Yes indeed, HMOS just had lower Rds in their mosfets, so they were faster but used more power. Their output levels were high enough (~3.65 volts) for most CMOS devices operating at 5 volts, so long as they weren't loaded. I've tried HCT logic in place of the AC logic, but it didn't help.

It doesn't look alive when it's getting really hot, but I went back to the spec's, and find that everything's hooked up per the spec's. The original HMOS parts wanted an inverter between X1 and X2. The X1 (input) is getting the oscillator output, but X2 doesn't produce much of interest. ALE is not active beyond the first transition. (It triggers the single-sweep on reset). Oddly enough, it doesn't trigger the nCS or nOE on the EPROM, even once. The same board hits those signals just fine with the CMOS part.

This thing consumes about 45 mA when running at full tilt with the DS89C420. The heat generated in the HMOS MCU clearly suggests about 400-500 mA (~ a couple of watts). (it rises from 20 degrees C (room temperature) to about 85 degrees C in about 1 minute. (Temperatures are guessed, not measured, though I may take my temperature probe to the thing just to see what it says.)

My PSU isn't metered in this case, and it's not adjustable, either. I suppose I could put a precision resistor in series and measure the voltage across it on the unregulated side, if I had the urge, but, so far, it hasn't seemed necessary. The fact that it gets blazingly hot pretty quickly indicates "something" is really wrong. I just don't see it. I'm apparently going to learn something here ... I hope ...

RE