Coming in late here. I just read the whole thread and have a few comments...
First off, I agree with Ian that voltmeter readings across the anode resistors can not be taken literally. Meters do not handle the "odd" wave forms here very well. The only way to know for sure how many volts are dropped is to use the scope and read the peak voltage off the screen.
Your scope trace looks normal. The peak is right about 1.25 volts which matches the forward voltage spec of the EL817 opto. The frequency / duty cycle looks reasonable.
My question is this: has it been absolutely positively verified that the 1K resistors on the inputs of the optos are really 1K?? Remember that optos are not switches, i.e. at low input currents they behave like a transistor that's in its linear region - partially "on". If the input resistor is much more than 1K we'll be operating in that region and the opto's output current will be limited.
EDIT: Saw the new screen capture, does not change what I posted. Okay, 1.38V instead of the 1.25V I wrote but that's still well within the opto's specs.
Look into it later when the dust is clearing off the crater.