[Halld-tracking-hw] CDC feedthrus and end plate

[Fernando J. Barbosa]

Hi Gerard,

Press-fitting the socket into the Al plate looks like a good idea but I 
am concerned about the quality of the connection over time. Perhaps 
having a slightly oversized hole and using conductive epoxy, followed by 
Humiseal 1B31 or 1B73 on the surface would prevent corrosion on the 
interface (using the closed socket part #).

The ground wire needs to be insulated because the straw interconnect 
cables, although insulated with PVC, are at HV. However, it can be solid 
to plug directly into the socket but I have some reservations, and you 
point these out, as to the cut/finish of the wire. I favor a crimped pin.

Best regards,
Fernando


Gerard Visser wrote:
> Hi Fernando and Tim,
>     Please consider as I mentioned the use of a pressfit pin socket in 
> the endplate if space is at a premium. [Well, ok, I think we know that 
> is true...] I don't know if this is really practical in aluminum, it 
> certainly works in printed circuit boards. I think it would be 
> relatively easy for you to make some tests, if you can afford to delay 
> releasing the endplate drawings for a little while yet.
>     Consultation with Mill-Max (or other vendor) might give a more 
> accurate judgement of which socket to select and whether it will work. 
> But naively I think - trying to stay small - that for instance 
> Mill-Max # 0252-0-15-15-32-27,
> http://www.mill-max.com/pin_rec_catalog/productInfo.cfm?webpartnumber=0252&start=1&leaddiameterrange=.015-.025&pin_or_rec=&PartDescription=Receptacle%20With%20No%20Tail&taildiameter=NONE&tailtype=NONE&mountingfeature=PRESSFIT&mountinghole=0.057&bodylength=&search=&pr=Rec&stagecode= 
>
>  pressed into a (round) hole 0.057+/-0.002 x 0.150 min depth (not to 
> break through, of course), and located anywhere that there is 0.080 
> diameter clear area on the outside surface of the endplate, would do 
> nicely. If you can find 152 such appropriate locations for this, one 
> near each preamp connector. We can certainly find a ~0.020 diameter 
> pin to crimp or solder on the ground wire, to plug into this socket.
>     I just think that this might be easier to accomodate than tapped 
> holes, and also easier to actually install the ground wires. Of 
> course, a lug under a screw is probably a better connection, so if 
> there is room for that I certainly don't want to dissuade you from 
> that approach.
>     Use of the pressfit pins should certainly be conditional on some 
> tests to see how they go in bulk aluminum, which is not their usual 
> application. And also probably to optimize the diameter. Also of 
> course, this approach would add an additional setup and processing 
> step to the endplate production, must cost a little.
>     Please consider it, unless you have a reason to rule this out 
> immediately.
>     Sincerely,
>
>         Gerard
>
> ps. The ground wires can be solid bare wire (#16 - #22), no? That 
> seems the easiest to me. [And actually #22 solid wire can be plugged 
> right into this socket, assuming it is cleanly cut; if using 
> tin-plated wire then the contact should be changed from gold to tin, 
> also a standard option - but leave the body plating gold. Bare copper 
> might be preferable though. Or a slightly larger pin for #18, e.g. 
> 0334-0-15-15-34-27-10-0. Maybe that is the way to go, looks even better.]
>
> Fernando J. Barbosa wrote:
>> Hi Gerard and Tim,
>>
>>     I am in agreement with Gerard. I have always made the point that 
>> the shorter the connection (any) the better and distributed 
>> throughout the end-plate. Slava has also made the point that a direct 
>> connection (say, a stud through the gas plenum) would be impossible 
>> due to the density of the straws and the location of the preamp card 
>> connectors.
>>     Ground wire (16 AWG w/FEP insulation?) connections in a regular 
>> pattern will work better than the peripheral connection attachment. 
>> Of course, the quantity of these connections will need to be 
>> determined but I agree that we should make provisions for as many as 
>> possible even if not used in the final configuration. I have not 
>> looked at the latest drawings in detail but I requested a large 
>> number of blind taped holes throughout the endplate when we were 
>> considering different alternatives to the straw connections in the 
>> past - maybe these were dropped along the way.
>>     The grounded disk is supposed to be copper and there are no major 
>> constraints on the thickness, other than cost. But this is already an 
>> expensive disk due to all the machining that will be required.
>>     Therefore, if there are no mechanical strength issues on the 
>> endplate, let's add the taped holes throughout the endplate and make 
>> the connections as Gerard suggests.
>>
>> Best regards,
>> Fernando
>>
>> Gerard Visser wrote:
>>> Hi Tim,
>>>     Yes, I know. This (152 direct lines) is what I had in mind. 
>>> Maybe it is too much to ask for. Your question whether we can live 
>>> without it is a good one, wish that I (or Fernando) knew the answer 
>>> 100% for sure. *Perhaps* the answer is yes. If I had to guess, I'd 
>>> say better than even odds.
>>>     One the other hand if you could for instance add the 152 
>>> features (4-40 blind taped hole? or a blind hole to fit a Mill-Max 
>>> press-fit pin socket? or something like this?) to the endplate, then 
>>> if the eventual answer turns out to be no, it does not mean the 
>>> finished, and I imagine rather expensive, endplate is unusable.
>>>     It was the absence of any such features as far as I could 
>>> understand in your drawing, that stimulated this discussion.
>>>     If it is simply impossible to have 152, or even 152/2 or 152/3 
>>> or 152/4 by some sharing scheme, then we must risk doing without it. 
>>> If it is possible, my opinion is the features should be added, they 
>>> can always be left unused if it turns out that we don't need them, 
>>> i.e., if grounds around the outside will not suffice.
>>>     By the way, the boards will (I really hope) be sitting on a 
>>> ground plane, e.g., the upstream plenum is a sheet of thin aluminum, 
>>> right? Because I can pretty much guarantee you that 152 preamp 
>>> boards with each a skinny 40 cm ground wire between the preamp and 
>>> the endplate, this will not work. If the connection is through a 
>>> common sheet of (preferably plated) aluminum, it stands a good 
>>> chance but I would still be nervous that the direct (152 5cm lines) 
>>> plan might prove to be necessary.
>>>     Well, I think we should consider this more and you'll have to 
>>> tell what you can really do. But the shorter, wider, more 
>>> distributed the ground connections are, the lower the risk and 
>>> potentially the better the performance. We must not forget this :)
>>>     Sincerely,
>>>
>>>         Gerard
-------------- next part --------------
A non-text attachment was scrubbed...
Name: barbosa.vcf
Type: text/x-vcard
Size: 188 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/halld-tracking-hw/attachments/20090913/0d60c58f/attachment.vcf