Re: [Discuss-gnuradio] Discuss-gnuradio Digest, Vol 116, Issue 24

[Josh Blum]

On 07/23/2012 09:31 AM, Kinal, George V. wrote:
> Interested in your advice: * You need to set the wire format to sc8
> (thats 2 bytes per complex sample) to achieve 50 Msps over gigabit
> ethernet.
> 
> Where is this documented ?
> 

It might be worth looking at the python code generated when using a USRP
source/sink in GRC.

Here is the new factory function. The old one you are calling into has
less options:
http://gnuradio.org/cgit/gnuradio.git/tree/gr-uhd/include/gr_uhd_usrp_sink.h#n112

And here is quite a bit of info about stream args:
http://files.ettus.com/uhd_docs/doxygen/html/structuhd_1_1stream__args__t.html

> 
> 
> The observed RF output appears to be irregular, as if some samples
> have been skipped or the sample file is incomplete or ?  Could we
> have an incorrect wire format or some other related problem ?   There
> are no buffer underrun problems except when the focus is removed from
> the terminal window, when we get the stream of L's effect, now well
> known   :-)  .
> 

I dont think you can have the wire format wrong, thats kind of hidden
from user access. You could be interpreting a binary file wrong when
passing the buffers to UHD for example.

It might be worth using an app like uhd_siggen to spit out some test
tones and prove that the overall system is working as expected.

-josh