June 5, 2014 Testing

Goal:

• Test fixes for ripple tests in manager. start with mode 1. Test 2 & 3 as well. Might also be worth testing wider bandwidth mode (mode4?). Also want to try the whole system. Also want to do all eight banks.
• run integrat fix tests. Start with UTCDELTA. Anish wants this run in all eight banks at the minimum integration time. I'm wondering if I should run this with modes 10-19 since the bof file for these modes is broken. The bof file for modes 20-29 is the same as the bof file for modes 10-19. If I don't go for the full run of modes today, I can also do some switching observations for modes 1-9

Info:

• Session: TGBT13B_502_65

Log:

• Scan 45: RippleQbandmode1. Manager configured to do MMCM calibration. Yay! Most of the ripples are gone!!!!
• Paul putting fixes into manager "for real"
• Tried scan and it failed with SOAP errors
• restart turtle & grail
• Scan 46: RippleQbandmode1. Trying again to make sure manager fix held. It did!!
• Scan 47: RippleQbandmode2. Trying a mode change. Error on Bank B (vegas-hpc2): "Difference in measured and stored MMCM Optimal Phase for zdok1 exceed tolerance with values of 8 vs. 46." Paul says that's because there can be "multiple" good MMCM phase results. He's going to look into it.
• mode 2 looked good except for one polarization on Bank B. This is the same one that the error message above was generated for.
• Paul made a fix to the calibration values and we're trying again.
• Scan 48: RippleQbandmode2. Trying again. No error message. result looks good.
• Scan 49: RippleQbandmode3. Trying a different mode. Works!
• Scan 50: RippleQbandmode25. Trying a mode that Paul had trouble finding values for. Activate failed. Paul is fixing it. Aborted scan.
• Scan 51: RippleQbandmode25. Trying the mode that Paul had trouble finding values for.
• Scan 52: RippleQbandmode22. Trying mode22 (RAMPS mode). Changed to 5min integrations. Most banks/pols looked fine except for Bank C, polarization 0 (sampler C1_0).
• Turning off MMCM calibration to see if the MMCM is making things better or worse.
• Scan 53: RippleQbandmode22. Trying again to see what MMCM is doing in this case.
• Turning back on MMCM calibration to see if it's repeatable.
• Scan 54: RippleQbandmode22. Trying to see if the structure in C1_0 is repeatable.
• Scan 55: RippleQbandmode24. Trying to see if the structure is repeatable in a related mode: same bof, same clock rate. Looks good. Some center features look like the calibration hasn't been applied so we want to double-check.
• Scan 56: RippleQbandmode1. Checking a mode that was good before to make sure the MMCM calibrations are still being applied. Looks good so calibrations look like they are being applied.
• Scan 57: RippleQbandmode4. Checking new bof and clock frequency. Results look good. Didn't see as much of a comb structure.
• Scan 58: RippleQbandmode7. Checking new bof and clock frequency. Looks horrible. There's a step with signal in the middle.
• Scan 59: RippleQbandmode6. Checking new bof and clock frequency. Looks good.
• Scan 60: RippleQbandmode9. Looks horrible. There's a step with signal in the middle.
• Turned out that the OGP and INL calibrations hasn't been applied to the cards. So Paul copied the calibrations over and applied them to banks A-D.
• Scan 61: RippleQbandmode22. Running again so see if OGP + MMCM calibration helps.
• Scan 62: RippleQbandmode22. Running again back to back to see if something is bad.
• Scan 63: RippleQbandmode20.
• Scan 64: RippleQbandmode24.
• Scan 65-67: RippleQbandmode22.
• Turn off MMCM.
• Scan 68: RippleQbandmode7. looks fine.
• Turn on MMCM.
• Scan 69: RippleQbandmode7. Wait. The MMCM were not on.
• Turning on the MMCM for real this time.
• Scan 70: RippleQbandmode7
• Scan 71: RippleQbandmode1. Double-checking that it's on.