[pups] roll-your-own-unibus boards?

[Michael Davidson]

----- Original Message -----
From: Brad Parker <brad at heeltoe.com>
>
> Does anyone have any thoughts on how hard it would be to make a unibus
> board which is an IDE controller?

It should be very simple.

Several people have built IDE controllers for the QBUS.

ftp://digital.dp.ua/DEC/ata/README.txt

http://www.chd.dyndns.org/qbus_ide/

A unibus controller will be very similar except for the obvious
differences in the bus interface.

> I've never looked at unibus controlleqr schematic, but plan to.  I'm
> assuming much of the old ttl can be sucked into something like a Xilinx
> coolrunner CPLD...

Yes, but assuming that you are only implementing programmed i/o
there will be so little logic involved that you might as well just use
discrete TTL.

In fact, DEC used to make single height modules which did address
selection and interrupt control (M105 and M7821). If you could find
a pair of those, your work would be almost completely done for you.

> I also assume it's reasonably straightforward TTL, and at (by today's
> standards) slow speed... true?

Yes - very straightforward.

>
> Any hints, or gotcha's as far as fabrication or interface?  Has anyone
> done this (in the modern day, that is :-)

Yes, make sure that you use appropriate bus transceivers, In particular,
make sure that your bus drivers conform to the unibus spec, can sink
sufficient current, and do not have rise times that faster than the spec
allows.