Complete information on working SM style device.

[MarkSnoswell]

for everyones info:

You can not fill 3D space with regular tetrahedra. For those interested tetrahedra can be stacked linearly and they for a tetrahelix.  google tetrahelix for more information.

[aleks]

I have a piezo delay line from an old color TV--the delay is 63 microseconds.  I might want to try that instead of the iron coil...

This is a much more realistical thing. Piezo delay should be a broad-band delayer for the most part. Iron wire just delays the highest frequencies only (but for 3.5MHz should work, of course), maybe working as an all-pass filter.

[aleks]

konduct:  I've got a juno 106 synth but it don't go higher than about 22khz...how are you planning on doing the frequency doubling?   

Thanks

If that helps, you may try using a signal multiplier chip (valves can be used for that, but that's a complex biz). You may shift the 22kHz signal to any frequency by multiplying it by the carrier frequency sinusoid (which may be even 1MHz sinewave) - well, some "dirt" will be left as negative frequencies in the original waveform signal will be shifted as well.

Actually, it can be a nice approach overall since you can use PC soundcard to define waveforms and explore different waveforms in an automatic way. Well, if the multiplier is precise you may even left your new square wave generator alone. (on the other hand, no, it's not useful due to negative frequencies shift).

If you can find a "complex" number multiplier chip (which should take 4 inputs and produce 2 outputs) you can shift any waveform precisely. You then just need to create a waveform in quadrature (it's pretty easy to do in a computer by using a linear-phase Hilbert transformer). Other 2 inputs of the chip should carry sine and cosine waves of the required shift frequency.

I think such "complex" multiplier can be created out of 2 normal multipliers if you care about 1 output only, or 4 multipliers if you need all 2 outputs.

http://www.directindustry.com/prod/analog-devices/signal-mixer-multiplier-22009-194871.html if that helps. AD8343 sounds like a good variant as its inputs can go from DC: depending on the  performance of the chip (it's not given on the page in enough detail) and your circuitry you can have gigahertz square or sawtooth (or any other form) waves.

For some really versatile base signal generation you could tie some multi-channel PC sound card with ASIO drivers that is able to run at 192kHz samplerate. For example, if you get a 8 input/8 output card you can produce at least 2 complex signals (on 4 outputs) + have 4 outputs for various aux signalling (just make sure sound card's inputs and outputs are not high-passed: high-quality audio cards do not use DC filters because of phase-linearity considerations). 8 inputs can be used for various low-frequency measurements (you may tie another multiplier with a sinewave oscillator to move any gigahertz oscillations to the low-freq range, so this whole approach is very scalable and tunable if you can get the multipliers and oscillators to work).

You may then build any signal I/O environment you like with such pro audio software like SynthEdit or SynthMaker http://synthmaker.co.uk/ well, at least you'll be able to search for resonances and create required waveforms quickly. For some CVS spreadsheet'ing it's not suitable - on the other hand you can create your own C modules that may help collect required statistics easily. Just to let you know, modern audio cards have extremely low harmonic and IM distortions, and high SNR: 100 dB is a common number, much better than what valves can have.

(example schematic below, 'complex quadrature' is a base waveform signal with its quadrature counterpart synthesized by software means - complex quadrature signal does not have negative frequencies, 'freq voltage' is a DC voltage used to define oscillator frequency)

NOTE

Hrm... all that work just to realise that multiplication shifts frequencies. So, if you have F1, F2, F3 in the signal, multiplication by FF will yield F1+FF, F2+FF and F3+FF frequencies. This won't maintain signal's shape as harmonic relationships will not be maintained. Sorry for a misleading post. On the other hand, sound card can be used for quick prototyping and various measurements.

[poynt99]

So far the frequency numbers that keep coming up from the users here are sub-harmonics of the NMR of iron.

egs. 3.5 MHz, 35 kHz, etc.

spherics clearly states "The frequency should be a harmonic of the NMR of iron".

@spherics, this needs clarification please.

[konduct]

Hrm... all that work just to realise that multiplication shifts frequencies. So, if you have F1, F2, F3 in the signal, multiplication by FF will yield F1+FF, F2+FF and F3+FF frequencies. This won't maintain signal's shape as harmonic relationships will not be maintained. Sorry for a misleading post. Smiley On the other hand, sound card can be used for quick prototyping and various measurements.

It could if you wanted it to but the mix can be 100% dry (orig signal) or wet(higher signal).  It splits it into two separate signals. Again...overcomplicating things...if the audio gear had all this noise or wasn't suitable for something like this, it probably wouldn't sound good coming out of the speakers either. My rig sounds nice. The most distortion I may get would probably come from the amplifier.  But I think that the ability to adjust things on the fly or even better, I can automate changes of any of the parameters...I can set up sweeps...cool down intervals...anything. I really think this is going to be like taking a Ferrarri to a gokart track!