Re: [Discuss-gnuradio] USRP producing samples larger than 2^8

[Eric Blossom]

On Wed, Mar 14, 2007 at 07:31:33PM -0700, Chris Stankevitz wrote:
> Hi,
> 
> Should I get samples with a magnitude > 128 from the USRP no FPGA gain?
>  I believe the answer is no.  Take a look at these plots of the data I
> am getting (plot of real portion only):
> 
> http://img157.imageshack.us/img157/8648/plot1gi6.gif
> 
> http://img245.imageshack.us/img245/8528/plot2fp1.gif
> 
> I collected the data using this method:
> 
> ./usrp_rx_cfile.py -d 16 -f 1575420000 -g 0 -N 4000000 file.dat
> 
> Furthermore, I created a test jig which wrote USRP output to stdout and
> verified these spikes are not in the binary file or the binary file
> reader.  I'm using the DBSRX with a GPS antenna and a low noise amplifier.
> 
> Thank you!
> 
> Chris

Chris, 

Does it only occur once, and does it occur in the first 1024 samples?
That is, if you look at, say 1M points, is there only a single
occurrence?

If yes to all 3 questions, then I think you're seeing a known problem
where for some not fully understood reason (but probably related to
failing to properly reinit the FX2 buffers and/or the FPGA) you get
the sinc looking spike.

Eric