[Hackrf-dev] HackRF-dev Digest, Vol 21, Issue 4

[Michael Ossmann]

On Thu, Sep 04, 2014 at 07:01:58PM +0000, McDonald, J Douglas wrote:
>
> The idea is to know what the aliasing properties of HackRF are.
> That is, say I set it to receive 100 MHz at the center of the bandpass
> and set the sample rate at 10 megasamples per second, I&Q. 
> 10 megasamples per second I&Q says Nyquist frequencies are at += 5 MHz 
> from 100 Mhz or 95 and 105 MHz. Isn't that right?

Yes, except that there is no bandpass filter.  There is a lowpass filter
on the analog baseband signals.  (I suppose you can think of it as a
bandpass filter centered at 0 Hz, but it is implemented as a pair of
lowpass filters.)

> I then send in a signal that sweeps, very slowly, from 80 to 120 MHZ,
> constanmt amplitude. When at 94 MHz the signal seen will not be zero,
> at 80 MHz it probably will be. I just do the scan and not how many
> dB it is down versus frequency and plot that. 

That would test the frequency response of the baseband filter.

> By vary with the RF frequency I am asking this: Apparently HackRF
> has a somewhat different internal signal flow at various
> frequency bands (as implied by different output powers
> for different frequencies, implying similar things for receive.)
> Could this difference result in different results(i.e response versus
> input frequency)  if I substitute
> 1 GHz or 2GHZ or 2.4 GHZ or 5GHZ for the "100 MHz" above,
> keeping sample rate fixed?

You should not see any appreciable variation in the apparent baseband
filter response at different tuning frequencies.

There are several different baseband filter bandwidth settings, however.
In general you should choose the narrowest one that is suitable for your
application.  By default, a passband slightly smaller than the digital
bandwidth is selected.  For details on the baseband filter, see the
MAX2837 datasheet.

Mike