Qbus split I&D?

[Ethan Dicks]

On Tue, Mar 17, 2015 at 5:02 PM, Noel Chiappa <jnc at mercury.lcs.mit.edu> wrote:
>     > From: Ethan Dicks
>
>     > As much as I would like one (or two), what's the actual market for such
>     > a thing?
>
> Oh, probably almost as large as all the other custom UNIBUS/PDP-11 cards
> people are building these days... :-)

I kinda have to disagree there... the RAM card John is talking about
works in *lots* of machines, most importantly, 11/20 (there are a few
out there and expansion opportunities are limited), 11/34 (there are
lots of them out there) and anything where you want to use the device
capabilities (it's a disk controller + RAM).

A PDP-11/70 RAM card works in one type of machine that was not
terribly common and there can't be more than a few dozen total in
hobbyist hands.

>     > I'd rather have a faster machine while I'm using it.
>
> Somehow, I'm thinking that if it's speed we're after, a hardware PDP-11 isn't
> where I'd turn... :-)

In my case, I already have 4MB of MOS, in 8 cards loaded with 4116
DRAMs... I don't need to run it 24/7 so, yes, even though a single
card sucks up less power than a cage of RAM cards and is cheaper to
run, I'm unlikely to ever lay hands on a PEP70 card and I'd really
like the speed bonus of fast SRAM in a single card that I _can_ get.
I'd also like a modern disk controller, but Unibus SCSI is quite rare
(and not the problem we are addressing here).

There's clearly no reason to make a slow 4MB card.  Worst case, add a
toggle switch for 250ns memory cycle time for that 1978 feel.

-ethan