[Discuss-gnuradio] Re: USRP2 FM TX and FM RX working together

[Patrick Strasser]

schrieb Jorge Miguel am 2010-09-22 13:10:
> I forgot to mention, but I get a lot of messages like:
> .....SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS
> SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSaUSSSSSSSSSSSSSSS.....
>
> I did the calculations and with the 104 interpolation USRP sink and 208
> decimation USRP source and 16 bits per sample means that less than
> 100Mbs are taken from the 1Gbps ethernet wire. So there souldn't be any
> overload.

Connection capacity is not the problem, but consuming the data.
GNU Radio has no intrinsic concept of sample rate. It processes data in 
pipes as far as possible and blocks when sources do not provide data or 
sinks do not consume data. The blocks are connected by buffers that are 
not infinite in size.
If your input rate (Ethernet) does not match your output rate (Audio), 
at some point either the buffers are all empty (underrun) or all buffers 
run full (overrun). This is the point where you get notifications in 
form of S, aU, aO, etc.
It is your task to calculate the data/sampling rates correct and convert 
rates with resamplers.
Maybe you want to have a look at the example FM receiver and compare 
your rates with the ones used there.

> Furthermore I still don't know why is necessary the multiply cont block
> in the modulator.

This is your volume/gain knob. In GRC you can connect it to a slider and 
try different values in real-time.
Depending on your sink data type (complex, short) you have different 
ranges for full scale signals.
Generally you would want to scale your signal nearly to the maximum 
amplitude to use the full precision, especially when working with 
non-floats.

--
Engineers motto: cheap, good, fast: choose any two
Patrick Strasser <patrick dot strasser at student dot tugraz dot at>
Student of Telemati_cs_, Techn. University Graz, Austria