billdeg
Technician
I was not planning to fix the G114 right yet, but I can compare and contrast with the working card, makes it a lot easier to pinpoint the failure (with or without a schematic).
I figured out that if I put a piece of shielding behind the core memory H217D card that would solve all of my weird RAM inhibit problems. I have since been able to load BASIC via papertape and it now runs flawlessly. I am using only the first 8K of RAM so I can put something else in the remaining 8K available (H217D is non-parity 16K RAM).
Yes "the bag" experiment. Not really wrapped around everything, but the way things are laid out on the backplane one could speculate to shield both the H217D from the G114, as well as the M7261. So, I just folded the bag over and between the two target areas. I probably don't need the shielding between the back of the M7261 and the front of the M930.l
Thanks for the update. I don't understand why what is intended as electrostatic shielding would work in a situation where magnetic coupling would presumably be the problem. I wouldn't think that there would be enough effect compared to a good plane-conductor. Can't argue with success, 'tho. It looks like your bags are the metallized type? Ref: http://www.esdjournal.com/techpapr/ryne/esdbags.htm
Note that, aside from the magnetic shields used in the 8/e-f-m, DEC also included shields in the MF11 (and variants; up to 3x 8K stacks). They were placed between the stack module-sets, and not between specific module-pairs. The ED references "electromagnetic shield (1700021-02) required between slots 03 and 04, and slots 06 and 07 when optional memory segments are added. Shields are positioned in module sections C, D, E, and F and held in with module holders (1209856-0-021)".
So the shield is not so much to keep the core arrays from affecting other logic, but to keep all the noisy digital logic out of the sensitive analog core array circuits.