The other day, I posted about repairing some Macintosh II motherboards.
Two of the 'dead' motherboards had bad 31.3344 MHz oscillators. I thought it would be interesting to open up the oscillators and try to determine the cause of failure.
I cut them open with a Dremel with a cut-off blade. Some damage and dust in the photos below was caused by this method.
Frankly, everything looks intact. The little cut and brown burn mark near the black blobs is a laser trimmed resistor. However, the SMD capacitor in the image below does not look securely bonded. This is conductive epoxy -- not solder. It doesn't look like they used enough of it or didn't press down on the SMD part.
Underneath the crystal is a semiconductor die with bond wires. I soldered and shaved in order to get readings of some of the components. So, ignore all that damage. Instead, take a look at the pads on the left and the right where I snapped off the arms that held the crystal. On the left, you can see that the conductive epoxy held to the pad. On the right, it came off cleanly.
Here's my failure theory. I think there was a manufacturing 'defect' in that they did not use enough epoxy or the epoxy was defective (old? not runny enough). It just didn't bond securely. With time, the epoxy degraded further and then I put it in an ultrasonic cleaner which vibrated everything. That was enough for at least one junction to become loose and no longer conduct adequately.
The use of conductive epoxy would explain why so many crystals and oscillators can be damaged in ultrasonic cleaners.
Here is a closer shot of the die.
- David
Two of the 'dead' motherboards had bad 31.3344 MHz oscillators. I thought it would be interesting to open up the oscillators and try to determine the cause of failure.
I cut them open with a Dremel with a cut-off blade. Some damage and dust in the photos below was caused by this method.
Frankly, everything looks intact. The little cut and brown burn mark near the black blobs is a laser trimmed resistor. However, the SMD capacitor in the image below does not look securely bonded. This is conductive epoxy -- not solder. It doesn't look like they used enough of it or didn't press down on the SMD part.
Underneath the crystal is a semiconductor die with bond wires. I soldered and shaved in order to get readings of some of the components. So, ignore all that damage. Instead, take a look at the pads on the left and the right where I snapped off the arms that held the crystal. On the left, you can see that the conductive epoxy held to the pad. On the right, it came off cleanly.
Here's my failure theory. I think there was a manufacturing 'defect' in that they did not use enough epoxy or the epoxy was defective (old? not runny enough). It just didn't bond securely. With time, the epoxy degraded further and then I put it in an ultrasonic cleaner which vibrated everything. That was enough for at least one junction to become loose and no longer conduct adequately.
The use of conductive epoxy would explain why so many crystals and oscillators can be damaged in ultrasonic cleaners.
Here is a closer shot of the die.
- David
Last edited:
