The following questions are 'informal' questions I've got continously and this is an attempt to explain the reason the design look like it does.

Q: Do we really need HV-grouping since there is a continous ground-plane on top?

A: There have been long and intense discussions about this issue.
The decision is to wait and see the outcome of the measurements done at Indiana and Duke.
You can read Indiana's preliminary report in postscript or framemaker source
and Duke's preliminary report at postscript or MS-WORD


Q: Why the fuzz over HV-grouping? Why not just keep the grouping and keep everyone happy!?

A: Easier said than done! The mechanical constraints on the tp-designs are quite severe.
It's simply not possible to just place the stamp-cards directly on top of its HV-group though this solution would probably be the best and simplest.
As it is now, tp#1 and tp#2 are possible to design with respect to HV-groups, but tp#3 can not be designed using only one signal-layer if HV-groups should be kept strictly.
It is of course possible to use more signal-layers in a multi-layer board, but that increases the cost and also introduces via-holes which will be possible gas-leaks.
It's also possible to use non through-going (=blind) vias or vias with the drill-holes filled with epoxy (=plugged) but these techniques are quite expensive and also have a longer delivery time that simple 2-layer boards.
And even though reliability of PCB's are extremely high, the reliebility decreases with the number of layers and the same goes for withstanding HV-surges - the thinner the isolating layer the lower maximum voltage.


Q: Is it necessary to use protection-resistors on the tension-plate?
Can't they be placed on the read-out electronics instead?

A: On the old read-out electronics consisting of two different boards piggy-backed, there simply wasn't enough room left for resistors.
However, on the new updated flexible stamp-card, there is a possibility to put resistors directly on the stamp-card.
There are three different plots available for tp#1 design, one with HV-grouping intact + resistors, one with HV-grouping but without resistors and finally one without either HV-grouping OR resistors.


Q: Won't a re-design of the tension-plates to exclude resistors take too long?
Won't the schedule be impacted?

A: Since most work is already done on the designs, the total design time will be around a week for all three designs.
The re-designed tp's must of course be validated by Duke, Indiana and Penn too see that they are in accordance with the mechanics and the old designs - except for resistors of course!
The delay for the PCB submission will be around 2 weeks, but this time is gained when we don't have to mount resistors.
The mounting time offered by companies are quoted to be minimum 2 weeks, so in total the time before manual production will actually be shorter since Christmas Holiday impact will be less!


Q: What about the pin-socket pairs for the stamp-cards - they haven't been qualified!

A: It's true that the pins-sockets haven't been qualified, but no qualified connector fits the design without severe, difficult and long re-designing involved.
The pins and sockets are both gold-plated to minimize risk of oxidation and minimize contact-resistance and they are machined in stead of 'stamped' pins which are much cheaper but of less mechanical strength.
This type of pins are meant to be used on PGA-packages and the sockets are similar to those found in high-quality IC-socket. Both types are available in MIL-spec, so they should not have too big problems being officially qualified. Unfortunately an official qualification takes too long for it to be available until production of tp's, so a decision has to be made whether they seem good enough or if we should postpone production.
What looks like surface-mount pads on the stamp-card are actually blind pin-holes which increases mechanical strength enormously.
My own tests on scrap-boards have shown that it's easier to break a pin in two or break of the solder-joint than to break off the pad. This shows the pin itself to be almost equally rugged as a hole-mounted pin.