Back in the early '80's, Intel and its employees cared a good deal about whether their customers, then, and they have different customers now, were successful in implementing their products. Today, you can't even contact an Intel employee who can read and write. Neither Intel nor AMD have any interest in these components, which they haven't manufactured in over 15 years.

These AMD components are clean, unused, and the sealed antistatic bags in which they were received in '84 or so were not opened until last Wednesday. The Intel parts were in plastic boxes with antistatic foam, as they, the two of them, were samples, but the boxes had never been opened either. The Intel parts have gold pins, so they weren't corroded or otherwise degraded.

We have a very dry climate here in Colorado, and these parts have been "lost" for the 20+ years since I got them. I don't believe environmental conditions have degraded them.

As I've said before, the circuits appears to function properly with the DS89C420 or DS89C450 in the socket. It's just that my interest in the DS-parts is in using them as a drop-in for the standard part, and, on a lark, I wondered whether the Intel or AMD parts would drop in a functional DS-application. Now I'm puzzled as to why they don't work interchangeably.

I am, of course curious as to whether the considerable capacitance (I normally put <100 uF on a small board such as these.) which is well in excess of 600 uF, might have some effect on the relationship between Vcc and the oscillator startup. I'll try this with a 7805 as time allows, but the 7805T is barely capable of handling 1 ampere. If this thing is dissipating about 2.5 watts, it probably has other problems, and the 7805 may not be up to the task. It's presently being fed from a regulated external power supply.

Now I could put on a small switcher (one of those modules from NSC) and regulate the +5 from the +12. That's also a 20-amp supply. Most of my prototypes are built on a quad regulated supply. That way I can, ultimately, plug them into a Multibus or PC or VME backplane, or any of a number of others that use a globally regulated supply.

One other thing you might find of interest, Kai, is that, while the running DS89C420 produces about 50 mV of Vcc-gnd noise, the noise level on the not-running AMD and Intel parts is about 2 volts, which, of course, is not tolerable.

RE