On Tuesday 06/Mar/2012, we pulled one of D1000223 cables (S1104079, used for picomotors, connecting the feedthrough and the connector bracket on the ISI table) and replaced with S1104077, which is from a different bake batch. Now all picomotors work.
We tested the bad cable using a cable tester. It tested fine at first, but when I wiggled the cable near the connector I was able to repeatedly make it fail.
Richard wants to test this cable using HV mode of the tester.
-
Originally pin1, 6, 7 and 13 were all connected (in-vac side) on Mar 02.
-
It was divided into two groups, 1-13 and 6-7 when I wiggled the cable on the cable bracket on ISI on Mar 02.
-
Then somehow pin 6 and 7 were disconnected from each other when I disconnected and reconnected the cables on ISI on Mar 02.
-
Through this process and other measurements, we identified that pin1-13 problem was in D1000223-S1104079, but could not identify the source of pin6-7.
-
On Mar 06 morning, before the cable was swapped, we repeated the test. We've found that pin6 and 7 were again connected. Something apparently crept back, though we don't know what.
-
After swapping the cable, everything was fine on Mar 06.
-
There's still some chance that pin6-7 will creep back to bad status (if the problem was not in the swapped cable). Pin 6 and 7 are used for red QPD picomotors, which are not used for one arm test.