> I'm looking for any information on the KA630-CNF -- a panel used to
> plug into the same KA630 connector as the normal terminal breakout
> panel does, but it is used to configure the KA630 in such a way that
> it is NOT bus master and has a separate 'doorbell' register.

Every KA630/620 has a doorbell register; configuring it as a non-master
cpu merely changes the address of that register.  (It has other effects
too, but not directly on the doorbell register.)  After all, the slave
CPUs have to be able to interrupt the master, too... :-)

> I'd like to do a little playing with it as a coprocessor...

You've found a source of KA630s that don't have Microvaxen around them?
Congratulations.  Back when we were looking, that bordered on
impossible; we ended up using a KA620 instead.  (Since *all* the slave
cpu software was our own, and the KA620 is different from a KA630 only
just barely enough to keep DEC marketing happy - IIRC the only
difference is that the P0 and P1 page table addresses are physical
addresses on the 620 vs kernel virtual addresses on the 630 - this was
not a problem for us.)

What do you need to know about the -CNF?  I've worked with them (on
620s, but in this respect there I'm fairly sure there is no difference
between a 620 and a 630).

If you're doing anything that needs high reliability out of the
doorbell interrupt, there's also an ECO you likely want: on *some*
620/630 boards, somewhere on the order of 0.1% of the doorbell
interrupts will simply be lost.  This turns out to be something that
never should have worked, and it came as close as it did only because
DEC overdesigned the thing by a factor of about three. :-)  The
capacitance of a cross-board etch run is high enough that some signal
will occasionally not rise far enough fast enough.  The fix is to just
reduce a resistor.  Last I heard, DEC would do it for you for free, but
you have to know about it and specifically ask for it.  If you want, I
can probably track down the relevant numbers to throw at your local
field circus people....

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B