Page 183 - Troubleshooting Analog Circuits

I 70 13. Letters to Bob

Dear Bob,

You have been writing an excellent series of articles, as you always do. This series
reminds me of the old days when I used to be able to fix things. It’s necessary to pass
this type of wisdom along to each generation to keep common sense in the field of
engineering development.
However, as with all us old-timers, you slipped on a few points, to wit: Eyelets are
out, especially in multilayer boards. They are still OK for 2-sided experimental
boards and single-sided boards where changes in components are desired. In the
latter case, an eyelet socket is applied to the finished circuit to hold everything stable.
Using an eyelet in a multilayer board will distort the plated barrel, frequently causing
separation from the inner layers of copper. This shows up only as a temperature-
induced, intermittent problem during testing: It will become a hard failure only when
it gets to the customer or is in the field. Good plating shops can create excellent
plated-through holes at a reasonable cost today, so give them a try.
The use of lock-washers on printed wiring boards is fraught with danger. We spent
a lot of money at one company in hying to find out why our screws kept backing out
of their holes, causing their boards to come loose in their mounts. The problems
happened during a thermal/vibration test. Running the same boards through the vi-
bration test without the thermal test did not loosen the screws. Thermal cycling was
the culprit. It caused the board to expand, and near the glass transition temperature
(+125 “C) it caused the board to deform to relieve the stress. When the assembly
cooled, there was no incentive for the material to flow back into the gap; thus, the
board was loose. A spring washer, be it a star or a Belleville type, will eventually
loosen to the point that the desired electrical connection is poor, if not lost altogether.
It’s better to solder if you can. If a screw is needed, make the pad very large and the
washer under the screw wide, in order to spread the load and maximize the joint’s life.
Recently I discovered a potential problem with surface-mount capacitors, when
soldered onto a PC board. The capacitance can be increased beyond the specifica-
tions of the capacitor, if any flux is left hidden under the capacitor. If all the flux is
not washed away, it can cause the capacitance to appear out of spec. Washing with
solvent and then washing in the dishwasher can solve that problem.
Keep the good practical articles coming, and we will all benefit.

Richard T. Lamoureux
Hawthorne, CA.

RAP’S reply:
Well, I believe you are correct, that you are more discerning and have more experi-
ence in these areas than I do. Thank you for the tips. You are saying that any bolt on a
PC board is likely to loosen unless in an air-conditioned, constant-temperature office
with no significant self-heating. H’mmm. Thanks for the tips.

RAP

Dear Bob,

I really enjoyed your recent series on troubleshooting. Although I’m not much of
an analog designer, I have gained experience over the years in digital-system design,