Kyle, Gerardo 

1140 hrs. local -> Spun up LD, Turbo and QDP80  

1330 hrs. local -> 7.1 x 10-8 torr*L/sec with cal-leak open -> closed cal-leak -> < 10-11 torr*L/sec with cal-leak closed -> Calibration is OK

1335 hrs. local -> Valved-in Turbo to Y-mid volume -> Y-mid background initially found to be 1.8 x 10-8 torr*L/sec -> Fell steadily to 8 x 10-9 torr*L/sec over the next 60 minutes of pumping

Begin Testing

1435 hrs. local -> Began leak testing aLIGO GNB spool welds -> "Bagged" flange-to-spool welds in 180 degree sections -> Applied 5 LPM helium flow at bottom of bag and pumped at top of bag via O2 sensor/pump -> Started 100 second timer when 0% < O2 < 7% -> No response for either flange (4 bagged sections) -> Tested butt welds of cylinder spool roll -> Bagged butt weld in three sections (from flange to stiffener, between stiffeners and from stiffener to other flange) -> Background remained steady at 8 x 10-9 torr*L/sec during entire test period until a few seconds before the conclusion of the last weld test at which point it began to climb -> Stopped helium flow -> signal continued to climb -> removed bags from two most recent welds tested -> signal continued to rise -> Signal peaked at 1.2 x 10-7 torr*L/sec (tens of minutes after stoppage of helium flow) -> Isolated LD from Turbo exhaust -> signal immediately fell to < 10-11 torr*L/sec -> Recombined LD to Turbo exhaust -> Opened up building exterior doors to air out room.  

1645 hrs. local -> Valved-out turbo from Y-mid volume and shut down LD and pumps with Y-mid background at 4.6 x 10-8 torr*L/sec. 

Conclusion:  In retrospect, we should have reduced the helium flow rate once a minimum O2% was achieved.  As is, we put more helium into the room that was necessary (calculate a total of 85 liters over 90 minutes).  As such, we cannot conclude that the last weld tested was the source of the signal response.  We did demonstrate, however, that there is a leak(s) at the Y-mid (as opposed to other explanations) when we isolated the LD from the Turbo exhaust and the signal immediately fell off.

I propose that for the next session we pump the Y-mid with the turbo for a few hours to get the background back into the 10-9 range then revisit the last two butt welds tested.