GENERAL: Jonathan flying. Over Lauder around 21:51 AO2: Sample flow started while fueling 23:00ish, AEROS not working. Disconnected,reconnected ethernet cable. Should have pinged to make sure, but I was in a rush. Accidentally have hit little button on back outboard side, but hit it again. AO2 data seemed fine afterwards on AEROS, so didn't change. 00:38 Went back into VNC to check on health of AO2 system 00:41 Recorded diagnostic vals from earlier in check sheet 02:30 Noting that AO2CO2 has distinct upwards trend from beginning of flight to now, roughly 380-405. MEDUSA: Pos #23 (1007) has crunchy threads Pos #22 (and possibly #21) shows a leak Pos #31 has broken valve stem on, I believe, both sides. Sample will be lost. (All notes until 23:39 are written after the fact... this is the best I can recall of what has happened so far--some details can be checked on AEROS, some not.) 21:10ish Rushed preflight led to not checking flows until just before takeoff. Flows were not ok. Pbyp. was reading low, and wandering btwn 710 and 740, not achieving the intended 783 (By P sensor is reading roughly +23). Flow seemed ok. 21:13:24 Upon takeoff, byp is still issue, but I decide to go through prepurge anyway because we need to start it 21:30 Pos 31 seems to have issue with P--flow drops. Will have to check 22:00 Flow is decreasing in Pos 1. Trap thought to be to blame. Upon checking Pos 1,2,3,5, and eventually 19, 30, it becomes obvious this is not just in bypass or in one pos, so is universal, and separate (tho possibly indirectly related to) from low bypass pressure issue. 22:10 or so, removed upstream trap after shutting off, VB, then Pump. Restarted pump, turned on VB, and flow returns to 9.98V in bypass. Inspection of inlet 5 mins later shows that it WAS very blocked. Pb is still low, though. 22:30, roughly, shut down VB, Pump, and set PC1,2 OPEN and recorded values: Pu=266 Pd=232 Pb=229 and these rise as we are descending. Always Pu is greater than other by 30-50 torr as the P must go past pumps to get to d,b This was intended by Britt to be a meas of cabin P somehow to see whether Bypass P sensor is just reading wrong... impossible to do this test w/o disconnecting a line and putting in sample loop. 22:36 Xchat goes down 22:41 Pu=351, Pd=299, Pb=299 22:44:30 increased Pdset to 450. Now controls Pb to 782, which is desired val with +20 offset, so this will be dst setpt for now 22:50+? Concnered with whether system will recognize samples, tried to find program to download flask table. Could not find. So, did not reset flask table. 23:10 Decide to take one additional flask at beginning to ensure sampling protocol is working on MED computer. This worked fine. NOTA BENE: values for "sample duration" in MED Comp flask table will be representative of the total time flushed for various positions, including trouble-shooting. Thus, total time of sample flushing will disagree with deltaT between samples. 23:16 Take this "extra" flask and sample protocol works on comp. Second flask, roughly identical is taken at 23:19:37 as the 28K flask on 23:?? Reset Vlv2 to pos 1. VLVset2, though, still says 16, even though I chnged w/ hand link. CAN THIS BE CHANGED IN PROGRAM? Noted in AO2_MEDUSA_TODO_090121 00:50 Upon box switching to pos 17, checked that valve hadn't rotated to 2 after my corrective actions. It had not. Valve was in correct place. 00:55 Upon switch to box 2, flow drops to 9.8V, ~3050 sccm. 01:05-08 CLOSED Flask 1-16. Inspection of Pos 31, Flask 1205, reveals that the valve stem (within barrel) is broken. 01:21 It becomes apparent that pilots and scientists have miscommunicated on how many dips there will be, so initial sampling was a bit more frequent/close together than needed. 01:02 Pbypass starts steadily climbing to 827 by 01:30. This seems considerably faster than P increase over other samples. Causes me to worry that parker valves may not be switching as cleanly after switching out rotor and depositing additional debris in, upst and downs of the 34-Port 01:34 Continued increase in P to bypass loop (now 835) seems to suggest, in addition to a larger leak, that the Bypass P was NOT representative of the sample pressure. As long as leak is, indeed, across Parker NO/NC, the bypass pressure should try to equil with the sample P. It seems likely, then, that this gradual increase is due to the fact that by increasing Pdownset, we increased the P in all of the flasks. 01:37 Pb=842 01:41 Pb=854, then took sample, and it dropped to 720 02:02 Realizing that pilots are using PALTF, and we are using GALFT. ALL VALS NOTED for alt EXCEPT 28 are in GALTF. Samples taken at tops of dips are all at 28K by pilots'/PALTF's measurement. ADDITIONALLY, PALTF and GALTF agree less and less as we get farther poleward, with the value of GALTF dropping relative to PALTF. At 65/66 degrees South, PALFT read 28K, and GALTF read ~26.5. In contrast, at 48 degrees South, PALTF read 28K and GALTF read ~28.5. Steve says this is due to geoid effect. 02:53 Flow falls a bit, for sample 30, but I don't notice that this is due to passing 39Kft/inlet P setpoint. Sample 30 is thus also compromised. 02:59 After flushing Pos 31 anyway, so as not to disrupt Sample Sequence, I realize P set is not controlling, and switch to 90/670 setpoints. 03:04-07 Closed flasks 17-32. Closed 32 first, in case of small leak to bypass. Closed 32 only about 1 minute after finishing sampling it. Managed to close flask 31. I believe what happened to flow during prepurge is that flask stem sheared in barrel, with the tip still plugging the out port, and that P in the flask built until the tip of the stem popped out. At this point, the flow started to return to normal. As long as flask can be opened successfully in lab, this may still be usable, though it needs repair afterwards. 03:07 Leaving Pump on for the time being to see how flow does with time, and what Pbypass does. 03:58 Noted that WT P is only 686 vs 792 on sample. Believe this has been the case for a good while. CHECK. May be due to pulling out WT black reg. knob during preflight?