Note, the initial tests demonstrated that an MCM can short out without affecting readout of the MCMs closer to the TEM. However, the current draw was higher than expected and also depended on the GTRC address. The problem was identified to be current flow on the logic-1 address lines through the GTRC address pad input protection diode. To rectify this we are adding resistors to the flex-circuit cables to limit the current flow (see the latest drawing of LAT-TD-02049). Below is a test of this, showing that in the end the short-circuit current is consistent with the expected polyswitch hold current.
I made the measurements we talked about and everything works like you
Board on pos.3 - DVDD shorted - 1.44 A to the ground
Board on pos.1 - DVDD shorted - 1.18 A to the ground
Board on pos.0 - DVDD shorted - 0.62 A to the ground
Board on pos.1 - 227Ohm in each A0 line - DVDD shorted - 0.60 A to the
I tested also the functionality. When the board on pos1 is shorted, then the
board on pos 0 is still operable.