After Kyle found the leaking ESD style feed-thru on BSC8 (aLOGs 1595,1596,1604) we pulled another out of C&B and it was checked on the leak detector.  KKyle found that to be good.  Since we are flanging BSC6 at this time, it seemed reasonable to check for leaks pre-install.  This feed-thru did not pass and has a 2e-7 Torr-L/sec leak.  So that is two of three feed-thrus that have gone through Clean & Bake that have failed.  We grabbed the one RMcCarthy had in EE and leak checked it as well.  While this unit had not been through clean and bake it too failed at 3e-7Torr-L/sec.  These leaks we've found to be associated with the ceramic portion of the seal rather than the welding.
SO leaking on 3 of 4 units so far--we'll check more certainly soon.
The leaks would not reveal themselves until we put the He source right down in the ceramic zone.  It was not sucked it at the surface near the welds.  Also it only showed up at 1 of the 5 conductors--they are all separate, see image.