Zanzal's MultiTapped Transformer Test Platform

  • Last Post 28 November 2018
Zanzal posted this 20 November 2018

So here is a little thing I shared with Chris earlier. He thought others might benefit from it and asked me to share it with a larger audience. I'm not making any claims at this time, but this device has done some strange things. I managed to destroy several 12-15W loads with it in under a second. I would urge caution and remind would be experimenters that this device will generate alternating currents that can be hazardous to one's health...

Here is my circuit summarized:

JS Circuit

I'm not sure what to call my device or even if its worth naming at this point. The long string of inductors is simply a multi-tapped transformer wound on two stacked "nanoperm" toroids. I'll be experimenting with different configurations for the transformer piece so there may be better designs, but I'll include a picture to make it more clear.

The transistors shown in the schematic are IGBTs that do not have a reverse diode..They are controlled by an ESP-32 MCU. 

There is a special firing order involved with the IGBTs. It is:

3 -> 4 -> 2 -> 5 -> 1 -> 6 -> 0 -> 7

When 4 is turned on 3 gets turned off, etc. For a given frequency f, a complete cycle would be completed. So at 60Hz this cycle is completed 60 times in one second. The on time for each step is not equal. The inductors close to the voltage source are on the least time. The timings are:

0 = 0.20, 1 = 0.15, 2 = 0.10, 3 = 0.05, 4 = 0.05, 5 = 0.10, 6 = 0.15, 7 = 0.20

I may also experiment with linear timings, but I don't think they will give better results. There is also an oscillatory version I am using that cycles back:

3 -> 4 -> 2 -> 5 -> 1 -> 6 -> 0 -> 7 -> 0 -> 6 -> 1 -> 5 -> 2 -> 4

To correct for the extra steps each timing is divided by 1.6. To explain this with integer math, lets say you were using an 8Mhz CPU and you could control timing by delaying for a certain number of clock cycles, if you ran this at 1000Hz, you would complete a full cycle in 8000 clocks. So you would turn 3 on for 0.05 * 8000 = 400 cycles. but when using the oscillatory version 3 would be on 0.05 / 1.60 * 8000 or 250 cycles. Anyway this is pretty much all there is to this thing. 

Feel free to build it and experiment. If you do this I recommend starting with the oscillatory pattern, it seems to be the better one. I also have tried other patterns but they are not giving good results. Still there are a lot of permutations possible so feel free to play around with other ideas.

Order By: Standard | Newest | Votes
Zanzal posted this 20 November 2018

Some pictures:

Attached Files

Zanzal posted this 20 November 2018

One difference between the picture and the schematic. I switched to an AC microwave oven bulb since my 12V DC bulbs were getting destroyed. The hardware specifics don't really matter all that much IMO. The main element here is: Timing (and maybe the transformer core). I'm still trying to figure out how important the transformer design is, and also if IGBTs without the anti-parallel diode (but turn off slower) work better than mosfets.

The device has some quirks. I can actually render it non functional by using certain firing patterns. It seems to cause some sort of issue with the transformer cores. If I leave them alone for a bit they reset. Very strange stuff. I may go into detail about some other patterns, but for now I'll just put out helpful information. I thought it worth mentioning though in case people decide to experiment, permanent breakage could be possible, and may be in line with other experimenters who encounter issues after long durations.

Jagau posted this 20 November 2018

Hi Zanzal
Congratulations very nice setup
Just a question, I did not understand what the device does
when you have time, can you explain to me again?



Vidura posted this 20 November 2018

This is a interesting device Zanzal. I just want to comment that I had a similar setup in mind for a solid state version of figueras part G(not the switching orders you used) . As Aetherholic used this type of material in his built and had very good results. It might be worth to make a test with equivalent switching patterns(make before break, at least two switches active simultaneously), this would produce a coilshorting in a sequencial manner. I will not build it for the instance, as I have to many proyects for complete now . Regards Vidura

Zanzal posted this 20 November 2018

Hey Jagau,

For me this is a good platform for research. I can swap out transformers and test different ideas on a single board. I think perhaps my answer to Vidura may help make it more clear.

Hey Vidura,

In my firmware I created 5 basic programs and derived 5 oscillatory versions. By oscillatory I mean the sequence reverses once before stopping at the place it started. The 4 sequences I didn't include above are:

0 -> 1 -> 2 -> 3 -> 4 -> 5 -> 6 -> 7 (Sweep)

3 -> 7 -> 2 -> 6 -> 1 -> 5 -> 0 -> 4 (Interleave+Sweep)

3 + 4 -> 2 + 5 -> 1 + 6 -> 0 + 7 (Twin)

3 + 7 -> 2 + 6 -> 1 + 5 -> 0 + 4 (Twin+Sweep)

All my patterns are reversible, can be oscillated, and can be given linear or weighted timings. I believe weighted is the better approach and so far no testing has revealed differently. I believe what you are describing with coil shorting may be similar to the pattern I named "Twin" above. It causes power to flow through both branches at the same time. This is not necessarily the same as shorting though. It may be possible to implement shorting through bipolar transistors but I don't see this as possible using mosfets or IGBTs since they turn on based on a voltage differential between the source and the gate. I could be wrong though. Something to experiment with later...


Chris posted this 21 November 2018

My Friends,

Thank you all for supporting Zanzal's work as you have! Seven Thumbs Up is the most I have seen.

This is very important work! Massive potential lays in this device and Zanzal deserves all the credit for his genus!

@Zanzal, perhaps a few Parallel Caps may increase the output some more. Reactive Power can be used to increase the Current well beyond the standard Current Level by applying a 12 volt potential across the winding.

I will do a post on this later on.

Awesome Work Zanzal and thank you so much for sharing! Awesome!


  • Liked by
  • Zanzal
Zanzal posted this 21 November 2018

I added another variable to the mix, a duty cycle. If 100 the device operates as described, otherwise it stops the flow of current early for each step. This led to a new discovery. The operating mode twin, actually is potentially more powerful than I had thought, at 100% twin is easily the worst mode, but at between 50% and 75% it appears to outperform the other modes in some respects.

This is a very strange result. I am not sure why this is. I was not entirely surprised when I first started playing with this thing that twin seemed bad since for AC one might expect the voltage potentials to be equal since current is flowing symmetrically.. But something is happening when the duty cycle is reduced. There is of course a slight timing difference because my microcontroller doesn't turn them both on at the exact same time there is a small offset perhaps 1.5us or so.

  • Liked by
  • Chris
  • cd_sharp
Chris posted this 21 November 2018

Zanzal and all,

Youre right, I cant prove this but have suspected this before several times.

It does appear a timing delay, where the switching is not 100% perfect, can increase the chances of coils becoming more advantageous!

I have used a H-Bridge as you know, and the old one gave me better results most of the time, and I believe that's because the switching was not clean and smooth, leaving somewhat of a delay on one or more Coils.

Excellent observation Zanzal!


  • Liked by
  • Zanzal
  • Aetherholic
alohalaoha posted this 21 November 2018

Hey Zanzal

Good work man. You have reinvented the bycles. In positive mean of course. This configuration has remembered me of one very old extraordinary device invented by Japanesse EE Shinichi Seike "Gravity strain energy absorber". Same cyclotrone configuration setup like yours, but he used only 3 transistors switching cycle. He used 3 indpendent signal generators for each transistor with tunable phase shift adjustments.

Maybe you could find some new ideas from his work

There very probably could be some unknown process of energy extraction. Simillar configuration used Stiven Maprk in his TPU and Inogda in his TPU replication. From Inogda's work (i had knew him) i know he had succeed to release very big amount of infra red radiation at output of his TPU with very small current consumption. Simply incandescent light bulb had gave very weak ligh but enormous IR radiation. Then he has simply disappeared several month later. Never heard any news from him because he has lived in Donbas (DNR republic). In that time there was a civil war made by American Cabal Neocones. One one their bloody games in Matrix's virtual reality whre we all live.

Best wishes


  • Liked by
  • Zanzal
  • Chris
alohalaoha posted this 21 November 2018

About the strange cycling of energy flow i have just remembered some beautiful words from Arktika Arvet (Arctic Vedas).

Quote"When the the AR penetrate the darkenss OM, the vortexes are born, which tear from the darkness, many stagnant substances from the frozen Fire, from which were produced all worlds of matter and throw them outside (of darkness). This movement is accelerated by the aspiration of the AR, attracted by infinite abyss. This is how the new center of AR take place in the infinite abyss.".

Maybe one the best remoteview pictured by Charles Leadbeater and Annie Besant.

Show More Posts