I am experimenting with reverse biased Zener and Avalanche breakdown and noise generation. I also have just started testing a handful of junk box transistors as well as some new assortments.

I'm using a pretty simple setup;


I'm not using any amplification. It works fairly well on a breadboard. I've tested dozens, at least, of devices and separated and labelled the ones I find interesting. I find the onset of the breakdown noise and look for a sensitive "sweet spot" just past the knee.

I leave it running quite often over night with various pn devices and bias voltages. Then I run the results through Audacity software, check the spectrograms and the waveform and the frequency spectrum.

I have been noting in the waveform some extremely high (relatively) fast spikes, sometimes to clipping.

These spikes seem to be randomly occurring, clustered or isolated, day or night etc. They are definitely discreet, fast events caused by something. Could it be cosmic rays, muons, gamma radiation, etc. Or are they just some kind of artifact of the apparatus?

 

You may want to google shot noise:

http://en.wikipedia.org/wiki/Shot_noise

 

You might also want to look at 1/f noise:

http://en.wikipedia.org/wiki/Flicker_noise

 

You might also want to monitor the output of your power supply.

 

You might also want to monitor the output of your power supply.

Ideas for a good circuit to do this?

 

Whatever you're using to monitor the breakdown noise should do; just taking its input from the +ve rail.

 

Whatever you're using to monitor the breakdown noise should do; just taking its input from the +ve rail.

Here's two hours of the waveform for the effect rendered by Audacity software. The full file is 9 hours. The small ripples in the main sections are the breakdown noise of the junction just barely past the knee. They look very "noise-like" when zoomed in but they are not amplified in hardware or software here.

But my question was about the tall spikes within the breakdown noise. It's highly intermittent, just a few, if any, per hour. Would that frequency of occurrence be compatible with shot noise or 1/f noise.

 

Have you monitored the power supply rail to see if that is the source of the spikes?

 

Should I use the ground on the sound card with the ground on the power supply. I assume you mean run it through a capacitor so should my reference be the power supply ground, the sound card ground or both ? Same circuit as above, but with the transistor removed and that post grounded?

 

g

But my question was about the tall spikes within the breakdown noise. It's highly intermittent, just a few, if any, per hour. Would that frequency of occurrence be compatible with shot noise or 1/f noise.

That can happen when an appliance, say a refrigerator, turns on or off.

 

Should I use the ground on the sound card with the ground on the power supply.

Yes.

I assume you mean run it through a capacitor

Yes. Post #1 circuit with everything except C1 removed. Power +ve rail > C1 > sound card input.

 

But my question was about the tall spikes within the breakdown noise. It's highly intermittent, just a few, if any, per hour. Would that frequency of occurrence be compatible with shot noise or 1/f noise.

Any chance to say their duration?

 

I will, but I've got to wait till I get home in a couple hours to post it.

 

Any chance to say their duration?

I attached an image. The time scale is at the top. It looks like the spike itself is something like 0.0002 second
with a precursor of about 0.001 second. A hysteresis period of about 0.003 second follows.

 

Here's what I came up with for a power supply monitoring circuit, again run into my sound card. I'm running it now and no spikes yet, but I'm going to run it for > 9 hrs.

 

You might also want to monitor the output of your power supply.

There are numerous examples in hobby magazines that use a series pair of PP3 type 9V batteries. The current limiting resistor is usually pretty high.

 

Yes.
Yes. Post #1 circuit with everything except C1 removed. Power +ve rail > C1 > sound card input.

Here's what I got from the capacitor only circuit. They seem to persist, though the rate seems reduced. I supposed that indicts the power supply or the mains.

 

Yes, looks like interference on the power supply rail. Or it could be interference getting through from the PC to the sound card; or even a sound card fault.

 

This use of the sound card for data acquisition is probably not ideal. Unfortunately, it remains the lowest cost option. Buying batteries for an application like this is a wallet buster. And I absolutely have to have at least +18V (+-9) or more variable down to 2V or so, or even lower. The power supply is nothing special. "Bog-standard" as they might say in the UK.
​

 

Maybe the circuit should have been:

I'll try that a while.