I ran two tests, on 9/5/12 and 9/27/12, while keeping as much similar between the two tests as possible. The RGA was kept at it's default settings, for the head 70eV electron energy and 90V focus . The CEM was operating at a gain of ~30,000. This test should probably be re-ran with a lower ionizing voltage of 50eV where Rai worked out a good range for Ne sensitivity but lower Ar++ rates. The GS-13 being tested was repaired and loaded with trace Ne here at LHO and loaded into a modified VBO-A (represented by the block diagram). 

The same GS-13 pod was tested in VBO-A without having removed it from the chamber. The first set of graphs show two 10 minute accumulations, the first peak being a 4.88x10-9 torr-L/s calibrated leak + a GS-13 pod, the second peak being a 10 minute only pod accumulation. The second set of graphs show the overlay of the two relevant peaks from both the tests and their corresponding fits. Unfortunately, for the fit I've used some normalization on the data to get the fitting algorithm to work (if anyone has some good suggestions for getting matlab to fit this without that I am all ears). The two tests calculate out to 1.22x10-9 torr-L/s 1.54x10-9 torr-L/s pod leak rates respectively.

The air leak into the chamber can be sort of fickle as my skill with sealing the copper wire is not the greatest and some fluctuation can be expected in between tests. Since these two tests were run without the chamber being opened a greater variation could be expected in between tests than this shows. Hopefully we can have some confidence in the podded seismometers going into H1 using VBO-A without having to continuously ship to and fro LLO.