Test control and set up for coherent de-dispersion modes
Details
M&C version switch began at 09:30
Submitted c0200x0128 at 09:54
OBSFREQ is correct but OBSNCHAN and OBSBW are still a factor of 4 too small
Ray is diagnosing
Ray fixed the shared memory parameters. Will try again
Submitted c0200x0128 at 10:30
OBSFREQ was incorrect on Bank C. Ray is fixing.
Ryan stepped out for a telecon and lunch at 10:50
Ryan returned at 13:05. Ray fixed OBSFREQ issue.
Taking a c0200x0128 scan at 13:09
Scans look OK. Tone is in right place, parameters are correct.
At first glance, c0100x0512 also looks OK
Going to step through all coherent LBW modes
Significant packet loss in c0100x0256 search. tint was 20.48 us
Significant packet loss in c0100x0512 search. tint was 20.48 us
NETSTAT blocked on Bank D during c0200x0512. tint was 40.96 us
Significant packet loss in c0200x0512. tint was 40.96 us
Tested frequencies for HBW coherent mode. Look OK.
Tried to go to an incoherent i0100x0512 mode to check switching signal control but Bank A manager seemed to crash.
Ray had accidentally made a change that set obs_nchan to zero for non-coherent modes. This is fixed now.
Ryan had to run to a telecon at 14:50, returned at 15:55
Ray and Joe are in process of switching back.
Checkout scans begin at 16:00. All looks good.
Conclusions
Coherent modes are looking good. Shared memory parameters are now correct. There are issues with dropped packets in certain modes. Should look into ways of mitigating this.
Didn't get a chance to test switching signals, but I think we are pretty close to being ready for regression testing.