Re: [Discuss-gnuradio] Receiver sensitivity of wbx, TVRX.

[Matt Ettus]

On 07/17/2010 02:54 PM, naveen nischal wrote:
> Hi All,
>
> Guess it has now become a cliche question. but we have been trying to
> make a link budget for the usrp and the AO-51 satellite and are looking
> for some help finding the receiver sensitivity of either the WBX or the
> TVRX. I have seen the discuss GNU radio archives and found like 50
> questions about receiver sensitivity but did not find any answer. so if
> any of you who have been trying to know the same or do know how it works
> can help us get some information about it would be greatly appreciated.


The reason you didn't find an answer is that there is no answer.  It's 
kind of like asking what the temperature is without specifying when and 
where the measurement is to be taken.

Sensitivity is only a meaningful measurement when you state the type of 
modulation, error correction coding, data rate, and acceptable error 
rate.  100 kbits/sec will have better sensitivity than 1 Mbit/sec.  BPSK 
will have better sensitivity than QAM64. Rate 7 convolutional coding 
will perform better than hamming code.  A required BER of 10e-3 will 
have better sensitivity than a required BER of 10e-6.

The ONLY true universal measure of receiver quality is noise figure, and 
that is what we measure.  If you have all the above info plus the noise 
figure, you can calculate the sensitivity.

The TVRX will have a noise figure of 7-10 dB.  The WBX noise figure is 
shown in great detail vs. frequency and vs. gain in the plots at:

        http://code.ettus.com/redmine/ettus/documents/16

At low frequencies the noise figure is 3 dB.  At higher frequencies it 
rises to about 6 dB.

Matt