• Please review our updated Terms and Rules here

MicroPDP 11/73

I read a little about all this and think I got the jist now. BA23 had first three slots (Why do they number right to left ? Drives me nuts.) Q/CD rest Q/Q. I can plug in a dual card into 3Q though correct ? Just not 3CD without risking it. Any quad cards have to start in 4.

In a BA23 with the standard H9278-A backplane with 3 Q22/CD slots and 5 Q22/Q22 slots you can plug a dual wide card into the AB Q22 half of the first 3 slots or either Q22 half of the last 5 slots.

Some quad wide cards must be installed in one of the first 3 Q22/CD slots because they use the C-D interconnect as a private bus. For example:

M8190 KDJ11-B
http://www.bitsavers.org/pdf/dec/pdp11/1173/EK-KDJ1B-UG_KDJ11-B_Nov86.pdf

M8637 MSV11-J
http://www.bitsavers.org/pdf/dec/qbus/EK-MSV1J-UG_001_May85.pdf

M7606 KA630, M7608/M7609 MS630
http://www.bitsavers.org/pdf/dec/vax/630/EK-KA630-UG-001_Feb86.pdf
 
In a BA23 with the standard H9278-A backplane with 3 Q22/CD slots and 5 Q22/Q22 slots you can plug a dual wide card into the AB Q22 half of the first 3 slots or either Q22 half of the last 5 slots.

Some quad wide cards must be installed in one of the first 3 Q22/CD slots because they use the C-D interconnect as a private bus. For example:

I should have been clearer and said quad cards besides the cpu etc that use those slots when I said quad cards. I meant the add-on quad cards. I do have it right though that I could put a M7555 RQDX3 in a Q dual slot (ie the top) and be fine in the first 3 slots of a BA23 ? I only have one bus grant card and I sort of need it in the ba123.
 
First let me say that gslick is giving you good help here on the ABCD vs ABAB slot type issue, so I'll let you work with him on that. I don't get the sense that you've got it quite perfect yet, but he'll get you there. I was only concerned that you understand the potential for severe damage if cards are installed incorrectly.

The method DEC used to describe it's backplanes evolved over time. I looked around the internet to see if the Microcomputer Products Handbook 1985 was available anywhere for download, but didn't find it. Here's an excerpted page which makes it clear to me, see if it helps you. I may yet sacrifice one of my hard copies of this little book in the interest of a scanned PDF version. Hope it helps the discourse.

BA23%20H9278A.jpg


"A ribbon cable installed in J2 of the backplane assembly carries the CPU signals to the 20-pin connector of the control panel." is what the KDF11-B manual has. Its EK-245AA-MG-001 I downloaded from somewhere or other. It seemed odd to me too.

I downloaded it, and indeed that is what it says, so I can readily understand your confusion. If you look at the text of the previous page, you'll see a more correct rendition.

...I read a little about all this and think I got the jist now. BA23 had first three slots (Why do they number right to left ? Drives me nuts.) Q/CD rest Q/Q...

Perhaps it would be helpful for you to look over Volume 2 of that manual, which pertains more to the BA23 Box. You'll see images which depict the system "sideways" as if it were rack mounted. Perhaps everything will make more sense to you then.

The BA-23 Pedestal arrangement, while an interesting use of the system, is not really how it was designed to be.

That manual also has all 3 of the common PDP-11 CPU Cabinet kits depicted on the same page (2-13 PDF page 47). You'll see that the purpose and controls of each are different, making it necessary that each CPU type have it's own Cabinet kit to adapt it to the Box of your choice. At least, that was the thinking and it's why I say you should always store the CPU and it's Cabinet Kit (and cable) together when not installed.
 
Perhaps it would be helpful for you to look over Volume 2 of that manual, which pertains more to the BA23 Box. You'll see images which depict the system "sideways" as if it were rack mounted. Perhaps everything will make more sense to you then.

The BA-23 Pedestal arrangement, while an interesting use of the system, is not really how it was designed to be.

Oh man. Figure 1-2 is enough to make my brain explode. Looks like the backplane is part of the power supply and I'm not sure what's up with the back.
 
Last edited:
Oh man. Figure 1-2 is enough to make my brain explode. Looks like the backplane is part of the power supply and I'm not sure what's up with the back.

LOL - I seem to be having the opposite of the desired effect?


  • Figure 1-2 shows the BA-23 removed from any pedestal, and in a "horizontal" rackmount position. However, it is not a "rear" view, but a "Front" view - with the drives removed from the drive bay. They are taking you on a tour, to set you up to make sense of the next few Figures which are details of the variations of the control panel and the Signal Distribution Panel (Fig 1-5) which is located deep in the drive bay (as seen from the front view Fig 1-2)


  • Figure 1-6 Shows the cabling as you will need to do it once you're to the point of installing Drives in that front bay.


  • Figure 1-7 is a good 3D "Mechanical" view of the backplane assembly. However, to explain the Grant Continuity Chain, I usually prefer Fig 56 from my previous post instead (especially Fig 56-2), as it conveys not only the "Chain sequence" but the "Slot Type" as well. It "speaks to me" better, but use which ever works for you.


Personally, I don't flex too well any more, and bending over - wrestling cards and cables in the back of a Floor Pedestal isn't my idea of fun. Instead, I make it so my BA-23s slide out of the pedestal altogether, so I can set them on the bench (horizontally) and work on them. It's called "Full Access".

Now all I need to do is "pump-up" so it's easier to handle all that weight!
 
Last edited:
Personally, I don't flex too well any more, and bending over - wrestling cards and cables in the back of a Floor Pedestal isn't my idea of fun. Instead, I make it so my BA-23s slide out of the pedestal altogether, so I can set them on the bench (horizontally) and work on them. It's called "Full Access".

Now all I need to do is "pump-up" so it's easier to handle all that weight!

They certainly aren't easy to work on in the pedestal. Especially with big clumsy hands like I have. The mV2 board set is a real pain to get in and out with its additional ribbon cable covering the ejectors. I'd love to take the BA23 out and leave it out but, desk space is much rarer than floor space these days.

It actually looks like I finally managed to catch a M8190 KDJ11-B cpu on ebay for a price I could afford. Auction wasn't clear about revision or much of anything so, I guess I'll see when it gets here. Hopefully it works.
 
... Whats the difference between RSX-11m and RSX-11m Plus ? I have a bunch of different disks images of things and I'm not really sure which to use. As I understand it you need at least 4.6 for y2k . Booting simh seems to confirm this. ...

Previous post on the differences between M and M+ still withstanding, I found this explanation of Y2K specifics on the Mar, 2006 archive of Mentec-inc.com.

RSX-11M-PLUS Highlights

The primary change between V4.5 and V4.6 of RSX-11M+ and V4.7 and V4.8 of RSX-11M relates to the KDJ11-E and to the provision of support for the TOY clock hardware on the KDJ11-E beyond 2000. Upgraded TOY firmware was made available from DIGITAL in early 1999, the O/Ss were modified to support that firmware.

In addition to the above changes have been made to utilities, which allow the software to extend beyond some of the next failure points (i.e. a number of utilities have limitations which will cause them to fail after 2027, some are limited to 2099). Most date limitations in V4.5 of RSX-11M+ are for 2027 - changes in V4.6 will take most utilities to approx. 2155.

A new option will allow the system to display the year in four digit format rather than the traditional 2-digit format that has been displayed in the past.

A number of cosmetic display issues have been corrected which caused some utilities to display 2000 as 100. These are primarily in the layered products and have been corrected in the update for V4.5 for the layered products.
 
...It actually looks like I finally managed to catch a M8190 KDJ11-B cpu on ebay for a price I could afford...

Excellent. (Fingers are crossed)

I came to the conclusion some time ago, that although I like having the pedestal as an option, it's really only practical for "stable" configurations. Stuff I'm experimenting with, like systems with external drives, or odd module configurations [boards in the backplane] are reserved for a rackmounted system, preferably a "tall" rack that puts the BA-23 at a convenient height. This isn't too hard, given I have TSV, RL, RK, and other IO devices in the stable that are rack-only kinda stuff.

I know what you mean about the "desk space" premium. If you can deal with it as a temporary measure, "while working on it", and return it to the pedestal "for use", your system will stabilize quickly and you'll keep it safe from back-stress induced mistakes. You'll work it out - it's important to find a way to be comfortable so you can enjoy the experience.
 
I know what you mean about the "desk space" premium. If you can deal with it as a temporary measure, "while working on it", and return it to the pedestal "for use", your system will stabilize quickly and you'll keep it safe from back-stress induced mistakes. You'll work it out - it's important to find a way to be comfortable so you can enjoy the experience.

If the BA123 was a bit higher it would be perfect for working on these. Still have to get on the floor as it is. They ever put that backplane in a rackmount setup ?
 
The people of 2155 are going to be screwed when their rsx11 boxes crash.

I'll let you in on a secret - I keep my systems set to dates in the 80's and 90's. No date issues no matter what software I'm running.

After all, what does it matter what the year is?

Just pick a year where the days of the week and the numeric date align with the current one and you're all set.

This year, 2012 matches 1984 perfectly. It began on Jan 1 = Sunday, and It's a leap year. You can even re-use a 1984 Calendar if you wish.


Quick story:

In the days of RSX11M 3.0...

Our R&D department used to occasionally deal with a contractor. This contractor's development system would not handle "comments" in it's assembly files. So, whenever we got modifications back from this guy, all our comments were stripped from it, and his changes had none either. Over time, one product he habitually cared for grew to 10k+ lines of assembly code, and not a single comment in it!

We couldn't do anything about it, as management (the Director of R&D) was a friend of this dude. Eventually the time came when this contractor was busy for a set of changes that had to be made, and we in engineering had to take up the task of making and debugging the latest modifications.

Well, during this horribly hot summer, I helped one of our co-op students completely revise and comment the entire code set during "shutdown month". Poor Tom, he slaved like a dog to get it done! As my department was the only one in the building with the A/C on [ours was an independent system], he moved into my area and I got to hear all the cursing and foul language as he figured out what 10k lines of code was doing.

At the end, we had a very nice package that was maintainable, and better still - all the listing files were at their final addresses [no more relocation] so debugging in the hardware was a snap. Listings had a symbol cross-reference and everything. We could make real, predictable progress with changes from there out. FINALLY!

Well, a couple months later, management wanted more changes... and their favorite contractor was available again.

Tom was in a panic... he came to me - angry as if someone had doused him with a bucket of fish guts. "What are we gonna do? I'll quit before I give this a_____e my commented code!". [yes - by this time we'd hired Tom, partly because of the good job he'd done with what had become known as an impossible product] A little brainstorming presented the answer. We modified a program I'd written to extract comments out of sources for documentation purposes, and made it also output the code (sans the comments, which it could already recognize) to a like-named file in another directory.

It took a few minutes to run, and we had a set of sources back in the same condition we'd last received them.

Why did I think of this now? As a little hint to the guy, I set the system date for this creation of files to "Sunday, December 7, 1941", and that's how they were all time stamped. RSX never checked and didn't care.

So unless they changed it, RSX will probably accept any date you care to set. Hi Tom... you reading this??
 
If the BA123 was a bit higher it would be perfect for working on these. Still have to get on the floor as it is. They ever put that backplane in a rackmount setup ?

Rackmount BA-123?

Can't answer. I've only seen the "Coffee table" model myself. [always wanted one actually] A quick dance through the BA-123 manual reveals no rack options. I'll keep an eye out next time I'm in the S&Os.

Update: This claims no BA-123 Rackmount
 
Last edited:
I should have been clearer and said quad cards besides the cpu etc that use those slots when I said quad cards. I meant the add-on quad cards. I do have it right though that I could put a M7555 RQDX3 in a Q dual slot (ie the top) and be fine in the first 3 slots of a BA23 ? I only have one bus grant card and I sort of need it in the ba123.

If the cards you have on hand today are the M8189 KDF11-B, the M8067 MSV11-P, and the M7555 RQDX3, then you should be just fine placing them in this order in your BA23:
[Slot 1, A-B-C-D] M8189 KDF11-B
[Slot 2, A-B-C-D] M8067 MSV11-P
[Slot 3, A-B] M7555 RQDX3
 
Hi!

I see this thread is rather old but my question might fit into this thread anyway.

I got a M8190-AB KDJ11-B CPU card and I might get my hands on an M8637-DF. I do have several M8067 cards and I started to wonder how to place them in a spare BA23 box which was housing a microVAX-II before.

What I understand from this thread is, that the CPU card M8190-AB is a kind of hybrid and must be installed into an Q22/CD (AB/CD) slot. Cards above the CPU card will see PMI on column CD while cards below will see QBus. Am I correct that the below QBus on columns CD is just there to support Quad Qbus cards? Meaning, I cannot put an Dual size card into CD. Sorry for manifesting, I can see on the CPU card that it seems to have too less pins to drive a full AB slot on CD. It is just the grant lines?

If I'm right I could build somthing like this with maybe the M8067 memory in slot 3; CPU in slot 2 and later a PMI type memory like M8137 into slot 1?

BA23.jpg

Btw. This is the CPU card I got:
M8190AB.jpeg.jpg

Can any quad size card be placed in Q22/CD and Q22/Q22 [AB/CD vs. AB/AB]

Cheers,
Peter
 
To investigate further I stepped over this PDP-11/83 configuration poasted by RICM http://www.ricomputermuseum.org/Home/equipment/dec-pdp-11-83

Is this config accurate? I would have expected the CPU being in slot 2 and memory in slot 1.

//Peter

We usually get donated equipment because it is broken and has been fiddled with. I documented the system as we received it, so it may be incorrect. If it is, I can swap the location of the boards.
 
Is this config accurate? I would have expected the CPU beeing in slot 2 and memory in slot 1.

I've run my 11/73 with the M8192 CPU in any of the slots 1-3 on my BA-23 backplane, with the QBUS memory boards in the other slots. For an 11/83, the PMI memory uses the CD interconnect for its signals, there is nothing magical about slot 3 other than AB is QBUS and CD is for the PMI.
 
We usually get donated equipment because it is broken and has been fiddled with. I documented the system as we received it, so it may be incorrect. If it is, I can swap the location of the boards.

Thank you for the clarification!

//Peter
 
Can any quad size card be placed in Q22/CD and Q22/Q22 [AB/CD vs. AB/AB]?

//Peter

I love responding to new posts to threads I followed nearly seven years ago.

The answer to your question is no. A specific example (and maybe a special case actually) is the quad-height RLV11. RLV11 is a set of two quad height cards (M8013, M8014) that pass signals between each other on the CD interconnect. If you put RLV11 cards in AB/AB slots, it does serious damage. I have a posting on alt.sys.pdp11 from probably ten years ago that explains the damage and the repair.

Most (but not any) quad height cards can however be placed in AB/AB or AB/CD backplanes. Usually there are jumpers on the boards to select passage the interrupt signals on either half of the card.

Lou

PS. It has been a few years since last I heard from the fellow who went by "RSX11M+". Has anyone heard from him within the past two years?
 
Back
Top