[Hackrf-dev] communication problem HACK RF - LINUX

[Dominic Spill]

On 31 July 2015 at 08:54, Simone Ciccia S210664
<simone.ciccia at studenti.polito.it> wrote:
>
> I'm experiencing issues with my HACK RF just feeding samples in a
> simple gnuradio flowgraph, HACK-RF SOURCE -> NULL SINK, in Linux 12.04LTS
>
> For sample rate above 10MHz I get continuous Overrun ("O").

This is a sign that the HackRF is producing samples at a faster rate
than the host can process.

> Hack RF is interfaced with the PC through the USB 3.0 port, and the tests on
> it (i.e. with hackrf_transfer) indicates that works correctly at 20MHz.

Some USB 3.0 ports have poor USB 2.0 performance, but as
hackrf_transfer works correctly, I would expect that your port is
working correctly at high speed.

> Searching for the problem I found it could be the Linux buffer size,
> thus I set the file sysctl.conf as follows,

These are network buffers, so they are not related to the HackRF
(these adjustments probably solve a problem for USRP2 or other network
attached SDR hardware).

> Anyone have already encountered/solved such issue?

My normal suggestion would be to use a more simple flowgraph, but
yours is about as simple as it gets.  How powerful is the host?  Using
gnuradio for "HackRF source -> null" is more heavyweight than using
hackrf_transfer to achieve the same, so perhaps that is making the
difference?

Have you tried other sample rates?  e.g. 16MHz?