3

We have a PCB with a detector that must be positioned at a particular height above a mounting surface. Our plastic parts are slightly undersized, so we must add some space (0.4 mm) between our PCB and the mounting surface in order to position the detector in the proper place.

The geometry is such that changing the board thickness doesn't solve the problem. We will eventually change our plastic mold to make up for the incorrect dimension, but we need to keep producing parts in the meanwhile.

We are looking for cheap and easy ways to create this space without making our assemblers fiddle around with shims, which can be dropped, misplaced, forgotten, etc.

I'm thinking that perhaps we can solder a washer-like part onto the PCB surface around the screw holes, if such a product exists. Or perhaps there is a way to recess the detector into the PCB somehow. Do such techniques or PCB surface shims exist? Can anyone recommend another method to tackle this problem?

rothloup
  • 408
  • 2
  • 12
  • 1
    Could you use some self-adhesive shim stock in a large enough size that it would have visible un-covered areas in the finished assembly? If it's a contrasting color then it might be a simple final inspection to check. You should perhaps have some sort of go/no-go to verify the resulting height anyway. – Chris Stratton Jul 26 '19 at 15:26
  • Maybe have a big custom die-cut shim made that is obvious if it's forgotten and will cover all of your mounting holes. Steel rule dies can be relatively inexpensive. – Spehro Pefhany Jul 26 '19 at 15:27
  • Would 0.4 mm thick sticky tape be too fiddly? – Andrew Morton Jul 26 '19 at 15:27
  • I understand why you asked this here, but it's really more of a mechanical design question and the people over at [Engineering.SE](https://engineering.stackexchange.com/) might have more-or-less "standard" solutions for this sort of thing. Shall I migrate it for you? – Dave Tweed Jul 26 '19 at 15:31
  • 1
    @DaveTweed The PCB milling part of the question is unique to here, so it might be better to leave it. I'm pretty sure there are some possible solutions with a milling a board layer before a multilayer stackup but don't have any direct information. – Spehro Pefhany Jul 26 '19 at 15:39
  • 1
    I just came across this hack, but it doesn't look like it's at all standard practice yet: https://hackaday.com/2019/01/18/oreo-construction-hiding-your-components-inside-the-pcb/ – rothloup Jul 26 '19 at 15:46
  • Is it the detector a surface-mount component? – pericynthion Jul 26 '19 at 15:55
  • @pericynthion: Yes, it is SMT. – rothloup Jul 26 '19 at 16:04
  • 1
    How many pieces are we talking? @Spehro's answer of a custom shim is good if it's a large volume. For small volumes, maybe you can find some self-adhesive plastic sheeting in small circles or strips that could be stuck to the bottom of the detector. Done off-line, this would become a "permanent" part that those assembling the PCBs wouldn't have to do anything extra with. – spuck Jul 26 '19 at 16:05
  • 1
    Most board shops don't like to promise better than 10% accuracy in getting the board thickness. Maybe you can get 8% by paying extra, or 5% by choosing the right shop and paying extra. So if you have a 1.6 mm board thickness, you have probably at least 0.1 mm of thickness variation from lot to lot to deal with. You might end up needing to have a selection of shims available to deal with this variation anyway, once you go into volume production. – The Photon Jul 26 '19 at 16:28
  • @spuck We could be living with this assembly for about 2k pcs. – rothloup Jul 26 '19 at 19:28
  • Could you sacrifice some SMD components that is the right height and place one (or a few) around each of the relevant mounting holes to make the seating plan higher. Perhaps just the leads of a components would be the required height. – KalleMP Jul 26 '19 at 20:32
  • If you're doing the assembly yourself you could solder on some flat brass washers. I'm not sure if an assembly house would want to do this, but it's worth asking. – Jeanne Pindar Jul 26 '19 at 20:52
  • Sounding more and more like your plan for eventually changing your mold should be updated to design more tolerance into this aspect of the device. – Scott Seidman Jul 28 '19 at 14:38
  • @ScottSeidman Indeed, our next revision will take sensitivity out of various aspects of the design. As each iteration should! – rothloup Jul 28 '19 at 14:49

3 Answers3

2

In the past, when I wanted a certain distance through hole parts I have specified a cut teflon tube. I put this as a line item on a BOM and instructions in the line item of what pin(s) of the component I wanted to place it on. I've never had a problem with an assembly house, except with a few that needed to double check the intention of the line item.

In your case maybe a kapton washer would do the trick.

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
  • 3
    This would not remotely work for the short distance specified as needing to be made up, nor would it have the *accuracy* required as indicated by the fact that such a small difference needs to be corrected. Further, even if it did, teflon cold-flows, so even if it starts out in spec it may not stay there. – Chris Stratton Jul 26 '19 at 15:47
  • Chris is correct that the size and tolerances needed would make the teflon tube idea not work for us. The Kapton washer, however, could work except that I can only find it readily available in 5 thou thick. I'll search some more for thicker kapton washers, thanks for the lead. – rothloup Jul 28 '19 at 03:06
  • I didn't see the thickness when I wrote the first edit. There are suppliers that do make thick kapton sheets in many different sizes. I had one die cut to make spacer for a heatsink to keep it from making contact with the PCB – Voltage Spike Jul 28 '19 at 04:48
1

I wonder if you can just put some solder pads with the correct dimensions to generate a a ring of 0.4mm high blobs of solder reliably. I suppose it would depend upon your tolerance. I would think this approach might require a stencil hole larger than the solder pad to have a bit of a surplus.

Scott Seidman
  • 29,274
  • 4
  • 44
  • 109
  • Scott: This is interesting but we need tolerances that are on the order of molded plastic parts. So I don't think this will work in our case. – rothloup Jul 28 '19 at 03:03
0

In our case, we ended up using a die-cut plastic shim stock that is added to the assembly while the board is being mated to the plastic parts. It wasn't ideal (as I mentioned in the OP, we didn't want to "fiddle" with shims) but it was the only economical method we could find given our geometry, timeline, and volumes.

rothloup
  • 408
  • 2
  • 12