TGBT15A_915_61 Spectral line balancing in VPM branch

Goals

• Run the CheckModeN scripts in TGBT15A_915 in the new branch, and the released branch
• Check using Ryan's regression script that the results agree

Details

New Version

• Justin switches to VPM branch in the morning
• 1:15pm Get logged in and set up.
• We are getting flood with "lost connection: RPC connection from BankNMgr on vegas to IF Manager" messages
• VEGAS built agains wrong version of IF Manager
• Astrid cannot establish connection to Scheduling Block Executor...
• Scan 1 aborts because antenna data being logged to wrong directory; remove Antenna from scan coordinator
• CheckMode1 - somehow we are on scan 5
• realize that it is already balanced, start over deliberately unbalancing before I start
• CheckMode1 - Scan 8
• CheckMode4 - claims to fail, but values look ok - and Converter rack still has attenuation available. (CM5 and 14)
• CheckMode5 - ok; go back to 4; maybe I was looking at the wrong attenuators?
• CheckMode4 - scan 13, balances ok, but was at correct value.
• CheckMode4 - scan 14, balances ok, from incorrect values; input and flitersnap values look ok
• CheckMode5 - scan 15 - ok from incorrect values. Stop changing them at this point, or we will run out of time: just do IF rack
• 7 through11 - scans 17 - 21 - ok
• Mode 12 Scan 22 - aborts due to invalid data, and then gets hung in aborting. Turn managers off and on again
• Mode 12 - scan 23 - abort again. Justin restarts all of the Managers
• First configuration after the restart causes a flood of SOAP "Bad device" messages
• First balance after restart gives "Exception can't convert power value to a float", and also failes to balance Bank E.
• Next attempt fails with abort due to invalid data - coincidence? Justin restarts them all yet again...
• Bank A would not turn on - needs a second try, then comes back.
• All get invalid data again (all banks, as before). Decide we have to switch back to release version at this point.

Released Version

• 2:45pm - Justin switches back - seems to take a long time
• I try to run from planck, which clearly doesn't work
• 2:57pm - get going
• Scan 26 CheckMode1
• Scan 27 - Modes 2 - 11
• End of observing at 3:15pm

Scan consistency checks

Release Version