OK, after seeing this link about repairing a bad main relay ( http://www.markl.f9.co.uk/howto/electrical/main-relay/main-relay.htm ) that was posted in another thread, I became curious as to why our main relays fail. I decided to take my bad main relay, which I had lying around, and have it microsectioned in the Failure Analysis / Reliability Lab where I work. (For those of you who don't know, which is probably all, I work as a Quality Engineer in a plant that manufactures printed circuit boards.) The microsection I had performed was a vertical microsection, where we cut the PCB (printed circuit board) and the component lead at the point of failure vertically (90 degrees) relative to the PCB. What we saw was what is called a circumferential crack in the solder joint around the component lead. The crack occurs at the edge of the through hole on the PCB where the component lead goes through. The failure is an intermittent one, because there is still electrical continuity at ambient temperatures. As soon as the temperature elevates in the car, the component leads, printed circuit board, and solder joints all heat up and expand at different rates, and the crack opens up and electrical continuity is lost.
Why did the crack develop in the first place? Because of something called a CTE (coefficient of thermal expansion) mismatch between the component leads, solder, and the PCB. With a CTE mismatch, the components involved all expand and contract at different rates with the temperature. The mismatch over time will crack solder joints, and affected components. This coupled with the fact that the printed circuit board used is a cheaper single sided (copper conductor traces on one side) as opposed to more expensive double sided (copper conductor traces on both sides) board exacerbates the problem because the solder joint between the lead and the PCB is miniscule. ( With a double sided PCB, the solder joint occurs along the length of the lead all the way through the PCB instead of just one side.)
Anyway, if anyone is interested, I have a Word file with photos taken during the analysis that shows the actual crack that I can e-mail. If someone wants to host my photos, feel free to - just let me know.
By the way, the repair indicated in the above link is a valid one, if anyone cares to undertake the repair. The dark "ring" shown is the actual circumferential crack.
Why did the crack develop in the first place? Because of something called a CTE (coefficient of thermal expansion) mismatch between the component leads, solder, and the PCB. With a CTE mismatch, the components involved all expand and contract at different rates with the temperature. The mismatch over time will crack solder joints, and affected components. This coupled with the fact that the printed circuit board used is a cheaper single sided (copper conductor traces on one side) as opposed to more expensive double sided (copper conductor traces on both sides) board exacerbates the problem because the solder joint between the lead and the PCB is miniscule. ( With a double sided PCB, the solder joint occurs along the length of the lead all the way through the PCB instead of just one side.)
Anyway, if anyone is interested, I have a Word file with photos taken during the analysis that shows the actual crack that I can e-mail. If someone wants to host my photos, feel free to - just let me know.
By the way, the repair indicated in the above link is a valid one, if anyone cares to undertake the repair. The dark "ring" shown is the actual circumferential crack.
