3 frequency methods

[starcruiser]

Really in the zone today.  Been thinking about the cryptic reference to "3-frequencies" and remembered something about harmonics in waveshaping circuits.

Looking into this, I find that waveshaping circuits combine the fundamental frequency with the harmonics to create the waveshape that is desired. 

For example, a squarewave may theoretically have an infinite number of only odd-order harmonics.

In the peaked wave, harmonics such as the third, seventh, and so forth, cross the reference line 180 degrees out of phase with the fundamental; the fifth, ninth, and so forth, cross the reference line in phase with the fundamental. So, not just any old harmonic will do!

Sawtoothed waves are comprised of odd and even such as the fundamental, 2nd and third on up to 30th or more depending on the shape - ringing all the way.

Very interesting.  I'm thinking that synthesized waveforms may not be up to the task.  Not that it will not work, just not the direct route.

So maybe a harmonic rich oscillator is the way..... Hmm Intersting idea

So is that why you started the other thread with th astable oscillator?

I am wondering if the oscillator might benefit from a bandpass filter (or 3), I was thinking about the video of the large SM TPU and thought the coils int he center looked like parts from a speaker crossover, Did Steven use these components because he had them on hand? Also he mentions the 5Khz frequency, this is in the Mid's for a speaker crossover isn't it?  More to ponder I guess.

[Grumpy]

filter out the crud that you don't want.

[supersam]

hey eveyone,

not to think i know everyting but, steven said, when you start to look at the worst possible scennarios then you will see all kinds of of things, including alot of hash and all kinds of frquencies."

lol
sam

[otto]

Hello all,
something for you to think about:
when I pulse my TPU I see my voltage adding and its OK. I see a lot of "hash" and its NOT OK. Why?
In the moment when my light bulb 60W is glowing, say at a half, I see on my scope the hash is gone!! I see only clear kicks at various frequencies. In this moment "PESE" helped me. To be short, I connected the 3 drains(collectors) from my MOSFETS TOGETHER.
And here is what I have now:
I still have a lot of hash in my signals BUT when Im pulsing my coils I have a adding VOLTAGE and ADDING CURRENT!!! Thats easy to see with my ANALOG amperemeter. I dont have to tell you the relationship of voltage and current.
The bad in this all is a lot of hash. As Grumpy and mramos said FILTER, again FILTER.
there was mentioned a tube amplifier.... Tubes have not so much harmonics, they are "cleaner". SS are very "dirty". They produce a lot of harmonics and this is the "dirt".
Again I got a lot of help from "PESE". Think, what will happen if we have on our 3 MOSFETs nice, clean signals (kicks) and with such kicks we pulse our coils. I want to say 1 MOSFET 1 frequency and this all driven with a Main oscillator. This main oscillator synhronises the 3 oscillators with the MOSFETs. I have NOT done this yet but I will.
In this moment, when I pulse my coils, say at 20 microsec. I can see suddenly a new pulse is forming and the frequency is then only 10 microsec. What I will say, it looks to me the signal is phase shifted at 180? and of course 90? and I dont like it.

Otto

[Grumpy]

Otto,

Thanks much for explaining this to everyone.

I suspect a relationship between the phases of the signals and the voltage/current, but I have not determined how it all works.  To start, I am trying to keep everything in phase going to the TPU.  Filtering is all passive. 

LCR filter will not induce a phase shift at center frequency, but will shift above and below -ask Pese!

I can see no reason, in any realm of theory, for three frequencies except to obtain larger pulses and I can get that through more direct means.

You may have "ringing" from the mosfets switching very quickly, try to dampen.  If you smoke the gate, use a zener diode to limit the voltage at the gate.

Hmm, never thought of differentiating (low pass filter) the mosfet output - probably blow the caps apart - he he he!

Voltage quadruple through mosfet to differentiator - should get a wicked pulse.

One last thing, anyone every tried a Marx or Fitch pulse generator?