<Tristate didn't exist in the early days of TTL; the only way to do
<busses was OC or with muxes. The way this affected the design of PDP-11
<CPUs is described at length in _Computer Engineering_.
Actually it did at a premium cost. The other thing is it was a lousy bus
driver. You had parts like 72125/6 and 8t97.
The PDP-8E also reflected the open collector heritage and that fact also
allowed some things like the ability to jam data to memory and "microcode"
instructions. That and it was cheap. The offset was the absolute need
for the bus loads module (terminated pullups).
There was another benefit, a board that was not supposed to be selected
and asserting a high (or not low) didnt fry the next board asserting a
low. I encounted this for the first time on s100 (a dozen 8t97s vaporized)
when a control signal failed and we had a bus jam.
Allison
Show replies by date