[sdiy] Handy DIY tip #213/noise color

[Ian Fritz]

Tim --

I just record the noise to hard drive and do a spectrum analysis using 
Sound Forge.  It's usually flat at high frequencies, so I compare the value 
at some low frequency (1 Hz, say) to the hf limit.

This isn't any kind of standard, but works well for selecting a good 
transistor.  I went through about 50 2N3904's this way and found just one 
really good one and a couple of not-to-bad ones.  The best ones actually 
had the lowest noise level.

   Ian

At 01:35 PM 1/15/2003, Tim Ressel wrote:
>Yo,
>
>Okay, since this got all serious (tsk tsk), I have a
>serious question: Just how do you test the Noise to
>Noise ratio? Or to put it a better way, how does one
>gauge the quality of noise? What is the standard?
>
>--tr
>
>--- Ian Fritz <ijfritz at earthlink.net> wrote:
> > Another example: Suppose you are trying to make a
> > pure white noise source.
> > If the junction you are using is not ideal, you also
> > get some undesirable
> > 1/f or popcorn noise at low frequency.  So that
> > "bad" noise would be the
> > "noise" and the "good" (white) noise would be the
> > "signal".