I saw you mentioned "sheet beam" tubes and it's interesting that I was thinking about them just the other day.
I have a friend who is a musician who is interested in using some obscure technology for his performances. I know there's one company that makes a commercial product which uses a sheet beam tube as some kind of distortion generator / waveform folder.
I was wondering what would happen if you fed white noise into such tube as the modulator, and then used a high-frequency square wave as the "carrier." So you're basically using it as a "ring modulator", and then if you filter the carrier and upper sideband you get noise out but with the distribution's mean shifted down in frequency, right? So adjusting the carrier frequency basically makes a low-pass filter for noise.
I wonder if you could then get resonance by injecting positive feedback into the accelerator/screen grid (the only one left.)
No, it's not. Er, to some degree of "correct" I suppose...
You'd get something like a 6BN6 if you strap the sides of a 7360 in parallel. Probably with a whole lot less "gating" gain though (you'd be using the parallel deflection plates as g3, which probably need -100V to reach cutoff; actually how much is hard to say, because the datasheet only specifies a bias of 25V, nothing else).
Anyway, using a switching mixer (i.e., square wave "LO", or a large enough sine that it behaves the same), you get the white noise flipped around each harmonic. If the square wave is 50% duty cycle, then it has only odd harmonics, and will put sidebands around f, 3f, 5f, etc.
If the white noise is band-limited to BW < f, then the resulting spectrum has blocks of white noise at f +/- BW, 3f +/- BW, 5f +/- BW, and so on.
If f = BW, then the baseband sideband (gee, that makes sense..), i.e., in the range 0-f, is inverted.
If BW > f, then everything is still true, but you get overlap at even harmonics, so it's 3dB stronger where they overlap (assuming the top and bottom frequencies are uncorrelated).
If BW > 2f, then you get even more overlap, but now you've got correlated frequencies overlapping, and noise doesn't go up any more.
A 6BN6 is simply a switched amplifier: you get proportional gain from g1, and g3 turns it on or off (gain = nominal vs. gain ~= 0). It would be a
1-quadrant (bilinear) device if g3 were particularly linear, but I don't think it's very suitable for that, so you'd use it at larger g3 amplitudes were it looks like a switch.
A 7360 is rather more powerful, because it is a balanced (two quadrant) device. The output (difference in plate currents) is given by Vg1 * gm * Vdiff, for Vdiff within the linear range.
Now, as for neato time-domain functions -- remember that cathode current is constant (controlled by Vg1 + Vg2/mu), so as you vary g3, you're diverting plate current into screen current. Indeed, g2 vs. p can be used as a shitty LTP (this application appears in RDH4).
It might also be desirable to have the plate voltage saturate, so that screen current dominates during part of the waveform, and plate voltage follows cathode voltage (if it were an op-amp, this would be in the phase-reversal regime!). Which might give you some ideas of applications, like a crude full-wave rectifier.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.