Second Hard Drive for UNIX-pc; HwNote10

[John B. Milton]

In article <7362 at chinet.chi.il.us> ignatz at chinet.chi.il.us (Dave Ihnat) writes:
>In article <420 at limbic.UUCP> gil at limbic.UUCP (Gil Kloepfer Jr.) writes:
>> [My input]

Thank you Dave, that about says it. There are a lot of us out here who have
ALREADY gone the big drive (P5.1) route and are still running out of space.

>>	2.  Because of (1), having two separate upgrades means extra
>>	    circuitry to decode the address bits which ultimately
>>	    define the same port.  Highly unnecessary.
Nope. All the decoding is already done. The FULLY decoded MCR2SEL* line,
previously unused would drive one clock input in your design, two in mine,
no problem. The bits for the MCR2 port can be latched anywhere, not just in
one place, as in Gil's design. This is done all over the UNIXpc.

Gil's design could be modified to NOT provide the P5.1 just by not adding
all the wires, so his idea would work for folks who already have the P5.1
installed.

Lenny,
When revlev is set to 2 (by detection of P5.1 OR tweeked):
The sign on message changes from P3..P5 to P5.1 (not by tweek)
The phone driver WILL execute a stub when 2400 baud is selected.
The gd driver will access the second hard drive.

If you want to run my board without the P5.1 (and use only <=8 head drives),
all you have to do is: boot, run tiny program in /etc/rc (revlev tweek),
mount 2nd drive. For the diag disk, you will have to apply a patch to let it
access the 2nd drive for testing, formatting or bad block entry.

>>To bring up an even more elementary point, the only reason why you'd want to
>>upgrade to P5.1 in the first place is for a 4th head select (for a big
>>disk drive) and/or the 2nd drive select.  You might as well kill two birds
>>with one stone than to keep opening the machine and fiddling with upgrades.

I am assuming that if you want to add both upgrades, you WILL kill two birds
with one stone. The first (P5.1) can be installed and tested with the machine
laying spread eagle, then my board added. Of course, the machine CAN NOT be
run for more than about 5-10 minutes open.

>>Again, the two upgrades are closely related,
>>and they're almost meant to be done together.
Not almost, they ARE two of the four for the P6 machine that got shot down.

>>                                               I already have someone
>>who wants to work on a PAL design which will likely eliminate 4 TTL ICs and
>>could fit right on the motherboard like the current P5.1/4th-head
>>upgrade

One of the major reasons I'm going for a mounted PC versus a buried perf board
is a solid place to connect drive cables. There are no less than 6 connectors
on my board, 5 of which go outside the system. I have doubts about all these
functions fitting in a 20 pin PAL. Since there are ONLY pre-powered 20 pin
spares in the back left corner, this is a problem. There is NO way if you
consider all four drive selects. I looked at doing a piggy back board first.
There is no clearence to get straight up headers in there, so you have to use
90 degree headers, so you can only use the edges of the board, which makes the
board huge if you have 2*34+4*20=148 pins=14.8 linear inches=4 x 4 board. I
shudder to think about what would happen if those cables coming in from outside
the machine were yanked out...

>>(although I would prefer to see the PAL on a separate piece of
>>perfboard with the appropriate disk driver chips, like John probably does).
But not the connectors?

>Why?  There are empty pads on the motherboard for the P5.1 upgrade that work
>marvelously; what's wrong with using them, if you already have to pull the MB
>and solder on it?
I can think of 6 reasons (layer 1, layer 2...)

Now for some things I forgot to mention in HwNote10. Yes, I will be including
all the PAL equation stuff (PALASM source, JEDEC file). It's looking like a
PAL16RP4 is the right one, and I will need the RST* line.

I remain confident in the lack of effect on tea prices.

Stay tuned, keep those opinions coming.

John
-- 
John Bly Milton IV, jbm at uncle.UUCP, n8emr!uncle!jbm at osu-cis.cis.ohio-state.edu
(614) h:294-4823, w:764-2933;  Got any good 74LS503 circuits?