A common way of getting a noise source is to use a back biased NPN junction, but the RMS output varies greatly from transistor to transistor. Does the RMS value depend on any other measurable/predictable value (beta, maybe?)

 

A common way of getting a noise source is to use a back biased NPN junction, but the RMS output varies greatly from transistor to transistor. Does the RMS value depend on any other measurable/predictable value (beta, maybe?)

See this article on shot noise.

 

You can measure it, typically noise is charactarized as x volts a a -3db bandwidth of y Hz. In other words, if you change the bandwidth, you change the RMS noise.

For your reading pleasure, here is a short but informative article on zener noise:
http://users.cosylab.com/~msekoranja/tmp/04236738.pdf

 

OK, late reply here - thanks for the help, shot noise and advice on measuring noise received, but what I want to do is to say - if i buy 100 of this particular zener diode (or transistor and use it back biased) what range of noise voltage can I expect between them? They are all different, if I want a type in which the noise varies very little from diode to diode, how do i know which ones to choose? There is no measure of it on the datasheet, so can I infer it from some other measurement that is on the datasheet?

 

OK, late reply here - thanks for the help, shot noise and advice on measuring noise received, but what I want to do is to say - if i buy 100 of this particular zener diode (or transistor and use it back biased) what range of noise voltage can I expect between them? They are all different, if I want a type in which the noise varies very little from diode to diode, how do i know which ones to choose? There is no measure of it on the datasheet, so can I infer it from some other measurement that is on the datasheet?

I doubt the semiconductor processing and testing for those devices controls for that characteristic, and I don't think the noise level can be inferred from anything else on the datasheet; to get repeatable noise voltages I think you'd have to resort to specialized components like noise diodes.

 

hhmm. interesting. thanks for the tip, i'll look into those diodes

 

More reading

 

how about a "noise AGC"?

 

I would think a CC source feeding a temp controlled junction device of your choosing would stabilize the noise amplitude sufficiently to allow one to "grade" the device and perform a crude bin sort of multiple devices.

 

Or a cermet trimmer pot?

 

yep, these are things I had considered -

grading the devices is time consuming and not really suitable as i will not be assembling this myself. it doesn't really improve on a trimmer which is the current solution.

the noise AGC is - i think - an improvement on the trimmer

 

the noise AGC is - i think - an improvement on the trimmer

This app note from Linear Tech shows one approach to doing that, in Fig. 17.