Installing core backplane into PDP 11/40

[billdeg]

http://vintagecomputer.net/digital/pdp11-40/coreplanes_installed.jpg

Over the past two days I put in a core backplane and 16K core in my 11/40 "short cabinet" system. I also added another SPC backplane to I can have do more stuff with this machine. I removed the 64K MOS memory for now.

http://vintagecomputer.net/browse_thread.cfm?id=635

One thing I wanted to pass along..when you put in the Grant Continuity cards, be sure the gold pins are facing to the left. Also, you don't need GC cards in empty core backplane slots.

CPU and 16K core
http://vintagecomputer.net/digital/PDP11-40/16K-Core-Installed.jpg

The system's higher functions are not working correctly. I can no longer boot to the console (via 9600b terminal) - both the M7856 serial card and the M9312 do not appear to be functioning like they were before I installed the new backplanes. I suspect power is the issue, or I fried something. I am confident I have termination, grant and NPG correct.

I CAN run programs from the front panel, store programs in core, cold boot and they're still installed in memory. No perceived faults.

I am guessing I am not getting enough power to the DD11C backplane. Note that there is a NPG card in the last slot of the DD11C, this backplane was altered to require this, in addtion to slot 2 which is comes that way. I have the M7762 RT-11 controller in slot 2. Quick tests showed some low +5V spots. I have more than one H744 to track down. typical stuff.

 

[Marty]

Hi All;

Congratulations, Bill !! on getting some Core to Work..

I am a Core Person my self, But, as You know, I have trouble keeping Core Power Supplies Going..
They Tend to Blow up on me (the Power Supplies)..

THANK YOU Marty

 

[billdeg]

I read your posts about the 11/45 from a while back. You had a lot of issues with serial cards then. Me too. They shut down the system if they're not working correctly. I am going to have to repair and get at least one running. I have both M7800 and M7856's to resuscitate.

Where there common issues with these cards or were they all just messed up? I plan to test each with extender boards, probe for faults starting with the 1488 and 1489 chips.

b

 

[Marty]

Hi All;

Bill, it was a combination of both.. It was mostly Power Supplies taking things out, but it was also starting with all unknown Boards..
Put all of that into the mix and as You have read, alot of trouble would brew up all over the place..
I would like to see Your progress on the Serial Cards, and the 'how' of Your troubleshooting of them..

THANK YOU Marty

 

[billdeg]

I learned that the m7856 is intended for a dd11-c backplane. Many 11/40s came with dd11-b's. That means if you're not sure it's safer to use the m7800. My point is just be sure what you'replugging serial cards into.

I am using both typee of spc backplane. You can't I don't think use a dd11-b for a RT11 controller. I f someone can say for sure that'd be useful.

More to come....

 

[wa2flq]

I learned that the m7856 is intended for a dd11-c backplane. Many 11/40s came with dd11-b's. That means if you're not sure it's safer to use the m7800. My point is just be sure what you'replugging serial cards into.

I am using both typee of spc backplane. You can't I don't think use a dd11-b for a RT11 controller. I f someone can say for sure that'd be useful.

More to come....

The M7856 and the M7800 should work in an SPC slot. I don't see anything in http://www.chdickman.com/pdp11/Notes/DD11.shtml or the respective manuals to suggest otherwise. The only difference is the LTC on the M7856.

The M7822 Controller (RL11) while hex size, only gets power from Rows A + B. The docs say it will work in a Unibus or MUD Slot. I would NOT put it in the DD-11B with the DF11 slots in Rows A+B.

 

[AK6DN]

I learned that the m7856 is intended for a dd11-c backplane. Many 11/40s came with dd11-b's. That means if you're not sure it's safer to use the m7800. My point is just be sure what you'replugging serial cards into.

I am using both typee of spc backplane. You can't I don't think use a dd11-b for a RT11 controller. I f someone can say for sure that'd be useful.

More to come....

The M7856 DL11-W (SLU with line clock) and M7800 DL11-A/B/C/D/E (SLU, configurable) are both SPC UNIBUS quad cards, so can go in any SPC slot (sections C/D/E/F) of any UNIBUS backplane with SPC slots. This would be any DD11-A, DD11-B, DD11-CK, DD11-DK, and other processor specific backplanes like for the 11/44 that has SPC slots or the DD11-PK for the 11/04-34 which as SPC slots in rows 3 thru 9. So anywhere you can use an M7800 you can use an M7856, and vice versa.

By RT11 controller I am inferring that you meant RL11 controller instead, which is a hex SPC compatible module. This means the A/B sections only have +5/GND power connections, and sections C/D/E/F are the standard UNIBUS SPC pinout. So the RL11 controller could go in any of the above backplanes in SPC slots 2 thru N-1 (ie, not first or last SPC slots, as those have UNIBUS IN/OUT in sections A/B). So your RL11 could go in a DD11-B slots 2 or 3, and a DL11-A/B/C/D/E/W could go in any of DD11-B slots 1 thru 4. According to the DEC wirelist for the DD11-B the DF11 pinout has +5/GND in sections A/B compatible with the RL11 controller (MUD) usage with +5 on AA2/BA2, and GND on AC2/BC2/AT1/BT1.

Summary:

DD11-A 4 slots: (1) UNI IN A/B; DEV C/D/E/F
                (2) unused A/B; SPC C/D/E/F
                (3) power A; unused B; SPC C/D/E/F
                (4) UNI OUT A/B; SPC C/D/E/F

DD11-B 4 slots: (1) UNI IN A/B; SPC C/D/E/F
                (2) DF11 A/B; SPC C/D/E/F
                (3) DF11 A/B; SPC C/D/E/F
                (4) UNI OUT A/B; SPC C/D/E/F

DD11-CK 4 slots: (1) UNI IN A/B; SPC C/D/E/F
                 (2) MUD A/B; SPC C/D/E/F
                 (3) MUD A/B; SPC C/D/E/F
                 (4) UNI OUT A/B; SPC C/D/E/F

DD11-DK 9 slots: (1) UNI IN A/B; SPC C/D/E/F
                 (2) MUD A/B; SPC C/D/E/F
                  |
                 (8) MUD A/B; SPC C/D/E/F
                 (9) UNI OUT A/B; SPC C/D/E/F

DD11-PK 9 slots: (1) UNI IN A/B; CPU C/D/E/F
                 (2) MUD A/B; CPU C/D/E/F
                 (3) MUD A/B; SPC C/D/E/F
                  |
                 (8) MUD A/B; SPC C/D/E/F
                 (9) UNI OUT A/B; SPC C/D/E/F

Note: MUD = Modified UNIBUS pinout
      UNI = Standard UNIBUS pinout
      SPC = UNIBUS Small Peripheral Controller pinout
      CPU = 11/34, 11/04 CPU board specific pinout
      DEV = DD11-A device control slot (appears to be same as SPC)
      DF11 = DF11 module pinout

DD11-A/B were designed for use in pre BA11-K UNIBUS chassis; DD11-xK were designed for use in the BA11-K boxes (and subsequent compatibles).

Don

 

[billdeg]

Thanks for the link to the DD11 notes page. I was having trouble finding this info. I still need to learn how they work, but I have found that I can look up the card and it usually tells you which backplane it's intended for. I don't need to understand everything to at least trust that the docs will know.

I went through all of my serial cards and found M7856 and M7800's. Of those I separated them by those that disrupt the processor when installed vs. those that do not. I have a 7800-YA that I can use with a teletype, and load papertape programs that way. I am trying to boot RT11, but I need a serial interface first. I have to review each card jumper by jumper to see what they're set to do, and then decide which will be compatible with the RT11 M7762. I am unsure if RT11 expects a specific baud rate or if it adjusts to whatever the interface card is using. Still learning.

b

 

[wa2flq]

....

I am unsure if RT11 expects a specific baud rate or if it adjusts to whatever the interface card is using. Still learning.

b

The RT11 OS will use the baud rate set/wrapped on the card, itself provided its at the standard console address 777560-566.

The details of the rate from the DL-11x type cards are not machine readable/controllable.

 

[billdeg]

The RT11 OS will use the baud rate set/wrapped on the card, itself provided its at the standard console address 777560-566.

The details of the rate from the DL-11x type cards are not machine readable/controllable.

Wrapped on the serial card you mean? i.e. uses whatever the serial card says for port 777560, this port being the gateway to condition the characteristics of the serial signal?

 

[Qbus]

Couple thoughts for what it’s worth, may not be worth anything but will propose anyway. First: The M7856 DL-11 is a great card, works in any slots and has a RTC built in. I have several and set the configuration of them as the CON and RTC on. If you start doing anything in assembler epically using the Gill book a working processor some small amount of memory and the DL-11 is all you need. I set all my spare cards to the same settings and if for some reason want to use another port can go back and do that but once you have a working card just go ahead and copy all the jumpers to the rest and it’s an easy matter to check all the cards first. You got a couple Unibus systems that you can check out the cards on first and rule out any issues there.
Second: looking at the picture I can see that you don’t have grant cards in all the open D slots (G727) for carrying the grant priority signals and I think the system will not run unless they are all jumped. There is also a NPG on slot C CA1 to CB1 that gets jumped under the chassis for anything that does not use the NPG signal, as far as my experience goes only the RL drive controller in my Unibus system uses that. The DL and other cards can care less.
Third: don’t waste time with paper tape, if you have the RL controller and you CON port is up and running just load the RT-11 system from a disk pack. Life will be much easier then playing around with the paper tape unless you want to load something like diagnostics or something.
It’s nice to have a system like that with a huge expansion chassis but think in the end that brings out a lot more issues if you don’t have things to fill in a lot of the SPC slots, also may be wrong about this but I always thought there was something different about the MUD slots between the MOS Ram and the Core in terms of voltages? I don’t know myself having never seen a core in a PDP-11 myself, but I do have a Data General system that I have been playing around with that’s got 32K of core and its scary stuff! You can turn the system off and come back later and turn it back on and everything is right where you left it, amazing.

 

[billdeg]

The slots with the missing grant cards are for core memory. The backplane does not require these to be installed.

I was told not to put a MOS ram card into a core slot, nor should I put it in the 11DD-B.

I have the expansion chassis set up to mimic the other systems' so I can test duplicate configurations in different machines. If you're ever in N Delaware area, let me know. I live just over the border in PA.

b

 

[wa2flq]

Wrapped on the serial card you mean? i.e. uses whatever the serial card says for port 777560, this port being the gateway to condition the characteristics of the serial signal?

Correct - all the signal characteristics are controlled on the DL11-xx cards by jumper, switches, wire wraps. This includes number of bits, parity, baud rate, stop bits, active/passive interface type (current loops, RS232..) Be sure the card has the correct address (777560) and interrupt vector (060)

RT-11 will use the card in a vanilla fashion. Basic interrupts for each input and output byte received/sent. No handshaking for modem or CTS/RTS or other control signal support (often called out of band controls). In-band control (in the serial data stream) exists for flow control keys (Control Q, Control S) or interruption (Control C) are provided.

 

[AK6DN]

On all my DL11-W serial cards I manually add a 30ga jumper from CA1 to CB1 so that it will bridge the NPG across the slot. On my UNIBUS systems I first strip all the backside CA1/CB1 wire jumpers on the backplane and use full height dual grant cards so that I never have to look at the backside of the backplane again, if I can help it.

Allows me to move cards around at will without having to remember which slots are NPG enabled, or not.

Don't know the history of why DEC didn't put the BRx/BGx and NPR/NPG signals in the same backplane block, so that the small grant card could have handled both, and the whole NPR/NPG thing would never have even come up in the field. Maybe it was a DEC field service make-work conspiracy? Or just one of those head-smack why-didn't-we-do-that after the fact observations?

Don

 

[Qbus]

I have the slide rails with the pivots installed so I can flip the system over to get to the undersides without much aggravation, that’s the benefit of having the system installed in a huge heavy rack. I cannot imagine having to flip over a system or trying to drag around a full size system to get to the back side of the backplane. So I don’t mind playing around with the NPG jumpers epically when it appears that just about nothing uses that signal, besides the RL controller cannot think of anything else? Maybe RX or other drives controllers?
Getting back to the system in question: So all of the empty sockets are MUD sockets for memory? And not SPC? Perhaps someone more knowledgeable then I can clarify but on a lot of the Unibus MOS RAM cards I have they have the jumpers installed for both the grant and NPG, just jumpers with the memory not addressing those busses. I know from first hand experience that until I had both chains complete that my system would do what your describing, allow you to access from the control panel but would not go into RUN until both chains were complete. If you get it going with the RL drive maybe I can come up and bring my RL XXDP pack, that’s got tons of diagnostics and test designed to frustrate. It may be worth the trip just to see how you got that system downstairs!

 

[billdeg]

Quick update - I have come to the conclusion that either ALL of my RAM is bad or there is an issue with the CPU memory management and parity RAM (MF11-UP). I am pretty sure power is ok and basic CPU operations are fine. I can save little programs like chase the lights into any spot in core memory, power off, power on, and run them. The problem is when I exam through 15-20 addresses or attempt something that manipulates memory. I get parity errors (parity light on the parity card M7259) , even though the data is in memory. The related manuals are pretty clear, but complex. I have a lot of things to check over (CPU, RAM, Options). I don't think the issue is with the M7259 falsely reporting parity errors. I am thinking there must be a CPU jumper expecting one kind of RAM or an option card different to what is actually installed. Not all of the components I have in the system are from the same place.

When I have something to report I will. Just so you know I have not given up, but for the past few days I have not made much progress, other than a lot of card swapping and notes.

b

 

[billdeg]

figured out the problem(s). I was starting the memory address toggle-in program from the wrong start location. Once I figured that out I was golden. Moving along, I hope to repair my serial card, or I will replace with another one. Ultimately I'd like to have two serial cards.

 

[m_thompson]

The RICM is picking up a donated PDP-11/40 next week. We may have questions...

 

[Marty]

Hi All;

Bill, as You already know, I have made that Kind of Mistake many a time before..
Glad You Found it !! If nothing else, it makes You feel good, that You are making progress..

THANK YOU Marty

 

[billdeg]

after giving up on the M7856 I had been using (it receives but does not send back to the terminal), and an M7800 I hoped might work (bombs when attempting to do much of anything), I found another M7856, jumpered that one and now I am back in business. I have now 16K core, serial comms and a M9312 ROM/Terminator with RL01/2 and CONSOLE ROMs. Basically I am back where I was before with MOS RAM. The project was to expand the backplane and replace MOS with core RAM. NEXT - get RT11 running.

As far as the original M7856 serial card goes, it appears to have been repaired more than once, some of the traces are bare, simply replacing the 1488/89 is not enough. I can compare the working card with the bad one, that should help diagnose/fix.