Technically your 1st reply was clear and comprehensive. And nowhere did I say that your reply was wrong.
I made an assumption based on my own experience, intelligent newbies(not buffoons) usually make system overview errors than simple electronic theory mistakes. That is within their box the design is theoretically good. But not having thought outside the box, the system level errors appear, ground looping through a grounded sensor, RFI from unshielded cable etc.
And he did mention that his PCB was ready, so likely he was at least right now in need of a paramedic with a bandage, to get on with his field trials. Of course decaps also are a part of the bandaid and final solutions.
Optos appear to be trivial components, but their internal construction is important, especially the I/O capacitance and breakdown voltage. BurrBrown (TI) have nice app notes on how they test them.