I had decided to just use the chamber as it was, but then Frank and Alastair both independently suggested that I at least spend a day or so trying to track down any egregious leaks that may exist. Alastair had bought 2 KF blanking flanges explicitly for this purpose, so it was hard to resist.
The procedure is pretty time consuming and so is still underway, but it basically goes like this:
Going through all possible combinations is intractably time consuming and probably pointless, so I have gone about it in a somewhat haphazard but reasonable way. First I removed one corner chamber, then another chamber (so, 2 total) and the tube between them, and so on. I've gone through three iterations so far while doing other things today, and the fourth one I've left running overnight is just one corner chamber (the one to which the pump is connected) and two hoses. I think the last two that I'll try are just 1) one chamber by itself and 2) one hose by itself.
The goal is to see if any combination is drastically better than the entire system. So far, all combinations get to ~7 uTorr within 2-3 hours. If there is no obvious "bad part", then I will chalk it up to the overall quality of the o-rings and/or residual water in the tank.
It should be reiterated that there is no "goal pressure" for the gyro so far. The original idea was just to evacuate the area within the cavity to reduce air noise from index of refraction fluctuations and buffeting. So, uTorr vacuum should be fine---and perhaps even the mTorr level we get without a pump attached---but it can't hurt to remove any gaping holes in the system.
1, do not use solvent to clean o-rings,try dry lint free wipers
2, use solvent to detect leaks on metal seals only
3, spray some gas or bag-fill suspected piece
4, remove water and solvent saturated o-rings and replace them with dry cleaned ones