zener diode plasma noise region?

[coppercone2]

So if you play with a zener diode of around 5-10V there is a 'noise knee' that it has, basically what I did a long time ago was wire in some resistor in series with a zener diode and connect that to a precision lab supply. I found that there was a small range (i think it was some number of microvolts) where the zener diode would have a much larger noise output (ac coupled into a scope), at least 10x larger then usual.

I did not have a spectrum analyzer at the time. I built a zener noise source before but usually its just biased by a high impedance divider and fed into a amplifier.

I attribute this to what is called plasma breakdown in the zener diode: on page 19
https://www.onsemi.com/pub/Collateral/HBD854-D.PDF

Am I correct in thinking this? The range was narrow, under it you would not get much noise and over it you get noise but of a much lower amplitude. I think I used a 4.5 digit calibrator to find this zone. I remember worrying about it on voltage rail monitors that used zener diodes. I did not have a SA at the time to see what the NSD looked like.

Based on what I saw in simple zener diode noise generator circuits, it does not appear that they would happen to bias in this microplasma region? IIRC it was different on different diodes and rather narrow.

[David Hess]

It is consistent with what I know about avalanche breakdown and one of the things which distinguishes standard rectifiers from avalanche rectifiers; the later are processed so that the entire junction breaks down uniformly preventing hot spots.

I suspect this mode of operation is not ideal for wideband noise generation because of thermal effects.

[T3sl4co1l]

I'd suppose it's not usually used because you have to trim the bias per component.

Unless the components are selected and tested for operating point and noise level (and hopefully stability as well), which is probably a very hard thing to find off-the-shelf (and expensive to test, in any case).

Statistically speaking, you want to use two diodes, and invert the signal from one.  Otherwise there's a small autocorrelation, where few impulses are expected following larger impulses (a consequence of the rampy waveform).

Same thing was done with old school, magnetically biased, gas discharge tubes (e.g., 6D4).  I have such a generator:
https://www.seventransistorlabs.com/Images/Gaussian_Noise1.jpg
https://www.seventransistorlabs.com/Images/Gaussian_Noise7.jpg
(Triode thyratrons with bias magnets on the right; the rest is amplifiers and filters.  Design seems to be mid-60s hybrid, including silicon diodes, a couple high speed mesa transistors (I forget if germanium or silicon?), and a few germanium power transistors in the supply (regulated filament power!).

Tim

[coppercone2]

hmm if I can find another one I will try to put it into a spectrum analyzer with the same setup.

I was wondering if it was maybe making noise into higher frequency areas, but like david hess said I doubt its  thermally stable.

How far does that machine push the noise?

what do you mean by rampy waveform? On page 19 you mean? its autocorrelating by heat bursts? that can effect a RNG? so it weeble wobbles ?

back when I saw it, I just compared the oscilloscope noise to the carefully biased one vs the one fed into a 10x gain amplifier, and I thought its cheaper to use the amplifier then to adjust a voltage divider, but I never considered what it would do at high frequencies.

Also IIRC it worked the best around 7.2 volts with the ebay zener diode kit I got, I thought maybe the higher voltages would make some kind of awesome noise that would blow out my windows but I think the effect stopped working.

[T3sl4co1l]

what do you mean by rampy waveform? On page 19 you mean? its autocorrelating by heat bursts? that can effect a RNG? so it weeble wobbles ?

Like this, though not as exaggerated as this 2N3904 in avalanche, because diodes aren't three-layer devices.

Tim

[Zero999]

Yes, I've noticed this before, when designing a noise source. Also think in terms of voltage, rather than current. I think you'll get more noise by using a current source, rather than a a voltage source and resistor.

Does anyone know if it's temperature dependant? Would making the current source NTC using transistors and diodes, rather than stabilised using a voltage reference, be beneficial?

[coppercone2]

you know pretty sure that when I just played with the filters on the rigol 1052e when I made this a long time ago, it had changed the NSD some what, but that thing is kinda iffy.

[T3sl4co1l]

Yes, I've noticed this before, when designing a noise source. Also think in terms of voltage, rather than current. I think you'll get more noise by using a current source, rather than a a voltage source and resistor.

Does anyone know if it's temperature dependant? Would making the current source NTC using transistors and diodes, rather than stabilised using a voltage reference, be beneficial?

You'd have to map out the e_n vs I_r vs T space.  Maybe it's consistent across production of a given part, or across parts, but who knows.

Regarding impedance, it's not very important.  The equivalent source resistance is whatever the dynamic resistance of the diode is.  As long as you aren't loading that heavily, you aren't losing much signal.  Or loading it by half to get ~maximum power point, whatever.

A lower impedance would be desirable to increase bandwidth -- rather than allowing the voltage to jump down randomly, keeping it up, making it more likely to spark soon after, again and again.  That would also allow higher bias current, to some extent.  Which is really to say that the Rz will be lower in turn, but probably about the same power noise density.

Tim

[ArthurDent]

About 25 years ago we needed a noise source up to 100Khz so I used a unbypassed zener with a fairly high value resistor as the source. I fed it through 2 amplifier stages and by varying the gain of the 2 stages managed to get somewhat white noise that had pretty consistent amplitude over that range. I don't recall that the resistor value or voltage being that critical but I did use a variable bench supply to choose the values.

I had a General Radio 1390-B random noise generator from 1965 that gave me some ideas.

http://www.ietlabs.com/pdf/Manuals/GR/1390%20Noise%20Generator.pdf