CICADAGreenBankSpectrometer2018jan23 < CICADA

You are here: NRAO Public Wiki>CICADA Web>GreenBankSpectrometer>CICADAGreenBankSpectrometerCommissioningLogs>CICADAGreenBankSpectrometer2018jan23 (2018-01-23, RyanLynch) Edit wiki text Edit Attach Print version

TGBT17B_501_04 - Additional VPM Acceptance Tests

Goals
Details
Conclusions

Goals

Test all search modes
Test all 32 and 64 channel modes
Test all c1500 and i1500 modes using B0329+54
3C343 will also be observed as a flux calibrator for the above

Details

Joe started switch over to v18.2 at 12:30
Switch over finished at 12:55
Will start off testing c0800 search modes on B1937+21
Submitted AcceptanceTestsPSRCODD800_search SB at 12:58
- Initial balance failed. Trying again.
- No issues on next attempt
- Scan #1 is c0800x0128 cal
  - Seems to have run OK
- Scan #2 is c0800x0128 search
  - File sizes seem appropriate
  - Can't see the pulsar in quick fold
- Scan #3 is c0800x0256 cal
  - Seems OK
- Scan #4 is c0800x0256 search
  - Again, data being recorded, but no pulsar
    - Possible that vegas.scale is too low
- Scan #5 is c0800x0512 cal
  - Looks good
- Noticed that GUPPI is hung up. daq_server seems to have restarted during previous scan. Seems to have recovered on its own, though.
- Scan #6 is c0800x0512 search
  - This time the pulsar is there but the pulse profile is very broad.
D'oh! vegas.dm was set to zero!!! Setting it to the DM of B1937+21
First, resetting GUPPI gpus
Submitting AcceptanceTestsPSRCODD800_search SB again with proper DM at 13:24
- Scan #7 is c0800x0128 cal
  - Looks good
- Scan #8 is c0800x0128 search
  - Seeing the pulsar, but profile is still very broad.
- Scan #9 is c0800x0256 cal
  - Looks good
- Scan #10 is c0800x0256 search
  - Profile still very broad
  - GUPPI data is the same. So it's not VEGAS specific. I don't think my DM could be off by that much. I'm using the same DM as in the NANOGrav ephemeris.
- Scan #11 is c0800x0512 cal
  - Looks good
- Scan #12 is c0800x0512 search
  - This one looks fine * Scan #13 is c0800x1024 cal
  - Looks good * Scan #14 is c0800x1024 search
  - Looks good * Scan #15 is c0800x2048 cal
  - Looks good
- Scan #16 is c0800x2048 search
  - Looks good
- Scan #17 is c0800x4096 cal
  - Looks good * Scan #18 is c0800x4096 search
  - HPC5 ge stuck in a blocked status. Other HPCs seems OK. Profile looks kind of broad again, though. * Trying c0800x4096 again * Scan #18 is c0800x4096 cal
  - Looks good. * Scan #19 is c0800x4096 search
  - No I/O issues this time. Pulse still broad. Weird.
Testing c0800x0032 and c0800x0064 modes.
Submitted AcceptanceTestsPSRCODD800_32_64_chan SB at 13:54
- Banks G and H aborted. HPC taking too long to be ready (2802 ms). Managers self-recovered.
- Resubmitted SB.
- Aborted again. Problem with 32-channel mode?
- Submitted one more time. This time aborted due to failed balance. Samplers not updated.
- One more try. Failed again on balance. Skipping 32-channel mode.
- Still taking a long time to configure for c0800x0064, but seems to be working.
- Scan #22 is c0800x0064 cal
  - Looks OK
- Scan #23 is c0800x0064 fold
  - Nasty RFI but pulsar is OK
- Scan #24 is c0800x0064 search
  - Profile is broad again. Very strange.
Moving on to i0800x0032 and i0800x0064
Submitted AcceptanceTestsPSRInco800_32_64_chan at 14:11
- Scan #25 is i0800x0032 cal
  - Looks OK
- Scan #26 is i0800x0032 fold
  - Looks OK, but there may be some baseline ripples...
- Scan #27 is i0800x0032 search
  - Same as above
- Scan #28 is i0800x0064 cal
  - RFI, but otherwise looks OK
- Scan #29 is i0800x0064 fold
  - Ripples again, maybe some ghosting? Hard to tell because pulsar is so bright.
  - Looks like an odd feature in lower 100 MHz. Pulse profile seems to be reversed. Some sort of aliasing?
- Scan #30 is i0800x0064 search
  - Same reversed pulse feature in lower 100 MHz...
Submitted AcceptenceTestsPSRInco800_search at 14:27
- Scan #31 is i0800x0128 cal
  - Looks OK
- Scan #32 is i0800x0128 search
  - Reversed pulse again, but this time it looks like its only in the bottom 50 MHz
- Scan #33 is i0800x0256 cal
  - Overflow artifacts. Otherwise OK
- Scan #34 is i0800x0256 search
  - Reversed pulse now in lower 12.5 MHz only
- Scan #35 is i0800x0512 cal
  - Looks OK
- Scan #36 is i0800x0512 search
  - Looks OK. Reversed pulse still there
- Scan #37 is i0800x1024 cal
  - Looks OK
- Scan #38 is i0800x1024 search
  - Looks good. Reversed pulse still in bottom 8 channels (it seems)
- Scan #39 is i0800x2048 cal
  - Looks OK
- Scan #40 is i0800x2048 search
  - No reversed pulse is obvious, but perhaps it fell out of band sensitivity
- Scan #41 is i0800x4096 cal
  - Looks OK
- Scan #42 is i0800x4096 search
Switching to C-Band at 14:51
Will check c1500 modes now
Submitted AcceptanceTestsPSRCODD1500 at 14:53
- Scan #43 is c1500x0032 cal
  - VEGAS is getting way too much power. Resubmitting to rebalance.
- Scan #44 is c1500x0032 cal
  - VEGAS does not seem to be initializing properly. Seems like there is still a problem with the 32 channel coherent modes.
- Skipping c1500x0032
- Scan #45 is c1500x0064 cal
  - Looks good
- Scan #46 is c1500x0064 fold
  - There is that baseline ripple!
  - Breakthrough doesn't see this, so what is the cause...?
- Scan #47 is c1500x0064 search
  - Ripple not as bad, but still there. * Scan #48 is c1500x0128 cal
  - HPC program taking too long to be ready. Aborted. * Resubmitted * Scan #50 is c1500x0128 cal
  - Looks OK
- Scan #51 is c1500x0128 fold
  - Looks OK other than ripple
- Scan #52 is c1500x0128 search
  - Pulse profile is totally weird.
- Scan #53 is c1500x0256 cal
  - Looks OK
- Scan #54 is c1500x0256 fold
  - Still see ripple
- Scan #55 is c1500x0256 search
  - Bad fold again. Is the time stamp wrong?
- Scan #56 is c1500x0512 cal
  - No cal. I think the levels may be too low. Also may have had some dropped packets on this bank
- Scan #57 is c1500x0512 fold
  - No pulsar
- Scan #58 is c1500x0512 search
  - No signal
- Scan #59 is c1500x1024 cal
  - Looks fine
- Scan #60 is c1500x1024 fold
  - No pulsar. Lots of dropped packets.
- Scan #61 is c1500x1024 search
  - Lots of dropped packets
- Scan #62 is c1500x2048 cal
  - Lots of dropped packets
- Scan #63 is c1500x2048 fold
  - Lots of dropped packets
- Scan #64 is c1500x2048 search
  - Lots of dropped packets
- Scan #65 is c1500x4096 cal
  - Lots of dropped packets
- Scan #66 is c1500x4096 fold
  - Lots of dropped packets
- Scan #67 is c1500x4096 search
  - Lots of dropped packets
Moving on to i1500 modes
Submitted AcceptanceTestsPSRInco1500 SB at 15:41
- Got a bunch of faults about invalid registers in i1500x0032. BOF file seems OK. Will give it another try.
- Second attempt seems OK
- Forgot I put an AutoPeakFocus in here. Oh well. Tsys is around 19 K
- Scan #73 is i1500x0032 cal
  - Looks good
- Scan #74 is i1500x0032 fold
  - Looks good aside from ripples
- Scan #75 is i1500x0032 search
  - Same as above
- Scan #76 is i1500x0064 cal
  - Looks good
- Scan #77 is i1500x0064 fold
  - Pulsar looks fine save for ripples. Bandpass looks wrong, though, with an almost sawtooth pattern.
- Scan #78 is i1500x0064 search
  - Same as above
- Scan #79 is i1500x0128 cal
  - Looks OK
- Scan #80 is i1500x0128 fold
  - Pulsar is there with ripple
- Noticing Bank C channel 2 is still about 5 dB too low
- Scan #81 is i1500x0128 search
  - Same as above
- Scan #82 is i1500x0256 cal
  - Looks good
- Scan #83 is i1500x0256 fold
  - Pulsar is there with ripples
- Scan #84 is i1500x0256 search
  - Same as above
- Scan #85 is i1500x0512 cal
  - Looks OK
- Scan #86 is i1500x0512 fold
  - Pulsar there with ripples
- Scan #87 is i1500x0512 search
  - Same as above
- Scan #88 is i1500x1024 cal
  - Looks OK
- Scan #89 is i1500x1024 fold
  - Pulsar there with ripples
- Scan #90 is i1500x1024 search
  - Same as above
- Scan #91 is i1500x2048 cal
  - Looks good
- Scan #92 is i1500x2048 fold
  - Small number of dropped packets
  - Pulsar there with ripple
- Scan #93 is i1500x2048 search
  - Same as above. Lots of dropped packets on Bank B, but only Bank B.
  - Same as above on Bank A
- Scan #94 is i1500x4096 cal
  - About 5% dropped packets
  - Cal still looks OK
- Scan #95 is i1500x4096 fold
  - About 5% dropped packets
  - Pulsar there but degraded
- Scan #96 is i1500x4096 search
  - 70% dropped packets
Submitted AcceptanceTestsFluxcalInco1500
- Scans #97/98 are i1500x0032 on/off pair
  - Looks good
- Scans #99/100 are i1500x0064 on/off pair
  - Looks good
- Scans #101/102 are i1500x0128 on/off pair
  - Looks good
- Scans #103/104 are i1500x0256 on/off pair
  - Looks good
- Scans #105/106 are i1500x0512 on/off pair
  - Looks good
- Scans #107/108 are i1500x1024 on/off pair
  - Looks good
- Scans #109/110 are i1500x2048 on/off pair
  - Looks good
- Scans #111/112 are i1500x4096 on/off pair
  - Looks good
Submitted AcceptanceTestsFluxcalCODD1500
- First c1500x0032 scan didn't start correctly
- Next scan got an HPC program taking too long to be ready
- Skipping c1500x0032
- Scans #115/116 is c1500x0064
  - Looks good
- Scans #117/118 is c1500x0128
  - Looks good
- Scans #119/120 is c1500x0256
  - Looks good
- Scans #121/122 is c1500x0512
  - Looks good
- Scans #123/124 is c1500x1024
  - Looks good
Have some extra time. Might as well do some FSW scans.
Submitted AcceptanceTestsSpecFSW SB at 17:40
- Samplers in VEGAS CLEO screen seem to updating with reasonable values but vegasdm chart recorders are not updating for any banks other than A. pol0 histograms are updating but not pol1.
- Scan #125 is Mode 1 FSW
  - Looks OK
  - Data for other IFs, pols, looks OK
- Scan #126 is Mode 4 FSW
  - Looks OK
- Scan #127 is Mode 7 FSW
  - Looks OK
- Scan #128 is Mode 10 FSW
  - Looks OK
- I realized that L-Band was not in focus. But we still saw the galaxy! Putting it back in.
- Scan #129 is Mode 1 FSW
  - Looks OK
- Scan #130 is Mode 4 FSW
  - Looks OK
- Scan #131 is Mode 7 FSW
  - Looks OK
- Scan #132 is Mode 10 FSW
  - Looks OK
Turned things over to Joe to switch back to v16.3 at 18:03
Ran a scan in Mode 29, FSW to check out
- Had to cycle VEGAS off/on a couple times, but it checked out after that.
Session ends at 18:30

Conclusions

Bank C channel 2 is not balancing properly in i1500 modes. It is about 5 dB too low
Need to figure out baseline ripples in 1500 MHz modes
Need to figure out 8-channel pulse reflection in incoherent 800 MHz modes
Need to figure out 1500x0032 and c0800x0032 mode configuration
64 channel modes still have bad bandpass
vegasdm chart recorders and histograms seem to not come back to life when coming out of coherent pulsar to spectral line

-- RyanLynch - 2018-01-23

Topic revision: r5 - 2018-01-23, RyanLynch

CICADA

Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback