Successful TPU-ECD replication !

[bob.rennips]

I've was playing around to see how fast the amplitude grows when combining waves. Unexpectedly I came across something that I hadn't considered.

Although I've used sine waves for the calculations the same principle still applies because your coils are physically spaced at different degrees apart around a circle. Becauase you are dealing with pulses into the kilohertz and even mega hertz which create magnetic waves that mix, the exact position of your coils, to the nearest 0.5mm will be critical to the phase required between your pulses. My guess is that 1st, 2nd and 3rd harmonics will work when the coils are identical AND they are positioned precisely around a circle equal distances apart. If your coils are slightly different and not quite in the correct place, your sweet spot frequencies don't lie on the 1st, 2nd, 3rd harmonics, resulting in only partial conversions.

Anyhow, back to the waveforms. The red and blue are 90 degrees out of phase. When combined they create another wave (green wave) that is 45 degrees out of phase. However when the green wave and the original red wave combine they form yet another wave but this time the phase is 26.56... out of phase.

Now that is unexpected for me. I had thought that combined waves would continue to divide the phase in two. What this means is that phase plays an extremely important part in the TPU. The continuous mixing of waves results in a continually changing set of phases as well as increasing amplitudes.

Note that when waves of the same frequency mix the resultant wave still has the same frequency, but its phase and amplitude change.

In Otto's working design, by tieing the secondaries back to the primaries in all the coils this will ensure a continued mixing of new pulses with combined pulses. I think Otto's circuit is in affect a circuit that continuously creates waves of different phases. The TPU effects, including the generation of the seed, being caused when the right combination of phased waves is created.

I note in SMs TPU snapshots there appear to be 2 or 4 coils. I haven't seen a setup that appears to use 3. Given that all coils are in parallel, it would not be difficult to add a fourth coil, and still pulse the four coils with the 3 frequencies.

[z_p_e]

Darren:

I totally agree with your analysis, especially the need for synchronization of the harmonics or at least some precision control of the wired OR common path. Did SM not say that there is a requirement of quality engineering equipment to be able to perform enginerring analysis? Maybe I mis-read that but I somehow remmebered the need for good tools.

Regards
chrisC

Chris,

Yes precision frequency adjustment would be necessary I would think. This is why I'm not sure using DIP switches for frequency selection (if I read that correctly) will be useful for capturing the desired frequency. It sounded like Roberto and Jason are going that route.

Please note that my last posted representation of the circuit, is merely that. It was not intended as a circuit to build. I was trying to illustrate how the primaries are being switched, and to make it easier to visualize what happens with different input scenarios.

One could build this circuit, and it would work, but it won't be the same as the original. I recommend everyone build a divider and drive it with a single (quality) oscillator with fine frequency adjustment. If there is a 3 frequency combination that makes this thing sing, then this is the only way you are going to find it. Also, stay with the 3 FETs and 3 coils the same as Otto's, then you'll be comparing apples with apples.

SM mentioned the use of expensive lab instrumentation to analyse and perhaps detect the overunity he talked about so much involving the piece of wire. This has always troubled me because he mentions that it is so simple and we should start with this in our research, yet we may never be able to detect the effect it supposedly achieves, especially when we are not certain what exactly the effect is we are looking for, i.e. the "kick".

Darren

[MeggerMan]

Some high voltage mosfets for people to consider as the IRF840 devices are becoming obsolete:

STB21NM50N
550V - 0.19R
18A
rise/fall 18/30ns

STF12NM50N
500V - 0.29R
11A
rise/fall 15/14ns

STB20NK50Z
500V -0.23R
17A
rise/fall 20/15ns

STP25NM50N
500V - 0.11R
22A
rise/fall 23/22ns

Rob

[ronotte]

@Rob,

I'm testing with good results the much more reliable (and cheap) IRFP460: 5ooV, 20A, Trise=59nanosec.

Roberto

[bob.rennips]

Darren:

I totally agree with your analysis, especially the need for synchronization of the harmonics or at least some precision control of the wired OR common path. Did SM not say that there is a requirement of quality engineering equipment to be able to perform enginerring analysis? Maybe I mis-read that but I somehow remmebered the need for good tools.

Regards
chrisC

Chris,

Yes precision frequency adjustment would be necessary I would think. This is why I'm not sure using DIP switches for frequency selection (if I read that correctly) will be useful for capturing the desired frequency. It sounded like Roberto and Jason are going that route.

Please note that my last posted representation of the circuit, is merely that. It was not intended as a circuit to build. I was trying to illustrate how the primaries are being switched, and to make it easier to visualize what happens with different input scenarios.

One could build this circuit, and it would work, but it won't be the same as the original. I recommend everyone build a divider and drive it with a single (quality) oscillator with fine frequency adjustment. If there is a 3 frequency combination that makes this thing sing, then this is the only way you are going to find it. Also, stay with the 3 FETs and 3 coils the same as Otto's, then you'll be comparing apples with apples.

SM mentioned the use of expensive lab instrumentation to analyse and perhaps detect the overunity he talked about so much involving the piece of wire. This has always troubled me because he mentions that it is so simple and we should start with this in our research, yet we may never be able to detect the effect it supposedly achieves, especially when we are not certain what exactly the effect is we are looking for, i.e. the "kick".

Darren

I suspect he's talking about a real time spectrum analyser. Darren, I'm sure you are aware of what these do but  for those that don't.... basically a digital spectrum analyser does an FFT of the incoming wave. The result is a moving bar graph where each bar represent a particular frequency and the bar  height the amplitude of that frequency. This is exactly like the animated graphic equiliser display on your hifi where the bars bouce up and down. Low down notes appear on the left, higher notes on the right. Except in a high end spectrum analyser these can go into the gigahertz range.

This is particularly useful in examining say an output signal, such as from Otto's device. You could start to observe patterns where the larger frequencies cluster. Then you could adjust your input to enhance those frequencies etc. etc.