Over on a web forum [1] someone asked about some of the older logic
chips that required multiple supplies and it's been bothering me that I
can't find out the answer.

Some NMOS chips like the 8080 needed an extra negative (-5) and more-
positive (+12) supply to function. But pretty soon there were chips like
the Z80, which ran from a single supply, but was still described as "NMOS".
And Wikipedia says that there was a Soviet clone of the 8080 (the
KR580VM80) that used the extra supplies but could (?) operate properly
with the -5v supply at 0v and the +12v supply at +5v.

So what were the extra rails for? Textbook illustrations of NMOS gates
always seem to show just one supply. What did the internal schematic
of the 8080's logic gates look like vs. the Z80's? Why were later NMOS
chips able to dispense with the extra rails?

My best guess is that manufacturing allowed closer control of the
transistors' Vt or other parameters, and so they could use smaller
biasing voltages and still have acceptable margin. But that's really
just a guess on my part. Surely someone in this group knows for sure!

  [1] http://forum.sparkfun.com/viewtopic.php?t=11874

--
   Wim Lewis <w...@hhhh.org>, Seattle, WA, USA. PGP keyID 27F772C1