27 Sep
2003
27 Sep
'03
4:03 p.m.
After several hours with the 11/84 last night, it appears that I'm in
the same boat as Gunther with respect to the Unibus terminator.
First off, in *any* configuration with an M9302 Terminator in A-B of
the last slot, the CPU hangs at boot at LED 77. I have plenty of G7273
NPG-grant continuity cards, as well as non-NPG knuckle-busters. Which
they are.
I'm pretty sure my configuration of the card cage isn't right.
Here's the original layout (forgive any misalignment in the drawing -
Mozilla apparently translates tabs to whitespace, and not correctly):
A B C D E F
|| M7677 - 11/84 Monitor ||
|| M8190-AE - CPU ||
|| M8637-BC - 1MB ||
|| M8637-DF - 1MB ||
|| M8191 - Unibus PMI ||
|| G7273 - NBPG ||
|| M7860 General Dev Interface? ||
|| M7860 General Dev Int. ||
|| M7819 - 8-Line Async ||
|| M7819 - 8-Line Async ||
|| TCU-150 - Custom serial splitter, NPG unkn. ||
|| M8526 - RX211 ||
|| UniB OUT || M7258 - LP11 Printer Int. ||
The TCU-150 is entirely undocumented except for a couple of invoices
for etching and assembly. It's supposed to be part of a system of
splitting an available serial line into multiple comm channels, but
that's all I know. For whatever it's worth, the original configuration
map on the lid of the system unit is accurate, and appears not to have
been altered, so I'd guess that the backplane is in its original state, too.
What I want is some variation of this:
A B C D E F
|| M7677 - 11/84 Monitor ||
|| M8190-AE - CPU ||
|| M8637-BC - 1MB ||
|| M8637-DF - 1MB ||
|| M8191 - Unibus PMI ||
|| M7486 - UDA50-A ||
|| M7485 - UDA52 ||
|| M7521 - DELUA ethernet ||
|| M8526 - RX211 ||
...although the DELUA isn't strictly needed. Even without the DELUA,
with the last 5 C slots populated with non-NPG grants or with G7273 NPG
grants, I don't get a reasonable response from the RX02. And as I said,
with the terminator installed, the box just hangs at powerup.
Can anybody tell me what I'm doing wrong?
Doc
27 Sep
27 Sep
7:37 p.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 2017 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20030928/793b017f/at…
28 Sep
28 Sep
12:32 a.m.
Tony Duell wrote:
>Doc wrote:
> What I want is some variation of this:
>
> A B C D E F
>
>|| M7677 - 11/84 Monitor ||
>|| M8190-AE - CPU ||
>|| M8637-BC - 1MB ||
>|| M8637-DF - 1MB ||
>|| M8191 - Unibus PMI ||
>|| M7486 - UDA50-A ||
>|| M7485 - UDA52 ||
>|| M7521 - DELUA ethernet ||
> || M8526 - RX211 ||
>
>
> ...although the DELUA isn't strictly needed. Even without the DELUA,
>with the last 5 C slots populated with non-NPG grants or with G7273 NPG
>grants, I don't get a reasonable response from the RX02. And as I said,
>with the terminator installed, the box just hangs at powerup.
Part of the problem was that since the NPG jumper goes on pins CA1 to
CB1, I thought the G727A goes in slot C. After discovering the mistake,
further thought and a look at a G7273 tells me that that logic was
screwed anyway. :(
I thnik the problem is that you've moved the RX211
into a slot where the
NPG jumper isn't cut. Put it back in the next-to-last slot, and put a
grant continuity card (either type should work) in the last slot. Have
all other slots dilled with G7273s for the moment (leave out the DELUA
and the UDA50). See if you can get the RX02 to behave like that.
Well, with only the RX211 installed in its original slot, the M9302
terminator properly installed, and G7273's everywhere else, the CPU
still hangs. With G727A cards instead of G7273's, the RX02 will boot an
RT-11 installation floppy.
Once you've got that working, you'll need to
check the backplane to
remove the NPG jumpers on the slots where you put the UDA50 and the
DELUA. Both are NPR/NPG devices. IIRC, the UDA50 is a 2-board set. You
have to remove the jumper from the 'digital' card's slot, it doesn't
matter about the other one, so you might as well take out both jumpers to
avoid confusion.
The above, and the fact that installing the UDA50 pair and trying to
boot DU0 hangs the box, tells me I'll have to pull the backplane and do
some snipping.
Thanks *very* much for the help, everyone. Just getting a normal
boot off the RX02 validates the system and the backplane, and warms the
cockles of my heart!
In other news, the PDP11V03-L that came with the 11/03 boots and runs
perfectly. It's not nearly as slow as I had expected, either.
The friend who is helping with this and providing space for it all
was quite impressed with it. The 11/03 is several years older than he
is....
Doc
12:01 p.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1035 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20030928/35643549/at…
10:07 p.m.
Tony Duell wrote:
Yes. Specifically, with G272A grant cards in slot D of rows 5-10 and
row 12,
an RX211 in C-F of row 11, and an M9302 terminator in AB of row 12, the
11/84 will boot from an RT-11 floppy in an RX02 drive. Replace the G727A
cards with G7273 dual-height cards, or add other boards, and the boot
aborts
with a Controller error.
Well, with
only the RX211 installed in its original slot, the M9302
terminator properly installed, and G7273's everywhere else, the CPU
still hangs. With G727A cards instead of G7273's, the RX02 will boot an
RT-11 installation floppy.
You are telling me it will work with the single-height grant cards in
slot D, but not with the dual-height ones in C and D? This is impossible -- the only difference between the
2 types of grant
card is that the dual-height ones also complete the NPG jumper. If it
works with the single-height ones, then the NPG chain must be complete
along the backplane. Adding a second 'jumper' (on the dual height card)
in parallel with the wire-wrapped on on the backplane can't make any
difference.
That's actually good to hear. It doesn't make sense to me either.
I've pulled
the backplane (H9277-A), and there are no broken traces or any other obvious
damage, but there are no jumpers between the first pins of slot C in
*any* row.
There is a lot of wire-wrap back there, but I'll be a couple of days
just mapping
it, much less figuring out the purpose.
It sounds like you have backplane problems. Maybe some
of the dual-height
cards aren't making proper contact. Maybe the slots you're putting them
in aren't satnadard Unibus slots (unlikely, but...) and the 'extra'
jumper is shorting a couple of signals that it shouldn't...
This machine was in production till about 2 months ago, and was
reportedly in
good health when decomissioned. However, it's not acting in a manner
that I'd
expect from anything I've read.
At this point, I'm looking at the 11/44 next to it an thinking that's
gotta be better.
:)
As bad as it sounds, the primary goal is to validate a couple dozen
RA60 disk
packs, not nw\ecessarily to bring up the 11/84. Whatever gets us there
quickest is best.
Doc
29 Sep
29 Sep
11:59 a.m.
Anyone have a spare M8191 Unibus Adapter for an 11/84 for sale?
Curt
----- Original Message -----
From: "Tony Duell" <ard(a)p850ug1.demon.co.uk>
To: <cctalk(a)classiccmp.org>
Sent: Sunday, September 28, 2003 1:01 PM
Subject: Re: PDP11/84 continued
Well, with
only the RX211 installed in its original slot, the M9302
terminator properly installed, and G7273's everywhere else, the CPU
still hangs. With G727A cards instead of G7273's, the RX02 will boot an
RT-11 installation floppy.
You are telling me it will work with the single-height grant cards in
slot D, but not with the dual-height ones in C and D?
This is impossible -- the only difference between the 2 types of grant
card is that the dual-height ones also complete the NPG jumper. If it
works with the single-height ones, then the NPG chain must be complete
along the backplane. Adding a second 'jumper' (on the dual height card)
in parallel with the wire-wrapped on on the backplane can't make any
difference.
It sounds like you have backplane problems. Maybe some of the dual-height
cards aren't making proper contact. Maybe the slots you're putting them
in aren't satnadard Unibus slots (unlikely, but...) and the 'extra'
jumper is shorting a couple of signals that it shouldn't...
-tony
7761
days inactive
7763
days old
5 comments
3 participants
participants (3)
-
ard＠p850ug1.demon.co.uk -
curt＠atarimuseum.com -
doc＠mdrconsult.com