[time-nuts] Re: electronics question or how not to fry my raspberry pi

[Hal Murray]

folkert at vanheusden.com said:
> the scope software says 2MHz but output is really 10MHz).

That's one of the joys of digital scopes.  You are seeing the interaction of 
the sampling rate and the signal frequency.

The chart on the right says the sampling rate is 12 MS/s  Try it with a 
sampling rate of 100 or 200 MS/s

There are 2 ideas tangled in here.  The first is Nyquist.  If your signal has 
a bandwidth of X and you sample at more than 2X, then you can reconstruct the 
signal perfectly.

The second idea is aliasing.  We normally think of the "bandwidth of X" as 
being from 0 to X.  But it also works if that bandwidth is centered on some 
high frequency.  The signal gets aliased down to baseband.  You can 
reconstruct the correct signal if you know the offset which will be some 
integer multiple of your sampling rate.  DSP geeks do this all the time.

You can probably work out a few examples with pencil and paper.  Try a signal 
rate of 1 and a sampling rate of slightly less than 1/10.

Does anybody have a good URL that explains this?


-- 
These are my opinions.  I hate spam.