A rather strange effect puzzled us today...
We use a MOS optocoupler to switch a RS485 terminator. The production test sometimes failed and indicated the terminator was on while it should be off. When probing the drive to the optocoupler, it sometimes switched on or off, depending on when or what we were measuring. Very erratic behavior. Comparing the production with the prototypes - the proto was fine, but it had another brand of MOS switch. Checking the datasheets of the old and new part - should be OK. Measuring again: sometimes we even found a slow drift of the ON resistance (while the LED drive clearly was off).
Suddenly an idea hit us - the new coupler (Avago ASSR-1410, SMD) package was white, the older one was black. Now I've used white packages a lot, never had a problem with Sharp etc. But it just might be... OK, pulled over a halogen lamp: MOS went on. Cardboard over the coupler: MOS went off.
Aha! It was just the interception of the outside light (cloudy sky) by our hands when measuring and probing that changed the operation! And the production test was done in broad light, not in the closed cabinets.
Has anyone ever seen this behavior in MOS switches? I've use many types, never a problem. Must be something in that Avago design with not enough filler in the white epoxy.
Regards, Arie de Muynck