The TPU uncovered? (A PROBABLE technique.)

[turbo]

With an additional coil as we see them in the videos, the little one in the middle, we have nice, clear signals of a lot of volts.
May I say like with tubes??? Yes.

Otto

The tubes are more easy because of two things:

1. You have a high voltage signal.
2. They are fast.

When you do use Mosfets use them in combination with capacitors charged to high voltage and the saturable inductor switches this sharpens the pulse and the caps deliver the needed high voltage so you will end up with something that looks like the tube setup.

M.

[pauldude000]

@buzz

I did just a little while ago, and replied.

@marco

Good advice.

@all

I need to clear one thing up concerning tubes.... Bluntly stated, they are not as fast as SS.... No joke, they are not. They have their uses, and some aspects which are usefull which are not necessarily shared with SS, but if you are considering using their speed/VS/Solid-State, you are barking up the wrong tree. Show me any tube capable of working in the low Ghz range.... I CAN SHOW YOU TRANSISTORS....

Mosfets are just more finicky, and a little harder to select and use, but the rise/fall time of your average good mosfet blows away the rise/fall time of a comparable tube.

Otto, if I were you I would follow marcos advice, and use mosfets for the oscillation circuit, and your tubes for the amplification. VERY good advice, which saves on mosfets.

I am going to do some research later, and give you all some ideas for using mosfets to make an oscillator.


Paul Andrulis

[buzz-ard]

1. Resonance is not enough by itself.
2. Rotation is not enough by itself.
3. The output characteristics from every coil tried so far increase dramatically with higher voltage. 9v by itself is not enough. It must be raised to a higher potential.
4. >>>>>>Rotation is an absolute necessity<<<<<<, not merely a possible.
5. Pulsing of the collector in an isolated leg of a bifilar should be tried by all!
6. Using a Steel core is worthless. If anyone has had luck with steel or ferrite cores, give me the info. They detract from performance as far as I can see.
7. Frequency of operation for any given coil is IMPERATIVE. All of the coils tested have "sweet spot" frequencies. They may not even produce much of any output too far from a "sweet spot" (natural resonant frequency of the unit).
8. Pulse length is important.

If they were full sized, pulsed with say 50v..... OUCH! (Who am I kidding, if I were to pulse them with 50v NOW I have no clue what the field would do, or how strong it might get. It tries to explode at 12v, and feedback kicks in my B&K's safety circuitry if I turn to 15v for more than a second or two.)

Great post, Paul! You and I are parallel in many of our observations. I'll throw in my two cents here, at least until I get my page built:

1. Resonance is indeed not enough. All of the results I've had so far have been strictly through resonance, but quite interesting nonetheless.
2. Rotation is, I believe, part and parcel of the effect. By itself alone it's not enough, I agree.
3. I'm not certain that initial high voltage is required. In nearly all the good results I've gotten, 3 Vpp from the function generators has been my reference level. When I increase voltage I get the same effects, just with higher peaks. I've gotten 200 volts clean AC sine from 20 V DC square offset negative out of my B&Ks, but after thinking it through it's meaningless because I'm relying only on resonance - I've really only been playing with basic transformer effects, I think. Seems voltage and current build in SM's units, so a minimal input should suffice to get it going (9V battery, two AA batteries, etc.).
4. YES!
5. I've had much fun with the bi-filar approach. This is definitely part of the working design. Use the spare leads for feedback into the control coils - it may also drain off some of the feedback getting to your B&K. I've had surprising results using feedback.
6. Iron/steel in the core does not seem to help, I agree. It makes a huge difference in the Tesla replica, though. I have a build planned using a steel core, where intend to saw it multiple times and leave a 1/4 inch gap or so, making it segmented so as to create separate poles within each control coil region. I will wind this build in both Tesla/AC and TPU/DC fashion to see the effects of the split core, but I expect it to benefit the Tesla configuration the most.
7. The sweet spot has been easy to find and easy to tune in my builds. I do believe there's a cascade/avalanche effect that will occur when it's all lined up. But I've found that there's usually a really narrow range where things happen, so precise control of the frequencies is imperative.
8. Pulse length is important. I've had the best results at 50% duty so far, but again I've been tuning for resonance. I've also seen square waves superimposed on the sine traces (really strange-looking indeed), but these are just noisy artifacts that disappear with more tuning.

A thought for all: Turn your function generator way up high then scope the output - there will be what we used to call "switch-bounce" as the output drain stabilizes, showing squiggly traces that settle down after 6-8 oscillations. You can also see switch bounce in stereo gear, like when you turn on the power (the "kick" and thump from the inrush of current, which stabilizes after a few cycles). This squiggly bounce is supposedly not a desirable thing in the TPU, as the squigglies would cause ripple in the resulting mag field and distort the effect. Squigglies are more pronounced at higher frequencies, illustrating one of the trade-offs of using SS devices. Tubes generally don't do this as badly, nor do continuous power sources like batteries. Spark gaps do something similar as I recall, but it's much wilder and even more destructive - this is why I had so little confidence in the reed-switch-as-commutator idea that was floating around in another thread.

[buzz-ard]

@Jon - How coincidental. I just found your page yesterday and looked over it briefly. Looks good - I intend to go back when I have a few more minutes.

I haven't graduated to the MOSFET world yet - I'm still gnawing away with my 3 function generators. But I think that's about exhausted and it will soon be time to move on up. I can tell you that adding feedback from the collector changes the dynamics of the coil, and I can also say that my setup is working best with a multi-turn collector, Mobius-looped at the ends (as opposed to making a twist mid-coil).

About all I can offer until my new page is built is my last reply to Paul. I'll post a link to my site soon as I can get the HTML built to show it all and share all I've got.

[pauldude000]

@buzz

Check your junkbox again! (at least yours are entering in through the hotmail inbox now!)

@all

I am going to try to link the logical thoughts here, for both you AND my own mind.

The first input signal creates a field. The second harmonic resonant signal attracts the lines of force before collapse from the first, drawing it toward the second in train, giving apparent motion, IE rotation. Thinking about this, all frequencies should be a narrow but powerful pulse to spin the field up.

A third stabilizing frequency would be good at this point, to maintain and speed up rotation. The rotation then is resonant with the applied signals, but probably at a much lower frequency (a sub harmonic). Our coils are wound wrong on the collectors.......

They should probably be laid literally one on top of the other, each wound fully around the circumference of the TPU and pulsed sequentially. (inner coil -> 0 degrees, second coil -> 90degrees, outer coil -> 180 degrees phasing). This will give a massively rotating field.

With my "wagon wheel" TPU, I learned some VERY valuable insights on this, and it matches precisely with SM's own description of the build of his coil.  (I am glad I tried the "wild hare" approach on this what I thought of as non-sensical wind.)

I think sharp spike cap discharge through mosfets is the way to go frequency regulated by square wave oscillators.

Find three "sweet spots", for your coils. Write down the center frequencies. Build one oscillator for each coil at the specific frequency, lowest to highest. Have all three switchable. Fire the first low frequency, then fire the second, (if I am right, you should notice a small output here) then fire the third..... Manually firing the frequencies should do the trick, as it is almost guaranteed to be out of phase this way. (Being IN phase would be sheer accident.)

If my thoughts are right.... MAN OH MAN I THUNK I HAVE HAD ANOTHER MENTAL BREAKTHROUGH!

Paul Andrulis