Re: [Discuss-gnuradio] Need some answers about FM TX

[Matt Ettus]

Robert McGwier wrote:
> Thanks for all the comments but I still have my doubts.  I am more like
> to believe in clock leakage and DAC nonlinearities causing this than I
> am with square wave harmonic activity.  I will investigate further with
> the spec-an if I have time.
> 

When the aliasing happens with an ADC it can be though of as an impulse
train (impulses at the sampling points) being multiplied by the signal.
 This is the same as convolution in the frequency domain, resulting in
the aliasing behavior.

With a DAC, it is not a pulse train, but a zero-order hold (aka boxcar
filter, sinc, etc).  Thus the aliases (more properly, alternate images)
are weighted by a sinc function, causing them to get weaker as they get
to higher Nyquist zones.  It is about 18 dB down in the second zone, for
example.

Matt