PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2020-08-06, 10:21:38
News: A feature is available which provides a place all members can chat, either publicly or privately.
There is also a "Shout" feature on each page. Only available to members.

Pages: 1 2 3 4 5 6 7 8 9 [10] 11 12 13 14 15 16 17 18 19 20 21 22 23
Author Topic: Graham Gunderson Energy conference High COP demonstration  (Read 157706 times)

Group: Elite Experimentalist
Hero Member
*****

Posts: 2339

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
   
Group: Guest
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
   
Group: Guest
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.

   

Group: Professor
Hero Member
*****

Posts: 2578
Everyman decries immorality
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.


---------------------------
Everyman Standing Order 01: In the Face of Tyranny; Everybody Stands, Nobody Runs.
Everyman Standing Order 02: Everyman is Responsible for Energy and Security.
Everyman Standing Order 03: Everyman knows Timing is Critical in any Movement.
   

Group: Professor
Hero Member
*****

Posts: 2578
Everyman decries immorality
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)


---------------------------
Everyman Standing Order 01: In the Face of Tyranny; Everybody Stands, Nobody Runs.
Everyman Standing Order 02: Everyman is Responsible for Energy and Security.
Everyman Standing Order 03: Everyman knows Timing is Critical in any Movement.
   
Group: Elite
Hero Member
******

Posts: 3639
It's turtles all the way down
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.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Jr. Member
**

Posts: 79
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
   

Group: Professor
Hero Member
*****

Posts: 2578
Everyman decries immorality
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 ?


---------------------------
Everyman Standing Order 01: In the Face of Tyranny; Everybody Stands, Nobody Runs.
Everyman Standing Order 02: Everyman is Responsible for Energy and Security.
Everyman Standing Order 03: Everyman knows Timing is Critical in any Movement.
   
Jr. Member
**

Posts: 79
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
   
Newbie
*

Posts: 29
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
   
Group: Guest
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 ...
Quote
NVE’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?

   
Group: Guest
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.
   

Group: Professor
Hero Member
*****

Posts: 1809
Are there semiconductors available that are fully isolated?
Yes there are, but they require isolated power supplies, too.
   
Group: Guest
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?

   

Group: Elite Experimentalist
Hero Member
*****

Posts: 2339
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
   
Group: Elite
Hero Member
******

Posts: 3639
It's turtles all the way down
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


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Elite
Hero Member
******

Posts: 3639
It's turtles all the way down
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.
« Last Edit: 2016-08-02, 17:09:45 by ION »


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Jr. Member
**

Posts: 79
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
   
Jr. Member
**

Posts: 79
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
   

Group: Professor
Hero Member
*****

Posts: 2578
Everyman decries immorality
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.


---------------------------
Everyman Standing Order 01: In the Face of Tyranny; Everybody Stands, Nobody Runs.
Everyman Standing Order 02: Everyman is Responsible for Energy and Security.
Everyman Standing Order 03: Everyman knows Timing is Critical in any Movement.
   
Jr. Member
**

Posts: 79
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
   
Group: Elite
Hero Member
******

Posts: 3639
It's turtles all the way down
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.

Quote
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.

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.



« Last Edit: 2016-08-02, 21:36:31 by ION »


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Professor
Hero Member
*****

Posts: 1809
Trying to measure the BH curve of my used 10mH/100uH transformer...
What are the inductances of one winding while you short the others?
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 2339

The 100uH secondary becomes 8uH and the 10mH primary becomes 3.5mH.

Itsu
   
Group: Elite
Hero Member
******

Posts: 3639
It's turtles all the way down
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?
« Last Edit: 2016-08-02, 23:55:17 by ION »


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Pages: 1 2 3 4 5 6 7 8 9 [10] 11 12 13 14 15 16 17 18 19 20 21 22 23
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2020-08-06, 10:21:38