Ah. The secret that I cant find anyone else mention. The Veb numer in the spec sheets to look up, and exceed. Now I am getting noise out of this npn :) Thanks for being detailed...
The noise from a reverse junction is 'shot' noise, either from zener or avalanche effect (under 7 volts is zener). White noise is made by thermal noise, or johnson/nyquist noise. Obviously this circuit will have both these, but mainly shot noise and 1/f noise or flicker noise, also made by all active components.
If it's flat in the frequency response for the requirement, it's white. Even though it's not johnson noise. You are definitely right just that's all depends on the application.
Interesting. Wonder how temperature sensitive it is? You could probably use your spectrum analyzer function on your scope to see how correct the noise is.
+HillOrStream Considering the "noise" transistor on its own, I'd think the noise would extend well into RF and be limited by the junction capacitance. I'd need a wide band OP amp to get a good measurement.
@@murrrr8288 With a regular opamp you do that by tying the two inputs together an no feedback from the output. This leaves the opamp in max gain and the smallest noise at and in the input stage will appear max amplified at the output. But it does require a bubble (+ & -) power supply. You can test if tying the inputs of the 358 together (so only to each other and non to ground) on a single 9v battery has the same results. If so, I would love to see it and hear the noise in a Yt video.
Many simple microcontrollers have an 8 or 10 bit A to D converter. With amplification and filtering, the signal could be processed then fed into the micro and a random number generated.
