Simulation and test of the EField behaviour around the Power Supply Pad
One of the big problems in designing the anode readout plane is the routing of the power supply lines.
The Saclay people told me that there should be any problem of cross talking through the VIAs if the power supply is clean but you should be careful to place the power pad far enough from the readout pads to avoid sparks. They gave me som empirical numbers to calculate the minimum safe distance: 1 cm per 10 KV times 2 because of the Argon filling, plus some safety.
The maximum voltages we may want to apply is of the order of magnitude of 8KV for the worst case scenario which implies a safe distance of 1.6 cm at least which is not acceptable becausw it would mean losing almost three rows.
The grid can be of help in this case because, if it is to cover completely the power supply pads it will act as a insulant between the power supply pad and the readout pad. For this procedure to be most efficient I'm thinking of setting the power supply pad in some inner layer of the PCB, cutting away the above insulating layers and setting the grid over it. The electrical connection will be guaranteed through a conductive paste that should not overflow from the internal side of the grid itself. On the lower side of this first grid element there will be a groove that will allow for the overflow of the conductive paste preferentially on the external side. Any way we should put some insulating glue on the internal area.
Because of all this complex features we may even want to avoid this first stage to be made by a ceramic element and we could try to design a small base which can be modeled this way with reduced costs. One idea would be to build this thing out of the same material the PCB is made, maybe even together with the PCB itself.
The easiest way would be to put the grid on between the power supply pad and the readout pad. This will waste a little bit more space (of the order of 1 mm). We may anyway recover that space if we use a slightly more complex design for the grid. For example we may extend the grid over the ecteral border in the screwing areas
We may anyway test the sparking properties in a particular test anode board