• Please review our updated Terms and Rules here

So I got my first PDP-11

Haven't messed with the PDP-11 in a while (been saving cash for other stuff), but I did get my hands on a M9301-YF, since the M9302 I got had the wrong ROM. From what I read, the /04, /34 cpus have pull up resisters and W1-W5 do not need to be jumpered right? I set the dip switch to ON,OFF, and all the rest ON. My system looks like this:

M7263 M7263 M7263 M7263 M7263 M7263
M7891 M7891 M7891 M7891 M7891 M7891
M9301 M9301 M7856 M7856 M7856 M7856
EMPTY EMPTY M7856 M7856 M7856 M7856
EMPTY EMPTY EMPTY GRANT EMPTY EMPTY
EMPTY EMPTY EMPTY GRANT EMPTY EMPTY
EMPTY EMPTY EMPTY GRANT EMPTY EMPTY
EMPTY EMPTY EMPTY GRANT EMPTY EMPTY
M9302 M9302 EMPTY GRANT EMPTY EMPTY

The memory's power led is green, and the parity error light is off. Now, I only have an operator's panel on this so it really makes it more difficult to debug the machine. If I flip the right switch to HALT, would it shut off the RUN light? Also, are there any guides on making a serial cable for the M7856?
 
Correct, W1 through 5 out. Your switch settings are also correct to dump you into the console emulator.

The cable you need to make is a BC03-L. The diagram is in the 11/04 and 11/34 field print sets.

Does the run light go out when you throw the halt switch?

Lou (I also have an 11/04, which runs well. I hope you get yours running.)
 
Last edited:
So it seems that you’ve been having an on-going issue with the run light. (The light should be off when the HALT/CONT switch is down.) The latest backplane configuration you posted looks good to me, but I’m still thinking that the inability to halt the processor is most likely a grant problem. I’ve found that the G727A grant continuity cards fit great in either direction but don’t seem to work as well when backwards, so you may want to double check that none are reversed. The blank side should be closest to the processor (facing up in your machine); the side with the traces should be closest to the M9302 (facing down). If nothing obvious is detected there, you might try completely removing the M9302 to see if that temporarily rectifies the problem. If so, it would then appear that one or more of the NPG backplane jumpers is missing, cut or otherwise open, and you will need to replace the jumper(s) or install a G7273 double grant (or other module that completes the grant chain) in the appropriate slot(s).
 
Lou - N2MIY said:
Correct, W1 through 5 out. Your switch settings are also correct to dump you into the console emulator.

The cable you need to make is a BC03-L. The diagram is in the 11/04 and 11/34 field print sets.

Does the run light go out when you throw the halt switch?

Lou (I also have an 11/04, which runs well. I hope you get yours running.)
Click to expand...

Found it, a normal IDE cable shoul work right? Just pull apart the ribbon cable and solder on wire? No, the run light stays on.

KM11 said:
So it seems that you’ve been having an on-going issue with the run light. (The light should be off when the HALT/CONT switch is down.) The latest backplane configuration you posted looks good to me, but I’m still thinking that the inability to halt the processor is most likely a grant problem. I’ve found that the G727A grant continuity cards fit great in either direction but don’t seem to work as well when backwards, so you may want to double check that none are reversed. The blank side should be closest to the processor (facing up in your machine); the side with the traces should be closest to the M9302 (facing down). If nothing obvious is detected there, you might try completely removing the M9302 to see if that temporarily rectifies the problem. If so, it would then appear that one or more of the NPG backplane jumpers is missing, cut or otherwise open, and you will need to replace the jumper(s) or install a G7273 double grant (or other module that completes the grant chain) in the appropriate slot(s).
Click to expand...

Hmm, not to get my hopes up but I am pretty sure I put it in backwards, traces facing up towards the processor. Will try it in the other direction. I did check the NGP while back and all pins were wired together.
 
Well, flipped the grant cards around so the pins face down instead of up. The parity light now goes red on startup, and clears when the INIT switch is thrown. The RUN light is active regardless if the HALT switch is thrown or the M9302 is present. I did have a weird case where when the halt switch was thrown, the parity light started to flicker red until it stabilized to a full red. Stopped when I removed the M9302 to do my test.

Seems we are getting close, at least the memory is responding to the INIT/BOOT switch and is actually running.
 
I was hoping that correcting the orientation of any reversed grant cards would be the answer to your run light issue. Even though that isn’t the case, it sounds like you needed to fix a problem there anyway so at least it wasn’t wasted effort.

Since removing the M9302 had no impact, I think it may be a good idea to check the connection between the console and backplane. The two are connected by a short ribbon cable that has a red stripe along one edge. The end that plugs into the backplane should be keyed so it only plugs in one way. On the console side, the cable should be plugged into connector J1 so that the red stripe is farthest from the rotary power switch (or on your left as you face the front side of the panel with the rotary switch on your right).
 
KM11 said:
I was hoping that correcting the orientation of any reversed grant cards would be the answer to your run light issue. Even though that isn’t the case, it sounds like you needed to fix a problem there anyway so at least it wasn’t wasted effort.

Since removing the M9302 had no impact, I think it may be a good idea to check the connection between the console and backplane. The two are connected by a short ribbon cable that has a red stripe along one edge. The end that plugs into the backplane should be keyed so it only plugs in one way. On the console side, the cable should be plugged into connector J1 so that the red stripe is farthest from the rotary power switch (or on your left as you face the front side of the panel with the rotary switch on your right).
Click to expand...

Checked the ribbon cable.... red strip on the backplane is facing down to pin 1, and on the panel it is left (when facing the front)... plugged into J1. Interestingly, when J1 is disconnected, the parity light still clears when INIT is thrown. From diagrams it seemed that one of the signal lines was transported through J1.
 
Hope I am not wasting your time but I spent a week trying to find out why my reset switch did not work on my 11/23, turned out it was a blown gate that was connected to the switch. Did you check the gate that the switch connects to or the gate that’s driving the RUN light? It’s a stupid things but easy to test.
 
Qbus said:
Hope I am not wasting your time but I spent a week trying to find out why my reset switch did not work on my 11/23, turned out it was a blown gate that was connected to the switch. Did you check the gate that the switch connects to or the gate that’s driving the RUN light? It’s a stupid things but easy to test.
Click to expand...

Not at all, any suggestions are helpful. I am going to have to grab my logic analyzer and do some tests. Hate wiring it up tho.
 
So I’m just catching up on this thread and it sounds like most of the simple fixes I would have expected might solve your problem didn’t. Qbus makes a good point in that a bad gate on the console could result in the same symptoms. Not sure how far you may have gotten checking that, but I have a few last suggestions before breaking out all the test equipment and schematics. Some revisions of the KY11-LA console when used in a BA11-L required a jumper connecting the two GND tabs at the top of the board. Without that, you may have a floating ground on the console panel. Also, inspect the backplane one more time for any bent pins or obvious dangling wires – anything that might interrupt or alter the signals.

Effectively, the processor sees a halt from the front panel as an interrupt request. The sequence of events that should happen when you move the CONT/HALT switch to the HALT position are:

  • The console panel asserts HALT REQUEST L on the bus
  • The processor acknowledges the halt request by asserting HALT GRANT H – this is the signal that turns off the run light (BUT there is a caveat here)
  • The console acknowledges the halt grant by asserting SACK – in this state, the console owns the bus until it drops SACK when you move the CONT/HALT switch back to CONT
So when HALT GRANT H is high, the run light should be off but, glancing at the schematic, here is the caveat – BUS INIT L must also be high. If you disconnect the J1 ribbon cable from the console, I would be curious as to what the voltage levels are for BUS INIT L (pin 1) and BUS DC LO L (pin 9) coming from the bus, because if either is low, it sounds like an upstream problem.
 
KM11 said:
So I’m just catching up on this thread and it sounds like most of the simple fixes I would have expected might solve your problem didn’t. Qbus makes a good point in that a bad gate on the console could result in the same symptoms. Not sure how far you may have gotten checking that, but I have a few last suggestions before breaking out all the test equipment and schematics. Some revisions of the KY11-LA console when used in a BA11-L required a jumper connecting the two GND tabs at the top of the board. Without that, you may have a floating ground on the console panel. Also, inspect the backplane one more time for any bent pins or obvious dangling wires – anything that might interrupt or alter the signals.

Effectively, the processor sees a halt from the front panel as an interrupt request. The sequence of events that should happen when you move the CONT/HALT switch to the HALT position are:

  • The console panel asserts HALT REQUEST L on the bus
  • The processor acknowledges the halt request by asserting HALT GRANT H – this is the signal that turns off the run light (BUT there is a caveat here)
  • The console acknowledges the halt grant by asserting SACK – in this state, the console owns the bus until it drops SACK when you move the CONT/HALT switch back to CONT
So when HALT GRANT H is high, the run light should be off but, glancing at the schematic, here is the caveat – BUS INIT L must also be high. If you disconnect the J1 ribbon cable from the console, I would be curious as to what the voltage levels are for BUS INIT L (pin 1) and BUS DC LO L (pin 9) coming from the bus, because if either is low, it sounds like an upstream problem.
Click to expand...

Well I got a hold of a KY11-B. First time making ribbon cables, seems I screwed up because one end is backwards. Anyway once I flipped it around, the 7-seg display lit up and lo behold, the run light is OFF! It seems the issue was either the cable from the backplane to the operator panel, or the panel itself. Going to run a few tests to check the memory and cpu. So excited, seems that patience is paying off! When I do CTRL-INIT the parity light goes off, and RUN lights up and then turns off.

Added: Did a quick memory test.... values read out correctly!

Hmm, having am issue getting the CPU to run. I deposit 5007 into address 0. When CTRL+Start is pressed, the RUN light flashes and then stops.
 
Last edited:
Ioncannon said:
Well I got a hold of a KY11-B. First time making ribbon cables, seems I screwed up because one end is backwards. Anyway once I flipped it around, the 7-seg display lit up and lo behold, the run light is OFF! It seems the issue was either the cable from the backplane to the operator panel, or the panel itself. Going to run a few tests to check the memory and cpu. So excited, seems that patience is paying off! When I do CTRL-INIT the parity light goes off, and RUN lights up and then turns off.

Added: Did a quick memory test.... values read out correctly!
Click to expand...

Congratulations. Seeing that you wanted to upgrade anyway, it sounds like you solved two problems then.

Ioncannon said:
Hmm, having am issue getting the CPU to run. I deposit 5007 into address 0. When CTRL+Start is pressed, the RUN light flashes and then stops.
Click to expand...

What do you see in the display when it halts?
 
KM11 said:
Congratulations. Seeing that you wanted to upgrade anyway, it sounds like you solved two problems then.



What do you see in the display when it halts?
Click to expand...

All zeroes. The 5007 opcode is put into location 0 so maybe it is the address?.
 
Ioncannon said:
All zeroes.
Click to expand...

Curious. What you have there is the world's shortest loop but only when loaded at location 0. Here's another one that isn't as picky about where it's loaded: deposit 777 at a different location, say 300 for example, and run it from there. What happens then?
 
KM11 said:
Curious. What you have there is the world's shortest loop but only when loaded at location 0. Here's another one that isn't as picky about where it's loaded: deposit 777 at a different location, say 300 for example, and run it from there. What happens then?
Click to expand...

Do I run it from address 300? Same thing. Also, CTRL-BOOT also does the same thing.
 
Sorry, I guess I should have provided more detail in that last post. When a halt is encountered, the display would normally show the contents of the program counter. You said the first loop you attempted to run seemed to start and then stop, but seeing all zeroes in the display was inconclusive. By starting a loop at a different location, I was hoping to determine if it also stopped and, if so, where. (I was expecting you to tell me the display showed 300 or 302 if it halted but I understood you to say that it stops and you still see all zeroes.) Of course, both of those loops should run continuously until manually stopped so something is not right. I hate to say it, but you may be dealing with a different manifestation of the same issue you were troubleshooting earlier.

BTW, I assumed, maybe incorrectly, that the SR DISP light has been off, otherwise the display is showing you the contents of the switch register (which most likely is zero).
 
Last edited:
KM11 said:
Sorry, I guess I should have provided more detail in that last post. When a halt is encountered, the display would normally show the contents of the program counter. You said the first loop you attempted to run seemed to start and then stop, but seeing all zeroes in the display was inconclusive. By starting a loop at a different location, I was hoping to determine if it also stopped and, if so, where. (I was expecting you to tell me the display showed 300 or 302 if it halted but I understood you to say that it stops and you still see all zeroes.) Of course, both of those loops should run continuously until manually stopped so something is not right. I hate to say it, but you may be dealing with a different manifestation of the same issue you were troubleshooting earlier.

BTW, I assumed, maybe incorrectly, that the SR DISP light has been off, otherwise the display is showing you the contents of the switch register (which most likely is zero).
Click to expand...

Actually, I was wrong. When it is run from 300, the display shows 300 after it halts. Clearing the screen so it says all 0s, then pressing CTRL-START causes it to flash RUN and display 300. I tried this on other addresses and same thing happens.

BTW, where is the PC loaded from? Is it the current value on the display or the current loaded address?
 
Ioncannon said:
BTW, where is the PC loaded from? Is it the current value on the display or the current loaded address?
Click to expand...

I guess I don’t understand this question. At a high level, I’m assuming that you entered the following:

300
LAD
777
DEP
CLR
300
LAD
CNTL-START

(The SR DISP, MAINT, and BUS ERR lights should have all been off.) After you pressed CNTL-START, the program should have run until you halted it, but instead I understood that the run light blinked and you saw 300 in the display. On a working console, the 300 would be the current value of the program counter (which is R7 of the CPU).

If the CPU appears to consistently halt, I think I would enter a small program that traverses a few instructions and then single stepping through it to see if it executes correctly. (I believe there’s a good example of this process in the PDP-11/34 System User’s Manual which shares the same console.) This exercise might point to a specific problem and also serve to further test the console.
 
Not trying to steal your thread or anything, have been watching your progress with great interest being I am almost at the same place you are. This is where I am with my 11/34, will say one thing and don’t know if it means anything or not but my life got way better after stuffing my DL11 in the last slot on the board, just in front of the 9302 terminator. Yea I know it’s supposed to work in any SPC slot but that last slot was like magic with my system. Anyway this is where I am.


Ok, got my DL11-W chugging along at 9600 baud, 8 bit with the clock running also. Thing is I am use to working on Qbus stuff and on all the q systems I have if you exit the “boot” program you can use the con to examine and change memory locations. Looking at the 11/34 book it looks like my M9301 YA should allow me to dump out of the boot program and use the con the same way. Think you push the control and interrupt keys at the same time but if I do this my con port stops working. No matter what I do I can not get the con to change or modify anything, wont even echo. If I start running the boot program the con to do anything. Once I halt the system can use the front keypad to do all the normal junk, can enter addresses, change contents, step thru locations and run simple programs. Never having used a Unibus system is this normal for an 11/34 with the M7859 and keypad or dose my M9301 have a stupid ROM or is there something else I am missing? Would be nice to change and modify stuff via the con but maybe that’s just how Unibus systems are? Next step will be stuffing a M7762 up the frame and seeing if I can boot from a RL just hate to have to strip cables from the 11/23 to try to get the 34 up.
 
You must log in or register to reply here.
Back
Top