Quote Originally Posted by steveo View Post
i agree with your theory, but in reality, i have left it overnight programming in a loop with zero errors, and resistance between the two boards at the factory ground point is zero when just grounded through the connectors in this way. i did alligator clip the big ground trace between boards but when i took it away there was no difference.
Great that it's working for you. The resistance may be zero, but what about the impedance? As frequency goes up those longer grounds start to look inductive. Depending on activity it is possible the grounds can shift. When working there will be no difference other than noise immunity by adding grounding. Most things have a good margin though and will work (?)

The way I would set my test bench is just slightly different. The cards can easily be stood up in a fabricated wood frame such that solder sides face each other: just like in the stock case. This will place the grounds between the connectors close to each other. Then a short braid ground can be used. This is what I plan.

Another good thing about doing it this way is that you don't need to unbend the ribbon cable. Mine failed after just a little handling. IMHO IDC connections should not have been used...

-Tom