Magnetic Implosion Transformer Replication

[Spokane1]

Gentlemen,

I was at the conference as a guest of Aaron and what Graham presented was no less than a public demonstration of a potential age changing technology and few people realized it. I thought for sure that many attendees would be chasing him down to get his autograph but it didn't happen.

Dumb me, I thought he was going to do a Part II of the complicated transformer system he demonstrated at last years conference which was important in its own right. So, I didn't plan to take photos or notes. I was hoping that he improved the COP od that system from 1.02 to maybe 1.05.

I helped him carry in his set up. Just for good luck I plugged in the solder iron and had it on standby. His system came right up with only some minor tuning required.

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.

I have used up all my documentation space so I can't post the one photo I took just after he shut it off.

The device takes low voltage DC in and outputs DC so COP measurements for the rest of us can probably be done with Harbor Freight DVM's - at least at with those COP values.

Graham has no idea why this thing works as well as it does. He seemed to be in a daze or sort of a tranquil state of disbelief during the conference. I suppose the rest of us might be passing out cigars and doing cart wheels.

The best part is that Graham plans to make all of this "Open Source" so that others can build and improve upon these principals. He has a web site for this material but I don't believe there is much on it at the moment. I have offered to draft any schematics he would like to post. Right now he wants to take some time off and be with his family.

He showed all the various wave forms on the presentation screen and offered his best understanding as to how it worked (but no idea as to why). The source appears to be a 12.7 A-hr lead-acid gel battery that feeds an "H bridge" chopper that then goes to a step up transformer. The peak excitation voltage is around 800 volts. The input is a single sine wave followed by 1/3 cycle of zero voltage and then repeats. This is the input to his custom conversion transformer which is composed of two large "C" Ferrite core pieces. They don't have the same properties. One has a high permittivity and the other has a lower value. The cores are beefed up with about 20 Manganese Ferrite PM's (2" x 3" x 0.5". The primary is wound on the upper core while the secondary is on the bottom. The secondary is composed of four parallel turns of about #10 AWG Litz wire in three loops. (The good stuff). The output is rectified with a synchronous diode made of Si-C FETS and a custom snubbing driving network composed of zenor diodes so that no gate resistor was used.
There is also a timing network involved to coordinate the various cycles, but no microprocessor. The output is then filtered with a large capacitor network and then drives the automotive lamp.

If the instrumentation is correct he should be able to loop the system and have a self runner, but he ran out of time. This device only became operational last Wednesday (7/6/2016). On the week end he submitted a provisional patent.

-more to follow-

[Spokane1]

Additional Details and Speculations

-continued

The ferrite "C" cores also have a total gap of 0.010" which is composed to two separate spacers of 5 mil Mylar sheeting.

There was no attempt to measure the temperature of the cores given the time frame.

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.

This is not a cheap device to build (at the moment) you will need four (4) FETS that will cost $70.00 each. I would estimate a person would be into the whole set up for around $1,500 US. But that is far cheaper than Peter's jumbo Bedini motor that he demonstrated at around $3,500. Now the price of instrumentation is a different matter.

Graham reports that he got the idea last November when he couldn't sleep. The inspiration came to him to take the components of a two year old failed experiment and move forward with it. He retrieved the box of parts from his storage van and started in. (My dates could be all wrong)

Aaron has be a great help to Graham both in the documentation and the invitation to speak. Aaron provided a set of the expensive FET's in Grahams time of need a few months ago. Aaron has probably hit the jack pot as a far as a subject goes for publishing goes. I would hope the rest of the world would take an interest in this subject and buy his books and DVD's.

So get your order in for Graham's lecture so that Aaron will fix that one up first or second.

One reason for the muted response to Graham's demonstration could be all the other great presentations. John Bedini was disclosing his work with the Royal Rife technology. Peter Lindeman was displaying a huge Bedini engine/motor. We also had Morey King and his take on the Henry Moray technology. Plus several other great speakers. I was to busy with Graham to attend all of the lectures.

There was only one young man that appeared to take a serious interest in what Graham was doing. He was taking notes and asking important technical questions. Graham offered all the technical detail this individual could copy. By rights there should have been 20 people clustered around him. Maybe there will be next year.

What I find interesting when comparing it to my own interests in the E.V. Gray technology is that there is a lot in common. To me this points to possibility of a shared fundamental physics. The Gray technology was measured to have a COP of 282 at the 7.5 kW level so maybe Grahams system could be up scaled several times and the COP improved as well. The Gray system operates using huge current pulses, Graham's system is continuous. The Gray system's main focus was repulsion force (torque) while Graham's device seems to output classical electricity (but I don't know that for sure - the lamp seemed normal enough with no exotic glows). The fundamental frequency of operation of my Gray transformer is 46 kHz. Gram's system is running at 50 kHz but he is using ferrite. Gray used common steel transformer laminations, so there is hope that the high u core can be replaced something more accessible if commonality hold true.

I suspect that the operational theory may be something like what McFreey proposed, at least the excitation of the conversion transformer has these elements in it. That is a huge magnetic pulse followed by a specific RF burst. what happens after that is any bodies guess. Graham's system follows what McFreey said how the OU is contained in a short but powerful output voltage pulse from the secondary. Who knows, but it is a place for me to start.

I have know Graham for about 10 years since we live in the same city. I'm trained as a classical EE, but I know very little compared to his knowledge of magnetics and electricity. I can barley ask descent questions when we get together. He is an excellent teacher and can sketch with the ability of a cartoonist.

I know of no other Free Energy researcher who would discover such a profound and powerful process and then give it away.

Mark McKay, PE

[BroMikey]

Thanks Mark

That is very exciting especially when you sited the split
permeability of the "C" cores of Ferrite (if I understood you)
There is so much to say about your post and the details of
this invention as delivered.

I am glad someone was there who was able to see this jewel
on display and take great notes.

[Reiyuki]

some goodies

OK, here's a few initial goodies, as I documented plenty this time

Imgur: The most awesome images on the Internet


* Input is an H bridge set up as a charge pump, 200-300v constant current source, and I'm sure lots of efficient circulating current.
* Input is not driven like a normal ZVS, there is a rest period. The input drive is closer to 33% A, 33% B, 33% off.

* Output is a synchronous rectifier with HV carbide MOSFETs, normally closed but opened for a very small part of the total cycle (somewhere around/near the A/B phase transition of input circuit). The EMF spike can reach the 1-2kv, and the circuit is re-engaged near the peak of this spike.
* The filtered output is constant DC with a spike at every interruption.


* System operation is around 20-50kcps, with constant output range around 5-10 watts. Input can be tuned allowing COP to undefined or negative, with reduced total output. There are cutoff regions above and below this freq range where the effect stops manifesting.
* The demonstration (I believe) was about 1.5w in and around 8.5w out, and he mentioned it can be tuned for different results. The power level seems to be limited to saturation of the ferrite.
* Control circuitry appeared to be a ring circuit of hex inverters for pulses and timing.
* Output is wound to maximize L and minimize R (RF style winding)
* All wire is Litz. It handles well at RF but is unsure whether it is required for operation.
* There are magnets placed above the primary to partially saturate the primary ferrite to the 'knee'.


Since *he* doesn't really know why it works, I'm not going to bother speculating either. Just posting the details I'm pretty sure about. Hopefully this and the attached images can give you a better idea on the setup. It's all still pretty bleeding-edge stuff.


Anyone know how it works? I'm sure we'd all like to know (including him!)

[k4zep]

I will follow this with interest and attempt to duplicate when enough information becomes available!

Ben

[Reiyuki]

I will follow this with interest and attempt to duplicate when enough information becomes available!

Ben

Anyone working on TK/Kapagen/Akula systems or homemade Tesla coils is going to have most of the circuitry needed for this, though it may need modification.
HV inputs with large resonant circulating currents and sharp HV kicks that often make for expensive learning experiences on how to drive them. The bane of SSTC's everywhere.



Magnetically, input ferrite is high permeability, output ferrite is low permeability. Input coil is partially saturated and surrounded by ceramic magnets all pointing inward.


Electrically, input is a resonant tank circuit controlled with an H bridge, configured into something that looks like 'Push-Pull-Sleep' (rather than push/pull).
Output circuit is a synchronous rectifier that is normally closed and is opened just before the push/pull transition above. I hope the timing diagram made a little sense from what I posted. Output is DC with some kicks (SM's ears must be burning)



If you ask Graham, he will say outright he doesn't know why it works, but will go into detail over what principles he thinks are important to the process and how it *might* work. But again, it's all bleeding edge and he did not want to accidentally make a modification that breaks the device just before the conference.

He named it a 'magnetic implosion transformer', which by itself should be a big clue as to how he thinks it works.

[Spokane1]

Sourcing Parts

Dear Reiyuki,

Are you the young man that asked Graham all those technical questions at the conference. You are about the only person that would have this much current knowledge of the technical detail of this device. I'm glad you got started I chose not to hound Graham for all those details at the time since I think it is going to take hours to document all the needed specifications to reproduce this.

By the way, do you plan to build one of these yourself? I figure it will cost at least $1,500 to do it the way he did IF we can get all the right parts (like the "U" cores). Would you consider collaborating on sourcing some of the parts.

I'm looking into having some of those parts fabricated, like the acrylic end blocks . Once the setup fee is paid making more copies is all most just the cost of material. Would you like me to keep you in mind should I put in an order for some of these parts.

I'm not going to go out for quotes just yet until I get all the dimensions drafted.

I plan to get a hold of Graham this week and hopefully get a partial list of the major components needed so that I can get them coming. I shall post those part numbers as soon as I get them.

I really appreciate those nice annotated photos you took and shared.

Do you know of anybody else that took a serious interest in this? I didn't see anyone. I suppose every one else is waiting for all the information to show up on the web site that Graham said he would place it on. Do you recall what that web site was? It started with a "p" and ended with .green. I don't think there is anything on it right now, but I can't check it since I failed to write it down.

If there are only three of us (at the moment) taking this technology seriously then I doubt that Teslagenicx will invest the money to fabricate PC boards for the ring timer, the H bridge, and the synchronous diode any time soon.

Mark McKay, PE

[Reiyuki]

Dear Reiyuki,

Are you the young man that asked Graham all those technical questions at the conference. You are about the only person that would have this much current knowledge of the technical detail of this device.

Yep, guilty. I'm the one you're thinking of that was asking all the detailed replication-specific questions. I have a lot more information than I've posted so far, but like him I also have a day job, and I wanted to start with the material I am 99% confident is valid and accurate.

Do you plan to build one of these yourself? I figure it will cost at least $1,500 to do it the way he did IF we can get all the right parts (like the "U" cores). Would you consider collaborating on sourcing some of the parts.

An exact replication, probably not. I intend to follow the same concepts, using variable saturation in dissimilar ferrites with external flux fields. . If you're doing a 100% replication, the most expensive parts will be the actual circuitry (there is a lot of it), while the hardest-to-find parts will be those specific ferrites. ~1-2k sounds about right, along with several months of construction.

Do you know of anybody else that took a serious interest in this? I didn't see anyone.

At the conference, the work I do was a close match to what he's been doing. Most of the other tinkerers there were in the motor and HHO fields, which is simply a different line of work.

Graham doesn't know it, but on the internet several of the Russian experiments (TK/Akula/Kapagen stuff) have electronics that are VERY similar to his. They have the same problems keeping MOSFETs from blowing up, and using sharp high voltage gradients for results. The people replicating in that field are in the best position to replicate Graham's stuff.


The control circuitry is IMHO the most novel part of Graham's system. The Push/Pull/Sleep driver on input, and the interrupted synchronous rectification on the output. Genius. Those TK/Kapagen groups may be able to apply these concepts to their circuits to great effect.

I suppose every one else is waiting for all the information to show up on the web site that Graham said he would place it on. Do you recall what that web site was? It started with a "p" and ended with .green. I don't think there is anything on it right now, but I can't check it since I failed to write it down.

grahamgunderson.com is the only one I'm aware of, the other one is floating in a notebook out of my reach at the moment. I'll update this post later if I find the site you're talking about.

If there are only three of us (at the moment) taking this technology seriously then I doubt that Teslagenicx will invest the money to fabricate PC boards for the ring timer, the H bridge, and the synchronous diode any time soon.

It's all cutting-edge anyway so we don't even know the best circuits to use. Some parts of Graham's system might end up being 1000% overengineered; we might be able to get some of the same effects with a $10 premade H bridge driver and the proper control circuitry.


IMHO, the best thing for people to play with in the short term is with 'interrupted synchronous rectifiers' on the output of various transformers, along with a 'sleepy H bridge' on the input.

[Spokane1]

Dear Reiyuki,

I also have a day job. Ben is retired but only has a 2' x 4' work space.

I'm delighted that you have such a broad knowledge of those Russian technologies. I don't get out that much. Also I suspect that you are quite technically qualified yourself in order to begin to ask those kinds of questions and to consider other solutions and methods for the circuit development.

I hope that some of those sub-systems are 1000% over engineered because as it is I don't think many garage scientists are going to attempt this technology right now.

So you plan to go in your own direction using the principles that you gained from the conference. If that is the case then I don't suppose you will be in much need of exact fabricated parts for the conversion transformer. That's fine. But if you happen to know where we can get those "UU" cores with the right permittivity please feel free to clue us in.

It will be interesting to see what you find in your own area of exploration. It is always good to get a different perspective on what is going on.

Mark McKay

[k4zep]

Anyone working on TK/Kapagen/Akula systems or homemade Tesla coils is going to have most of the circuitry needed for this, though it may need modification.
HV inputs with large resonant circulating currents and sharp HV kicks that often make for expensive learning experiences on how to drive them. The bane of SSTC's everywhere.

Good Morning Reiyuki,

Ben here. I have followed the T/Kap/Aukula systems gentlemen for years
so have lots of their schematics and general ideas of how their systems operate. I also susupect that This system, Bedini's systems all, when they work, are working for the same reason.



Magnetically, input ferrite is high permeability, output ferrite is low permeability. Input coil is partially saturated and surrounded by ceramic magnets all pointing inward.

So one half of the C is on the top side where the primary coils is wound (high Perm) and biased with the magnets the move i'ts knee to where the pulsed waveform can probably kick it over and under that point and the as you call it "Push-Pull-Sleep" or "Forward-Reverse-Off providing that big kick back pulse as you said should be a no brainer to reproduce with a minimum of parts when resonant frequency of the input coil is known. Then the bottom one half is a lower perm material. Do you remember if it was also resonant, I doubt it due to the synchronous rectification? It would appear to be a higher "C" circuit, that is less inductance, more Cap. if it is resonant. The most important thing here is he has found a system that appears to work and is open about it!!!!!


Electrically, input is a resonant tank circuit controlled with an H bridge, configured into something that looks like 'Push-Pull-Sleep' (rather than push/pull).
Output circuit is a synchronous rectifier that is normally closed and is opened just before the push/pull transition above. I wonder if this allows a sort of reset in the residule field by removing the load for a short time. I hope the timing diagram made a little sense from what I posted. Output is DC with some kicks (SM's ears must be burning)

If you ask Graham, he will say outright he doesn't know why it works, but will go into detail over what principles he thinks are important to the process and how it *might* work. But again, it's all bleeding edge and he did not want to accidentally make a modification that breaks the device just before the conference. He is a true "Tec"...don't muck with a working circuit!

He named it a 'magnetic implosion transformer', which by itself should be a big clue as to how he thinks it works.

Reiyuki, I meander a bit as I tend to think out loud as I comment on you information. If and when you get part numbers of those H bridge drivers it would be a great help. As I was not there, Frequency, current etc. into the input coil would be great information. One thing I question is when the input power decreased over time but the output remained the same, any thoughts as to what was happening?

Respectfully
Ben

[k4zep]

Gentlemen,

Mark, a observation below.

I was at the conference as a guest of Aaron and what Graham presented was no less than a public demonstration of a potential age changing technology and few people realized it. I thought for sure that many attendees would be chasing him down to get his autograph but it didn't happen.

Dumb me, I thought he was going to do a Part II of the complicated transformer system he demonstrated at last years conference which was important in its own right. So, I didn't plan to take photos or notes. I was hoping that he improved the COP od that system from 1.02 to maybe 1.05.

I helped him carry in his set up. Just for good luck I plugged in the solder iron and had it on standby. His system came right up with only some minor tuning required.

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.

I have used up all my documentation space so I can't post the one photo I took just after he shut it off.

The device takes low voltage DC in and outputs DC so COP measurements for the rest of us can probably be done with Harbor Freight DVM's - at least at with those COP values.

Graham has no idea why this thing works as well as it does. He seemed to be in a daze or sort of a tranquil state of disbelief during the conference. I suppose the rest of us might be passing out cigars and doing cart wheels.

The best part is that Graham plans to make all of this "Open Source" so that others can build and improve upon these principals. He has a web site for this material but I don't believe there is much on it at the moment. I have offered to draft any schematics he would like to post. Right now he wants to take some time off and be with his family.

He showed all the various wave forms on the presentation screen and offered his best understanding as to how it worked (but no idea as to why). The source appears to be a 12.7 A-hr lead-acid gel battery that feeds an "H bridge" chopper that then goes to a step up transformer. The peak excitation voltage is around 800 volts. The input is a single sine wave followed by 1/3 cycle of zero voltage and then repeats. This is the input to his custom conversion transformer which is composed of two large "C" Ferrite core pieces.



They don't have the same properties. One has a high permittivity and the other has a lower value. The cores are beefed up with about 20 Manganese Ferrite PM's (2" x 3" x 0.5". The primary is wound on the upper core while the secondary is on the bottom. The secondary is composed of four parallel turns of about #10 AWG Litz wire in three loops. (The good stuff). The output is rectified with a synchronous diode made of Si-C FETS and a custom snubbing driving network composed of zenor diodes so that no gate resistor was used.
There is also a timing network involved to coordinate the various cycles, but no microprocessor. The output is then filtered with a large capacitor network and then drives the automotive lamp.

If the instrumentation is correct he should be able to loop the system and have a self runner, but he ran out of time. This device only became operational last Wednesday (7/6/2016). On the week end he submitted a provisional patent.

-more to follow-

So if I understand correctly, the H bridge chopper is in the LOW voltage side of the circuit that is....Battery--Chopper---Step up transformer---into the Special transformer? Assuming the output of the step up transformer and input side of the "implosion" transformer is resonant, I wonder if when it is dead on resonance and the current into the input chopper transformer drops to almost zero or greater than than that, it is because there is feedback of extra power backwards that causes it to drop even further than normal? If that assumption is correct, it would appear that the demonstration was initially tuned slightly off resonance but over time drifted into almost perfect resonance. A simple micro where you could adjust the frequency very precisely along with the "Off" time could quickly optimize it

Does anyone remember if that step up transformer is a 12VAC to 800VAC? When I originally read the start of this thread, I thought the H bridge had to work at 800 VDC which would have been a real bear!

Ben

[aniccame]

h bridge

This circuit may be of interest:
DC Transformer

[Reiyuki]

So one half of the C is on the top side where the primary coils is wound (high Perm) and biased with the magnets the move i'ts knee to where the pulsed waveform can probably kick it over and under that point and the as you call it "Push-Pull-Sleep" or "Forward-Reverse-Off providing that big kick back pulse as you said should be a no brainer to reproduce with a minimum of parts when resonant frequency of the input coil is known.

Yes. I like the 'forward-reverse-off' description better as well and will use that term from now on.
He said on this setup the effect would show up in a fairly wide band, 20-50kcps.

Then the bottom one half is a lower perm material. Do you remember if it was also resonant, I doubt it due to the synchronous rectification?

From the photos I have, the only output connections I see are to those massive electrolytic caps on the left, so you're probably right about that.

It would appear to be a higher "C" circuit, that is less inductance, more Cap. if it is resonant. The most important thing here is he has found a system that appears to work and is open about it!!!!!

The input has HV capacitors on it. It looks vague like a ZVS like in a Tesla Coil, except the FETs are driven manually.

Output circuit is a synchronous rectifier that is normally closed and is opened just before the push/pull transition above. I wonder if this allows a sort of reset in the residule field by removing the load for a short time.

Or, maybe it reflects the flux back and the high perm ferrite sucks it up and gives it back the rest of the cycle? IE: using the field, then returning it. Graham says he doesn't really know either. Since it's 20-50khz, magnetostrictive effects might also be responsible for it.

If and when you get part numbers of those H bridge drivers it would be a great help. As I was not there, Frequency, current etc. into the input coil would be great information. One thing I question is when the input power decreased over time but the output remained the same, any thoughts as to what was happening?

C2M0025120 (1200v with VERY low ON resistance and VERY fast switching).
Amazing carbide MOSFETs. Fragile though, so you'll want lots of protection on the gate to keep the magic smoke inside. Look for the HV russian schematics or Solid State Tesla Coil guys for ways to drive and protect the FETs.

Input: the push/pull is tuned resonant and the OU effects manifests around 20-50khz. It is an H bridge running at 200-300v with a big inductor in series so it becomes a constant current driver.

The total output was always around 5-10w, and COP would vary depending on what point you triggered the interrupter.

Imagine the entire cycle as 360deg, and lets say the Push/Pull pushes from 0-120 and pulls from 120-240.

In that example, the output interrupt occurs around 100-140deg and lasts 100nS-2uS (circuit reengaged just after the pulse peaks, which is ~1200v or as far as the FETs will safely handle).
If I understand him correct, varying the phase would cause changes the COP and the total outputs, so changing the phase interrupt would cause the I/O to look something like this:
100deg 5w in 4w out
110deg 4w in 8w out
120deg 3w in 6w out
130deg 0w in 4w out
140deg -1w in, 3w out

IIRC on his demo, he was interrupting just before 120deg, resulting in ~3.5w in and ~8.5w out.


We're still not sure what components of his system are critical and which are superfluous. It might be possible that the entire thing could be replicated with a single magnet and a split-core Joule Thief that is driven in this unique way. I have not seen anyone playing with a combination of interrupted outputs and PWM driven inputs.



Hope that answers everything, feel free to ask anything else. I may not have a working system but at least I documented one that did

[k4zep]

Yes. I like the 'forward-reverse-off' description better as well and will use that term from now on.
He said on this setup the effect would show up in a fairly wide band, 20-50kcps.

Excellent.

From the photos I have, the only output connections I see are to those massive electrolytic caps on the left, so you're probably right about that.

The input has HV capacitors on it. It looks vague like a ZVS like in a Tesla Coil, except the FETs are driven manually.

Did he have a normal 12VDC to 220VAC power supply rectified and then the output is fed into "special" transformer in a resonant mode via the H-Bridge?. I need a bit more clarification as to the basic input circuit. By "manually" do you mean that they were driven via the "ring" counter circuit? Also, is the inductance of the input side of the "transformer" being fed in a parallel resonance condition with a "C" across it, what allows it to appear in the constant current mode? I too wonder if magnetorstrictive effects of the cores are partially responsible for some of the effects seen. Another question, which seem to be coming along while I type. I have to assume the H Bridge is in a all fets off condition during the zero or off time which should allow the coil to sort of reset or settle in some way.



Or, maybe it reflects the flux back and the high perm ferrite sucks it up and gives it back the rest of the cycle? IE: using the field, then returning it. Graham says he doesn't really know either. Since it's 20-50khz, magnetostrictive effects might also be responsible for it.

If we are operating @ 20 khz, the period for one cycle or push/pull would be 50us. Then I would assume we would have a "Off" time of say another 25us.


C2M0025120 (1200v with VERY low ON resistance and VERY fast switching).
Amazing carbide MOSFETs. Fragile though, so you'll want lots of protection on the gate to keep the magic smoke inside. Look for the HV russian schematics or Solid State Tesla Coil guys for ways to drive and protect the FETs.

I looked the 25120 up and it is available anywhere between $65 and $70 bucks. There are several others that are rated the same voltages but less peak current like the CMF10120 24amps @$16.63 or a C2M01160120D 17.7 amps @ $8.32. with the rest of the parameters close enough.
Are these HV fets used in both the input and output circuits???

Input: the push/pull is tuned resonant and the OU effects manifests around 20-50khz. It is an H bridge running at 200-300v with a big inductor in series so it becomes a constant current driver.

The total output was always around 5-10w, and COP would vary depending on what point you triggered the interrupter.

Imagine the entire cycle as 360deg, and lets say the Push/Pull pushes from 0-120 and pulls from 120-240.

Understood

In that example, the output interrupt occurs around 100-140deg and lasts 100nS-2uS (circuit reengaged just after the pulse peaks, which is ~1200v or as far as the FETs will safely handle).The circuit could be made to detect that fast rising pulse and turn the output sync. detector back to load the circuit back down preventing over voltage across the input Fets. Do you remember if the HV showed up mostly in the Input side of the circuit FETS, and the clamping effect of turning the output back on held that pulse within reason, It dosn't seem it would be that high in the output side. So much interaction here. and the core is really "hit" if the voltage is at say 1000V Peak when you clamp the output (hence the input).

It would seem that the "Off" pulse to the synchronous detector is definitely less than the 50us period time of one full cycle in the transformer looking at the wave forms. It would also seem that the OFF time is very critical in that you turn it back on before the "spike" goes too high and blows the FET's!
If I understand him correct, varying the phase would cause changes the COP and the total outputs, so changing the phase interrupt would cause the I/O to look something like this:
100deg 5w in 4w out
110deg 4w in 8w out
120deg 3w in 6w out
130deg 0w in 4w out
140deg -1w in, 3w out

Varying the phase would simply determining how high the HV pulse got before you clamped it back down via the output network.

The waveform on the scope didn't show the width of the "Off" time in the output. so guessing here. I bet it was pretty short, basically determined on the rise time of the HV pulse. As soon as you re-clamp, everything drops to zero. I wonder why he doesn't clamp on the negative side of the sine-wave?

IIRC on his demo, he was interrupting just before 120deg, resulting in ~3.5w in and ~8.5w out.


We're still not sure what components of his system are critical and which are superfluous. It might be possible that the entire thing could be replicated with a single magnet and a split-core Joule Thief that is driven in this unique way. I have not seen anyone playing with a combination of interrupted outputs and PWM driven inputs.

You just might be correct! As the transformer is split, if you bought two toroid forms one high Perm., one low Perm, split them, you could make two complete transformers cores with the gap between the two half cores.

Hope that answers everything, feel free to ask anything else. I may not have a working system but at least I documented one that did

Thank goodness as time will show!

Enough for now!

Ben

[Reiyuki]

All: I posted a very rough block-model of how I think the circuitry is laid out at the bottom of this post. It is wrong, I'm sure, but probably close to the actual setup. Hopefully it should make things a little bit clearer before Graham smacks me upside the head with the actual one.

Did he have a normal 12VDC to 220VAC power supply rectified and then the output is fed into "special" transformer in a resonant mode via the H-Bridge?.

Yes. The input H bridge is being used as a charge pump, much like a ZVS.

I need a bit more clarification as to the basic input circuit. By "manually" do you mean that they were driven via the "ring" counter circuit? Also, is the inductance of the input side of the "transformer" being fed in a parallel resonance condition with a "C" across it, what allows it to appear in the constant current mode? I too wonder if magnetorstrictive effects of the cores are partially responsible for some of the effects seen.

Yes, they were driven by the ring. It looked like a string of hex inverters with RC delay lines, so you could tune 6 'phases' and trigger the interrupter.

Another question, which seem to be coming along while I type. I have to assume the H Bridge is in a all fets off condition during the zero or off time which should allow the coil to sort of reset or settle in some way.

This I don't know. It is either driven 33%-33%-off, or I misinterpreted the setup. It might also be driven 33%-66%, and it *might* be a plain old bridge. Whatever explains the 3 phases in the scope shot.

I looked the 25120 up and it is available anywhere between $65 and $70 bucks. There are several others that are rated the same voltages but less peak current like the CMF10120 24amps @$16.63 or a C2M01160120D 17.7 amps @ $8.32. with the rest of the parameters close enough.
Are these HV fets used in both the input and output circuits???

Maybe? He implied that the bigger the spike you could get, the better, and explicitly said those 1.2kv rated FETS were actually good up to 1.7kv. It implies he was trying to maximize the interrupt spike.

If we knew the actual principle and critical components, we could probably just use IRF840's. I figure we start with the cheap stuff and look for changes in efficiency that do not fit the standard models.

In that example, the output interrupt occurs around 100-140deg and lasts 100nS-2uS (circuit reengaged just after the pulse peaks, which is ~1200v or as far as the FETs will safely handle).The circuit could be made to detect that fast rising pulse and turn the output sync. detector back to load the circuit back down preventing over voltage across the input Fets. Do you remember if the HV showed up mostly in the Input side of the circuit FETS, and the clamping effect of turning the output back on held that pulse within reason, It dosn't seem it would be that high in the output side. So much interaction here. and the core is really "hit" if the voltage is at say 1000V Peak when you clamp the output (hence the input).

A damn fine question. I wish I had asked him that one. The answer to that is actually pretty important, because it would tell us what is happening to the flux inside the circuit. If we get spikes back to the input, we'll know there is flux being counter-induced in the primary. If we only get spikes on the HV side, we'll know there is a one-way action and that the output portion is the critical part.

It would seem that the "Off" pulse to the synchronous detector is definitely less than the 50us period time of one full cycle in the transformer looking at the wave forms. It would also seem that the OFF time is very critical in that you turn it back on before the "spike" goes too high and blows the FET's!

He reengages just after the pulse peaks, with the entire spike and recovery taking less than 1uS. I assume the input is carefully dialed up to reach a safe peak level.

The waveform on the scope didn't show the width of the "Off" time in the output. so guessing here. I bet it was pretty short, basically determined on the rise time of the HV pulse. As soon as you re-clamp, everything drops to zero. I wonder why he doesn't clamp on the negative side of the sine-wave?

More than 10nS, less than 1uS, that's all I remember. It's in the scope shot but not visible as it is too narrow. It's at the 'knee' where you start seeing vertical lines on the input and output current.

You just might be correct! As the transformer is split, if you bought two toroid forms one high Perm., one low Perm, split them, you could make two complete transformers cores with the gap between the two half cores.

A counter-question: Do we have any ways to artificially decrease permeability of a ferrite? Maybe heating red-hot with a blowtorch? That would make it a lot easier to source the materials, as we'd just need 2 U cores.