"Failed to set frequency to ..." message in Spectrum scan

GNU Radio
1

hello everyone,

I 'm trying to collect some samples from my spectrum sensing code
running on
USRP2, which is a slightly revised version of ‘usrp_spectrum_sense.py’.
My
problem is that after some time during the scan of some band, this
message
appears: “Failed to set frequency to …”. For instance I let the board
to
sweep my band 1000 times but after 100 or more scans (it’s really
variable)
it stops its normal operation and can’t lock on a frequency anymore.
I tested this with both WBX and XCVR front-ends in 400M and 2.4G. But
the
problem persists. Since I need to let the USRP2 work and collect samples
for
quite a long period, I really need to come up with a solution. I was
wondering if someone has already encountered this problem and has solved
it.

Rahman

2

On 04/30/2010 08:02 PM, Rahman Doost wrote:

the problem persists. Since I need to let the USRP2 work and collect
Discuss-gnuradio Info Page

Does the non-slightly-revised version do the same thing?

3

Well, the modification is for changing the code in such a way that it
works
on USRP2 since the original version is written for USRP. The rest is the
same. Also it does’t seem this is a programming bug since, it works for
some
time and then it crashes.

Rahman

4

On 04/30/2010 08:17 PM, Rahman Doost wrote:

Well, the modification is for changing the code in such a way that it
works on USRP2 since the original version is written for USRP. The
rest is the same. Also it does’t seem this is a programming bug since,
it works for some time and then it crashes.

Rahman

O
What happens when you exit the program, and then go back in, does it
recover?

–
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium

5

Again the same, works for some time, then crashes!

6

On 04/30/2010 09:59 PM, Rahman Doost wrote:

since,
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org

I suspect that some resource, perhaps file descriptors, are “leaking” on
each iteration of the
scan. Possibly because the device is being re-instantiated on every
iteration?

7

On 05/01/2010 04:40 PM, Rahman Doost wrote:

I checked and they’re not hundreds. In fact there does not seem to be
anything unusual about that.
Plus I see some ‘S’ sequences every 3 or 4 failed scans too. In USRP2
usage FAQ, it says this happens when packets are dropped. But I don’t
know which one causes the other one.

Rahman

The next thing to check for is a gross memory leak. There is clearly
some resource that is leaking,
probably on the host side. Because when you re-start the application,
everything is just fine for a while
after that, correct?

–
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium

8

Yes that’s correct. It always works fine at the start.