Re: [Discuss-gnuradio] (WBX+USRP) Unknown signal at the middle of any FFT plot in receive mode

[Marcus D. Leech]

On 08/23/2010 01:11 PM, jan acosta wrote:
> Hello,
>
> I am currently testing two USRP radios with 3 WBX transcievers.
>
> When I test the receive path, I would use gnuradio companion and go
> from USRP source to FFT plot.
>
> However, I see that at every center frequency of the FFT window there
> is a signal even though I am not inputting anything.
>
> The signal power is about 0 dBm.
>
> Also, if I inject a signal, the power of the signal is shown to be
> much higher than the true power I am sending through.
>
> For example, when I send a -50 dBm signal it shows up to be +30 dBm
> and I've tried using multiple signal generators.
>
> So now I'm stuck.  
>
> How do I fix this?
>
> Thank you.
>   
The FFT plot in Gnu Radio is *uncalibrated*--it's showing you *relative
power* and *NOT* dBm

You *could* perform some calibration and apply calibration constants to
the output of the FFT to turn it into dBm, but
  "out of the box", it *is not* dBm.  It's rather hard to make it be
dBm, as well, since each daughtercard type is different,
  etc, etc.

So, that "signal" at 0dB is actually quite small compared to your -50dBm
input signal.  it's an artefact of the down-conversion
  process. 

-- 
Marcus Leech
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org