• Please review our updated Terms and Rules here

Changing a Q18/CD backplane to Q22/CD

C

Crawford

Experienced Member
Joined
Oct 21, 2008
Messages
275
Location
Maryland
Folks,

So apparently my pdp 11/23 was once a pdp 11/03, and has a Q18/CD backplane (an H9273). This is all in a BA11.

Since I want to use this chassis for an 11/53, I was thinking of converting the backplane to a Q22.

BDAL18 (BC1)
BDAL19 (BD1)
BDAL20 (BE1)
BDAL21 (BF1)

Is it just a matter of connecting these lines down the bus?
 
Folks,

Has anyone done this upgrade before? I buzzed out the connections on the bus and yea verily there is no connectivity on the upper memory address lines.

Thanks, Crawford
 
I have done it before and posted about it here someplace, in connection with a VT103 backplane. I'd have posted to your question, but my experience was with an all AB-AB slotted backplane and I think yours is ABCD & ABAB. I'd have to resort to schematics to be sure how to advise you, or maybe ohm out one of my BA23 backplanes.

I was hoping someone else had already done it and would respond to you.

Update:

I didn't find the post, but here's the image that goes with it:
VT103-Q22_modification.jpg
 
Last edited:
Chris,

Thanks for the reply. Mine is indeed an abcd backplane. It looks like the vt103 is upside down from mine as the CDEF connectors on the B side are on the top row of connectors.

I am comparing my traces to a known BA23 backplane for comparison. It looks like you just added the upper 4 address lines.
 
Um... ok - there's something I need to make clear.

The ABCD markings on my drawing are not "Type" descriptions - I labeled them that way because of convention. The VT103 backplane's wiring is ABAB "type". Hence in my drawing - CD and AB can be treated the same for the purpose of modifications.

The BA23 backplane has ABCD for the first few slots and then ABAB thereafter. If you've got it out to confirm already - then check continuity for the pins I've added to the AB portion of the ABCD slots. There should be agreement. I've never checked the CD slot wiring - so you need to decide that for yourself.

Hope that helps - but since you have a BA23 there as a model - between that and my images I think you'll have what you need to decide. Is your backplane all ABCD or is it mixed like the BA23?

Please record what ever you do here for others to follow.
 
Last edited:
Q = = ab

Q = = ab

Chris,

Yes, my reference BA23 (H9278-A IIRC) is ABCD for the first 3 slots, but ABAB for the last 5. I thought that A+B = a complete qbus 'dual' slot, so ABAB was also referred to Q/Q.

But regardless, I think that expanding the bus lines is only relevant to AB (on the B slots actually) so your pictures showing your ABAB wrap the lines over to the other side. My 11/03 bus is ABCD down all the slots. Therefore I have just one side to deal with.

I'm not sure why DEC chose to do what they did with ABAB and ABCD. There are quad widht cards that can be damaged by plugging them into the wrong slots.

I also have a 'mystery chassis' that is ABAB for the entire chassis. I have to be careful what I plug into that bus (or start cutting traces.).

I will document what I do in pictures, etc.

? ; - )
 
Crawford said:
I'm not sure why DEC chose to do what they did with ABAB and ABCD. There are quad width cards that can be damaged by plugging them into the wrong slots.
Click to expand...

I am guessing that in that era there were two considerations: form-factor and level-of-integration.

The CD slots support dual-board-one-module configurations where you want to avoid an over-the-top inter-board connection; instead you have a private bus that can be used as-needed between adjacent boards. Like the PMI in the 11/83.

The ABAB configuration lets you hold more dual-height modules in the same vertical U-space in a 19" rack. Remember that the standard 11/03 form-factor was only 4 rows (1 SU) ... not the later 9 rows (2 SU).

DEC was trying to hit multiple customer sweet-spots with a cost-effective set of components that they could produce in volume to keep costs down (and profits up!).

It does seem that perhaps a bit more care could have been taken to avoid damaged boards when designing the CD slots and the boards, however we're talking 20:20 hindsight here and the DEC engineers at the time were probably responding to ongoing multi-board experience on the Unibus and existing customer requirements (and couldn't predict very far into the future ... like the width of the address bus). Putting new-boards into old-slots is always a bit risky, particularly if the pairing doesn't show up as a standard-configuration or well-known-mod. I'm sure that the Field Engineers knew what they were doing :->.

We're (or at least I'm) not nearly as conversant with the "fine points" ... so it's always RTFM and then caveat emptor ...
 
Let's see ah...
.

  • Yes - that was the thread. Apparently I didn't trip upon the correct search terms.
  • Yes - there are QBUS boards which will damage or be damaged by being placed in an inappropriate card slot.
  • Yes - this is even more true (and common) for board/slot combinations in so-called UNIBUS machines. A dance through Marty's 11/40 and 11/45 threads demonstrate what happens when one does not respect this.


Historical Perspective:
.
Coming from a Vacuum Tube training background - I already knew that completely incompatible tubes can use similar sockets. DEC had been using those backplane connectors as long as I can remember. In the earliest machines I came in contact with, uniform wiring between apparently similar slots was the EXCEPTION rather than the rule. In those days we wouldn't have considerd just moving things around - not that it was much of a problem because of labeling practice. (except by accident)

PDP-11 UNIBUS and then LSI-11 (later called QBUS) contained progressively more slot similarity. But both contain damaging pitfalls for novices of today who were not acquainted with such a world.
​

The prospect of extincting rare hardware through misadvice or misinterpretation of my advice, haunts me. I have nightmares that are more fun.

Destructability:
.
The advent of LSI and CMOS created a false sense of indestructability in electroinics. RCA specifically designed it's CMOS chips so that if one kept to all 5V logic and signals, it was nearly impossible to damage it's ICs. I've seen 40-pin CPUs inserted backwards heat to skin-frying temperatures, that were cooled off and reoriented correctly that worked.

Thruout my training I was only familiar with a world where electronics - when incorrectly connected - would be destroyed the instant power was applied. Sometimes without even a sound (the most difficult to find and fix) and more often with snap, crackle, pop, smoke or even flame with flying debris and toxic poison release.

Instructors gave specific lessons on what we would call "bio-rhythms" today, recounting experiences where they or colleagues damaged hundreds of thousands of dollars in equipment and lost lives through mishaps in the application of their craft.

Every new design we did had the possibility of self-destruction the instant it was powered the first time. This was the challenge, and heaven help me - fun, of being an electrical engineer.
​

Price vs Capacity:
.
The original Mac sold for over $2000 while the new Mac Pro recently offended the PC world with it's ~$3000 base price. My first minicomputer was $15,000 with no disk storage and cost an order of magnitude less than it's predecessor. It contained a few KB of memory and executed ~100K instructions per second.

Today that Mac Pro has Gigabytes of memory, Hundreds of Gigabytes of Disk storage (SSD) and at least 4 CPUs executing +3Ghz.
​



So when dealing with "legacy electronics" you're stepping back in to a different world. Please tread lightly - you're walking on our heritage.
 
Hi All;
Here, Here, RSX11M, I concur with You on every point !!!! THANK YOU !!!!
I was brought up in the Tube Era as well..
Yes - this is even more true (and common) for board/slot combinations in so-called UNIBUS machines. A dance through Marty's 11/40 and 11/45 threads demonstrate what happens when one does not respect this.
Knowingly or unknowningly.. It still can cause lots of problems, and not know how to elevate or fix it..
PDP-11 UNIBUS and then LSI-11 (later called QBUS) contained progressively more slot similarity. But both contain damaging pitfalls for novices of today who were not acquainted with such a world.
I am acquainted with Old Computers, but not DEC's particularities, so Yes, I am a novice.. My first Computer was an Altair 8800b.. Which I bought New and in Kit form..
The prospect of extincting rare hardware through misadvice or misinterpretation of my advice, haunts me. I have nightmares that are more fun.
I would be curious of Your nightmares ?? THANK YOU for Your Information and insight.. Please keep on passing on MORE !!!
THANK YOU Marty
 
RSX11M+ said:
So when dealing with "legacy electronics" you're stepping back in to a different world. Please tread lightly - you're walking on our heritage.
Click to expand...

Chris,

Good point. Part of why I'm building PDP-11s is the nostalgia for the systems I programmed in college. Bitsavers has a ton of info, but after reading what there is, I'd rather ask the questions here than fry something.

Thanks again for the info and the discussion.

-Crawford
 
You must log in or register to reply here.
Back
Top