6
\$\begingroup\$

I made a mistake when creating a symbol/footprint for a small (2.0 x 1.6 mm) SMD crystal and didn't notice until the prototypes didn't work. Unfortunately, the boards are tiny and expensive (0.4 mm WLCSPs, 0201 passives, etc). I need to use these existing boards for development and testing.

To fix it, I can rotate the crystal 90-degrees and solder it to the existing pads. The problem is that the metal case could now contact some vias and perhaps a nearby guard trace. These features are covered by solder mask, but I don't want to rely on the mask for electrical isolation.

So I want to raise the package up off of the board by 4 mils or so (~100 um) and solder it in place.

As of now, I'm envisioning cutting pieces of copper foil and placing them on the pads with solder paste, then placing the component on these risers (with additional paste). Although I have a (cheap) rework station, this will be difficult: the package covers the pads, the pads are only 0.55 x 0.65 mm, and they won't line up perfectly:

overlap

Seems like there might be a better way! Anyone have any good ideas for me? I'll need to modify 4-5 boards.

\$\endgroup\$
9
  • 2
    \$\begingroup\$ The best way is probably to respin the board. Something else that comes to mind, though, is a tiny interposer board, just a PCB with pads on both sides and tracks and vias that correct the mistaken footprint. \$\endgroup\$
    – Hearth
    May 2, 2019 at 2:05
  • 7
    \$\begingroup\$ kapton tape may do the insulation \$\endgroup\$
    – analogsystemsrf
    May 2, 2019 at 2:17
  • 1
    \$\begingroup\$ Was also thinking of kapton tape and solder paste. Hopefully you dont blow off the 0201 with the heat hun. \$\endgroup\$
    – Huisman
    May 2, 2019 at 4:18
  • 2
    \$\begingroup\$ Tangential: you imply you have a guard trace covered in solder mask - what purpose is the guard trace supposed to achieve? Usually they need to be exposed. \$\endgroup\$
    – pericynthion
    May 2, 2019 at 5:14
  • 2
    \$\begingroup\$ Solder mask is pretty tough, just remember how much effort does it take to scratch it off when needed with an x-acto knife. Just solder the damn thing. A tiny square of kapton tape on the bottom of the crystal will help as well. \$\endgroup\$
    – Ale..chenski
    May 2, 2019 at 5:35

8 Answers 8

Reset to default
10
\$\begingroup\$

Adding some carefully-cut strips of Kapton tape to protect the vias you're concerned about is an option, but personally I'd just rely on the solder mask for insulation in this sort of prototype situation. You can probe it after the rework to check for shorts, I bet there won't be any.

Additionally, isn't the actual crystal package smaller than the footprint? In your image you've overlaid the same footprint on itself rotated, but the actual component ought to be smaller and might fit entirely within the footprint despite the rotation. Or if it doesn't, perhaps you can find a smaller crystal with the same frequency and similar capacitive loading requirements.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ If properly reflowed the crystal shouldn't touch the solder mask anyway. Adding some excess solder might help that. \$\endgroup\$
    – Wesley Lee
    May 2, 2019 at 5:51
  • \$\begingroup\$ I vote for Kapton tape also. \$\endgroup\$
    – CrossRoads
    May 2, 2019 at 12:56
  • 1
    \$\begingroup\$ Thank you. I suspect you're correct; I should just solder it and check afterward. As far as the package size, it entirely covers the pads. I'm using the "least" classification in the IPC-7351b, which makes everything really tight. In fact, the actual package pads are the inset rectangles you can see in the footprint pads. \$\endgroup\$
    – bitsmack
    May 2, 2019 at 16:24
8
\$\begingroup\$

In the past I've done something similar with crystals.

The solution I used for lack of any kapton tape, was to place a big blob of solder on the pads to form essentially the standoffs you describe. Then also tin the pads of the crystal with solder.

Then place the crystal on top of the solder blobs (which have by now cooled solidified) in the correct orientation - it will be sitting a short distance off the board now.

Re-heat the solder blob for one corner to allow the solder to flow and connect the tinned crystal pads to the board. Make sure to hold the crystal steady for the first join so that it sits level.

Now that the crystal is held in place proud of the board, you can reheat the remaining corners in turn to complete the connections.

I wouldn't do this on a production board, but it is easy enough to do for a few test boards.

\$\endgroup\$
6
\$\begingroup\$

Ok, this maybe weird but how about soldering some other SMD resistors (very high value) or capacitors(very low value) underneath? It greatly depends on which pads are active and which are passive.

\$\endgroup\$
4
\$\begingroup\$

I think you may be overthinking a bit; Solder mask /is/ insulation, it's thin and doesn't hold up well to mechanical abrasion, but it does insulate. Probe out the case and the pads on the crystal, I suspect it may actually be isolated from the workings anyway and just there to provide an RF shield.

\$\endgroup\$
2
\$\begingroup\$

Tom Carpenter has a pretty good solution, I've done that in the past too, but getting those tiny pads individually soldered with a crystal that small is going to be really difficult without reflow, even if you have some extra exposed pad around the crystal. I have two suggestions you could try -

  • Do the solder balls on the pads, but put a couple layers of kapton in the middle to hold the component off the board before reflowing the balls with a hot air gun. If the balls are big enough, they'll make contact anyway
  • If you don't have kapton, the crystal is low frequency and you have 0ohm 0201 jumpers, you could solder 4 of them vertically and very carefully hand solder the crystal on top to form a table with resistor jumpers for legs. I've done this with 10k resistors when I've forgotten to put in a pull up/ pull down resistor and soldered magnet wires to the tops of the resistors. It's easier than you'd think.
\$\endgroup\$
2
\$\begingroup\$

Make a bump stickers from kapton and locate them between crystal's pads. Place crystal on these bumps and solder.

\$\endgroup\$
1
\$\begingroup\$

Liquid tape is magical stuff, you do have to apply several coats, but is should be able to provide an insulation from what you don't want the crystal case to touch.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ It's probably not going to be an improvement on the solder mask, and likely to react more poorly to soldering heat \$\endgroup\$
    – Chris Stratton
    May 2, 2019 at 2:56
0
\$\begingroup\$

Buy a 2 pin THT crystal, bend the pins and solder them on to the 2 relevant pads

\$\endgroup\$
2
  • \$\begingroup\$ That's a good idea. I'm embarrassed that it didn't occur to me :) \$\endgroup\$
    – bitsmack
    May 3, 2019 at 18:52
  • \$\begingroup\$ I just checked, nobody makes a through-hole crystal with the specs I need... Good answer, though. \$\endgroup\$
    – bitsmack
    May 3, 2019 at 18:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.