sheet for the 74HC123 chip says that the pins 6 and 14 of the chip should be hard-wired to ground (in addition to being connected to the timing components) for noise rejection. This does not appear to have been done on the Gunderson board, as far as I can tell, but I did it on the U7 chip on my breadboard (but not the U5 chip.) Go figure.

Quote from: TinselKoala on 2016.07.29, 12:28:47
OK... after a rather frustrating all-night session I have finally gotten some results from the U5-U7-U8 circuit.

The 74HC123s do not oscillate on their own. Apparently there must be an input from an external clock for the system to work. I supplied a clock signal from one of my function generators to Pin 10 of U7. I could not get it to work with a signal from the 556 circuit that I tested earlier. The result is shown in the screenshot below.

The CH1 Yellow trace is the Pin 13 output of U5 which goes to the Synchronous Diode. The CH2 Light Blue and CH3 Purple (Cyan and Magenta?) traces are the complimentary outputs of the 7414 U8 which go to the H-bridge. The CH4 Dark Blue trace is the input from my FG to Pin 10 of U7.

The two pots on U5 control the Yellow trace pulse width and pulse delay. One of the pots on U7 controls the pulse width of the two complimentary outputs of U8 to the H-bridge. I couldn't see any effect of the other pot on U7 but that may be due to where I'm injecting the clock signal. Too burned out at the moment to track that one down.

I'm not completely sure if the U7 chip is working on both sides. I'll try replacing it with a fresh chip to see if anything changes.

There may be a mistake on Spokane1's schematic because 1) there apparently needs to be a clock input to the 74hc123 system for it to work at all,  and 2) otherwise there is no synchronization between the 556 signal to the H-bridge and the signals from the 74HC123 system to the H-bridge and the synchronous diode.

Also the data sheet for the 74HC123 chip says that the pins 6 and 14 of the chip should be hard-wired to ground (in addition to being connected to the timing components) for noise rejection. This does not appear to have been done on the Gunderson board, as far as I can tell, but I did it on the U7 chip on my breadboard (but not the U5 chip.) Go figure.

Morning TK,

Try this circuit:

Thanks! That actually works!

Now I don't need the external FG clock, and the previously  non-functional pot now controls the frequency. Now we just need to play with the capacitor and resistor values in the various timers to get the right duty cycle and frequency range. I already changed the capacitor in the frequency timer section to slow it down to around 70 kHz minimum but should probably add even more capacitance to get down to the 30 kHz range.

The Dark Blue trace in the shot below is now the Frequency output at Pin 12 of the rewired U7 chip.  But does it actually correspond to what's visible on Gunderson's board?

And what about the synchronization problem? How to synch with the signal from the 556 that goes to both the H-bridge and the synchronous diode?

Anyhow, time for me to take the rest of the day off and get some sleep.

Thanks again, nice work K4zep !!    O0

Dear TK,

I don't know how you can carry on with this kind of work with out some sleep. Maybe things will be better for you this weekend. My wife also suffers from a sleep disorder. She lost her job at the VA because of  it.

Here is my first pass at a logic diagram.

I was wondering if the "System Clock" U3 has anything to do with the actual timing of the power MOSFETS?  It might be the master clock for the switch mode power supplies on the H-Bridge and the Backend Diode. This would allow that frequency to be adjusted to avoid harmonic interference between the main logic frequency and/or the instrumentation.

It appears that U7, U8 and U5 alone can supply the basic timing signals.

Have a good weekend.

Spokane1

Some further adjustment of cap values allows me to tune to the 50Khz range with considerable margin on either side, and to get exactly 50 percent duty cycle, again with a nice range. And the short Yellow pulse can be adjusted from very narrow to about half the width of the main pulse and can be positioned anywhere on a nice range from before to after rising edge of the main pulse.

Quote from: Spokane 1 on 2016.07.29, 14:54:52
Dear TK,

I don't know how you can carry on with this kind of work with out some sleep. Maybe things will be better for you this weekend. My wife also suffers from a sleep disorder. She lost her job at the VA because of  it.

Here is my first pass at a logic diagram.

I was wondering if the "System Clock" U3 has anything to do with the actual timing of the power MOSFETS?  It might be the master clock for the switch mode power supplies on the H-Bridge and the Backend Diode. This would allow that frequency to be adjusted to avoid harmonic interference between the main logic frequency and/or the instrumentation.

It appears that U7, U8 and U5 alone can supply the basic timing signals.

Have a good weekend.

Spokane1

Looks great, except that I've rewired the U7 according to the information from K4zep so that it does not need an external clock any more.  I don't know if this corresponds to Gunderson's board version, I'll leave that part up to you!

Don't worry about my sleep cycle, it's partially an adaptation to the heat down here in Northeast Aztlan. Unseasonably cool this morning at 10 am, it's "only" 84 degrees F. with relative humidity about 80 percent. So I tend to sleep during the hottest part of the day and am more active evening-night.

Visualizing the timing delay introduced by running the basic clock signal through multiple gates in the 74C14 hex Schmitt trigger inverter:

Again, the Blue trace is the raw pulse signal generated by the oscillating 74HC123. The Magenta and Cyan traces are the result of running this signal through 2 and 3 gates in the 74C14.

Dear TK,

You need to take a break and go have a beer, soda, or protein smoothie.

I really appreciate all the work you are doing on this circuit. With your assistance we are already at least couple of weeks ahead in this investigation.

I have a couple of old 74XX123's that I shall attempt to breadboard this weekend, depending upon how many honey-do's I have on the weekly list.

Spokane1

Finally i was able to replicate GotoLuc's scopeshot of his version of the GG scope signals using a very simple circuit, see screenshot 1

It uses 2 MOSFETs (IRF530) and an isolated FG (to avoid ground lead shorts) set at BELOW the LC resonance frequency (LC resonance is at 24.5KHz, to get this signal i had to go to 20.7KHz)
L=10mH, C=3.5nF, csr=50 Ohm  (using a 1 Ohm csr and/or my current probe results in very fuzze (weak) signals).
See picture of diagram.

Input signal from FG is 7Vpp square wave at 33% duty cycle, see purple trace in screenshot 2.

Regards Itsu

Quote from: TinselKoala on 2016.07.29, 15:42:33
Visualizing the timing delay introduced by running the basic clock signal through multiple gates in the 74C14 hex Schmitt trigger inverter:

Again, the Blue trace is the raw pulse signal generated by the oscillating 74HC123. The Magenta and Cyan traces are the result of running this signal through 2 and 3 gates in the 74C14.

Scope traces says it all, shows gate delays, etc!  Damn good work TK!

Ben K4ZEP

@Itsu: Nice work! You probably already realize this but if you move the common Scope Probe ground point to the top of the 50R resistor you can avoid including it in the Vdrop across the coil-capacitor tank. This works if your PSU has isolated negative lead (as usual)  and your FG is also isolated. Of course you will have to invert the CH2 signal in the scope to get the correct current direction.     O0

Two questions: What happens if the mosfets are reversed so that the Drains are connected together, instead of the Sources? And what happens if you wind a secondary on top of the primary windings of the 10mH coil, can you light up a light bulb from the secondary?


@K4zep: Thanks! Gate propagation delays are often neglected or forgotten about but when driving high-power H-bridge they can cause problems -- or sometimes even solve problems.
It's important to note that I'm using a 74C14N because that's all I have on hand, but the Gunderson breadboard appears to use a 74AC14PC which may be faster. I'll have to see if I can find one at my supplier to see if there is any significant difference in gate speed.

I was kind of struggling with the current signal as it looks too ragged to me compared to GG and Gotoluc's current signals, so it kept it as straight forward as possible.
But i can try to change the common scope probe ground point to where you showed.

Changing the MOSFETs to a common drain configuration, is a problem as the FG should be connected to the both gates and both sources (i think), but i can try.

The present 10mH coil is a small ferrit U-cores choke with little room to add a secondary (its just what came close to Luc's 10mH coil), so i have to wind/use another
transformer like toroid to check.

Itsu

Quote from: TinselKoala on 2016.07.30, 03:13:57
@Itsu: Nice work! You probably already realize this but if you move the common Scope Probe ground point to the top of the 50R resistor you can avoid including it in the Vdrop across the coil-capacitor tank. This works if your PSU has isolated negative lead (as usual)  and your FG is also isolated. Of course you will have to invert the CH2 signal in the scope to get the correct current direction.     O0

Two questions: What happens if the mosfets are reversed so that the Drains are connected together, instead of the Sources? And what happens if you wind a secondary on top of the primary windings of the 10mH coil, can you light up a light bulb from the secondary?


@K4zep: Thanks! Gate propagation delays are often neglected or forgotten about but when driving high-power H-bridge they can cause problems -- or sometimes even solve problems.
It's important to note that I'm using a 74C14N because that's all I have on hand, but the Gunderson breadboard appears to use a 74AC14PC which may be faster. I'll have to see if I can find one at my supplier to see if there is any significant difference in gate speed.

Morning TK,

The 74AC is about 10 times faster than the 74C......At 5VDC, Into a 50pf load, the AC has a min/mas rise time of 3-10ns with typical 8ns.  The C is much slower.  At 5VDC, it is an average of 220ns.  74AC is dam better chip if speed is the need!

Ben K4ZEP

Quote from: Itsu on 2016.07.30, 09:16:59

I was kind of struggling with the current signal as it looks too ragged to me compared to GG and Gotoluc's current signals, so it kept it as straight forward as possible.
But i can try to change the common scope probe ground point to where you showed.

Changing the MOSFETs to a common drain configuration, is a problem as the FG should be connected to the both gates and both sources (i think), but i can try.

The present 10mH coil is a small ferrit U-cores choke with little room to add a secondary (its just what came close to Luc's 10mH coil), so i have to wind/use another
transformer like toroid to check.

Itsu

Oh, yes, I forgot about the FG's connection to the sources. So it probably won't be possible to use a common Drain configuration. I was thinking how cool it would be to just strap the mosfets together back-to-back with a bolt. Oh well, nothing wrong with common-Source.

I used a 10 mH choke that I pulled from an old CRT TV chassis, with an extra winding over the outside of it for my favourite Joule Thief:

Quote from: K4zep on 2016.07.30, 11:35:33
Morning TK,

The 74AC is about 10 times faster than the 74C......At 5VDC, Into a 50pf load, the AC has a min/mas rise time of 3-10ns with typical 8ns.  The C is much slower.  At 5VDC, it is an average of 220ns.  74AC is dam better chip if speed is the need!

Ben K4ZEP

Well, thanks for looking that up! I need to try to get my hands on a few 74AC14 chips to play with then. But it does appear that I'm getting a bit better risetime from my C version, perhaps slightly less than 100 ns as shown on the scopetrace above. Of course the only load is the scope probe and input, so somewhat less than 50 pF.
But I'm actually also interested in the gate delay rather than just the risetime.
Unfortunately my supplier isn't open on weekends so I won't even be able to find out if he has them in stock until Monday. I guess I should have asked him when I was there on Thursday. Oh well.
The risetime of the pulses from the 74HC123 is impressive. I haven't measured it yet but even at 200 ns/div the rising edges are almost vertical. (Dark Blue and Yellow traces)


By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.

You've wandered into the Deep end of the pool....over most heads...mine included .

respectfully

Chet K

Quote from: TinselKoala on 2016.07.30, 13:45:43
Well, thanks for looking that up! I need to try to get my hands on a few 74AC14 chips to play with then. But it does appear that I'm getting a bit better risetime from my C version, perhaps slightly less than 100 ns as shown on the scopetrace above. Of course the only load is the scope probe and input, so somewhat less than 50 pF.
But I'm actually also interested in the gate delay rather than just the risetime.
Unfortunately my supplier isn't open on weekends so I won't even be able to find out if he has them in stock until Monday. I guess I should have asked him when I was there on Thursday. Oh well.
The risetime of the pulses from the 74HC123 is impressive. I haven't measured it yet but even at 200 ns/div the rising edges are almost vertical. (Dark Blue and Yellow traces)


By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.

Hi TK and all,
If you have time, go to:

http://teslaenergysolutionsllc.com/our-research/coil-shorting-experiments/

Using those H bridge switchers.  Switches 100Khz effortlessly!

Other good stuff there too! 

It looks like using GG's logic circuits and those output devices, home free on the input side.

Now if we can find some high speed super high efficiency Rectifiers........
And of course figure out the "core's"

Ben K4ZEP

Quote from: TinselKoala on 2016.07.30, 13:45:43
By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.

I care ...but the electronic implementation is not so interesting to me.  I can think of several different ways to achieve such input waveforms.  I think the magic smoke is in the magnetics (a la Smudge's thoughts) or faulty measurements. In any case don't let it out.
If the output waveform contained high and narrow spikes, then I would suspect McFreey's Modus Operandi.
Without accurate schematics, nice scopeshots, probe positions and proof of calibration, I suspect the ORBO MO ;)

Quote from TK

QuoteBy the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.

Doubt that I am a usual suspect, but for now I am reserving (and have been asked to reserve) my opinions.

I may break silence in the future, after the film debut.

BTW,
We  used the 74xx123 dual monostable back in the early 80's as a watchdog timer for our first 6500 based uP product.

On that note, I might add that the use of TTL logic is oftentimes not subject to the lockup conditions that can occur  in a  uP based design when it loses it's stack pointer and winds up in a tight loop, no longer servicing other routines.

Hence the need to "feed or kick the dog".

IMHO there isn't anything further to comment on if and until the exact measurement protocol is illustrated. And following that (if it indeed is shown in sufficient detail) no doubt there will be several questions that most likely won't be answered.

In the mean time I would suggest that folks study up on real power vs. imaginary power, and what constitutes an input power source vs. something that is part of the device. Getting this straightened out would be an important technicality to consider before such an exercise took place.

I created a 100uH secondary over the 10mH (5 Ohm) now primary.
I made the common Scope Probe ground point to be the top of the 50R resistor and inverted the ch2 (current) channel.

Without a load the voltage / current signal looks like as in screenshot 1

With a 6V / 0.6W incandescent light bulb (glowing dimly) as load to the secondary these signals become as in screenshot 2 (vertical settings changed!)
I am unable to obtain similar signals as in screenshot 1 with this load.

Measuring the load shows the signals as in screenshot 3.
Yellow the voltage across the bulb, green the current throught the bulb, red the math trace yellow x green.
Be aware that the current probe controller was set to 100mA/div. so the reported current and power needs to be taken times 10

The diagram is the updated circuit


Itsu

Quote from: poynt99 on 2016.07.30, 19:35:23
IMHO there isn't anything further to comment on if and until the exact measurement protocol is illustrated. And following that (if it indeed is shown in sufficient detail) no doubt there will be several questions that most likely won't be answered.

In the mean time I would suggest that folks study up on real power vs. imaginary power, and what constitutes an input power source vs. something that is part of the device. Getting this straightened out would be an important technicality to consider before such an exercise took place.

Agreed, especially so the second paragraph.

Itsu: nice work, but I must have missed the claims for gotuluc's circuit.

Quote from: Itsu on 2016.07.30, 20:03:49
I created a 100uH secondary over the 10mH (5 Ohm) now primary.
I made the common Scope Probe ground point to be the top of the 50R resistor and inverted the ch2 (current) channel.

Without a load the voltage / current signal looks like as in screenshot 1

With a 6V / 0.6W incandescent light bulb (glowing dimly) as load to the secondary these signals become as in screenshot 2 (vertical settings changed!)
I am unable to obtain similar signals as in screenshot 1 with this load.

Measuring the load shows the signals as in screenshot 3.
Yellow the voltage across the bulb, green the current throught the bulb, red the math trace yellow x green.
Be aware that the current probe controller was set to 100mA/div. so the reported current and power needs to be taken times 10

The diagram is the updated circuit


Itsu

This is great work, Itsu. The next step, I suppose, would be to create the "synchronous rectifier" which would "hide" the load from the transformer most of the time and only connect it for a small part of the full cycle, just when the peak power occurs. The output of this SR would be used to pump up a capacitor bank. Then you could light up the bulb from the smoothed DC output of the cap bank.

Quote from: poynt99 on 2016.07.30, 19:35:23
IMHO there isn't anything further to comment on if and until the exact measurement protocol is illustrated. And following that (if it indeed is shown in sufficient detail) no doubt there will be several questions that most likely won't be answered.

In the mean time I would suggest that folks study up on real power vs. imaginary power, and what constitutes an input power source vs. something that is part of the device. Getting this straightened out would be an important technicality to consider before such an exercise took place.

Yes, that's right.  Not only that, but we don't really have enough detailed information about Gunderson's H-Bridge and  Synchronous Diode circuits to be able to proceed intelligently with deconstructing-reconstructing the driver part of the device. Making something that produces the correct pulse train (the breadboard) is really trivial and as verpies says it can be done in several (I'll even say many) ways. I really don't think that "OU" lives in the breadboard, except perhaps by virtue of obfuscation and mismeasurement.



As far as "effortless" switching of the high-power H-bridge goes... I've had some Intersil HIP4080A H-bridge drivers in my box for a while. They are pretty darn sophisticated.

The HIP4080A is a high frequency, medium voltage Full Bridge N-Channel FET driver IC, available in 20 lead plastic
SOIC and DIP packages. The HIP4080A includes an input comparator, used to facilitate the "hysteresis" and PWM
modes of operation. Its HEN (high enable) lead can force current to freewheel in the bottom two external power
MOSFETs, maintaining the upper power MOSFETs off. Since it can switch at frequencies up to 1MHz, the HIP4080A
is well suited for driving Voice Coil Motors, switching power amplifiers and power supplies.
HIP4080A can also drive medium voltage brush motors, and two HIP4080As can be used to drive high performance
stepper motors, since the short minimum "on-time" can provide fine micro-stepping capability.
Short propagation delays of approximately 55ns maximize control loop crossover frequencies and dead-times which
can be adjusted to near zero to minimize distortion, resulting in precise control of the driven load.

Quote from: ION on 2016.07.30, 20:44:37
Agreed, especially so the second paragraph.

Itsu: nice work, but I must have missed the claims for gotuluc's circuit.

Hi Ion,

no claims from Luc, he just managed to replicate the scope signals from Graham Gunderson's presentation with a simple setup.
I am trying to replicate Luc using a similar simple setup, but i only managed to "kind of" replicate Luc's signals under "no load" up till now.


Guess i have to create the "synchronous rectifier" which would "hide" the load from the transformer most of the time like Tinselkoala mentioned.

Itsu

Quote from: Itsu on 2016.07.31, 09:43:12

Hi Ion,

no claims from Luc, he just managed to replicate the scope signals from Graham Gunderson's presentation with a simple setup.
I am trying to replicate Luc using a similar simple setup, but i only managed to "kind of" replicate Luc's signals under "no load" up till now.


Guess i have to create the "synchronous rectifier" which would "hide" the load from the transformer most of the time like Tinselkoala mentioned.

Itsu

Seems to me that if you edge triggered a timer that started when the small "trigger" pulse went high and adjusted its  time to time out at the start of the next cycle, that would be the "time element" to pull off any extra energy off the output transformer.  The "Q" output would turn on a FWBR of a ultra high speed, fast recovery diodes like the 1N4454 or better, Straight forward. Not quite as efficient as GG's circuit but if it is really OU, would work.   See representative schematic attached.

Ben K4ZEP

Quote from: TinselKoala on 2016.07.31, 02:55:07
Yes, that's right.  Not only that, but we don't really have enough detailed information about Gunderson's H-Bridge and  Synchronous Diode circuits to be able to proceed intelligently with deconstructing-reconstructing the driver part of the device. Making something that produces the correct pulse train (the breadboard) is really trivial and as verpies says it can be done in several (I'll even say many) ways. I really don't think that "OU" lives in the breadboard, except perhaps by virtue of obfuscation and mismeasurement.

My only interest are the power measurements, and I'm hoping that all those discussing them can and will be on the same page in terms of real/imaginary power and what constitutes a power source vs. being "part of the device".

At present we are not in agreement.

Quote from: poynt99 on 2016.07.31, 12:05:12
My only interest are the power measurements, and I'm hoping that all those discussing them can and will be on the same page in terms of real/imaginary power and what constitutes a power source vs. being "part of the device".

At present we are not in agreement.

I'm a little confused by your statement above and could use some clarification for my own sake. Would you please expand your thoughts on this in reference to say my MEI circuit previously posted?

Edit: No need to respond! I read your post on OU dot com and I think that explains it.

Thanks,
pm

Has anyone read this book ?

http://grahamgunderson.com/ou/

Maybe it contains a basic philosophy how to proceed, and perhaps the basic circuit, as it seems to claim in the advert.

Quote"A REAL OVERUNITY TRANSFORMER DEMONSTRATED TO AN AUDIENCE AND DETAILS ON HOW TO REPLICATE "

Note the word "details"

Comes with a 100% money back guarantee if not satisfied. How can you lose.

It is one of a collection of books being sold at johnxxxxxx.net

hint: change the xxxxxx's to something that sounds like "houdini"

I realize current printing is from the conference last year, but will probably be revised soon to include the  most recent conference.

Meanwhile, thanks to K4zep for the oscillator details, here's the revised/corrected wiring for U7, the 74HC123 monostable multivibrator configured as the master oscillator for the Gunderson breadboard "Simplified Presentation Logic Wiring" diagram that Spokane1 drew out for us.

When I first built it according to Spokane1's diagram I had to feed it with an external clock signal to get the system to make pulses, and one of the pots on U7 didn't have any effect. Using this revised wiring it is now self-contained, with the frequency and duty cycle pots both working and the approximate ranges indicated on the drawing. I've put in the actual values of the timing components I wound up using to get the ranges shown.

(Also posted at O U dot com.)

Quote from: partzman on 2016.07.31, 13:09:21
I'm a little confused by your statement above and could use some clarification for my own sake. Would you please expand your thoughts on this in reference to say my MEI circuit previously posted?

Edit: No need to respond! I read your post on OU dot com and I think that explains it.

Thanks,
pm

Part one of the discussion is about whether the H-Bridge input should be measured in GG's setup. No one other than myself and TK have weighed in on that (AFAIK), but I have made my argument, and as far as I am concerned it is a no-brainer; the input power must be measured after the H-Bridge in this case.

Part two involves the issue of real vs. imaginary power. I posed the question about being able to measure real power from a non-OU device that outputs a sine wave where its current and voltage are precisely 90º out of phase. I believe that you, tinman, and TK have all stated that yes it is possible. ION and myself have stated the contrary. Obviously one side must be correct, and the other not.

I'd be happy to see a simulation supporting your assertion. :)

Quote from: poynt99 on 2016.07.31, 20:05:51
Part one of the discussion is about whether the H-Bridge input should be measured in GG's setup. No one other than myself and TK have weighed in on that (AFAIK), but I have made my argument, and as far as I am concerned it is a no-brainer; the input power must be measured after the H-Bridge in this case.

Part two involves the issue of real vs. imaginary power. I posed the question about being able to measure real power from a non-OU device that outputs a sine wave where its current and voltage are precisely 90º out of phase. I believe that you, tinman, and TK have all stated that yes it is possible. ION and myself have stated the contrary. Obviously one side must be correct, and the other not.

I'd be happy to see a simulation supporting your assertion. :)

I am in agreement with Poynt with the exceptions noted below, i.e. the H bridge and associated drive circuitry represents a power function generator, and since the DUT is the transformer, to vet just the transformer, we do not need to include the drive circuit power dissipation.

Later, in the real world practical system, we could include all losses including drive circuitry i.e if we were claiming OU for a black box system that took pure DC in / DC out. Gunderson's claim is for an OU transformer, not a system, so in my mind that is the first DUT. The synchronous rectification may or may not be included. This depends on whether it's role is simply rectification or some other interaction that helps to produce the claimed OU effect.

If the claim is made that the H bridge interacts with the transformer in such a way due to feedback from the transformer to the drive circuitry or the H bridge itself that sustains the OU effect, and that OU effect would not be possible with an equivalent benchtop power function generator, then of course we must include the circuitry including H bridge. Then it becomes an OU system, of which the transformer is one component.

Do I understand the question? What are the claims? Perhaps some of the confusion stems from G.G.'s mistitled device "OU Transformer".

Quote from: K4zep on 2016.07.31, 11:53:17
Seems to me that if you edge triggered a timer that started when the small "trigger" pulse went high and adjusted its  time to time out at the start of the next cycle, t

His lab generator has 2 channels and one can be triggered from the other in a pulse burst mode so he does not even need to wire a monostable, nor an astable to generate these pulses.  All adjustable digitally and versatily .

Quote from: poynt99 on 2016.07.31, 20:05:51
Part one of the discussion is about whether the H-Bridge input should be measured in GG's setup. No one other than myself and TK have weighed in on that (AFAIK), but I have made my argument, and as far as I am concerned it is a no-brainer; the input power must be measured after the H-Bridge in this case.

Part two involves the issue of real vs. imaginary power. I posed the question about being able to measure real power from a non-OU device that outputs a sine wave where its current and voltage are precisely 90º out of phase. I believe that you, tinman, and TK have all stated that yes it is possible. ION and myself have stated the contrary. Obviously one side must be correct, and the other not.

I'd be happy to see a simulation supporting your assertion. :)

OK, somehow I missed the non-OU qualification when I submitted my MEI example.

pm

Is your MEI OU?  :D

What does your sim result of the MEI say?

Is the circuit running at 3MHz as per the scope yellow channel? If so, then the 4" scope leads themselves are going to introduce significant inductance in your measurement path, and that often spells trouble.

Quote from: poynt99 on 2016.07.31, 23:18:51
What does your sim result of the MEI say?

Is the circuit running at 3MHz as per the scope yellow channel? If so, then the 4" scope leads themselves are going to introduce significant inductance in your measurement path, and that often spells trouble.

May I respectfully disagree with that thought?  I am a ham radio operator and have regularly used equipment that operates in the 3 to 4 MHZ range.  The wave length for a 3MHZ signal is approximately 80 meters.  I really don't think a 4" scope lead is going to have much of an impact at that frequency.

Quote from: CITFTA on 2016.08.01, 00:47:22
May I respectfully disagree with that thought?  I am a ham radio operator and have regularly used equipment that operates in the 3 to 4 MHZ range.  The wave length for a 3MHZ signal is approximately 80 meters.  I really don't think a 4" scope lead is going to have much of an impact at that frequency.

Ground lead and probe leads can add significant ringing when taking measurements, especially on transient signals. DUT wiring, especially from the CSR will have significant effects on a current measurement not only from phase effects, but from reactance as well, throwing off not only the phase relationship between the voltage and current measurements, but the CVR value can be artificially increased from its actual value. Sometimes the probe cable itself can become susceptible to noise pickup. Handling the cable of a current or voltage probe while monitoring changes on the scope can reveal these effects.

Once one ventures into above 1MHz territory, the game changes quite a bit. At 3MHz, accurate power measurements will be a real challenge (even with a current probe).

Quote from: verpies on 2016.07.31, 22:44:19
His lab generator has 2 channels and one can be triggered from the other in a pulse burst mode so he does not even need to wire a monostable, nor an astable to generate these pulses.  All adjustable digitally and versatily .

Agreed!
Ben K4ZEP

All,

I'm treating this as simply a R&D adventure.  Going to build what I can, knowing what I know
and try to find the "source" of the energy if there is one and the go from there.  With my junk box,
my power supplies, etc, investing maybe $150. bucks.  I'm sure I'll let some expensive smoke out
of some little innocent parts but that's the way I learn.  We have a lot of clues, pictures, advice,
enough to give it a shot!

We can talk till the cows come home exploring the finer points of instrumentation but until we
have a "best guess device" to experiment with, we are as we say in the south, "running around, making a lot of noise and throwing
dust in the air" or really doing nothing.

There are those that "do" and there are those that BS, I'm in the "do" group. We have a lot of "doers" here! it will be interesting to see what comes out of our endeavors. Most of the "doers" have been at this for years and understand instrumentation" and the vulgarities therein!  For the newer experimenters
on the list you have a steep learning curve in building and instrumentation.  You will smile looking back at this in years to come!
If I don't find anything, so what?  If I can't find that elusive source of "excess" power, been there done that before and still enjoyed it!

Parts are in the mail, from several sources.  It will take time, several weeks but what an interesting way to start the week.  I'm looking forward to another
whack at this OU "thing".....

Ben K4ZEP

What's "BS" to some, is "thinking things through before diving in" to others. I'm in this group, and not ashamed to admit it.

Quote from: poynt99 on 2016.08.01, 12:29:32
What's "BS" to some, is "thinking things through before diving in" to others. I'm in this group, and not ashamed to admit it.

Morning P99,

Thinking before diving is good, but eventually you gotta dive in!  When I see page after page after page of "thinking", it becomes BS to me.
Personally, I have been thinking on this for years already!  I'm just thinking the gun has sounded, time to dive!

Ben K4ZEP

Poynt,

I'll respond to your questions in one reply if you don't mind.

First yes, I think certain of my MEI circuits taken in their context are OU. My opinion comes from hundreds of bench experiments and tests including more failures than successes. If you or anyone else disagrees, then prove my test results are incorrect with bench results. I've given in the true spirit of open source all the needed info for the circuits to be replicated and I don't mind if I'm found to be in self delusion.

LtSpice simulation of the circuits does not yield OU.

In regards to the scope probes, they are rated at 10Meg with 3.9pfd capacitance.  The only 4" ground leads used are on the voltage probe measurements and not on the CSR. The probe used on the CSR has the ground lead and squeeze probe removed and a small Tek supplied spring ground clip attached that is designed for RF probing.  The typical inductance of the Caddock MP-820 series CSR used is 10nH at .2" from the body.  No reliance is given to the AM503/A6302 amp/probe combo as it is used only for comparison.

Yes I have experienced probe cable placement variations particularly with the Tek current probes and amp connections to the scope. Not so much if any variations with the Tek TPP0500B probes.

In summary, as long as I believe these circuits are OU, I will continue to search for solutions to provide a prime source that will apply the negative energy to a DC source. It was my goal in posting these circuits that some clever individual would come with a solution but if these circuits are not OU, then all is in vain.

pm

Ben
The water looks cold ...I always was a chicken about that , Grew Up in the NE USA and we had a chance as kids to move to warm and sunny
southern California,  my brothers and I spent most of our young lives griping about never getting there .

years went by, and one day we took a trip to see our Kid brother in Los Angeles   and went down to the beach for a swim ...
my younger brother ran down Jumped in popped up and  yelled a curse to our parents for Denying us such Nirvana ..

without hesitation I gave it the Big Brother "all in" cannonball...

felt like ice picks at every point in My Body...COLD !! :D

came up from that ICY shockwave to see My half frozen Kid brother on the beach laughing hysterically at my suffering and offering up praises to our parents for keeping us in the NE USA
where at least we could swim in the summertime...

Ben
while I have great appreciation and respect for Poynt and all who crunch the numbers here and elsewhere,

your enthusiasm is contagious ,I wish I had your skills and equipment ...I would jump right in too
but I am keeping busy with another open source project where I do have the skills and tools .
the work of Rob Mason [member Evolvingape here]

@ Partzman
we must investigate this further .. and perhaps see what you are seeing on the bench which does not manifest in a Simulator  !!
thx for sharing

with respect and gratitude
Chet K

Dear All.

This isn't my area of expertise as most know.... I have just read a post TK made at OU.Com regarding the synchronous rectifier. Is there a possibility of the drive signal being added to the output? Are there semiconductors available that are fully isolated?

As I see it a fully closed system can only achieve 100% efficiency. Tesla used mechanical interrupters on many of his devices effectively allowing the system to be both open and closed at the right point in time.

Tying everything down at once puts the whole device in a single dimension, may be by allowing it to float separately is where the " magic " is ?

Ok it reads daft.... But I know what I'm trying say.   :)

Cheers Grum.

Quote from: K4zep on 2016.08.01, 12:35:56
Thinking before diving is good, but eventually you gotta dive in!

Not necessarily. That is the whole point of thinking it through first. ;)

Quote
When I see page after page after page of "thinking", it becomes BS to me.

Each to his own. Good luck with your experiments.

in the past I would dive right in and start building, only to find the lake was full of crocodiles and snakes (replace with swindlers and fraudsters). I did my due diligence painfully after the dive.

Now I am more cautious, and do a lot of research on the background and integrity of the presenter.

Due diligence in researching the whole picture, not just the hardware and claims will save you lots of time, disappointment and heartache.

If you have an infinite amount of time, disregard the due diligence part and by all means, dive right in.

For me time, is short and I have to choose my endeavors with care and forethought.

As Poynt said :  "each to his own", and as an added note, be careful not to confuse discussion or points of view by thoughtful and experienced individuals with BS.

Regards, ION

K4zep and verpies,

thanks for the comments, i presently use my cheap chinese FG which is isolated from ground.



I added two IRF530 MOSFET in a self triggered synchronous rectifier setup behind the secondary, and now it seems that the load (6V / 0.6W bulb)
is hidden from the input as it glows dimly and at the same time i could create Luc's input voltage / current signals, see screenshot 1
I can "tune" the glow of the bulb by adjusting the frequency and/or duty cycle.

Screenshot 2 are the signals across the bulb (yellow CH1 voltage, green CH2 current, red math trace CH1*CH2 (current controller is 10mA/div. like the display says))

Picture is the updated diagram.

Regards Itsu

Quote from: TinselKoala on 2016.07.29, 15:42:33
Visualizing the timing delay introduced by running the basic clock signal through multiple gates in the 74C14 hex Schmitt trigger inverter:

Again, the Blue trace is the raw pulse signal generated by the oscillating 74HC123. The Magenta and Cyan traces are the result of running this signal through 2 and 3 gates in the 74C14.

OK, FWIW,  now I have obtained a 74AC14 instead of the 74C14 I was using initially. The risetimes are much faster now and the inter-gate delay is also much smaller. The risetimes are now fast enough to induce some ringing which wasn't there before.

Yellow: Output of U5
Cyan, Magenta: Outputs of 74AC14
Dark Blue: Raw output of U7

Quote from: Itsu on 2016.08.01, 16:09:13
K4zep and verpies,

thanks for the comments, i presently use my cheap chinese FG which is isolated from ground.



I added two IRF530 MOSFET in a self triggered synchronous rectifier setup behind the secondary, and now it seems that the load (6V / 0.6W bulb)
is hidden from the input as it glows dimly and at the same time i could create Luc's input voltage / current signals, see screenshot 1
I can "tune" the glow of the bulb by adjusting the frequency and/or duty cycle.

Screenshot 2 are the signals across the bulb (yellow CH1 voltage, green CH2 current, red math trace CH1*CH2 (current controller is 10mA/div. like the display says))

Picture is the updated diagram.

Regards Itsu

This is really excellent work Itsu. You are not far from reproducing the whole "Gunderson Effect" ! Just a little more twiddling and you too will have attained "zero input power" while lighting up a small load from your secondary output.

The next thing I would suggest is to see what happens when you apply a magnet or two to the core of your transformer. The idea here is to play with the saturation level of the core to see if it has any effect on the input measurements.

Quote from: ION on 2016.08.01, 15:55:58
in the past I would dive right in and start building, only to find the lake was full of crocodiles and snakes (replace with swindlers and fraudsters). I did my due diligence painfully after the dive.

Now I am more cautious, and do a lot of research on the background and integrity of the presenter.

Due diligence in researching the whole picture, not just the hardware and claims will save you lots of time, disappointment and heartache.

If you have an infinite amount of time, disregard the due diligence part and by all means, dive right in.

For me time, is short and I have to choose my endeavors with care and forethought.

As Poynt said :  "each to his own", and as an added note, be careful not to confuse discussion or points of view by thoughtful and experienced individuals with BS.

Regards, ION

This post is worth reading again.

Quote from: TinselKoala on 2016.08.01, 17:08:35
The next thing I would suggest is to see what happens when you apply a magnet or two to the core of your transformer. The idea here is to play with the saturation level of the core to see if it has any effect on the input measurements.

Is this what you are talking about ?

Saturation current

https://en.wikipedia.org/wiki/Saturation_current

The saturation current (or scale current), more accurately, the reverse saturation current, is that part of the reverse current in a semiconductor diode caused by diffusion of minority carriers from the neutral regions to the depletion region. This current is almost independent of the reverse voltage. (Steadman 1993, 459)

Quote from: evolvingape on 2016.08.01, 17:47:49
Is this what you are talking about ?

Saturation current

https://en.wikipedia.org/wiki/Saturation_current

The saturation current (or scale current), more accurately, the reverse saturation current, is that part of the reverse current in a semiconductor diode caused by diffusion of minority carriers from the neutral regions to the depletion region. This current is almost independent of the reverse voltage. (Steadman 1993, 459)

Like this: https://en.wikipedia.org/wiki/Saturation_%28magnetic%29

An applied permanent magnet to a ferromagnetic material biases that material closer to it's saturation point, without the need for the increase in bias current.

Dear All,

Here is another implementation (Bare Bones) of a Logic controller for the Gunderson Device. This one uses 100K trim pots and 1nF timing capacitors. It also gives a wide range of control over the timing parameters. I used 74LS123 and a 7414 (obsolete). The newer HC CMOS versions should give even better performance and can run at a higher voltage. This arrangement appears to have no starting problems.

What is missing are the trim pot saver resistors (typically 200 Ohm), more bypass capacitors, the push button switches that Graham used (for testing?) and the extra wiring shields around the ends of the IC's (see Graham's photo to see how this was done).  I don't know if any of this extra stuff is needed since the fast transition times are controlled by the Power MOSFET drivers.

It appears to me that all of the timing parameters in this sub-circuit are manually controlled. I thought there would be some logic derived control signals, but apparently not.

Many thanks to TinselKoala and k4zep for the development of this circuit.

Up next. The development of an LTSpice VII simulation to see how those timing pulses control the H-Bridge to generate that discontinuous 2/3's sine wave excitation input.

Spokane1

Quote from: ION on 2016.08.01, 18:07:35
Like this: https://en.wikipedia.org/wiki/Saturation_%28magnetic%29

An applied permanent magnet to a ferromagnetic material biases that material closer to it's saturation point, without the need for the increase in bias current.

Does this mean that the bias (input energy) is provided by a constant permanent magnetic field generated by a solid occupying space, and not by a variable electromagnetic input source (electrical energy) ?

This would increase system efficiency up toward 100% without having to account for bias energy input 'by the user' in the final COP calculation ?

Gentlemen,

Now we are getting down to the challenging aspect of this reverse engineering effort.  How does an H-Bridge switch control a parallel resonate tank circuit to dead line an oscillatory process and then start it up again 180 degrees out of phase?

I thought there would be some novel timing trick found in the logic controller, such is not the case. So far we only have a simple pulse train and its complement. When you hit a tank circuit with a single saw tooth current pulse it does just what it is suppose to do. It rings like a bell. Even the application of a second pulse of reversed polarity only complicates the response (at least so far).  This leads me to consider two situations:

1. There is additional control logic to be found on the H-Bridge card (I doubt it because experimental tuning and adjustment would be difficult)

2. The response of the secondary tank on the primary tank circuit is more profound than we are generally exposed to. This is complicated by the novel magnetic circuit.

The syntheses of the proper waveform from multiple signal generators is probably not going to offer much of a solution. We need to have the forcing function develop inside the transformer.

Any Ideas?

Spokane1

This is all I've come up with so far

Quote from: Spokane 1 on 2016.08.01, 18:22:15
Dear All,

Here is another implementation (Bare Bones) of a Logic controller for the Gunderson Device. This one uses 100K trim pots and 1nF timing capacitors. It also gives a wide range of control over the timing parameters. I used 74LS123 and a 7414 (obsolete). The newer HC CMOS versions should give even better performance and can run at a higher voltage. This arrangement appears to have no starting problems.

What is missing are the trim pot saver resistors (typically 200 Ohm), more bypass capacitors, the push button switches that Graham used (for testing?) and the extra wiring shields around the ends of the IC's (see Graham's photo to see how this was done).  I don't know if any of this extra stuff is needed since the fast transition times are controlled by the Power MOSFET drivers.

It appears to me that all of the timing parameters in this sub-circuit are manually controlled. I thought there would be some logic derived control signals, but apparently not.

Many thanks to TinselKoala and k4zep for the development of this circuit.

Up next. The development of an LTSpice VII simulation to see how those timing pulses control the H-Bridge to generate that discontinuous 2/3's sine wave excitation input.

Spokane1

Good work Spokane1!!!!

Ben K4ZEP

Love that LeCroy!

The pushbuttons aren't really needed, at least in my breadboard setup. It's fun to push them though.

The MM74HC123A that I am using has an operating voltage range of 2-6 volts with absolute maximum of 7.0 volts.
The MC74AC14N has the same operating voltage range and absolute maximum. These chips are designed to work at nominal Vcc of 5 volts.

There is a broad latitude for selecting the timing components, obviously. I like to use relatively larger capacitances and smaller resistors because I think this minimizes effects due to stray breadboard capacitances, but... use what you've got if it works. That's what breadboards are for!

Extra wiring shields around the ends of the ICs? Whaat?  Are we talking about the orange links next to either end of the ICs? The pins 8 and 9 of the 74HC123s are connected together and to ground, and this is just the easiest way to make this connection. On the Pin 1-Pin 16 side the connection is made to the timing components, Vcc and ground as well. It just turns out that Gunderson wired his breadboard with a short link in both places. On my implementation I used a short purple link for 8-9 similar to Gunderson's link, but it was easier for me to wire the 1-16 ground and Vcc connections to separate top and bottom ground rails. I don't think these short links are meant as "extra wiring shields", they are simply connections to ground that are made in the most convenient way for a particular layout.

Don't forget that Gunderson also has the 556 pulse generating circuit sending pulses to the H-bridge and the Synch. rectifier. These pulses may have something to do with the discontinuous 2/3 sine wave excitation.

I'd like to identify the components on the mosfet board.   (ETA: I see now that this is the synchronous rectifier board.)
I see IL710-2, which is a high-speed digital isolator ...

QuoteNVE's IL700 family of high-speed digital isolators are CMOS
devices manufactured with NVE's patented* IsoLoop® spintronic
Giant Magnetoresistive (GMR) technology.

Oooh.... sounds very impressive and mysterious.

Also IXDD614PI high-speed mosfet drivers, one per mosfet.

There is also a DIP 14 chip whose number I can't read. I think it is a 74AC something. Any guesses? Could this be involved in synching the 556 pulses with the outputs of the U7 oscillator system?

Oh, this is the Synchronous Rectifier system, not the input H-bridge.

Ah, I see it is another 74AC14. So it's not combining the 556 pulses.

But what are the other components here?

I see NEC 2505 LL807, which appears to be another optoisolator of the AC-input type (two emitting diodes in anti-parallel)
But why three sets (per mosfet) of whatever this is?

And the transistor-like packages are unknown. I can't find a clear picture of them showing the part number.

Quote from: Grumage on 2016.08.01, 15:02:50
Are there semiconductors available that are fully isolated?

Yes there are, but they require isolated power supplies, too.

Is this the best picture of what might be the H-bridge driver system? There is certainly a lot going on there. Pulse-shaping network?

Quote from: TinselKoala on 2016.08.01, 17:08:35
This is really excellent work Itsu. You are not far from reproducing the whole "Gunderson Effect" ! Just a little more twiddling and you too will have attained "zero input power" while lighting up a small load from your secondary output.

The next thing I would suggest is to see what happens when you apply a magnet or two to the core of your transformer. The idea here is to play with the saturation level of the core to see if it has any effect on the input measurements.

Trying to measure the BH curve of my used 10mH/100uH transformer using this website setup:  http://meettechniek.info/passive/magnetic-hysteresis.html
under "Measuring arrangement with a digital oscilloscope" only shows a circle on the screen, see screenshot 1.

Guess the 10mH/100uH transformer relationship instead of the needed 1:1 transformer setup is causing the circle instead of the expected BH curve.

Other 1:1 transformers do show the expected BH curve, see screenhot 2 and i can manipulate this BH curve with permanent magnets.

Itsu

Quote from: Spokane 1 on 2016.08.01, 21:43:37
Gentlemen,

Now we are getting down to the challenging aspect of this reverse engineering effort.  How does an H-Bridge switch control a parallel resonate tank circuit to dead line an oscillatory process and then start it up again 180 degrees out of phase?

I thought there would be some novel timing trick found in the logic controller, such is not the case. So far we only have a simple pulse train and its complement. When you hit a tank circuit with a single saw tooth current pulse it does just what it is suppose to do. It rings like a bell. Even the application of a second pulse of reversed polarity only complicates the response (at least so far).  This leads me to consider two situations:

1. There is additional control logic to be found on the H-Bridge card (I doubt it because experimental tuning and adjustment would be difficult)

2. The response of the secondary tank on the primary tank circuit is more profound than we are generally exposed to. This is complicated by the novel magnetic circuit.

The syntheses of the proper waveform from multiple signal generators is probably not going to offer much of a solution. We need to have the forcing function develop inside the transformer.

Any Ideas?

Spokane1

This is all I've come up with so far

You asked so here is a thought.

There is the possibility that the H bridge is not responsible for clamping the ringing, rather it is the synchronous rectifier that clamps the discharge of magnetic energy, shuttling it to the load, thus preventing the first cycle from continuing to "ring like a bell" i.e. no visible damped oscillation.

When all the magnetic energy is discharged via the synchronous rectification, it would be no problem  to start the next cycle in either direction, positive or negative going.

This might account for the lack of ringing after the first full cycle.

It would seem that a fairly accurate timing diagram could be derived by back engineering from the available scope shots.

Playing with the timing of multiple function generators could work if they are synced to each other and the  synchronous rectifiers are activated properly during the dead time for the required duration.

If it can be found that there is some type of feedback from the magnetic circuit that is required to trigger the drive signals, it gets a bit more complicated.

Regards, ION

Quote from: TinselKoala on 2016.08.02, 01:44:36
Oh, this is the Synchronous Rectifier system, not the input H-bridge.

Ah, I see it is another 74AC14. So it's not combining the 556 pulses.

But what are the other components here?

I see NEC 2505 LL807, which appears to be another optoisolator of the AC-input type (two emitting diodes in anti-parallel)
But why three sets (per mosfet) of whatever this is?

And the transistor-like packages are unknown. I can't find a clear picture of them showing the part number.

There appear to be three sets of wires coming from the lower coils on the magnetic assembly.

Could it be that each of these sets  of wires feeds an NEC opto-isolator  for feedback timing or gating purposes, with the IL710-2  NVE opto-isolator being used for overall control from the timing circuit?

Seems overly complex to me, but I don't know what he had in mind.

Quote from: TinselKoala on 2016.08.02, 01:27:42
The pushbuttons aren't really needed, at least in my breadboard setup. It's fun to push them though.

There is a broad latitude for selecting the timing components, obviously. I like to use relatively larger capacitances and smaller resistors because I think this minimizes effects due to stray breadboard capacitances, but... use what you've got if it works. That's what breadboards are for!

Dear TK,

You are right that chip can handle a wide range of timing components.

Spokane1

Dear TK,

Have you seen this photo? This was taken by Reiyuki and posted via a Google drop box location. The link was posted in the first couple pages of this thread. These include photos from his note book and eight shots of scope photos from the presentation. Let me know if you need a copy of these 35 photos.

I don't think I'm going to attempt to reverse engineer those circuits from what photos we have. Graham said that the synchronous diode board has many components underneath it. About the best I could do is to get an idea of the kinds of components were used and cobble together something within my price range. Fortunately the purpose of these circuits is to provide extremely fast isolated switching. I understand that the same driver topology is used in each circuit.

You noticed that IXI616 chip (or something like that) on the Synchronous Diode board. I don't believe that is used for driving the CREE SiC Power FET's. Graham said that he had a fast separate surface mount driver attached to the back of each FET which is unseen in the photos.

Also notice the six (6) each surface mount components soldered directly to the top of the FET (2 Zeners, 2 diodes, 2 capacitors). This is a gate snubber circuit all intended for the elimination of the FET gate resistor. As I have said Graham drives his FET's with +12V rather than the recommended +10 Volts.

If I recall correctly, the over all design intent was to provide an isolated power supply for the driver components. So the transformer seen with each circuit system is a custom switch mode power supply unit. Graham said that he used three voltages  +12 and -5 for the actual driving functions (negative bias for the off state) and +5 V for his driver chip logic. So there must be a power switching transistor or MOSFET under the board to support the function of the driver power supply.

Apparently there is a single voltage that supplies the two subsystems (+20V ?) that comes from the logic board.

I'm confused about the purpose of the photo isolators. There are six of them along with their support components. If these are logic inputs then not all of them are used. Graham said something about how these are Zener devices and were used for regulataion of the isolated supply voltages. I could have that wrong. (most likely)

Anyway you look at it these circuits are a work of art. I rarely see such craftsmanship with so small of components. With my shaky hands I doubt if I could reproduce that level of quality.

Perhaps you can get a general idea as to what was going on there and that will be the best we can do.

Thanks again for your time you are spending on this project.

Spokane1

All that matters to me at this point is COP.. Black box.. Pin / Pout.

I do not invest energy in pursuit of system COP<1 efficiency performance increase, especially when it is murky and unnecessarily complicated, without a necessary justifying statement / claim from the inventor.

Quote from: ION on 2016.08.02, 14:31:23
You asked so here is a thought.

There is the possibility that the H bridge is not responsible for clamping the ringing, rather it is the synchronous rectifier that clamps the discharge of magnetic energy, shuttling it to the load, thus preventing the first cycle from continuing to "ring like a bell" i.e. no visible damped oscillation.

When all the magnetic energy is discharged via the synchronous rectification, it would be no problem  to start the next cycle in either direction, positive or negative going.

This might account for the lack of ringing after the first full cycle.

It would seem that a fairly accurate timing diagram could be derived by back engineering from the available scope shots.

Playing with the timing of multiple function generators could work if they are synced to each other and the  synchronous rectifiers are activated properly during the dead time for the required duration.

If it can be found that there is some type of feedback from the magnetic circuit that is required to trigger the drive signals, it gets a bit more complicated.

Regards, ION

Dear ION,

I agree that the function of the backend diode is important for this process, However the timing just doesn't add up.

Graham specifically pointed out (more than once) that the operation of the backend diode was an extremely short event.  I believe that you can discharge a capacitor quickly but an inductor (the conversion transformer secondary) has specific time limitations.

If there is some shorting process that effectively removes all the circulating energy from the input tank circuit then I believe it will take a few cycles to boost the 200 VDC input voltage back to the 800 VAC steady state circulating tank voltage.

I still don't have a solution, but I did notice something interesting in my LTSpice XVII simulation.  That large backend capacitor is not electrically part of a secondary parallel resonate circuit. Even though it is wired in parallel with the coils the resulting currents cancel each other out. In the world of perfect components of the simulation program there is no voltage drop across the backend capacitor (if the FETS remain closed). Reducing the capacitance value from 6618 uF down to 100 uF has no impact on the operating frequency. However this capacitor impacts the effective inductance as seen by the primary and the operating frequency doubles from what it was with a simulated single secondary.

This is going to be interesting but I'm no spice expert.

Thanks for your stab at this. I shall include it to the list of possibilities.

Spokane1

Dear Spokane1

Thanks for your review of my post. I will rethink this a bit more. Perhaps trying to match up the proposed timing diagram with the scope shots , if not then perhaps try to devise the timing diagram from the scope shots.

QuoteIf there is some shorting process that effectively removes all the circulating energy from the input tank circuit then I believe it will take a few cycles to boost the 200 VDC input voltage back to the 800 VAC steady state circulating tank voltage.

Looked at another way, if 200 VDC were applied to a tank circuit  for x time, then released, the first ringing cycle could easily be over 800 VAC (depending on the time constant and stored energy of L and C), and  subsequent cycles would progressively reduce depending on the  "Q" of the tank or external loading.

In this case there are no subsequent ringing cycles as a new charge is initiated quickly.

Does anyone know the source of the labeling on the attached image, as it does not jive with what I understand of how such a circuit should behave.

Yellow trace is labeled as input voltage, but to me looks like the ringing voltage after the input current ramp is completed and the H bridge "lets go". (at the end of the dead time?)

Green and blue also seem to me to be reversed. Green is labelled as output current, but seems to me it should be input current. A charging inductor produces a ramp (like the green trace), a discharging inductor is a current source, and would produce something more like the blue trace.

Blue it seems would be correct for output current, but is labeled as input current.

Perhaps I need to study the scope shots more thoughtfully, as they do not make sense to me as labeled, or maybe I am color blind, or a too old switchmode guy that never got into resonant topologies and associated waveforms.

Kind Regards, ION

Edit: The labeling was supposedly supplied by Reiyuki.

Quote from: Itsu on 2016.08.02, 10:10:18
Trying to measure the BH curve of my used 10mH/100uH transformer...

What are the inductances of one winding while you short the others?

The 100uH secondary becomes 8uH and the 10mH primary becomes 3.5mH.

Itsu

Can anyone take a guess at a timing diagram that matches up (lines up) with the attached Gunderson scope shot, perhaps graphically just below it?

Quote from: ION on 2016.08.02, 20:41:21
Can anyone take a guess at a timing diagram that matches up (lines up) with the attached Gunderson scope shot, perhaps graphically just below it?

Dear ION,

During the presentation Graham made a comment about the waveform. He said (something to the effect) that the Hz notation only applies to pure sine waves. The CPS designation applies to periodic non-sine wave functions. This reference was to the yellow tank voltage trace that he describe as the input voltage. This was 50k CPS or a period of 20 microseconds. The sine wave part of this forcing function is running at about 78 kHz (since it is a sine wave - or close to it). This makes each third of the voltage form about 6.666 microseconds in duration.

At his shop he said that he was going to make another bridge that only used two switches, perhaps this is to be a half H-Bridge. This makes sense since the logic circuit only provides ne timing signal and its complement. There presently appears to be no logic signal provided for the other half of the H-Bridge. Maybe he used the full H-Bridge at one time.

This would be nice since it would reduce the cost of the H-Bridge by 50%.

For starters in my simulation I'm only using one current pulse to excite the input circuit tank. The value of the parallel capacitor is 0.008 uF (8X .001 uF capacitors in parallel). I'm pretty sure that the inductance of the various coils are non-linear The input choke has a gap (about 0.020") as does the transformer (0.010"). This means that the K factor (coupling factor) is less than 0.99. I'm using k=.8553 since I read somewhere that this was the optimal coupling for a coupled coil (as in induction coil) to achieve 3rd harmonic ringing. So far I haven't seen any higher harmonics appear, just lower frequency beats.

For 99% of the time the transformer output is effectively shorted. As I have said the back end capacitor is not there since the applied currents cancel out each other due to the topology of the split secondary. From my count there are about 144 turns on the primary and 24 turns on each of the secondary coils that is a 6:1 ratio (step down) for each secondary coil. That would imply that the backend peak voltage would be about 800/6 = 133 VAC. Yet somehow he is still blowing 1.2 KV rated FETS. Probably because the body diodes take a dump on the account of excessive current. After all those huge secondary conductors are there for a reason.

I would not recommend lumping the two secondary's into one circuit element for a simulation, at least not yet.

As Graham visualizes it the energy sloshes back and fourth through the transformer. At the proper moment he just briefly opens the gate at the right time to take a tiny snip of the huge reactive oscillation power. He also said that the H-Bridge gives each cycle a little push to maintain the steady state. I visualize it like push a playground swing - which is probably a gross over simplification. Graham said that there was about 1.6kVA of reactive power circulating in his apparatus. Classical rough calculations would yield about 2A of circulating current at 800 VAC. The actual values would have to be done with RMS data.

Maybe the presentation will cover some of this tomorrow (or the next day)

Spokane1

Spokane1

Thank you for the response. Have you posted your sim anywhere or a rough schematic of the transformer input /output and capacitor placement?

I would be very interested in this.

BTW, it seems to me that hitting a parallel resonant tank circuit that is "discharged" will put huge stresses on the semiconductors in the H bridge due to the peak currents generated by driving into a capacitor. There is nothing to limit or absorb the peak currents except for the semiconductor junctions and tiny inductance of the connecting wires.


If the tank was ringing and hit at the right time in the cycle e.g ZVS or ZCS or if the tank was series resonant the semi's could enjoy a long life.

Thanks, ION

Dear ION,

Here is my perception of how the circuit is wired and one possible implementation of an attempted simulation. I've only been into this a few hours. Also attached are some response traces to single and double current pulses.

Spokane1

Thanks for that Spokane1.

Good to see that there is an inductor in series with the tank, that should reduce inrush current.

Now the input (yellow) waveform makes sense, it is the input to the tank, not the output of the H bridge.

Quote from: Spokane 1 on 2016.08.03, 14:56:56
Dear ION,

Here is my perception of how the circuit is wired and one possible implementation of an attempted simulation. I've only been into this a few hours. Also attached are some response traces to single and double current pulses.

Spokane1

Bunch of work there Spokane1,
Wow.

Ben K4ZEP

Quote from: Spokane 1 on 2016.08.03, 14:56:56
Dear ION,

Here is my perception of how the circuit is wired and one possible implementation of an attempted simulation. I've only been into this a few hours. Also attached are some response traces to single and double current pulses.

Spokane1

I would add some high speed diodes or synchronized switches to the output circuit. Presently there is no rectification of any sort, just a near dead short and this will adversely affect the input waveform.

To add another dimension to the argument of whether the H bridge can be replaced with a power function generator, it would only be true if the function generator could provide an open condition to the tank circuit during the 33% dead time. Normally, AFAIK, a FG will present a low impedance output to a DUT when programmed for zero output e.g. during the dead time. So the H bridge may be unique by "letting go" of the tank during this period.

  I am not familiar with power function generators capabilities, or whether it can provide the open condition so I tried driving the tank from a Voltage source executing a PWL (piece wise linear function)using a PWF file in LTSpice. The results do not match up with the same  pulse sequence driving an H bridge.

I used the same values and timing as Spokane1's sim except only 5 cycles at +/-200 volts. The 200 volts was so that I could get some juice into the rectifiers 1N5817. I realize I may need more than 5 cycles.

So this is a single ended drive from a pseudo FG. Next I'll show an actual  FET H bridge.

This is a first rough shot at it. I'll post all my results and more detail as time permits. forgive the errors.


green =drive
blue = tank voltage
red =tank current

Quote from: ION on 2016.08.03, 19:25:37
I would add some high speed diodes or synchronized switches to the output circuit. Presently there is no rectification of any sort, just a near dead short and this will adversely affect the input waveform.

Dear ION,

Good that you noticed. The intent was to get a feel for the shorted state operation before attempting to harvest some energy on the back end. It is my understanding that the shorted condition is what is going on 99% of the time, but I may be incorrect.

Good work on you own simulation. Using a controlled forcing function (wave generator) is a good approach.

Attached is a system schematic that we can all Bleed on (make changes with red markings)

Spokane1

Thanks for the schematic, it looks fine.

Spokane 1 said:

QuoteGood work on you own simulation. Using a controlled forcing function (wave generator) is a good approach.

Yes but I think I need to add at least one switch to open and release the FG during the dead time to simulate the H bridge, and add the output synchronous switches. Working on that.

Looks like not much activity on this transformer anymore except for a very few.



Regards, ION

I have finished watching Graham's MIT video and IMO, he really does give enough info to replicate.

Several points of note, when the input current as shown on CH2 (blue) reaches the negative peak, the primary core is saturated. At the flat top portion of the waveform, the primary core is at a higher permeability.  The core bias is set of course by the PMs.

During the negative half cycle of the input voltage on CH1, the first pair of opposite H bridge switches are on and during the positive half cycle the second switch pair are used which is already known. However, during the remaining part of the input cycle when the core is in a high perm state, the second pair is left on and conduction is thru primarily the mosfet reverse conduction.

The synchronous mosfet switching uses both normal and reverse conduction and may have already been pointed out but not that I recall. While current is flowing thru the mosfet in the normal mode that is, when conventional current flow is from drain to source, the gate is switched off for a short period producing a large positive going voltage spike on the secondary. Also during this time, the secondary current and flux reverses
direction as is seen in the green channel and when the peak reverse is reached, gate voltage is again applied to the mosfet which now conducts in the reverse conduction mode that is, from source to drain with very low resistance. This action is what results in the output caps being charged negatively as shown. Graham also stated that the output current phasing is 180' from that shown.

Regarding the issue of the current connections on the input Clark/Hess power analyzer, he removed the original connections well after he had explained the operation of the device and all the input/output measurements. He made the change during a time he was explaining another aspect of the device.

The secondary core half is operated in a mostly linear mode while the primary core half is operating around the saturation knee. This seems to be a key prerequisite in the appropriate flux flows (or magnetic vector potential) along with the proper output synchronous timing.

I can now understand how a small capacitance added to the gates of the output synchronous mosfets could destroy the OU effect.

pm

Quote from: partzman on 2016.08.04, 18:24:06
I have finished watching Graham's MIT video and IMO, he really does give enough info to replicate.

Several points of note, when the input current as shown on CH2 (blue) reaches the negative peak, the primary core is saturated. At the flat top portion of the waveform, the primary core is at a higher permeability.  The core bias is set of course by the PMs.

During the negative half cycle of the input voltage on CH1, the first pair of opposite H bridge switches are on and during the positive half cycle the second switch pair are used which is already known. However, during the remaining part of the input cycle when the core is in a high perm state, the second pair is left on and conduction is thru primarily the mosfet reverse conduction.

The synchronous mosfet switching uses both normal and reverse conduction and may have already been pointed out but not that I recall. While current is flowing thru the mosfet in the normal mode that is, when conventional current flow is from drain to source, the gate is switched off for a short period producing a large positive going voltage spike on the secondary. Also during this time, the secondary current and flux reverses
direction as is seen in the green channel and when the peak reverse is reached, gate voltage is again applied to the mosfet which now conducts in the reverse conduction mode that is, from source to drain with very low resistance. This action is what results in the output caps being charged negatively as shown. Graham also stated that the output current phasing is 180' from that shown.

Regarding the issue of the current connections on the input Clark/Hess power analyzer, he removed the original connections well after he had explained the operation of the device and all the input/output measurements. He made the change during a time he was explaining another aspect of the device.

The secondary core half is operated in a mostly linear mode while the primary core half is operating around the saturation knee. This seems to be a key prerequisite in the appropriate flux flows (or magnetic vector potential) along with the proper output synchronous timing.

I can now understand how a small capacitance added to the gates of the output synchronous mosfets could destroy the OU effect.

pm

Excellent, where did you get your copy so fast and when is it available?  I usually get an email on new "DVD's".....

Ben K4ZEP

Quote from: K4zep on 2016.08.04, 18:41:04
Excellent, where did you get your copy so fast and when is it available?  I usually get an email on new "DVD's".....

Ben K4ZEP

I purchased it online as a download from Aaron's sticky on EF.

http://www.energeticforum.com/renewable-energy/20547-magnetic-implosion-transformer-graham-gunderson.html

They have a special price of $27.00 now thru 8/5 (tomorrow) and IMO, it is well worth the price!

pm

Free energy! Only 27 dollars, get yours today, before the price goes up !     :D


Yes, it's your money to spend as you like, I just really hate to see it going to those people. 

Quote from: partzman on 2016.08.04, 19:12:05
I purchased it online as a download from Aaron's sticky on EF.

They have a special price of $27.00 now thru 8/5 (tomorrow) and IMO, it is well worth the price!

pm

Are you not concerned with your credibility on this forum pm ?

Quote from: evolvingape on 2016.08.04, 19:50:22
Are you not concerned with your credibility on this forum pm ?

I am not quite sure what you mean EA? If my comments and purchase of G's video thru EF affects my credibility here, then no, I don't really care and each would be better off without the other!!!

pm

There are many fellows who have strong opinions about "selling" _free energy_
Here I would appreciate you discuss that topic elsewhere and leave this thread for an investigation
of Graham's claim.

and  Honestly at this point the video is quite important for a proper investigation!

as far as Partzman's credibility at this forum , In my book it will never be in question.
he is a most valued asset to the open source community and is striving to be of service to this cause in any way he can.
and I must add ...
he can do a lot !!!

with respect and gratitude
Chet K

Quote from: TinselKoala on 2016.08.04, 19:41:39
Free energy! Only 27 dollars, get yours today, before the price goes up !     :D


Yes, it's your money to spend as you like, I just really hate to see it going to those people.

IMO, Graham's device works as advertised but there is a major issue which will reveal itself as time goes on. That issue is the high ratio of reactive input power to real output power as I've stated previously.  I'm sure Graham is aware of this and I believe he will overcome this problem as he is more than capable but it must be solved before the device can operate OU in a closed loop.  This could possible if the operating frequency is lowered, the VAR/real power ratio is reduced, and the complexity perhaps simplified. All this of course is just my opinion!

pm

It appears that Graham has no claim, the COP>1 violation claim came from Spokane 1, a proxy.

What is left is Graham's data, and is being examined by qualified and interested minds of this forum.

G's machine has not done a Watt of work demonstrated and there is no solid theory to suggest it could be so.

What is left is the system itself and the challenge in understanding, a goal to which qualified and talented minds of this forum seem to have applied themselves, for reasons known only to them..

So what's the problem ?

I will let Spokane1 address/qualify that.
I was not present at the conference ,I have not seen the video
and anything from my mouth [or keyboard] would be hearsay .


respectfully
Chet K

Quote from: Chet K on 2016.08.04, 20:55:45
I will let Spokane1 address/qualify that.

Smart move.

Quote from: Chet K on 2016.08.04, 20:55:45
I was not present at the conference ,I have not seen the video
and anything from my mouth [or keyboard] would be hearsay .

That's the information environment these day's, it takes committed hard work to verify sources and build a credible model of reality.

Quote from: evolvingape on 2016.08.04, 20:49:24
It appears that Graham has no claim, the COP>1 violation claim came from Spokane 1, a proxy.

Dear evolingape,

I made the over unity and COP>1 claims last July 10 at the end of the convention. Graham didn't make such a claim during his presentation at the conference. When he presented the operation of his machine and I could see the numbers (with the probes connected to the input C-H analyzer) I asked him from the front row "That looks like a COP of 6" he replied "Somewhere around there" and that is about as much as he said on the subject of OU or COP. However Graham did talk a lot more about the performance of his equipment in his shop.

I thought that making measurements with two independent expensive measurement systems that agreed is a far greater burden of proof than I have seen in this community - ever. 

I was expecting his machine this year to produce a COP of 1.05 or so. A COP of 5 is something to write home about and so I did.

I take full creditably responsibility should this new technology fall on its nose and turn out to be some kind of exotic instrument error, or an error in how the measurements were taken. So be it.

Extraordinary claims (which this is) require extraordinary proof (which we don't have).  We as a community are only going to get to the bottom of this is by building similar apparatus and exploring the measurements for our selves.  There is absolutely no requirement that you do this. Personally I think there is real merit in this technology based upon my years of research in a similar field.

I find your comments relevant and to the point. If this technology is to have half a chance at becoming mainstream then it will have to face all of this forums critical analysis and more.

Thank you for your input and keep those perspectives coming.

Spokane1
Mark McKay, PE

Quote from: Spokane 1 on 2016.08.04, 22:24:07
Dear evolingape,

I made the over unity and COP>1 claims last July 10 at the end of the convention. Graham didn't make such a claim during his presentation at the conference. When he presented the operation of his machine and I could see the numbers (with the probes connected to the input C-H analyzer) I asked him from the front row "That looks like a COP of 6" he replied "Somewhere around there" and that is about as much as he said on the subject of OU or COP. However Graham did talk a lot more about the performance of his equipment in his shop.

I thought that making measurements with two independent expensive measurement systems that agreed is a far greater burden of proof than I have seen in this community - ever. 

I was expecting his machine this year to produce a COP of 1.05 or so. A COP of 5 is something to write home about and so I did.

I take full creditably responsibility should this new technology fall on its nose and turn out to be some kind of exotic instrument error, or an error in how the measurements were taken. So be it.

Extraordinary claims (which this is) require extraordinary proof (which we don't have).  We as a community are only going to get to the bottom of this is by building similar apparatus and exploring the measurements for our selves.  There is absolutely no requirement that you do this. Personally I think there is real merit in this technology based upon my years of research in a similar field.

I find your comments relevant and to the point. If this technology is to have half a chance at becoming mainstream then it will have to face all of this forums critical analysis and more.

Thank you for your input and keep those perspectives coming.

Spokane1
Mark McKay, PE

Hi Mark,

I spent the bucks, downloaded it, just watched it, will have to watch it many more times and all BS aside,
well worth the money in my simple opinion!  My head is spinning, got a lot to digest!  Cleared up a lot of
misconceptions, cleared up waveforms, Tesla would sure like the switches we have now!!!!!!

Ben K4ZEP

I guess in a year or two we will have a better idea if the video is a 27 dollar instruction in OU, or a 27 dollar guide of distraction. Time will tell. From Thane Heins to UFO politics ect. one thing holds true, time proves them all to be full of it. And mucho dollars get wasted.

Good luck to those who investigate it and bad juju to those who deceive. So many similar past claims and not one single claim has been legit.

..

FarmHand

Been watching the same claims come and go for years, mostly claims which don't survive the Rigors of a proper investigation
or unsubstantiated claims which were never allowed a proper investigation.

although there have been a few claims which drifted into obscurity and remain as an itch which just can't be scratched ......

here however,  we have a fellow doing his absolute best to clear the air around his claim with a very good test protocol , in a public venue
along with a tutorial ....

this is a first for me !!

it is true , a proper investigation of any such claim Begs further investigation and qualification by Peers ...[at the very least]
I hope Graham decides to go this route !

respectfully
Chet K

Quote from: Spokane 1 on 2016.08.01, 21:43:37
Gentlemen,

Now we are getting down to the challenging aspect of this reverse engineering effort.  How does an H-Bridge switch control a parallel resonate tank circuit to dead line an oscillatory process and then start it up again 180 degrees out of phase?

I thought there would be some novel timing trick found in the logic controller, such is not the case. So far we only have a simple pulse train and its complement. When you hit a tank circuit with a single saw tooth current pulse it does just what it is suppose to do. It rings like a bell. Even the application of a second pulse of reversed polarity only complicates the response (at least so far).  This leads me to consider two situations:

1. There is additional control logic to be found on the H-Bridge card (I doubt it because experimental tuning and adjustment would be difficult)

2. The response of the secondary tank on the primary tank circuit is more profound than we are generally exposed to. This is complicated by the novel magnetic circuit.

The syntheses of the proper waveform from multiple signal generators is probably not going to offer much of a solution. We need to have the forcing function develop inside the transformer.

Any Ideas?

Spokane1

This is all I've come up with so far

This is still a very good question. If anyone has purchased the MIT video, are there any hints regarding how this is done?

To me the transformer design has very high leakage inductance primary to secondary due to the topology (magnetically a widely separated primary and secondary, gaps etc.)therefore, even if the secondary were shorted during the dead time it would not quench the oscillation on the primary so effectively, we should see a ringdown due to the leakage inductance.

This leads me to believe that there is nothing happening on the secondary that can quench the oscillation so effectively. It may assist, but would not alone produce the tiny damped oscillation we see in the scope shots.

Could it be the H bridge that is shorting the tank + swinging choke with a third cycle on the timing diagram where both the grounding Fets in the H bridge are kept on during the dead time cycle?

It is my belief that the basic circuit can be back engineered from a very careful analysis of the scope shots. Additional circuit info is a bonus and fills in some missing pieces. The elusive factors will be the exact magnetic circuit and materials.

Your important observation and question needs a good answer.

I might add that the DC output from a uniquely wound transformer is what Steve Marks achieved in several stand alone units of very high power.

Consider this: The ordinary magneto ignition system on a old style gas engine lawn mower motor uses a fly by magnet to induce a current into a coil which remains shorted by the points. At the appropriate moment, the points open and a HV DC pulse appears on the secondary. Normally this pulse rings with the distributed capacitance and the capacitor across the points to turn the DC output pulse into an HV oscillatory damped wave,  but the DC  could also be captured with a flyback diode and dumped into a capacitor. Some folks used a DC restorer circuit on the primary to present a single DC burst on the HV coil output rather than an damped ringing wave. This supposedly created a hotter spark from a unidirectional pulse.

In GG's device it seems  (in comparison) that the fly by magnet is replaced with a primary tank  circuit to charge the magnetics. GG opens the (points aka FET's) on the secondary at the appropriate time to release some, but not all of the stored current flowing in the coils of the magnetic circuit.

Regards, ION

Quote from: ION on 2016.08.05, 12:22:41
To me the transformer design has very high leakage inductance primary to secondary due to the topology (magnetically a widely separated primary and secondary, gaps etc.)therefore, even if the secondary were shorted during the dead time it would not quench the oscillation on the primary so effectively, we should see a ringdown due to the leakage inductance.

This leads me to believe that there is nothing happening on the secondary that can quench the oscillation so effectively. It may assist, but would not alone produce the tiny damped oscillation we see in the scope shots.

I agree.
Itsu's transformer also has a large leakage inductance, because when he shorted the secondary, the inductance of the primary was far from zero.

Quote from: ION on 2016.08.05, 12:22:41
This is still a very good question. If anyone has purchased the MIT video, are there any hints regarding how this is done?

IMO from Graham's comments, this is what is occurring in the primary circuit. We know that PMs bias the primary core flux in the saturation knee of the B/H curve. Then laying aside for a moment how we accomplish the following circuit wise, we reference G's scope shot. The negative half sine wave cycle produces current in the primary that moves the core up the B/H curve further into saturation during the negative current excursion seen following the voltage by ~90'. This of course reduces the core permeability to a very low value.

At the peak of this low perm level in the core, we now apply a positive half sine voltage to the primary which reverses the H field and moves the core down the B/H curve below the sat knee resulting in a relatively high core permeability. During this high perm core state, the current/flux is clamped or frozen during the remaining portion of the cycle and then the entire cycle is repeated.

Graham has accomplished this with an H bridge in a current fed parallel resonant configuration plus reverse mosfet conduction.

Quote
To me the transformer design has very high leakage inductance primary to secondary due to the topology (magnetically a widely separated primary and secondary, gaps etc.)therefore, even if the secondary were shorted during the dead time it would not quench the oscillation on the primary so effectively, we should see a ringdown due to the leakage inductance.

This leads me to believe that there is nothing happening on the secondary that can quench the oscillation so effectively. It may assist, but would not alone produce the tiny damped oscillation we see in the scope shots.

Could it be the H bridge that is shorting the tank + swinging choke with a third cycle on the timing diagram where both the grounding Fets in the H bridge are kept on during the dead time cycle?

It is my belief that the basic circuit can be back engineered from a very careful analysis of the scope shots. Additional circuit info is a bonus and fills in some missing pieces. The elusive factors will be the exact magnetic circuit and materials.

Your important observation and question needs a good answer.

I might add that the DC output from a uniquely wound transformer is what Steve Marks achieved in several stand alone units of very high power.

Consider this: The ordinary magneto ignition system on a old style gas engine lawn mower motor uses a fly by magnet to induce a current into a coil which remains shorted by the points. At the appropriate moment, the points open and a HV DC pulse appears on the secondary. Normally this pulse rings with the distributed capacitance and the capacitor across the points to turn the DC output pulse into an HV oscillatory damped wave,  but the DC  could also be captured with a flyback diode and dumped into a capacitor. Some folks used a DC restorer circuit on the primary to present a single DC burst on the HV coil output rather than an damped ringing wave. This supposedly created a hotter spark from a unidirectional pulse.

In GG's device it seems  (in comparison) that the fly by magnet is replaced with a primary tank  circuit to charge the magnetics. GG opens the (points aka FET's) on the secondary at the appropriate time to release some, but not all of the stored current flowing in the coils of the magnetic circuit.

Regards, ION

I agree that the transformer has a relatively large leakage inductance. If we knew more about the physical dimensions of the core and windings plus the turns count, this could be reasonably calculated. IMO this leakage inductance is very important to the device's operation. Since the secondary is shorted for >99% of the time, the leakage inductance is storing energy and would therefore be involved with the secondary in the flux and current reversal seen in the scope shot during the synchronous mosfet "off" time.

During the dead time when the primary flux is frozen by the reverse conduction mode in the appropriate bridge switches, we do see a slight drop in this current during the synchronous mosfet "off" time in the secondary. It is slight due to the poor coupling between the primary and secondary or the high leakage inductance. IMO, there is no continuous ringing of the leakage inductance as would normally expected but rather one differential part of a ring cycle during the secondary "off" time when the current and flux reverses. Other than this brief "off" time, the transformer secondary circuit is continually loaded.

So, we must consider the fact that the secondary current and core flux reverse during this short "off" time interval and are then quickly switched back into the output circuit. The current and flux are then seen on the scope shot to recover or reverse again over a longer period as determined by the secondary magnetic core and leakage inductance time constants. It is this action which creates the overall DC negative output current in the secondary winding and was explained in this way by Graham in the video.

What is puzzling to me is how the circuit 'maintains' the negative DC current in the secondary? Normally periodic functions, even transient ones, eventually average to zero. Perhaps this is what Graham is referring to when he states that he really doesn't know how it works. 

pm

         

That's a very thoughtful analysis pm.

You also said:

QuoteWhat is puzzling to me is how the circuit 'maintains' the negative DC current in the secondary? Normally periodic functions, even transient ones, eventually average to zero. Perhaps this is what Graham is referring to when he states that he really doesn't know how it works.

I agree.

It appears to me that what partzman and ION are describing is a loss mechanism, dissipating power in the core and H-bridge components, rather than one which could produce any gain in energy.

Oh, where is MarkE when he's needed. I'm sure he would have some very interesting things to say about this.

Here's a Tek pdf primer on Probe Skew.

Did Gunderson mention corrections for probe skew in his presentation?

Quote from: TinselKoala on 2016.08.05, 17:16:21
Here's a Tek pdf primer on Probe Skew.

Did Gunderson mention corrections for probe skew in his presentation?

Thanks TK this is a "must read" for serious power analysts.

Where common sense cuts to the core (no pun intended):

All probe skewing and Clark-Hess glitches aside, I still would have liked to see a $20 "Kill-A-Watt" meter on the AC mains side of his device. With a claimed 5x or more gain in power, surely we could have looked at mains Watts in and Watts out (even guestimating the automobile lamp wattage by brightness) and had a much better reading on the veracity of the claim. Why this was not done speaks volumes (to me anyway).

I find it hard to believe that his main supply and driver circuit would be soaking up so much power that it would mask the gain, and furthermore, if it required such a dissipative circuit in order to work, of what use is it?

Is anybody getting my drift?

Quote from: TinselKoala on 2016.08.05, 15:28:31
It appears to me that what partzman and ION are describing is a loss mechanism, dissipating power in the core and H-bridge components, rather than one which could produce any gain in energy.

Oh, where is MarkE when he's needed. I'm sure he would have some very interesting things to say about this.

TK,

As I understand Graham's input measurement points that is, at the transformer primary leads, he is not accounting for any losses in the H-bridge components nor the drive circuitry. Therefore, his apparent OU is measured between the transformer primary leads and the lamp load. Core losses are obviously going to be there but should be accounted for in his pin and pout measurements so I guess we are back to the question of his measurement accuracy.

If we do assume his measurements are reasonably accurate and as they indicate apparent OU, the question of where the gain is coming from is obviously very important. I certainly don't have the answer!

pm

Quote from: TinselKoala on 2016.08.05, 17:16:21
Here's a Tek pdf primer on Probe Skew.

Did Gunderson mention corrections for probe skew in his presentation?

Not that I recall but I'll have to watch the video again to be absolutely positive.

pm

See my post #280. Sorry I was writing while you were posting, pm.

Quote from: ION on 2016.08.05, 17:44:59
Thanks TK this is a "must read" for serious power analysts.

Where common sense cuts to the core:

All probe skewing and Clark-Hess glitches aside, I still would have liked to see a $20 "Kill-A-Watt" meter on the AC mains side of his device. With a claimed 5x or more gain in power, surely we could have looked at mains Watts in and Watts out (even guestimating the automobile lamp wattage by brightness) and had a much better reading on the veracity of the claim.

I find it hard to believe that his main supply and driver circuit would be soaking up so much power that it would mask the gain, and furthermore, if it required such a dissipative circuit in order to work, of what use is it?

Is anybody getting my drift?

ION,

I think so. C.C

IMO Graham's device is consuming ~500 watts reactive input power (Graham puts the figure at ~1200 watts) to produce ~8 watts real output. This is a ridiculous ratio but is obviously required for his device to work and I agree "what use is it".  At this point in time it has no use at all other than to prove he has a working device that produces real power from reactive power with his measurements indicating OU. With the high reactive power on the input and assuming the H bridge will operate at 98% efficiency, we would still be OU<1 and we are still ignoring the power required for any of the drive circuits.

So why pursue it?  From my viewpoint I find great value in G's work after working my own MEI circuits as he has accomplished the same results at a much lower operating frequency but with some drawbacks ie complexity, high ratio of VAR/VA, and relatively low power output. These all can be improved upon given time and effort IMO.

Now for one caveat I would like to mention regarding the high input reactive power ratio to the output real power. The higher this ratio, the higher the possibility for measurement error as Smudge, TK, and possibly others have pointed out regarding phase error. In Graham's device this ratio is 40-100:1 which honestly is a bother to me. Also, I would have preferred a non-inductive resistive load verses an incandescent lamp, all leads as short as possible, and although I am a fan of Tek current probes, CSRs would be better. In spite of all this, I will still attempt to replicate the device with my own interpretation.

pm

Gunderson is certainly a good presenter and seems to be a very likeable fellow. Congratulations for that; he's head-and-shoulders above any other presenter I've seen from the EF- endorsed crowd. My impression is that most of what he says is right-on correct, and the rest is plausible, even if it is based on the assumption that the measurements are not mistaken and so are in need of an explanation.

However... he has said that he has often encountered situations where one of his instruments tells him that he is getting an overunity COP but when the other instrument is put into play he sees disagreement. It seems to me that now he may have made a system where he is able to fool both sets of instruments at the same time, but not necessarily by the same manner.

The issue of probe skew is very important. He did make a point of applying one kind of correction to the scope and probes (degaussing and setting zero-points, and temperature compensation) but this is not compensating for probe skew, which is a timing (phase) issue.

Can anyone tell me the part numbers of the differential voltage probes and the current probes he used, and also the model number of the Tek scope? Apologies if that information has been covered somewhere else already, I just don't recall seeing it.

Quote from: TinselKoala on 2016.08.05, 18:48:12
Can anyone tell me the part numbers of the differential voltage probes and the current probes he used, and also the model number of the Tek scope? Apologies if that information has been covered somewhere else already, I just don't recall seeing it.

Dear TK,

Will these photos help?

The differential voltage probe is a Tektronix P5205 100 MHz model.

The current probe is also (I believe) a Tektronix device, but I cant see the model number. Perhaps there is enough here that you can cross it with a visit to an old Tek catalog. I know it is a hall chip device.

Spokane1

Quote from: partzman on 2016.08.05, 18:41:08
ION,

I think so. C.C

IMO Graham's device is consuming ~500 watts reactive input power (Graham puts the figure at ~1200 watts) to produce ~8 watts real output. This is a ridiculous ratio but is obviously required for his device to work and I agree "what use is it".  At this point in time it has no use at all other than to prove he has a working device that produces real power from reactive power with his measurements indicating OU. With the high reactive power on the input and assuming the H bridge will operate at 98% efficiency, we would still be OU<1 and we are still ignoring the power required for any of the drive circuits.

So why pursue it?  From my viewpoint I find great value in G's work after working my own MEI circuits as he has accomplished the same results at a much lower operating frequency but with some drawbacks ie complexity, high ratio of VAR/VA, and relatively low power output. These all can be improved upon given time and effort IMO.

Now for one caveat I would like to mention regarding the high input reactive power ratio to the output real power. The higher this ratio, the higher the possibility for measurement error as Smudge, TK, and possibly others have pointed out regarding phase error. In Graham's device this ratio is 40-100:1 which honestly is a bother to me. Also, I would have preferred a non-inductive resistive load verses an incandescent lamp, all leads as short as possible, and although I am a fan of Tek current probes, CSRs would be better. In spite of all this, I will still attempt to replicate the device with my own interpretation.

pm

Yes, when the ratio of reactive to real power is so high, just a tiny phase measurement error would account for the real power delivered to the load.

Since no one is saying it I will: a Kill-A-Watt meter on the input showing 1200 Watts reactive and a PF of 0.03 or so would have pretty much "blown" the presentation.

I'm not saying this device should not be pursued, I just have deep reservations but will continue on.

I don't expect that a sim will show the OU factor, provided it exists as it would be very hard to simulate all the magnetic oddities in the device, which are still "unknowns".

As far as the presentation goes, Graham is addressing a very wide audience so has to spend large bulk of time bringing people up to speed on Fourier Analysis, harmonics etc. The tiny bit of real beef gets buried in the ramble, but I agree, he has far less technobabble than other presenters.

OK guys,

Listen up.  Those who have watched the video are fortunate.  Those who have not are just whistling Dixie and spinning their wheels.  Remember a good hamburger meal with fries and drinks cost almost as much as the price of the video, a good movie with popcorn and
drinks for two cost more.  .  I stand absolutely astonished that some would bitch about that $27.00 price of the video.

Remember the beautiful picture of that waveform showing the input voltage, current
and the output current waveform. It is most important to realize that the output current waveform is inverted in that scope shot! He inverted it to show that the phase relationship of that output current wave to the input current wave which allowed the output to push back on the input and push the input to and including pure reactive operation.  Hence, only the losses in the input circuit has to be replaced.  The closer to pure reactive operation the higher the COP.
 
It is somewhat >180 degrees out of phase with the input waveform,  That basic transformer induced waveform is NOT OU, it has an additional component that the synchronous rectifiers add to the basic waveform and resultant voltage at the caps.  That does make it OU and causes it to push back. That comes from the burst of pulses immediately following the short cut off pulse to the primary coil. As emitted, these burst waves cancel out as wired in the CT output coil. (theory, could be different, have to build to be sure) but if rectified in the double coil CT full wave circuit, then add to the final output it does work.  These pulses I suspect come from the core and those pulses along with the 2nd, 3rd etc harmonics that are emitted and added to the output to make it work via very high speed rectification including low RF frequencies (harmonics).  Again, the rectified DC is added to the non symmetrical AC waveform that is rectified in the Caps (a very old method of rectification). Hence you have two sources of energy going into the caps.  Remember he said that the output circuit is VERY weird and he had to draw the circuit 3-4 times to be sure of it and had originally decided not to discuss it but at the last min in the presentation discussed it just a little!  There  probably is a very narrow window on the saturation curve of the core material where this happens. I also suspect he will never show that circuit! He really never showed us any waveforms of how that board worked internally. 

The neatest thing is that the circuit works at all!  He even admitted it.  He said it horrifically inefficient! I believe him.  But he has shown us something that will work if we understand it correctly.  He, in the video shows NO attempt to deceive, BS, etc.  I fully believe he could have talked for hours if permitted.

This is just my partial understanding of how this beast works. I have just thrown out my ideas and observations, that's my theory.  Each person that has seen the video can and will add up
what they see.  Maybe someone will hit everything and make it work! 

Now, I'll shut up and get on to building as time goes by.   What the heck is that primary core made of exactly!?
Lots of work here to do.

Ben K4ZEP

Excuse my stupid question , but I'm trying to extract some basic information from this bunch of electronic details. If there is purely reactive circuit and in some place when current is in correct place in circuit we have magnetic field and if that magnetic field is very very quickly shattered into pulses and those pulses induce current quickly rectified and stored in capacitor then the original magnetic field is returned back to other parts of circuits almost untouched - would that describe the operation of this device ?

A horribly inefficient free energy device, open source   that only costs two million dollars?     And whose "OU" effect can be completely cancelled by a couple of degrees of phase error in measurements?    :D

It's not the cost of 27 dollars that upsets me, it's where the money is going and how, and by whom, it is promoted. Send 27 dollards.... oops, dollars ... directly to Gunderson and I'll have no objection. Make it 27 hundred dollars if you like,  as long as Err-on doesn't get a penny of it.


The ferrite core is supposed to be composed of two parts of different material, one of low permeability and one of high permeability. From the description I think a square-loop material with a sharp "knee" on the B-H curve would probably fit the bill for the high permeability part.

(ETA:  Maybe even just the one, high-perm material could be used, with part of it saturated or partially saturated by the permanent magnets to result in a lower effective permeability in that part. )


In none of the various descriptions of how the device might work, do I find anywhere that some extra energy could be entering the system or created within it.

Quote from: Spokane 1 on 2016.08.05, 19:29:06
Dear TK,

Will these photos help?

The differential voltage probe is a Tektronix P5205 100 MHz model.

The current probe is also (I believe) a Tektronix device, but I cant see the model number. Perhaps there is enough here that you can cross it with a visit to an old Tek catalog. I know it is a hall chip device.

Spokane1

Yes, thanks, I saw the diff V probe photo and that helps. The P5205 has a typical propagation delay of 17 ns, which should probably cause a negligible phase shift at the more or less 12-20 microsecond per period range of the input signal waveforms.

Now if I had some information on the current probes (no, I can't ID it from the photos, except to note that it is black) and the scope model number itself I could tell you more.

TK said:

QuoteIn none of the various descriptions of how the device might work, do I find anywhere that some extra energy could be entering the system or created within it.

Smudge has attempted a possible explanation in a couple of his papers posted earlier.

Forest asked:

QuoteExcuse my stupid question , but I'm trying to extract some basic information from this bunch of electronic details. If there is purely reactive circuit and in some place when current is in correct place in circuit we have magnetic field and if that magnetic field is very very quickly shattered into pulses and those pulses induce current quickly rectified and stored in capacitor then the original magnetic field is returned back to other parts of circuits almost untouched - would that describe the operation of this device ?

Ben, Spokane1 and partzman have all pretty much broken down and nailed the basic operation of the device.
Ben and Spokane1 also have a lot of good dialogue on this at the EForum.

In the real world there is no purely reactive circuit, all will have a tiny bit of resistance if carefully constructed, even capacitors will have some dissipation factor. All losses can, however, be taken into account.

"almost untouched" is the problem, it's like being a little bit pregnant.

Many years ago I built a circuit that would charge a large inductor off the AC line up to 90 deg, then quit and dump the stored energy into a capacitor, repeat for the first 90 deg. of the negative half cycle. (the circuit worked inside of a bridge rectifier, so the output polarity was always in the same direction) My $20 Kill A Watt meter showed me that it didn't work as planned. Although the circuit was reactive, the power I thought was freely gotten showed up as an expense on that little meter.

Hopefully GG has found a more exotic method that actually works.

mmm..I wonder if a 60Hz resonant version of his device could be built, being charged directly from the mains? It would definitely require large low resistance inductors.

Quote from: ION on 2016.08.06, 12:01:00
Smudge has attempted a possible explanation in a couple of his papers posted earlier.

And her's another thought.  It seems that the high perm core (or parts of it) are magnetically biased at the saturation knee by appropriately placed PM's.  Now it is known that the BH curve for highly non linear core material has a significant difference between the usual input energy density (area to the left of the curve) and the so called co-energy density (area to the right of the curve).  The input energy is mathematically the integral of H with respect to B while the co-energy is the integral of B with respect to H. Aspden made much of the difference, using P to represent one energy and Q to represent the other, then used the remark "mind your P's and Q's" as a humurous way to emphasise the difference.  So maybe it is possible to supply a small amount of input energy then extract a larger amount of co-energy.  That leads to the question "how can you extract co-energy"?  If you have an air gap that is mechanically closed while the flux though it remains constant the mechanical energy you extract is that co-energy.  The flux is constant and only the mmf drop across the reducing reluctance is changing, and that energy is given by int(B.dh) which with constant B is simply B*deltaH.  That is not an electrical output because there is no voltage.

So how might this apply to Gunderson's transformer?  Well there does seem to be an air gap, or rather a gap filled with a soft plastic.  So maybe he has hit upon a mechanical resonance where that plastic gets repeatedly compressed and decompressed.  The compression represents co-energy stored ready to be fed back into the system where it can then be reclaimed as electrical energy.   Just something to think about.

Smudge   

It's still energy that you have to put into the device in the first place, in order for it to be extracted and used.


Here's an example of why I find this upsetting. Look at the way this is being promoted on EF:

QuoteThis year, Graham Gunderson showed the Magnetic Implosion Transformer that he invented and it measured out at 570% more output than the input he paid for!

(emphasis in the original)

But that isn't true at all, regardless of the measurement validity and accuracy issues that we have been discussing here. The "input he paid for" or rather that somebody has to pay for, is the input power from the mains, to all the different power supplies used to make the apparatus work. This is very different from a claim that the transformer itself is "OU" due to inputs measured directly to the transformer.

Again,

QuoteAs a matter of fact, he can dial the input down until it goes negative, which means the COP (coefficient of performance) is infinite! Coefficient is the ratio between desired work done in a system compared to the work that we have to pay for.

(emphasis mine)

http://www.energeticforum.com/291117-post1.html

Quote from: Smudge on 2016.08.06, 13:45:10
And her's another thought.  It seems that the high perm core (or parts of it) are magnetically biased at the saturation knee by appropriately placed PM's.  Now it is known that the BH curve for highly non linear core material has a significant difference between the usual input energy density (area to the left of the curve) and the so called co-energy density (area to the right of the curve).  The input energy is mathematically the integral of H with respect to B while the co-energy is the integral of B with respect to H. Aspden made much of the difference, using P to represent one energy and Q to represent the other, then used the remark "mind your P's and Q's" as a humurous way to emphasise the difference.  So maybe it is possible to supply a small amount of input energy then extract a larger amount of co-energy.  That leads to the question "how can you extract co-energy"?  If you have an air gap that is mechanically closed while the flux though it remains constant the mechanical energy you extract is that co-energy.  The flux is constant and only the mmf drop across the reducing reluctance is changing, and that energy is given by int(B.dh) which with constant B is simply B*deltaH.  That is not an electrical output because there is no voltage.

So how might this apply to Gunderson's transformer?  Well there does seem to be an air gap, or rather a gap filled with a soft plastic.  So maybe he has hit upon a mechanical resonance where that plastic gets repeatedly compressed and decompressed.  The compression represents co-energy stored ready to be fed back into the system where it can then be reclaimed as electrical energy.   Just something to think about.

Smudge

I agree and would again point to the use of a magnet biasing a small toroid core by SM (first TPU) and the resultant vibration and gyroscopic properties of the device windings being attributed to an electro-acoustic phenomena.

Verpies CARA project..

http://www.overunityresearch.com/index.php?topic=2751.0

;)

I am glad someone remembers.

I wish I could send Mr.Gunderson a suggestion to verify whether the OU effects disappears when the cores are immobilized and prevented from moving/vibrating in respect to each other.  A simple experiment like that would verify if the relative motion of the cores is essential to this OU effect.

If the output waveform contained high and narrow spikes, then I would suspect McFreey's Modus Operandi.

Without accurate schematics, nice scopeshots, probe positions and proof of calibration, the ORBO MO remains a possibility, too ;)

Below is data from an LtSpice simulation representing my interpretation of Graham's MIT.  It is greatly simplified using series resonance to drive the primary and uses only one synchronized secondary. The transformer consisting of L2 for the primary and L5 for the secondary is operating in a linear mode with no PM bias at an operating frequency of 33.33kHz.

I chose to use a coupling factor of K=.9 as a best guess for the transformer and the turns ratio is 1:1 with a dcr=.2 ohms for each winding.

The Infineon mosfets are not ideal as they have a rather large gate capacitance but are otherwise fast with low Rdson. The gate capacitance does not matter here as we are driving the gates with ideal generators.

The sim was run for 50ms and the stabilization profile is shown in the first pix.  The settling oscillation is created mostly by C1 in the secondary.

The second pix is the schematic with a 20 cycle plot starting at 48ms showing the average reactive power of Ecap x IL2 for an input of .561 watts.

The third pix is the same plot showing the dc output voltage of 9.27v across C1 and the 50 ohm load which equates to 1.72 watts. The resulting COP is 3.06.

The fourth pix is a zoom of the same plot to more clearly see the waveform relationships and magnitudes.  Although not as pronounced as G's the output current waveform does exhibit a slight flux and current reversal. This reversal is mirrored in the primary because this sim does not utilize different core materials, gaps, or any B/H curve bias.

Although not shown, the input reactive peak power is 105 watts and the input current thru L2 is .169 amps resulting in a power draw from the 20vdc supply of 3.38 watts. The VAR/VA ratio is 105/1.72 = 61.

The next step is a little bench testing to arrive at a more accurate K factor for the transformer and then attempt a non-linear gyrator based simulation.

IMO, G's concept can be built to operate at lower frequencies with more efficiency depending on available cores and materials which brings us closer to a working OU black box!

pm



 

Very nice work, pm. I had my doubts, but you seemed to have demonstrated that a sim can show OU.

I can eat my hat now, got any salt?

It would be neat to try a sim of a 60 Hz. version. That might simplify some of the power supply circuitry and might also reduce switching losses.

Regards, ION

Nice work PM.

Would you mind posting your sim file?

Quote from: ION on 2016.08.06, 17:45:54
Very nice work, pm. I had my doubts, but you seemed to have demonstrated that a sim can show OU.

I can eat my hat now, got any salt?

Regards, ION

I think Pin is being measured across the input fet/diode/cap, not the primary. How much difference it will make I won't know until we see Pin done with the voltage across the primary.

Quote from: poynt99 on 2016.08.06, 17:57:06
Nice work PM.

Would you mind posting your sim file?

No, find it attached below as a standard text file.

pm

Partzman's sim should be dead easy to loop. One diode to decouple the voltage source and another diode from the output to the input.

You could try with ideal "lossless" diode models, but with a COP of >3 it should work with any diodes.

Alternately a momentary switch could turn of and decouple the voltage source allowing the system to go into runaway.

You could also adjust the turns ratio slightly to allow higher secondary voltage to overcome diode drops.

Thank you ION and Poynt for your comments!

Inconsideration of Poynt's comment about measuring the power using the voltage across the primary rather than C4, I've attached a plot with the input reactive power taken with the 20 vdc supply subtracted from the voltage across L2.

pm 

Quote from: ION on 2016.08.06, 19:10:14
Partzman's sim should be dead easy to loop. One diode to decouple the voltage source and another diode from the output to the input.

You could try with ideal "lossless" diode models, but with a COP of >3 it should work with any diodes.

Alternately a momentary switch could turn of and decouple the voltage source allowing the system to go into runaway.

You could also adjust the turns ratio slightly to allow higher secondary voltage to overcome diode drops.

Well, not quite yet ION but I wish. O0 Our dc input power is 3.37w with our output power at 1.72w so our dc COP is .51. The COP >3 is real/reactive so we have some innovating and improving to do before we can loop dc to dc IMO.

pm

Quote from: partzman on 2016.08.06, 19:36:13
Well, not quite yet ION but I wish. O0 Our dc input power is 3.37w with our output power at 1.72w so our dc COP is .51. The COP >3 is real/reactive so we have some innovating and improving to do before we can loop dc to dc IMO.

pm

Hi PM,

Enjoying your work with the simulator.  Excellent work, It adds ideas for me too.  Keep it up!!!
I'm going to have to download that simulator and get up to speed on it!  I see it's free!

Ben K4ZEP

Quote from: K4zep on 2016.08.06, 21:52:11
Hi PM,

Enjoying your work with the simulator.  Excellent work, It adds ideas for me too.  Keep it up!!!
I'm going to have to download that simulator and get up to speed on it!  I see it's free!

Ben K4ZEP

Thanks Ben. The version I used for the sim I posted was LtSpice IV but you might try the new version XVII as it is available in 32 and 64 bit OS but I think it has problems running on XP however.

pm

Quote from: partzman on 2016.08.06, 19:06:19
No, find it attached below as a standard text file.

pm

Thanks. btw the forum should allow .asc file type attachments.

Quote from: partzman on 2016.08.06, 19:22:09
Thank you ION and Poynt for your comments!

Inconsideration of Poynt's comment about measuring the power using the voltage across the primary rather than C4, I've attached a plot with the input reactive power taken with the 20 vdc supply subtracted from the voltage across L2.

pm

Is that 2.81W? If so, there goes the transformer OU.  :(

Quote from: partzman on 2016.08.06, 23:23:58
Thanks Ben. The version I used for the sim I posted was LtSpice IV but you might try the new version XVII as it is available in 32 and 64 bit OS but I think it has problems running on XP however.

pm

Thanks PM,

I downloaded MAC version to my big MAC, can't get it to run.  Will download it to an old laptop running Win 8, work on it from there.
Something new to learn how to use!  A most useful tool I think. 

Ben K4ZEP

Quote from: partzman on 2016.08.05, 17:46:32
TK,

As I understand Graham's input measurement points that is, at the transformer primary leads, he is not accounting for any losses in the H-bridge components nor the drive circuitry. Therefore, his apparent OU is measured between the transformer primary leads and the lamp load. Core losses are obviously going to be there but should be accounted for in his pin and pout measurements so I guess we are back to the question of his measurement accuracy.



pm

QuoteIf we do assume his measurements are reasonably accurate and as they indicate apparent OU, the question of where the gain is coming from is obviously very important. I certainly don't have the answer!

Only the very closed minded will have trouble working out where the gain is coming from--those that are apposed to PMs doing useful work.



Brad

Quote from: poynt99 on 2016.08.06, 23:46:35
Is that 2.81W? If so, there goes the transformer OU.  :(

With the cursor in the box it is hard to tell that it is -2.81w so I've attached a plot view taken in LtSpice XVII that is clearer to read.

For some reason when I tried to attach the *.asc file directly to the post it refused. I think I have done this before so I dunno! ???

pm

Yeah, I figured it was -2.81W.

But no big deal it being negative (it all depends on probe polarity or how/where you subtract the 20V). The fact is this accounts for the missing power and indicates the true Pin for the transformer.

Sorry, but the sim is not showing OU on the transformer, and thankfully, ION won't have to eat his hat!

Quote from: K4zep on 2016.08.07, 01:21:58
Thanks PM,

I downloaded MAC version to my big MAC, can't get it to run.  Will download it to an old laptop running Win 8, work on it from there.
Something new to learn how to use!  A most useful tool I think. 

Ben K4ZEP

Ben,

There is a good LtSpice group on Yahoo that is free to join and they have a lot of Mac users that have worked out the various problems one might encounter. Loads of files on circuits, models, and explanations on the help files which can be rather difficult to understand at times especially when just starting.

Good luck. O0

pm

Quote from: TinMan on 2016.08.07, 12:38:12
Only the very closed minded will have trouble working out where the gain is coming from--those that are apposed to PMs doing useful work.



Brad

Brad,

I believe that it is possible to extract energy from PMs. I would reference Smudge's theories on electron spin and possibly neutron spin for example for those who might question. For G's device, it might be nuclear spin interaction or it could be some type of flux or magnetic vector potential interaction or something else. I hope the answer comes to light!

pm

Quote from: partzman on 2016.08.07, 13:29:35
Brad,

I believe that it is possible to extract energy from PMs. I would reference Smudge's theories on electron spin and possibly neutron spin for example for those who might question. For G's device, it might be nuclear spin interaction or it could be some type of flux or magnetic vector potential interaction or something else. I hope the answer comes to light!

pm

I believe all that is needed is to make the linear force of a PM -non linear,and it would seem that that is what the outcome would be in GGs transformer.

Also pm,do you believe that all that fancy switching is really needed in GGs setup,or is it a rouse to complicate things beyond replication?.

I believe that my FG is able to create the input voltage wave form that GG shows,simply by using the 1 shot mode,and determining the rest phase between each shot relative to the frequency required. This signal could then be used to drive an audio amp. Then all we need is to work on the transformer and circuit to produce the input current wave form.

Looking at GGs scope shot's,it would appear that he is measuring the output voltage across the load,and the current into the smoothing cap,after the mosfets used as the rectifier ?-im not sure,this may have already been discussed,and i have not been following that closely,due to other commitments.


Brad

Quote from: poynt99 on 2016.08.07, 13:13:44
Yeah, I figured it was -2.81W.

But no big deal it being negative (it all depends on probe polarity or how/where you subtract the 20V). The fact is this accounts for the missing power and indicates the true Pin for the transformer.

Sorry, but the sim is not showing OU on the transformer, and thankfully, ION won't have to eat his hat!

Poynt,

I have to respectively disagree so I'll give my logical explanation for this issue.

I've attached a plot that shows the Ecap voltage minus 20 volts for graphical reference. It is true that the Ecap voltage is the voltage across C4 but it also is the voltage across the primary L2 that is displaced by +20v above our reference point. Therefore logically it would seem that we should subtract 20 volts from the Ecap voltage to arrive at the actual voltage across L2 in reference to ground. The current and voltage polarities remain the same.

It is interesting to me that since the V8 supply has zero impedance at the op frequency, the equivalent circuit of L2 and C4 is a parallel resonant configuration instead of series as I originally described.

pm

Quote from: TinMan on 2016.08.07, 13:50:20
I believe all that is needed is to make the linear force of a PM -non linear,and it would seem that that is what the outcome would be in GGs transformer.

TM,

It would appear so.

Quote
Also pm,do you believe that all that fancy switching is really needed in GGs setup,or is it a rouse to complicate things beyond replication?.

With my understanding, Graham had little time to whip the device together before the seminar as is evidenced by the hacked circuitry. I also think he immediately began to re-work the device after the seminar (according to those close to him that have posted here) because he realized it is not in it's simplest form yet. So no, I'm not convinced the device needs to be as complicated as his original disclosure but I don't have one operating on my bench either. :-\

Quote
I believe that my FG is able to create the input voltage wave form that GG shows,simply by using the 1 shot mode,and determining the rest phase between each shot relative to the frequency required. This signal could then be used to drive an audio amp. Then all we need is to work on the transformer and circuit to produce the input current wave form.

With a generator that versatile your idea sure sounds doable. I wish there was more info on those cores!

Quote
Looking at GGs scope shot's,it would appear that he is measuring the output voltage across the load,and the current into the smoothing cap,after the mosfets used as the rectifier ?-im not sure,this may have already been discussed,and i have not been following that closely,due to other commitments.

Brad

I know the current waveform shows the secondary current but exactly where he has placed the probe I'm not certain!

pm

Quote from: partzman on 2016.08.07, 13:29:35
Brad,

I believe that it is possible to extract energy from PMs. I would reference Smudge's theories on electron spin and possibly neutron spin for example for those who might question. For G's device, it might be nuclear spin interaction or it could be some type of flux or magnetic vector potential interaction or something else. I hope the answer comes to light!

pm

So do either of you believe that it is possible to extract _more_ energy from a PM than was applied to the base material in order to magnetize it in the first place?

What happens to the magnet when you extract energy from it? Does it become less and less magnetic? If you extract energy from "nuclear spin interaction" does the material change properties, become more brittle, evaporate, emit radiation, or what? Is it channelling dark energy from the expansion of the cosmos?

Quote from: TinMan on 2016.08.07, 13:50:20
Looking at GGs scope shot's,it would appear that he is measuring the output voltage across the load,and the current into the smoothing cap,after the mosfets used as the rectifier ?-im not sure,this may have already been discussed,and i have not been following that closely,due to other commitments.
Brad

@tinman

If we all agree that @Spokane1's initial diagram is correct, then his output measurement is surely being juiced up by the logic controller that is bleeding into the output. Otherwise explain the dc flat line output while you see an h-bridge input of 1/3 pulse, 1/3 reverse pulse and a 1/3 pause. Not even a glitch can be seen on the output. My spider senses tell me GG realized this error but decided instead of owning up to it, he just mothballed the device and now you guys and others are left in the cold. Classic pattern of bait and switch. This cannot be a learning experience but another sleuthing series.

If there is reactive power involved then any lone diodes or diodes in mosfets would easily bleed backwards.

As for a $27 demo price, I don't pay for demos. That would be tantamount to Georgian TKism. If the demo is prime, I do give donations and $27 would be chump change. It's the principle. I never buy what I do not know I will get.

So I am on the sidelines on this one. I have great theories of operation according to spin conveyance, what he is doing wrong, what is right, what can weak magnets do to a core already moving good watts (nothing), but all this is pointless without the truth of the matter and that is only available from GG who I feel is not forthcoming. So again, for me such a deal is a bad deal. Better just spend time on other matters. Guys should stop weighing a pound of nothing. If the effort is purely a means to keep the knife sharpened, then it is a good FUN exercise otherwise thinking these outside efforts will replicate his device is not a smart move. Not until we know exactly the situation from GG himself. And I guess, as usual, that will never happen.

You do not need to replicate anything. You actually just need to understand the principle and once that is understood it can be tried and tested in small scale on any bench. But first, what is the principle? He even said it himself, he does not know why it works. That is the most ludicrous statement anyone can make that has obviously spent countless days and nights, probably has a box full of notes, knows 2000 times more about it then anyone else could to say such a thing. You can at least say "I THINK IT IS BASED ON................". Only someone who wants to mothball the problem would say what he said. So again, you can play games but you cannot hide human nature. That's pasted on our faces, our words and our choices of action. Mothball then ask to 2M$, then declare OU as being concrete to sell a demo video, all these words and actions do not spell true OU intent.

wattsup

Quote from: TinselKoala on 2016.08.07, 15:11:04
So do either of you believe that it is possible to extract _more_ energy from a PM than was applied to the base material in order to magnetize it in the first place?

What happens to the magnet when you extract energy from it? Does it become less and less magnetic? If you extract energy from "nuclear spin interaction" does the material change properties, become more brittle, evaporate, emit radiation, or what? Is it channelling dark energy from the expansion of the cosmos?

TK,

Good questions. Many, many moons ago I ran a bench experiment based on one of Smudge's papers using PMs in a coil arrangement and there appeared to be a slight gain. The problem is I didn't document very well at the time and I can't even remember what it was about. Could have been measurement error as well!

I was also the one who experienced OU in a MEG configuration way back when using MOVs for loads . Placement of PMs in the core did increase the output power and the measurements were taken on a Voltech PM3000 3-phase power analyzer which provided the means of measuring both input and output power simultaneously. The secondaries finally broke down with internal arcing and being they were so tedious to wind on the U cores, I quit the project. 

I would recommend Stefan Marinov's "Thorny Way of Truth Part3" which is available free online and read about his experiments using PMs with analysis using vector potentials. His devices and others he references are OU per their claims.

Basically, I can't give you any satisfying answers to your questions regarding energy extraction from PMs, but I still maintain the idea that it is possible.

pm

Quote from: partzman on 2016.08.07, 13:54:24
Poynt,

I have to respectively disagree so I'll give my logical explanation for this issue.

I don't understand your point/argument, sorry.

Respectfully, I have to say I stand by my statement.

A few salient points from "Power Measurement 101":

1) The power of a source, or a component (or device as a whole) is measured by taking the voltage directly across it and the current into (through) it and multiplying the two for power as a product. For scope type measurements, the average (MEAN) is taken of a minimum of 10 cycles of operation of this product. (There is obviously a bit more detail here I've not included for clarity).

2) The polarity of "power" when correctly measured for a source is negative. The polarity of "power" for all loads connected to a source are positive. Most often folks ignore the polarity however.

3) As a consequence of 2), power is conservative. Just as there is "KVL", there is also an implied "KPL", whereby all the "powers" in a circuit and source sum to zero.

Let's take your sim for example:

a) Power measured from the 20V source is 3.38W. Correct.

b) Power measured in the 50 Ohm load is 1.72W. Correct.

c) Power measured across MOSFET/DIODE/CAP is 0.561W. This is incorrectly associated with input power to the transformer. As per 1) above, it is not.

d) Power measured/computed when 20V is subtracted from the voltage measurement in c), is 2.81W. This is the actual input power measurement as per 1) above.

Surely you do not believe that by strange coincidence, 3.38W=0.561W + 2.81W ? I can assure you that it is not a coincidence. This is simply an illustration of the power balance in every circuit (in this case, between the 20V source, and the input side of the transformer).

What about the 1.72W measured in the 50 Ohm load? Since we know that 2.81W is what goes into the primary (and comes out the secondary), we know that the remaining 1.09W (2.81W-1.72W) must be getting dissipated in the output MOSFET.

Attached is a diagram illustrating the above points. Note, L5 is actually a source on the output side of the transformer, so its power is negative at -2.81W. There are two power loops, one on the input and one on the output side, and each sums to zero as per 3) above.

Quote from: partzman on 2016.08.07, 13:17:32
Ben,

There is a good LtSpice group on Yahoo that is free to join and they have a lot of Mac users that have worked out the various problems one might encounter. Loads of files on circuits, models, and explanations on the help files which can be rather difficult to understand at times especially when just starting.

Good luck. O0

pm

Thanks PM,

Got it to run on the MAC.  I'll look up that group on Yahoo!  It really is a good tool.
Thanks for the info!!!  Files, circuits, etc.  Good to load up on!! :) 

All the yea's and nay's on this group now!  Only way to know is to build, experiment,
come to our own conclusions. 

Ben K4ZEP

Quote from: poynt99 on 2016.08.07, 16:17:43
I don't understand your point/argument, sorry.

Respectfully, I have to say I stand by my statement.

A few salient points from "Power Measurement 101":

1) The power of a source, or a component (or device as a whole) is measured by taking the voltage directly across it and the current into (through) it and multiplying the two for power as a product. For scope type measurements, the average (MEAN) is taken of a minimum of 10 cycles of operation of this product. (There is obviously a bit more detail here I've not included for clarity).

2) The polarity of "power" when correctly measured for a source is negative. The polarity of "power" for all loads connected to a source are positive. Most often folks ignore the polarity however.

3) As a consequence of 2), power is conservative. Just as there is "KVL", there is also an implied "KPL", whereby all the "powers" in a circuit and source sum to zero.

Let's take your sim for example:

a) Power measured from the 20V source is 3.38W. Correct.

b) Power measured in the 50 Ohm load is 1.72W. Correct.

c) Power measured across MOSFET/DIODE/CAP is 0.561W. This is incorrectly associated with input power to the transformer. As per 1) above, it is not.

d) Power measured/computed when 20V is subtracted from the voltage measurement in c), is 2.81W. This is the actual input power measurement as per 1) above.

Surely you do not believe that by strange coincidence, 3.38W=0.561W + 2.81W ? I can assure you that it is not a coincidence. This is simply an illustration of the power balance in every circuit (in this case, between the 20V source, and the input side of the transformer).

What about the 1.72W measured in the 50 Ohm load? Since we know that 2.81W is what goes into the primary (and comes out the secondary), we know that the remaining 1.09W (2.81W-1.72W) must be getting dissipated in the output MOSFET.

Attached is a diagram illustrating the above points. Note, L5 is actually a source on the output side of the transformer, so its power is negative at -2.81W. There are two power loops, one on the input and one on the output side, and each sums to zero as per 3) above.

Poynt,

All you had to say was "PM, you need to reverse or invert the current thru L2 because the reference point is inverted" without all the condescending verbiage. Yes, you are correct in that no OU is present and I am sure with that you are pleased.

pm 

Quote from: TinselKoala on 2016.08.07, 15:11:04
So do either of you believe that it is possible to extract _more_ energy from a PM than was applied to the base material in order to magnetize it in the first place?

What happens to the magnet when you extract energy from it? Does it become less and less magnetic? If you extract energy from "nuclear spin interaction" does the material change properties, become more brittle, evaporate, emit radiation, or what? Is it channelling dark energy from the expansion of the cosmos?

Your rolling ball experiment should answer that for you TK.

Dose the magnet on a flywheel of a small engine loose it's stored energy to the ignition coil after millions of cycles ?.
What about the PMs in a PM DC motor--do they loose there stored energy to the rotor after millions of revolutions ?.


Brad

Quote from: TinMan on 2016.08.07, 22:44:32
Your rolling ball experiment should answer that for you TK.

Dose the magnet on a flywheel of a small engine loose it's stored energy to the ignition coil after millions of cycles ?.

No, because it isn't doing any work. Say the magnet is closing and opening a small reed switch on every cycle. The work to do this is coming from the flywheel, not the magnet. Replace the magnet with a cam and replace the reed switch with a microswitch that is hit by the cam on every cycle. Does the cam lose energy, or is the work coming from the flywheel? If the flywheel isn't powered, will it eventually slow down? If you break rocks with a hammer, is the hammer doing work... or are you just using the hammer to _transmit_ the work that you are doing?

Quote
What about the PMs in a PM DC motor--do they loose there stored energy to the rotor after millions of revolutions ?.

No, because they aren't doing any work. The work to drive the rotor comes from the electrical power applied to the coils.

Quote from: partzman on 2016.08.07, 22:30:24
Poynt,

All you had to say was "PM, you need to reverse or invert the current thru L2 because the reference point is inverted" without all the condescending verbiage. Yes, you are correct in that no OU is present and I am sure with that you are pleased.

pm

Partzman,

In my opinion, my posts below were all that needed to be said, but apparently they were dismissed without due consideration. As such, and since it appeared you were still convinced of your original assertion, I ventured to explain it in greater detail (much to my chagrin).

Am I happy about it, not really. Do I have tact, not really, especially after I've already made my point clear enough. :)

Quote from: poynt99 on 2016.08.06, 17:58:38
I think Pin is being measured across the input fet/diode/cap, not the primary. How much difference it will make I won't know until we see Pin done with the voltage across the primary.

Quote from: poynt99 on 2016.08.07, 13:13:44
Yeah, I figured it was -2.81W.

But no big deal it being negative (it all depends on probe polarity or how/where you subtract the 20V). The fact is this accounts for the missing power and indicates the true Pin for the transformer.

Sorry, but the sim is not showing OU on the transformer, and thankfully, ION won't have to eat his hat!

Quote from: TinselKoala on 2016.08.08, 00:18:27
No, because it isn't doing any work. Say the magnet is closing and opening a small reed switch on every cycle. The work to do this is coming from the flywheel, not the magnet. Replace the magnet with a cam and replace the reed switch with a microswitch that is hit by the cam on every cycle. Does the cam lose energy, or is the work coming from the flywheel? If the flywheel isn't powered, will it eventually slow down? If you break rocks with a hammer, is the hammer doing work... or are you just using the hammer to _transmit_ the work that you are doing? 
No, because they aren't doing any work. The work to drive the rotor comes from the electrical power applied to the coils.

So replace the magnet on the flywheel with a block of steel,and see how your motor runs--with there be spark-will the block of steel produce an electromagnetic field?

In regards to the DC motor-the same applies. Replace your PMs with steel keepers,and see how well your motor runs. You will find that it consumes a lot of power,and has very little torque--but why?. Well now you do not have the generating effect of the PMs to reduce the current flowing through the rotor coils--the BackEMF just is not there.

In time,i will be carrying out some calorimetry tests,and we shall see if PMs can actually do useful work  O0


Brad

Quote from: TinselKoala on 2016.08.05, 15:28:31
It appears to me that what partzman and ION are describing is a loss mechanism, dissipating power in the core and H-bridge components, rather than one which could produce any gain in energy.

Oh, where is MarkE when he's needed. I'm sure he would have some very interesting things to say about this.

While MarkE was a great man,and was well versed in EE-->we have Poynt  O0

It would be a tough bet,but my money would have been on the man that still walks among us.


Brad

Ah..... It seems Brad beat me to it!

Should I open a topic? " Do permanent magnets DO work "

Consensus?

Cheers Grum.

Quote from: Grumage on 2016.08.08, 10:27:56
Ah..... It seems Brad beat me to it!

Should I open a topic? " Do permanent magnets DO work "

Consensus?

Cheers Grum.

I think it is time Grum,to answer this question once and for all.

For every action,there is an equal and opposite reaction--and then there is the counter reaction  ;)

Open the topic,and lets get ION on the scene ,as he knows his way around calorimetry setup's,and i believe this is the way to show PMs doing useful work.


Brad

Quote from: TinselKoala on 2016.08.08, 00:18:27
No, because it isn't doing any work. Say the magnet is closing and opening a small reed switch on every cycle. The work to do this is coming from the flywheel, not the magnet. Replace the magnet with a cam and replace the reed switch with a microswitch that is hit by the cam on every cycle. Does the cam lose energy, or is the work coming from the flywheel? If the flywheel isn't powered, will it eventually slow down? If you break rocks with a hammer, is the hammer doing work... or are you just using the hammer to _transmit_ the work that you are doing? 
No, because they aren't doing any work. The work to drive the rotor comes from the electrical power applied to the coils.

With a magnet on the flywheel passing a coil to generate current, the magnetic field of the magnet is always there, it is not the movement of the flywheel, the flywheel just brings that field into proximity of the coil for a time thus generating current in the coil. That is the flywheel did not create the energy, the magnetic field yes, the flywheel was expended energy only to move that field too and from the coil.

Another way, a magnet held above a piece of steel and the steel rises up to the magnet when close enough, what created the energy to rise the steel? the movement of the magnet closer to the steel or the movement itself? ;) Or is it the magnetic field?

A perminent magnet does create energy, the problem is to extract that energy you may have to expend more than you get back.

Regards

Mike 8)

Dear All,

Here is a drawing of the Synchronous Diode Component Layout to help facilitate discussions about this sub assembly.

Enjoy!

Spokane1

Has anyone looked at the gate current injection into the drain circuits of MOSFETs in this device?

Quote from: verpies on 2016.08.09, 21:49:35
Has anyone looked at the gate current injection into the drain circuits of MOSFETs in this device?

Is that a data sheet parameter?

Spokane1

Quote from: Spokane 1 on 2016.08.10, 00:48:58
Is that a data sheet parameter?

No, it is a particular circuit's parameter.

The current supplied to charge the MOSFET's gate can make its way into the drain circuit through the gate-drain Miller's capacitance and sometimes through the gate-source capacitance.  It is an often overlooked input power path in high frequency circuits that is capable of injecting Watts of power into the drain circuit.

Quote from: verpies on 2016.08.09, 21:49:35
Has anyone looked at the gate current injection into the drain circuits of MOSFETs in this device?

Not that I know of, but I already brought it up here (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488771/#msg488771) over 2 weeks ago when this all started.

Quote from: poynt99 on 2016.08.10, 02:44:17
Not that I know of, but I already brought it up here (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg488771/#msg488771) over 2 weeks ago when this all started.

It is a real concern.  See a simple calculation here (http://overunity.com/16724/graham-gundersons-energy-conference-presentation-most-impressive-and-mysterious/msg489570/#msg489570).
For example, most MOSFET RF power amplifiers leak the input energy into the output stage even when that output stage is unpowered !.

Quote from: verpies on 2016.08.10, 03:46:09
It is a real concern.
For example, most MOSFET RF power amplifiers leak the input energy into the output stage even when that output stage is unpowered !.

Agreed. I see this whenever the PA is blown. The exciter power gets to the output/antenna making the radio appear to still be working albeit with very limited range.

btw, in the video there is a short section (only about 6 minutes unfortunately) when Gundy talks about the power measurements. At one point he has the input voltage and current traces up. It is easy to tell by eye alone that the measurement makes sense, i.e. 2V x 2A / 0.33 = 1.33W, assuming the half-sine parts are perfectly 90 degrees and can be ignored. I'm just not convinced that the wave forms being presented to the scope and power analyzer are truly correct and not skewed by phase shift, especially when he says it can be tuned for negative power.

The output side is troubling as well. Although I have not analyzed it like you are going to do, I am suspicious of the MOSFET Gate bleeds and by the fact changing one d-g cap can apparently kill the "effect".

Anyway, looking forward to your analysis.

Quote from: poynt99 on 2016.08.10, 12:29:26
Agreed. I see this whenever the PA is blown. The exciter power gets to the output/antenna making the radio appear to still be working albeit with very limited range.

Just a week ago that happened to me with this RF PA (http://www.ebay.com/itm/1-2-KW-LDMOS-power-amplifier-1-8-54MHz-HF-for-BLF188XR-/152093756341).

It was excited with 3W at its input and after the main 50V power supply went up in smoke, 1.8W was still leaking out of its output.

I could not find anything listed under the mosfet part number on Spokane1's diagram (C2M002510D).

However, looking at the photos I cannot see a full part number but I can tell that the part number starts with C2M and ends with 120.

Cree and DigiKey list the C2M0025120D mosfet, 1200V 90A 34mOhm Rdson, for only 69 dollars and 80 cents _each_. Data sheet attached below.

Free Energy is getting more and more expensive every time I look!   :D

Quote from: verpies on 2016.08.10, 14:19:58
Just a week ago that happened to me with this RF PA (http://www.ebay.com/itm/1-2-KW-LDMOS-power-amplifier-1-8-54MHz-HF-for-BLF188XR-/152093756341).

It was excited with 3W at its input and after the main 50V power supply went up in smoke, 1.8W was still leaking out of its output.

1.2kW?  :o

That ought to get you around the earth a few times.

Quote from: verpies on 2016.08.10, 14:19:58
Just a week ago that happened to me with this RF PA (http://www.ebay.com/itm/1-2-KW-LDMOS-power-amplifier-1-8-54MHz-HF-for-BLF188XR-/152093756341).

It was excited with 3W at its input and after the main 50V power supply went up in smoke, 1.8W was still leaking out of its output.

Ahhh TK, the best things in life are free!!!!! ;D

Ben K4ZEP

Quote from: poynt99 on 2016.08.10, 22:47:42
1.2kW?  :o

That ought to get you around the earth a few times.

For the life of me, I can't figure out what a blown line fed PA amp. and leakage through burnt circuits has to do with GG's design??? ;)
Anymore dust and dirt get thrown up in the air and we will forget what we are discussing on this thread!

Ben K4ZEP

It's about the energy used to drive GG's MOSFETs, which are connected to the output of his device, reaching the secondary winding of his transformer and being accounted there as this transformer's energy output.

If the RF energy leaks through MOSFETs' gates in RF amps, then it must be leaking through GG's MOSFETs' gates, as well.


The question is "How much" ?

Quote from: TinselKoala on 2016.08.10, 14:47:20
However, looking at the photos I cannot see a full part number but I can tell that the part number starts with C2M and ends with 120.

With some pseudo-color mapping. I could recognize ...25120 or ...26120 at the end of the number .

@Itsu

Could you put together a quick video showing the gate leakage of any N-Ch Power MOSFET in the circuit below ?
No power supply. Just program your FG to 1MHz square wave at 10V amplitude. Use a red LED (observe its polarity).

Quote from: verpies on 2016.08.11, 04:17:47
@Itsu

Could you put together a quick video showing the gate leakage of any N-Ch Power MOSFET in the circuit below ?
No power supply. Just program your FG to 1MHz square wave at 10V amplitude. Use a red LED (observe its polarity).

Beautiful demonstration!

I did it with  IRF530n and IRF3205 mosfets and a white LED, and tested a wide range of frequencies using a 0 to +9V pulse train of 50 percent duty cycle. 1 MHz produced a brilliant light. Down at 50 kHz the light was still fairly bright. Going over 1 Mhz and the light dims progressively. Also reducing the amplitude of the square wave causes the LED to go out somewhere between 4 and 5 Vp-p.
With 9V peak pulse train, the IRF3205 produced peak light at about 450 kHz. The IRF530n peaked at about 800 kHz.

No video from me, my camcorder battery is kaput.

Tinsel
Quote

No video from me, my camcorder battery is kaput.
-------

not good

we have to fix that...

please post a link for a replacement battery

Quote from: poynt99 on 2016.08.10, 22:47:42
1.2kW?  :o
That ought to get you around the earth a few times.

No, I'd need 1.21GW for that.
Actually, I was using only 1kW fed into a Litz coil for some nuclear resonance experiments and I hope that much RF never radiated out of the room.

Quote from: Chet K on 2016.08.11, 07:26:23
Tinsel
Quote

No video from me, my camcorder battery is kaput.
-------

not good

we have to fix that...

please post a link for a replacement battery

All,

All I'll say is Parasitic capacitances, capacitive reactance in Mosfet between two substrates. The test means nothing.
Take that blob of a MOSFET and replace it was a capacitor of the correct value in series with the LED.  Same effect.  C.C


Ben K4ZEP

You can't make accurate power measurements with LED brightness.

Sweep it in LTSpice or use this ckt.

You can choose to short the resistor on the leg not being tested for higher accuracy.

Quote from: verpies on 2016.08.11, 04:17:47
@Itsu

Could you put together a quick video showing the gate leakage of any N-Ch Power MOSFET in the circuit below ?
No power supply. Just program your FG to 1MHz square wave at 10V amplitude. Use a red LED (observe its polarity).

I followed that little diagram (10V pp  meaning +5V/-5V at 1MHz, red lead to gate) and used severall high power MOSFET's and IGBT's to show how they leak.

MOSFETs used in sequence are:

FQA11N90C
IRFP260N
IRFP4668
IRFP460
47N60C3    (i had to increase the amplitude to 13V pp to see a dimm light in the led)


IGBT's used:

G40N60
G4PH30K



Video here:   https://www.youtube.com/watch?v=av53ZMeWQ-k


Regards Itsu

Quote from: K4zep on 2016.08.11, 12:29:24
All I'll say is Parasitic capacitances, capacitive reactance in Mosfet between two substrates. The test means nothing.
Take that blob of a MOSFET and replace it was a capacitor of the correct value in series with the LED.  Same effect.  C.C

I agree except for, that it "means nothing".

The tests done by Tinsel, Itsu and myself clearly demonstrate, that these  "Parasitic Capacitances" exist in all transistors and are not negligible as they can transfer significant energy from the gate driver to the drain-source circuit.
As the dv/dt increases, the magnitude of this leakage, also increases. 
We tried it with only 10V/μs pulses and in the MIT device some MOSFETs are subjected to 600V/μs pulses, so the gate power leakage in the MIT can be expected to be 60² higher.

All of this does not mean nothing - it means that the energy delivered by the MOSFET gate drivers is not negligible and can contribute Watts to the output circuit, thus it must be accounted for in Graham's MIT device.

Quote from: ION on 2016.08.11, 15:11:03
You can't make accurate power measurements with LED brightness.

I know.  That test was only qualitative.  It merely demonstrated, that the gate leakage exists and at 10V/μs is significant to light up a LED.

The quantitative leakage test should be done by Graham at the actual dv/dt conditions that occur in his circuit.

Quote from: verpies on 2016.08.11, 16:52:36
I know.  That test was only qualitative.  It merely demonstrated, that the gate leakage exists and at 10V/μs is significant to light up a LED.

The quantitative leakage test should be done by Graham at the actual dv/dt conditions that occur in his circuit.

Yes, I'm sure you do know, I only posted that circuit so that Itsu or other interested persons could get some real numbers, and see the difference between source and drain injection on a two channel scope.

I'll check it in a simulation as time permits. Thanks for bringing it up.

One reason why I say a simple $20 Kill-A-Watt meter on his mains input could have helped to convince (or not) depending on the total input power reading.

Now you're getting somewhere with this analysis!

Excellent Sleuthing work gentlemen!

Once again, the value of "in depth knowledge" proves
its remarkable importance.

Here's an LTSpice file as a starting point for those that want to play with / modify   gate injection tests.

I lack the skill just yet to do really good power studies so I'll leave that to Poynt and partzman and any other interested parties.

Quote from: muDped on 2016.08.11, 20:23:36
Now you're getting somewhere with this analysis!

Excellent Sleuthing work gentlemen!

Once again, the value of "in depth knowledge" proves
its remarkable importance.

Agreed, muDped, in depth knowledge is a wonderful thing as is the common sense that does not choose ignore the "500 lb gorilla"* in the room.

(*lack of mains input power measurement)

Dear All,

Here is my take on one conceptual schematic for the Synchronous Diode sub-circuit. Please review and make constructive comments from this vast in depth knowledge base.

Spokane1

Quote from: ION on 2016.08.11, 21:45:07
Agreed, muDped, in depth knowledge is a wonderful thing as is the common sense that does not choose ignore the "500 lb gorilla"* in the room.

(*lack of mains input power measurement)

I think we all understand that a measurement of total mains input power would show much more than the measly 10 Watts or so output at the bulb.

And also, most of us probably understand that the claim made by the publishers of the video that this device gives a "measured out at 570% more output than the input he paid for" is just typical exaggerated marketing hype, since most people would think that "the input he (or someone) paid for" means just that: the mains power usage that shows up on the electric bill.

Whether or not someone may think it is legitimate only to consider as "input" the direct input at the primary of the transformer, the marketing claims made to sell the video are unsupported by any evidence.... and some of us know why that is.

It would be interesting to see the various input power measurements taken at different points and a final in to out figure based on an entire device that can run from a bank of batteries or similar.

If Graham is so clever then he must be aware of these power measurement nuances to look out for, which to me would indicate intentional deception, regardless of weather he actually spoke a claim of OU in clear terms the claim would still appear to be made by him in a roundabout way.

If it is shown to be well under OU with accepted accurate measurement protocols then I would say the behaviour of the people involved in the sales around the device- Video's ect, would be bordering on criminal. Without the OU hype I doubt they would sell much associated merchandise.

It fair dinkum looks like another con to me.

..

Quote from: Farmhand on 2016.08.12, 00:56:01
If Graham is so clever then he must be aware of these power measurement nuances to look out for, which to me would indicate intentional deception,

Indeed, he is well versed in these techniques but he might have trivialized the energy leaked by the MOSFET gates and did not even bother to measure it.
I've learned enough about his character over the years that I would not accuse him easily of intentional deception.  I'd rather lean toward TK's tag line:  "The easiest person to fool is yourself".

Quote from: Farmhand on 2016.08.12, 00:56:01
I would say the behavior of the people involved in the sales around the device- Video's ect, would be bordering on criminal. Without the OU hype I doubt they would sell much associated merchandise.

Note that these people are not the same as Graham G.  Vultures are abundant.

Quote from: Farmhand on 2016.08.12, 00:56:01
It fair dinkum looks like another con to me.

In my opinion the jury is still out.
There is no reason to be that hard on Graham G..

Quote from: Spokane 1 on 2016.08.11, 22:54:30
Here is my take on one conceptual schematic for the Synchronous Diode sub-circuit. Please review and make constructive comments from this vast in depth knowledge base.

Shouldn't the winding powering the ±5V regulators be center-tapped or have a different diode arrangement ?

Anyway, measuring the DC energy flow out of the outputs of these 12V and ±5V regulators to the gate driver would be more fair than measuring it at their inputs (or before the isolation transformer) because it would not include their losses, which are inherently high for such regulators.

It would be nice to do power measurement directly at the MOSFET gate terminals, but doing so would be very hard to do, due to the potentials involved, lack of stable ground and high crest factor (https://en.wikipedia.org/wiki/Crest_factor) waveforms appearing there.
This is probably the reason why Graham has not done these measurements.  They are hard to do after the isolation transformer and before the transformer, these measurements would be very unfair because of low-efficiency of the entire DC-DC converter including linear regulators downstream.

If you are in contact with Graham, ask him to measure the DC current coming out of these linear regulators with a calibrated isolated current probe or even a floating ammeter.
In case of a digital ammeter, I would be afraid that when the secondary of the MI Transformer is open-circuited, the entire secondary side of the DC-DC converter violently swings in less than microsecond by many hundreds of Volts ...and interferes with the operation of a digital electronics inside the meter, which becomes a transmitting antenna during that period.

...and please send Graham the link to this video:
https://www.youtube.com/watch?v=av53ZMeWQ-k

Quote from: Verpies
In my opinion the jury is still out.
There is no reason to be that hard on Graham G..

In the interest of Full Disclosure and Transparency it would
be interesting to discover what sort of remuneration each of
the participants in the operation will receive.

The "con" may be the work of others who enticed Graham to
participate as a demonstrator in that phase of the operation.
In other words, his work may be exploited for gain by clever
organizers.  It would not be the first time.

Having just now viewed the video there is something that no one has commented upon, and that is the part where he shows the transformer output voltage as a negative plateau over most of the cycle followed by a sharp positive spike.  He then integrates this to get the flux waveform as a sawtooth.  That flux is different to the primary flux obtained by integrating the primary voltage.  So it appears that there is flux leakage playing a major role in this.  More later when I have analyzed this new information.

Smudge 

Quote from: muDped on 2016.08.12, 06:33:21
In the interest of Full Disclosure and Transparency it would
be interesting to discover what sort of remuneration each of
the participants in the operation will receive.

The "con" may be the work of others who enticed Graham to
participate as a demonstrator in that phase of the operation.
In other words, his work may be exploited for gain by clever
organizers.  It would not be the first time.

To be enticed once to join the ranks of such men can be excused, however a long association that goes back to 2007 or earlier with A and P raises an eyebrow or two. Then there are the non-working patents assigned to Chava aka Mark G., another questionable character in the drama.

Of course there are those that know all this and continue to "soldier on" with the MIT. They know things that I am not privy to concerning GG and his character so I will trust their knowledge.

Therefore my attitude is wait and see, contribute where I can, get inside GG's head by reading all that I can find that he has written or spoken in videos. Body language can be very revealing.

For all I know at this point GG may be an honest guy with high ideals that has just been enticed and used by the wolves. Maybe a little or a lot naive to the fact that there is much trickery in the business world.

His heart seems to be in the right place. In the video he says something to the effect (56:22 to 57:11) "You gotta give it away, otherwise you are a schmuck", but admits to almost changing his mind about doing the conference.

He seems to strongly believe in total open source, yet has changed his position after the conference. Understood, since he had mixed feelings about doing the conference in the first place.

Quote from: verpies on 2016.08.12, 04:24:34
Shouldn't the winding powering the ±5V regulators be center-tapped or have a different diode arrangement ?

Dear verpies,

I agree with you. A center tapped secondary would make the reverse engineering effort simpler.  The problem is that I can't see a center tap connection on that torrid transformer. Perhaps Graham is using some king of forward converter topology that I am not familiar with.

Do you have any ideas that would match with the components shown?

Spokane1

Quote from: Spokane 1 on 2016.08.12, 15:49:01
Dear verpies,

I agree with you. A center tapped secondary would make the reverse engineering effort simpler.  The problem is that I can't see a center tap connection on that torrid transformer. Perhaps Graham is using some king of forward converter topology that I am not familiar with.

Do you have any ideas that would match with the components shown?

Spokane1

Typically the -5 volt supply is a light current draw bias supply so can be derived without a center tapped transformer by using the forward pulse. I have done this with success in the past.

If it has not already been mentioned, it appears the series Zener to the optoisolator on the -5 supply is reversed.

Since the enable input to the gating chip IXDD614PI wants to be high to allow pulses to the FET driving the transformer, it appears the circuit is configured to shut down should there be a excess voltage on one of the power supply outputs.

Good work on the schematic.

Quote from: ION on 2016.08.12, 15:46:56

His heart seems to be in the right place. In the video he says something to the effect "I want to share this with the world, and if I didn't I'd be a schmuck and I don't want to be a schmuck".

He seems to strongly believe in total open source, yet has changed his position after the conference. Puzzling.

Dear ION,

I find that sudden change in attitude (only one day after the conference) puzzling as well.  Your observation brings up a situation that I haven't considered before. Maybe someone had put the collar on him between the time the conference ended and when I called him the Monday evening after.

I'm sure all the reasons that he gave me were "The reasons that sound good - then there is the real reason".   If he had made the big breakthrough 3 weeks before the conference then he would have had lots of time to consider the degree of his intended disclosure. All the issues that he brought up about the landlord, family finances, and potential profit would have been the same before the conference as they were a day after.

At the conference, even after his presentation the next day he was freely sharing technical details with those who asked (mostly Reiyuki), then no less than 24 hours later he has this big turnaround. I'm starting to smell a rat - thanks to your comment.

All it would take would be one veiled threat via a phone call that mentioned his family and he would clam up in a heart beat. 

When I came over to his house a week later to drop off the Litz wire. He was glad to talk about his circuit, but I didn't have the understanding to really know what to ask.  He was going to take a call from his son and stepped  outside. I asked him if I could take some photos in his absence, he said OK. So for about 5 minutes I took as many close ups as I could. He came back and seemed nervous as if he was having to seriously reconsider what he had just offered. I quit taking photos as soon as he returned. I would have taken a hundred more.

He had been working on his circuit and did talk about certain improvements that he was planning to make. It was obvious that he was scanning different frequencies and making adjustments to the various pulse widths using signal generators.

Yes, I agree with you. Something is not right in this picture.  He spent several minutes making a public declaration about his intent to make all this stuff Open Source. He even had previously purchased a particular Internet site to place all this material on.

What would cause a person to make such a sudden turn around?

Spokane1

Spokane1 said:

QuoteWhat would cause a person to make such a sudden turn around?

Maybe he was reminded of the fine print in his contract with A and P, that they would own the rights to publish and also benefit from any subsequent disclosure information. (the collar you spoke of).

Genius technical types with high ideals rarely read fine print, tending to trust everyone is like them. They are often to busy with their work and a head full of ideas. Enter the jackals.

Just a guess, I seriously doubt MIB's are involved, as everything is still "for sale".

As for his open source website, I doubt anything new will be put there soon as that may violate his contract.
Maybe just some old stuff approved by A and P in order to "grease the skids".

I see another "advanced handbook" coming out soon.

GG is caught between a rock and a hard place, I don't envy him, especially if he really has the high ideals we see in the video.

Quote from: ION on 2016.08.12, 17:01:25
Spokane1 said:

Just a guess, I seriously doubt MIB's are involved, as everything is still "for sale".

Dear ION,

I agree with you. If the real MIB were involved then by the time I got there a week later everything in his shop would be gone and possibly his person to boot.

If someone did exert this kind of control it appears that it is in their best interests to have him continue as before - unfunded.

If A+P were claiming some kind of exclusive publishing rights, which is possible, then it would be no big deal - at least to me - to have said so.

What is different in this story is that Graham avoided discussing the actual performance measures of his machine before the conference presentation. He wanted to make it a surprise (which it was). The unknown parties (if they exist) would have had to act pretty quick. That suggests that whomever (if they exist) were probably in the audience.

Putting a price tag on his work was more my idea than his. We both know that nobody is going to purchase such an infant technology at this stage.  This was an attempt to reduce all the "joint venture" proposals that crop up with this kind of disclosure.

Spokane1

Dear Spokan1

Thanks for the extra insight and all your hard work on this project. If you have not already offered the photos you took at his home will you be doing that, or are those photos confidential?

I have attached a slightly revised version of your drawing with the Zener diodes on the minus five supply that I think should be reversed circled in red.

AS time permits I will work up a schematic of how the minus five should be derived. There are two methods, forward pulse and charge pump. I doubt that it is the charge pump method as that would require two diodes and two capacitors, whereas the forward pulse method uses one diode, one capacitor.

Dear Spokane1

Attached are two methods to generate a negative supply. Each has merits and drawbacks. Just roughed in (not optimized) and tested in LTSpice.

Regulators not shown for simplicity, but utilize the raw outputs. You should pick the one that best seems to match the photos.

Graham could have eliminated a lot of extra complexity by closing the loop and dispensing with the regulators, but I guess he had his reasons.

Quote from: ION on 2016.08.12, 19:19:44
Dear Spokan1

Thanks for the extra insight and all your hard work on this project. If you have not already offered the photos you took at his home will you be doing that, or are those photos confidential?

Dear ION,

The photos are available to anybody who has use for them. Its just that they are to big at 2 MB+ per photo to place here. I have 40 of them. I can only send about 6 at a time to most people's mail box.  I have posted them here in the 800 pixel resolution when a particular subject under discussion comes up.  I'm sure anyone who is actually attempting to build a replication circuit could benefit from looking at them.

Rieyuki set up a Google drop box, but I couldn't get it to work due to the lack of a password.

Send me your email  using the forum message service and I shall be happy to send them to you (6 at a time)

Spokane1

Quote from: ION on 2016.08.12, 20:12:03
Dear Spokane1

Attached are two methods to generate a negative supply. Each has merits and drawbacks. Just roughed in (not optimized) and tested in LTSpice.

Regulators not shown for simplicity, but utilize the raw outputs. You should pick the one that best seems to match the photos.

Graham could have eliminated a lot of extra complexity by closing the loop and dispensing with the regulators, but I guess he had his reasons.

Dear ION,

Hey that looks great, especially this first one. It seems to match the dual diode component layout we see on the circuit board.  It looks like I shall be redrafting that drawing next week.

As far as his reasons go I believe that voltage isolation was a big design consideration. In my Spice simulations I'm seeing some very high voltages develop in the tank circuit. I'm sure that Graham saw the same thing develop in his simulations, so he designed this high voltage separation approach, or so I think.

Also I'm not so sure all those little epoxy transistor looking things are regulators, that was just a WAG that seem to fit for the first schematic draft. If they were transistors I would expect more support components.

Spokane1

If they are 78Lxx series regulators (79L05 for the negative 5V) in TO-92 packages, they have a peak current rating of 100 - 140 mA depending on manufacturer.  Is this (78L12) enough to turn on a SiC mosfet through a gate driver at 50 - 75 kHz?

Quote from: Spokane 1 on 2016.08.12, 21:13:39
Dear ION,

The photos are available to anybody who has use for them. Its just that they are to big at 2 MB+ per photo to place here. I have 40 of them. I can only send about 6 at a time to most people's mail box.  I have posted them here in the 800 pixel resolution when a particular subject under discussion comes up.  I'm sure anyone who is actually attempting to build a replication circuit could benefit from looking at them.

Rieyuki set up a Google drop box, but I couldn't get it to work due to the lack of a password.

Send me your email  using the forum message service and I shall be happy to send them to you (6 at a time)

Spokane1

Dear Spokane1

Thanks for your kind offer, but you've got enough on your plate right now.

I think Chet or Slider remembers how to get to the ftp site that was setup here, maybe we can make a folder for the information repository of this device and include the photos there in one big upload (when you have more time).

Regards, ION

ION

The ftp site we had setup was on the previous server and no longer exists, i dont have a problem with him uploading them here, each post has a size limit on this forum so the way to do it is to upload 1 and then keep modifying the post gradually adding them 1 by 1.

Regards
Peter

I will assume this task if Spokane1 approves , perhaps in a separate thread in his bench section so he can moderate input delete , adjust and advise
as necessary.

As ION mentions, Spokane1 has quite a full plate already .

respectfully
Chet K

Quote from: ION on 2016.08.11, 20:03:10
One reason why I say a simple $20 Kill-A-Watt meter on his mains input could have helped to convince (or not) depending on the total input power reading.

Spokane1 only has three tasks that are worth pursuing and are relevant at this time:

1)  Replicate the device as accurately as he can.

2)  Integrate a $20 Kill-A-Watt meter into the circuit input.

3)  Analyse the results.

From GG's demonstration of a sawtooth flux waveform I have taken the primary current waveform then deduced the secondary current needed to create that sawtooth, assuming normal transformer action where the flux comes from the difference between pri and sec ampere turns.  Done crudely from my sketched waveforms.  And I get a secondary current very much like that shown by GG.  So I deduce that there is no significant flux leakage and my original estimate of the primary flux waveshape is wrong.

Smudge   

Quote from: evolvingape on 2016.08.13, 14:23:30
Spokane1 only has three tasks that are worth pursuing and are relevant at this time:

1)  Replicate the device as accurately as he can.

2)  Integrate a $20 Kill-A-Watt meter into the circuit input.

3)  Analyse the results.

All,

This could amount to 3-6 months work or maybe a lifetime!  I'm assuming most want to see if this is "OU".  So, take a large multi farad cap. bank and let the output charge it.  Run a simple high efficient inverter off that bank to power the front end via a isolation transformer, close the loop.  See what happens to the voltage on the Cap. Assuming there is some feedback to prevent running away in the circuits, the voltage on the cap should remain constant, and it should run on and on and on and on.  Even better, then add a small load to the output to show it is actually doing "Work".     ;D  I know of at least 3 people that will be working this in some form or another.  Myself, I'm more into research as to "why" it works and to understand the transformer part of the device. Building by rote simply does not work without understanding been there, done that. Time will tell!

Ben K4ZEP

Quote from: Smudge on 2016.08.14, 15:22:03
From GG's demonstration of a sawtooth flux waveform I have taken the primary current waveform then deduced the secondary current needed to create that sawtooth, assuming normal transformer action where the flux comes from the difference between pri and sec ampere turns.  Done crudely from my sketched waveforms.  And I get a secondary current very much like that shown by GG.  So I deduce that there is no significant flux leakage and my original estimate of the primary flux waveshape is wrong.

Smudge

What is your rationale for the statement "no significant flux leakage" considering the 10 mil mylar spacer between the "U" sections and the large physical distance between primary and secondary coils? What approximate level would be considered significant?

On another note, I am happy to try to simulate any simple drive circuit that we suspect could produce even some very small continuous DC on the secondary, without the use of rectifiers. We know from classical induction theory (and practice), this is supposedly not possible.

Ben said:

QuoteMyself, I'm more into research as to "why" it works and to understand the transformer part of the device. Building by rote simply does not work without understanding been there, done that.

Smudge has offered his thoughts on a possible mechanism, and I agree, we become another "cargo cult" if we only copy without understanding.
http://www.sjsu.edu/faculty/watkins/cargocult.htm

Working up a hypothesis with bench experiments seems a very worthy route. In GG's entire conference, he never hinted at the idea he had in the back of his mind to build such a device in the first place. He does say the discovery was accidental.

Regards. ION

GG shows the secondary waveform that he integrates (or rather the 'scope integrates) to get the flux.  This is not DC although there is almost a continuous negative DC plateau.  But that intermittent plateau has enormous positive spikes.  So the transformer does not produce DC.  Only after rectification is there a DC output.

As regards leakage I originally integrated the input voltage to get primary flux and that suggested considerable flux leakage because thaat waveshape could not produce the secondary voltage.  Now that the difference between pri and sec current creates the sawtooth flux we have normal transformer action in that regard.  So IMO the input voltage waveform shown by GG is not taken across the primary, but earlier in the circuit, perhaps ahead of some capacitance or inductance.  Of course there will be a small leakage flux but I can't determine its magnitude.

Smudge

Here is a different topological approach to Graham's device that utilizes a current fed parallel resonant inverter with a clamp phase and a pulsed rectifier. The uniqueness IMO is the common mode power consumption at Etap and differential power across the primaries L2 and L4 between Ecap and Ecap2. The transformer is linear in this example.

The first pix shows the busy plot with schematic. Some .meas calcs are shown that give the DC input power, the output power, plus the resulting COP.

The second pix shows one full cycle and the resulting plot maths for the cycle. The "if" statements are reasonably synchronized to V(Ecap) and V(Ecap2) respectively to be able to see the power of each branch. The power in at Etap is equaled by the sum of the power in each branch as can be seen while at the same time, the total power across the primary network branches of L2 and L4 is a net of ~-8mw.

Therefore it my conclusion that this circuit simulates OU in reference to the reactive power across the primaries compared to the resistive load output power.

Fire away! 8)

pm

PM, very nice!!

I like the idea of the current fed center tap p-p approach over the H bridge, it eliminates all the messy high side drive requirements, and is easier to test on the bench.

I'll study it further, but I'm a novice at LTSpice.

Regards
ION

P.S. I don't see the timing for pulse generators V1, V3 and V7, I guess they were cut off the schematic.

Also, new to me is the stacking  of the power supply sources V2, V8. plus I can't scroll the Spice error log to see the calculations because it is a screen capture.

If you are inclined, please post the .asc file so I can study it more deeply. Thanks much!

Input scope probe positions. The big Red and Black clips are the leads to the Tek P5205 High-Voltage Differential Voltage Probe. I still don't know the part number of the Current Probe.

Quote from: ION on 2016.08.14, 19:32:47
PM, very nice!!

I like the idea of the current fed center tap p-p approach over the H bridge, it eliminates all the messy high side drive requirements, and is easier to test on the bench.

I'll study it further, but I'm a novice at LTSpice.

Regards
ION

P.S. I don't see the timing for pulse generators V1, V3 and V7, I guess they were cut off the schematic.

Also, new to me is the stacking  of the power supply sources V2, V8. plus I can't scroll the Spice error log to see the calculations because it is a screen capture.

If you are inclined, please post the .asc file so I can study it more deeply. Thanks much!

ION,

Thanks.

I didn't include the generator signals to save from more confusion on the plots. I can show a plot with these but I've attached the asc file below so you can play with the device.

Sorry about V2 as it really is not needed. It is simply a lossless current sensor made with a zero voltage source but the current thru L1 is identical so V2 is redundant.

I did crop the error log so it would fit on the capture but you should be able to see the Iin, Pin, Eout, Pout, and COP numbers.

pm

Edit: The asc file attached just fine like Poynt said so I don't know what in the world I was trying to attach previously! 

Thanks pm.  :)

When analyzing the input circuit branches of the GG3, there appear to be some interesting details IMO over the time period shown in the plot that is attached.

The beginning current of L1 is 14.289ma and ends at 12.139ma for a loss of 2.15ma. This equates to an energy loss in L1 of 2.84uJ or 148mW over the 19.2us period. Note that during this same time period,  the average power output is 232mW.

What creates this loss of current in L1?  Observe the average voltage at Etap of 31.182v during this time period which is in opposition to the 20v DC of Vcc. Therefore due to di = E*dt/L, Etap forces a reverse current of (31.182-20)*19.2e-6/.1 = 2.15ma which is equal to the lost current in L1 during this period.

The currents in L2 and L4 are also interesting.  The current in L2 starts at 73.746ma and ends the period at 97.77ma with L4 following at ~14ma less.

Do these observations offer any potential for OU?

pm

Quote from: Smudge on 2016.08.14, 15:22:03
From GG's demonstration of a sawtooth flux waveform I have taken the primary current waveform then deduced the secondary current needed to create that sawtooth, assuming normal transformer action where the flux comes from the difference between pri and sec ampere turns.  Done crudely from my sketched waveforms.  And I get a secondary current very much like that shown by GG.  So I deduce that there is no significant flux leakage and my original estimate of the primary flux waveshape is wrong.

Smudge

Dear Smudge,

You seem to have a handle on the magnetic situation going on in the secondary. I was wondering if you would like to take a stab at what I have noticed in my simulations.

The operation of the primary tank circuit seems to follow what Graham showed in his presentation. The big difference is what is not happening in the secondary after the synchronous diode gets  a harvest command pulse.

In the simulation, using an ideal transformer with a K of .8588 (or so) and no provisions for saturation, much less the different split core parameters, there is hardly any impact on the secondary current except a hardly noticeable blip.

In Graham's presentation the operation of the synchronous diode on the secondary current is profound.

Any idea's as to what might be happening here and how the simulation can be changes to better match Graham's version?

Spokane1

from pm:

QuoteWhen analyzing the input circuit branches of the GG3, there appear to be some interesting details IMO over the time period shown in the plot that is attached.

The beginning current of L1 is 14.289ma and ends at 12.139ma for a loss of 2.15ma. This equates to an energy loss in L1 of 2.84uJ or 148mW over the 19.2us period. Note that during this same time period,  the average power output is 232mW.

What creates this loss of current in L1?  Observe the average voltage at Etap of 31.182v during this time period which is in opposition to the 20v DC of Vcc. Therefore due to di = E*dt/L, Etap forces a reverse current of (31.182-20)*19.2e-6/.1 = 2.15ma which is equal to the lost current in L1 during this period.

The currents in L2 and L4 are also interesting.  The current in L2 starts at 73.746ma and ends the period at 97.77ma with L4 following at ~14ma less.

Do these observations offer any potential for OU?

pm

Interesting observations, pm.

I was wondering what could be done to improve the efficiency of the circuit such as changing all DC resistance in the inductors to zero ohms and maybe replacing the FET's with pure switches or alternately modify the "Rds on" to zero ohms in the model.

Substituting pure switches might also eliminate gate injection, which may not be a major factor in this sim.

Do you have any ideas for optimization?

Unfortunately, I lack the skill to properly analyze and comment on the sim, and need to get further up the learning curve.

Keep the good work coming  O0

Regards, ION

Quote from: TinselKoala on 2016.08.14, 20:06:17
Input scope probe positions. The big Red and Black clips are the leads to the Tek P5205 High-Voltage Differential Voltage Probe. I still don't know the part number of the Current Probe.

Dear TK

I don't have the model number of that current probe, but if I ever get it I shall forward it to you.

About all we have to go on is that it cost $1,500 in 2000, is a Hall chip device,  and was purchased about the same time the Tek scope was.

Do you have any idea what could have been purchased back then for that kind of money? There can't be more than five or so manufactures that made that kind of stuff back then and they all probably offered the same kind of product.

Maybe this is the kind of device where vast differences between the manufacturer's offerings existed, and maybe still do.  Maybe each model has important configuration differences based upon current range and design of sensing amplifiers.

I know this is not much help for this kind of question.

Spokane1

Dear All,

I have secured some 3C90 ferrite cores from Surplus Sales of Nebraska. So far no luck finding the high permittivity versions of the same size.

Spokane1

Quote from: ION on 2016.08.15, 16:01:24
from pm:
Interesting observations, pm.

I was wondering what could be done to improve the efficiency of the circuit such as changing all DC resistance in the inductors to zero ohms and maybe replacing the FET's with pure switches or alternately modify the "Rds on" to zero ohms in the model.

Substituting pure switches might also eliminate gate injection, which may not be a major factor in this sim.

Do you have any ideas for optimization?

Unfortunately, I lack the skill to properly analyze and comment on the sim, and need to get further up the learning curve.

Keep the good work coming  O0

Regards, ION

ION,

Thanks for reading!

I used the IRF831s because I have then on hand for bench tests if needed and really wanted somewhat realistic losses in the sim.

You will find that if you try to use zero ohms for inductors, switches, etc, LtSpice will not start. If it does, it will not converge. So, one must have a small amount of resistance in the inductors and switches and yes the efficiency could be improved by reducing the values I used. However, disregarding the gate drive losses, the drain to source losses for M1 and M2 are 440uW and 394uW during each half cycle on time and 37mW and 6.5mW during the clamp time. I haven't calculated the losses for M3 at the moment.

Regarding other optimization, apart from your suggestions I really don't have any more to offer atm.

I do not presently understand how energy can be taken from the primary circuit branches to deplete the energy in L1 during the first 19.us of the cycle and see the energies increase in L2 and L4 while at the same time producing power in the load! ???

pm

Quote from: Spokane 1 on 2016.08.15, 16:01:56
Dear TK

I don't have the model number of that current probe, but if I ever get it I shall forward it to you.

About all we have to go on is that it cost $1,500 in 2000, is a Hall chip device,  and was purchased about the same time the Tek scope was.

Do you have any idea what could have been purchased back then for that kind of money? There can't be more than five or so manufactures that made that kind of stuff back then and they all probably offered the same kind of product.

Maybe this is the kind of device where vast differences between the manufacturer's offerings existed, and maybe still do.  Maybe each model has important configuration differences based upon current range and design of sensing amplifiers.

I know this is not much help for this kind of question.

Spokane1

Yes, indeed there are vast differences in current probes and indeed they all have their particular configuration differences, quirks and foibles. Bandwidth, DC offset and propagation delay characteristics are very important. Good ones cost a pretty penny today and always have done. Frankly, I get depressed and feeling faint whenever I look at current probes. The LeCroy AC-DC current probe I used in this demonstration cost over 3,000 dollars when it was new.

http://www.youtube.com/watch?v=KWDfrzBIxoQ

It is far simpler and generally more precise to use a properly-arranged Current Viewing Resistor and monitor the Vdrop across that item with ordinary passive probe(s) or a specially constructed Kelvin probe arrangement. If signal isolation is a problem during simultaneous measurements of V and I,  there are various ways around that.

As to Gunderson's Tektronix probe -- that's about all I can tell from the photos, that it is a Tek probe, of mid-recent manufacture. Maybe someone else can do the digging necessary to identify it. Here are the photos of the probe that I have found:

It could be something like this:
http://www.jensentools.com/tektronix-tcp0030-120-mhz-tekvpi-30a-ac-dc-current-probe/p/441-693

or this:
http://cpc.farnell.com/tektronix/tcp202/tektronix-current-probe/dp/IN02098

Quote from: TinselKoala on 2016.08.14, 20:06:17
I still don't know the part number of the Current Probe.

It may be either a TCP0030 (DC-120MHz) or a TCP202 (DC-50MHz) (a common choice).

Perhaps even a P6021 (120Hz-60MHz).

Quote from: Smudge on 2016.08.14, 15:22:03
From GG's demonstration of a sawtooth flux waveform I have taken the primary current waveform then deduced the secondary current needed to create that sawtooth, assuming normal transformer action where the flux comes from the difference between pri and sec ampere turns.  Done crudely from my sketched waveforms.  And I get a secondary current very much like that shown by GG.  So I deduce that there is no significant flux leakage and my original estimate of the primary flux waveshape is wrong.

Smudge

Dear Smudge,

You may have calculated no significant flux leakage from your analysis of the scope traces, but Graham was quick to point out to me that there is a lot of leakage where there shouldn't be. As I mentioned in a post a few weeks ago, he built little RF probes to explore the degree of leakage and was amazed at all the energy circulating around the outside of his core. Now that would be something interesting to see in itself.

I think it is vital to have on hand good classical analysis of this apparatus and then be able to compare those numbers to what we actually get. It is the differences that might help improve our search for the seat and the process of the non-classical conversion.

Spokane1

TK said:

QuoteIt is far simpler and generally more precise to use a properly-arranged Current Viewing Resistor and monitor the Vdrop across that item with ordinary passive probe(s) or a specially constructed Kelvin probe arrangement. If signal isolation is a problem during simultaneous measurements of V and I,  there are various ways around that.

I agree, the CVR can be built into a test setup and give good repeatable results unlike the TEK current probes I have used, where just a little speck of dirt in the gap or improper closure can throw the reading way off. Maybe the newer ones are better. I'm too old to keep up.  :-[

Spokane1 said:

QuoteAs I mentioned in a post a few weeks ago, he built little RF probes to explore the degree of leakage and was amazed at all the energy circulating around the outside of his core. Now that would be something interesting to see in itself.

I don't doubt that he would pick up a lot with an RF probe and at a good distance due to the fast switching transients. I'm sure there was a lot of radiated EMI. As you already know, straight magnetic leakage is a different animal, and a good magnetic probe will be shielded against RF.

Regards, ION

P.S. A good article on constructing RF sniffer probes: http://electronicdesign.com/components/simple-homemade-sensors-solve-tough-emi-problems

Quote from: Spokane 1 on 2016.08.15, 15:52:18
In Graham's presentation the operation of the synchronous diode on the secondary current is profound.

Perhaps the gate current adds to the secondary current.

Dear All,

Here is a source of inexpensive bias ferrite magnets.

Spokane1

Here is yet another circuit variation to Graham's demo device that is best described as an unbalanced current driven parallel resonant inverter with clamp phase. IMO, this would be an excellent alternative to a full H bridge circuit and the primary switching is ground referenced.

The first pix shows the schematic and plot with some .meas math to show the overall COP from supply to output.

The second pix is a rather busy plot with the various branch power measurements.

The third pix is a simplified plot showing the differential voltage across the primary L4, the power across L4, the current thru L4, and mostly resembles Graham's primary waveforms.

pm   

Good work, pm, and within reach of the average builder without exotic ferrites and magnets.

I do like this approach, and later, for fun, other ferrites and magnets can be tried once the basic driving circuit is built.

Regards, ION

Quote from: verpies on 2016.08.16, 02:03:08
Perhaps the gate current adds to the secondary current.

Dear verpies,

I'm sure there is some leakage we need to be aware of, but the current doesn't increase it decreases because it is being directed into that large backend storage capacitor, probably along with what ever energy escapes from the gate excitation signal.

In my simulations, with an ideal transformer, this sudden drop in secondary current doesn't happen and I wonder why.

Do you happen to know what the magnitude and the wave form would be for gate leakage energy into the Drain-Source Circuit? Are we looking at 30 milliwatts or 300 milliwatts? I suppose it would be similar to a parasitic capacitor of some fixed value. I also suppose that it could be simulated with a capacitor between the gate and the drain.

From the presentation, the green trace is the secondary current. It appears to take a sudden dump when the harvest pulse is triggered, and then takes some time to recover while the primary current is clamped at a constant value. Something interesting is happening there. I suspect this is the result of the saturated core condition.

Photo provided by Reiyuki from the presentation.

Spokane1

Quote from: partzman on 2016.08.16, 15:12:43
Here is yet another circuit variation to Graham's demo device that is best described as an unbalanced current driven parallel resonant inverter with clamp phase. IMO, this would be an excellent alternative to a full H bridge circuit and the primary switching is ground referenced.

The first pix shows the schematic and plot with some .meas math to show the overall COP from supply to output.

The second pix is a rather busy plot with the various branch power measurements.

The third pix is a simplified plot showing the differential voltage across the primary L4, the power across L4, the current thru L4, and mostly resembles Graham's primary waveforms.

pm   

Dear PM

Would you be so kind as to post a trace from you detailed simulation to show just the primary tank voltage (for reference) with the secondary current.  My tired eyes are having difficulty sorting out the important information (for me) from all the supporting traces.

Thank you in advance.

Spokane1

Quote from: Spokane 1 on 2016.08.16, 15:53:08
Dear PM

Would you be so kind as to post a trace from you detailed simulation to show just the primary tank voltage (for reference) with the secondary current.  My tired eyes are having difficulty sorting out the important information (for me) from all the supporting traces.

Thank you in advance.

Spokane1

Sure. Below is a plot that shows the differential input tank voltage across L4 and the output current thru the secondary L5. I also left the current thru L4 for reference.

I basically see the same results as your are getting on your simulation regarding the input and output currents. Any change in the secondary current is mirrored in the primary unlike Graham's circuit. IMO, only after we add the nonlinear core and PM bias perhaps we'll see the magnetic separation between the two cores. Also, my off time for M3 is 1us but if made long enough, the currents will reverse.

I will also add that during the clamp time in Graham's device when the primary current is basically frozen and the secondary current and flux reverse direction, there are bucking currents in the coil windings that will produce H fields between the core legs and outside the core legs. IMO, these are the fields that Graham has seen with the sensing coils.   

At present, I'm attempting to gain enough info to model the upper core half using the gyrator-capacitor approach. The sticky part is modeling the nonlinear B/H curve and the proper hysteresis loop for the material used.  BTW, are you an IEEE member?

pm

Edit: Changed "bucking fluxes " to "bucking currents in the coil windings".

Quote from: Spokane 1 on 2016.08.16, 15:44:59
From the presentation, the green trace is the secondary current. It appears to take a sudden dump when the harvest pulse is triggered, ...
...
In my simulations, with an ideal transformer, this sudden drop in secondary current doesn't happen and I wonder why.

I don't think it is a sudden drop but a current reversal due to half a cycle of high frequency LC oscillation. 

When the "synchronous diode" opens, the capacitance in series with the transformer's inductance drops precipitously and increases the LC frequency of whatever remains in that "open" circuit.
This is why adding a 10nF capacitor across the "synchronous diode" lowers the frequency of this LC oscillation, so the current can no longer reverse in the 1μs allotted by the "synchronous diode".

Quote from: Spokane 1 on 2016.08.16, 15:44:59
Do you happen to know what the magnitude and the wave form would be for gate leakage energy into the Drain-Source Circuit? Are we looking at 30 milliwatts or 300 milliwatts?

The current flowing due to the MOSFET's Gate Leakage is proportional to the voltages & frequencies appearing between the gate and drain/source terminals.  For example, the reactance of a 500pF capacitance @ 1MHz is 318Ω.

If you have the drain swinging by 600V with respect to the gate, then the instantaneous power transfer across that 318Ω reactance will be 1131W.
If that 600V swing happens only 0.5% of the total cycle time, then the average power transferred by that capacitive reactance will be 5.6W.

Quote from: Spokane 1 on 2016.08.16, 15:44:59
I suppose it would be similar to a parasitic capacitor of some fixed value. I also suppose that it could be simulated with a capacitor between the gate and the drain.

The MOSFET capacitances are not constant!  They vary with voltage... and the Miller's capacitance (C_DG) is subject to a transconductance multiplication. 
Because of these reasons these capacitances cannot be simulated with fixed capacitors.  It is better to build a mock-up of a secondary circuit and measure the currents empirically there (or in a sim with a good MOSFET model) where the MOSFETs are subjected to high dv/dt  conditions.

Quote from: verpies on 2016.08.16, 17:12:46
I don't think it is a sudden drop but a current reversal due to half a cycle of high frequency LC oscillation. 

When the "synchronous diode" opens, the capacitance in series with the transformer's inductance drops precipitously and increases the LC frequency of whatever remains in that "open" circuit.
This is why adding a 10nF capacitor across the "synchronous diode" lowers the frequency of this LC oscillation, so the current can no longer reverse in the 1μs allotted by the "synchronous diode".
The current flowing due to the MOSFET's Gate Leakage is proportional to the voltages & frequencies appearing between the gate and drain/source terminals.  For example, the reactance of a 500pF capacitance @ 1MHz is 318Ω.

[/color].

Dear verpies,

That sounds like a reasonable mechanism to explore. I shall ponder your process proposal. Thank you for the idea.

In my simulation I'm using a 150 nS harvest capture pulse. I can't expand the trace at the moment but if you could zoom into the transition zone of the secondary current you would see that it quits changing only after about 50 nS. In this imaginary circuit any capture pulse longer than 50 ns will not yield additional current for the backend filter capacitor.

So you are using a 1 microsecond capture pulse, If I understand you correctly?

Spokane1

Quote from: partzman on 2016.08.16, 16:51:37

I will also add that during the clamp time in Graham's device when the primary current is basically frozen and the secondary current and flux reverse direction, there are bucking currents in the coil windings that will produce H fields between the core legs and outside the core legs. IMO, these are the fields that Graham has seen with the sensing coils.   

Actually, the more I think about my statement above, I believe it is incorrect. The physical structure of G's core assembly would have bucking currents normally in loaded operation so H fields should be present in and around the cores most of the time.

pm

Quote from: partzman on 2016.08.16, 15:12:43
Here is yet another circuit variation to Graham's demo device that is best described as an unbalanced current driven parallel resonant inverter with clamp phase. IMO, this would be an excellent alternative to a full H bridge circuit and the primary switching is ground referenced.

The first pix shows the schematic and plot with some .meas math to show the overall COP from supply to output.

The second pix is a rather busy plot with the various branch power measurements.

The third pix is a simplified plot showing the differential voltage across the primary L4, the power across L4, the current thru L4, and mostly resembles Graham's primary waveforms.

pm   

I think it is very interesting (and hilarious!) that you have almost exactly duplicated the primary-side circuit of my microQEG now. 

Quote from: TinselKoala on 2016.08.16, 20:11:36
I think it is very interesting (and hilarious!) that you have almost exactly duplicated the primary-side circuit of my microQEG now.

Dear TK,

Would your front end circuit work in this application? Would it be cheaper? faster? more current capacity? easier to build?

I certainly am not concerned with what kind of circuit drives the Conversion Transformer. k4zep is considering using a fast SCR, which will probably work fine. I'm also not so sure that $70 SiC power MOSFET's are needed for the front end either.

But I shall have to get my setup up and running before I can definitively address these questions.

It is nice to see that great minds are moving in the same direction.

Spokane1

In confirming the branch currents and power levels in the GG6 device I posted in message #404, I discover that the device does not yield negative power in the primary as the measurements would suggest. Again when power polarities are carefully considered at each node, the power in the primary is positive and therefore the design is not a candidate for replicating Graham's device IMO.

Although I carefully checked the GG3 device, I'm presently in the process to confirm that it does have a net negative energy in the primary as the measurements seem to indicate.

Sorry for the distraction. :(

pm 

Quote from: partzman on 2016.08.16, 21:42:52
In confirming the branch currents and power levels in the GG6 device I posted in message #404, I discover that the device does not yield negative power in the primary as the measurements would suggest. Again when power polarities are carefully considered at each node, the power in the primary is positive and therefore the design is not a candidate for replicating Graham's device IMO.

Although I carefully checked the GG3 device, I'm presently in the process to confirm that it does have a net negative energy in the primary as the measurements seem to indicate.

Sorry for the distraction. :(

pm

As best I can tell at the moment, the GG3 circuit I originally posted on reply #386 does indeed have a net negative reactive input across the primaries over one complete cycle as it stands. I would certainly like confirmation whether my analysis is correct or incorrect so it can be determined if the device is worth pursuing.

pm

Quote from: Spokane 1 on 2016.08.16, 20:32:02
Dear TK,

Would your front end circuit work in this application? Would it be cheaper? faster? more current capacity? easier to build?

I certainly am not concerned with what kind of circuit drives the Conversion Transformer. k4zep is considering using a fast SCR, which will probably work fine. I'm also not so sure that $70 SiC power MOSFET's are needed for the front end either.

But I shall have to get my setup up and running before I can definitively address these questions.

It is nice to see that great minds are moving in the same direction.

Spokane1

Partzman's GG3 voltage sources V7 and V3, which drive the gates in his primary circuit, are rectangular-pulsed sources with different ON-time parameters. See the netlist items below.

My apparatus uses a symmetrical, sinusoidal gate drive, and is auto-resonating at the LC tank frequency rather than having its timing set by external gate drive pulses. Is this applicable to the Gunderson issue? Well, I don't know but maybe, since it can produce a "zero" or very close to zero "input power" at the primary coil of the air-core loosely coupled output transformer. I've also gotten some amazing results by assembling a CRT yoke ferrite core around/through the device's primary coil, with a many-turn secondary wound onto half of the ferrite yoke, and generated some very high currents that way, enough to sustain quite a hot arc. My apparatus runs on 12-18V DC input and is entirely self-oscillating, doesn't need any external clocks. Although I suppose an interrupter could be installed that would allow the "dead time" interval seen in Gunderson's waveforms.

As far as cost, this would be set by the price of the mosfets and the magnetics, mostly, I should think. Faster? Again, my circuit's frequency can be set by the LC parameters, and can be anywhere from, say, 20 kHz up to about 1 MHz. Of course as the frequency goes up more care needs to be taken with layout and construction. More current capacity? Depends on mosfets. Easier to build? Easier than the original Gunderson circuit, probably. No SMD components or unobtanium ferrites used, "printed" circuit board so simple it's actually funny. See the attached template.

You don't believe in the magic of 70 dollar SiC mosfets? That's funny.... neither do I !     :D

Quote from: partzman on 2016.08.16, 23:42:51
As best I can tell at the moment, the GG3 circuit I originally posted on reply #386 does indeed have a net negative reactive input across the primaries over one complete cycle as it stands. I would certainly like confirmation whether my analysis is correct or incorrect so it can be determined if the device is worth pursuing.

pm

How did you do your analysis? Walk me through it.

I'm estimating about 0.266 W average input power from the 20V source, and about 0.231 W average power dissipated in the 50R load resistor.

Quote from: TinKoa
You don't believe in the magic of 70 dollar SiC mosfets? That's funny.... neither do I !     :D

Back in the Old Days we used a Carborundum Crystal and a D Cell
for "bias" in some of our Crystal Radios.  Weren't any magic there
either... :'(  C.C  :o

Maybe the "magic" is in the selling of those unusual parts and related
"secrets of" manuals/videos to the eager "builders" in order to reap the
"free energy" of easy profits. 8)  ;)

Quote from: Spokane 1 on 2016.08.16, 17:59:50
So you are using a 1 microsecond capture pulse, If I understand you correctly?

Yes. This period was chosen based on statements made by GG  in his MIT video.

Quote from: muDped on 2016.08.17, 05:13:39
Back in the Old Days we used a Carborundum Crystal and a D Cell
for "bias" in some of our Crystal Radios.  Weren't any magic there
either... :'(  C.C  :o

Maybe the "magic" is in the selling of those unusual parts and related
"secrets of" manuals/videos to the eager "builders" in order to reap the
"free energy" of easy profits. 8)  ;)

You think?   :-X  :o  ;)

https://en.wikipedia.org/wiki/Silicon_carbide

Illegitimi non carborundum  ;)

https://en.wikipedia.org/wiki/Illegitimi_non_carborundum