PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2019-12-12, 14:16:31
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 154525 times)

Group: Professor
Hero Member
*****

Posts: 1479
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.
   
Group: Experimentalist
Hero Member
*****

Posts: 949
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

         
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
That's a very thoughtful analysis pm.

You also said:
Quote
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.

I agree.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Guest
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.
   
Group: Guest
Here's a Tek pdf primer on Probe Skew.

Did Gunderson mention corrections for probe skew in his presentation?
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
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?


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

Posts: 949
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

 
   
Group: Experimentalist
Hero Member
*****

Posts: 949
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
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
See my post #280. Sorry I was writing while you were posting, pm.


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

Posts: 949
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

 
   
Group: Guest
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.
   
Jr. Member
**

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

Posts: 3637
It's turtles all the way down
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.


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

Posts: 29
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
   
Sr. Member
****

Posts: 325
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 ?
   
Group: Guest
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.
   
Group: Guest
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.
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
TK said:

Quote
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.

Smudge has attempted a possible explanation in a couple of his papers posted earlier.

Forest asked:

Quote
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 ?

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.


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

Posts: 1289
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   
   
Group: Guest
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:

Quote
This 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,
Quote
As 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
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
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.


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

Group: Renaissance Man
Hero Member
*****

Posts: 2162


Buy me a cigar


---------------------------
Nanny state ? Left at the gate !! :)
   

Group: Professor
Hero Member
*****

Posts: 1479
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 and if it did not - I would expect CARA type MO.

Without accurate schematics, nice scopeshots, probe positions and proof of calibration, the ORBO MO remains a possibility, too ;)
   
Group: Experimentalist
Hero Member
*****

Posts: 949
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



 
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
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


---------------------------
"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 2019-12-12, 14:16:31