• Please review our updated Terms and Rules here

Restoring a DEC PDP-11/05

I would check the micrprogramcounter, MPC. It is in E101 and E111. The clock signal, the inputs and the outputs.

Is there any activity? If yes, what are the microsteps it is executing?

A logic analyzer would be helpful. But a KM11 could do as well.
 
E101 and E111 are both receiving a clock signal on pin 9.

To save having to upload numerous screen captures showing a ~4 volt signal, I'm just going to list the IC's and whether their input/output was a high or low.
Logic high for me seems to be between 3.75 and 4.00 volts according to my measurements. I do happen to have a logic analyzer built into my USB oscilloscope too.

E101:
pin 2 Hi
pin 3 Low
pin 4 Hi
pin 5 Hi
pin 6 Low
pin 7 Hi
pin 10 Low
pin 11 Hi
pin 12 Low
pin 13 Hi
pin 14 Low
pin 15 Hi

E111:
pin 2 Hi
pin 3 Low
pin 4 Hi
pin 5 Hi
pin 6 Low
pin 7 Hi
pin 10 Hi
pin 11 Low
pin 12 Hi
pin 13 Hi
pin 14 Low
pin 15 Hi

E13:
pin 1 Low
pin 2 Low
pin 3 Hi
pin 4 Low

E70:
pin 1 Low
pin 2 Hi
pin 3 Hi
 
I think that you either have to use a KM11 and single step it or use a logic analyzer so we can se any possible transitions until it ends up in this state. Trigger a DC LOW or AC LOW ( check which is reseting the CPU) and then capture all microaddresses from there.

Apparently it is stuck here for some reason and the steady state is not very interesting. We need to understand why it ends up here.
 
I think that you either have to use a KM11 and single step it or use a logic analyzer so we can se any possible transitions until it ends up in this state. Trigger a DC LOW or AC LOW ( check which is reseting the CPU) and then capture all microaddresses from there.

Apparently it is stuck here for some reason and the steady state is not very interesting. We need to understand why it ends up here.
I guess I’ll look into building a KM11. Compared to running test leads from the ic pins to the logic analyzer inputs, it seems a lot simpler. Just out of curiosity, how would I trigger a DC or AC low?
 
The KM11 has a toggle switch for this. There is lines on the bus DC LOW L. The signal is open collector so just add a driver that forces it low.
 
I guess I’ll look into building a KM11. Compared to running test leads from the ic pins to the logic analyzer inputs, it seems a lot simpler. Just out of curiosity, how would I trigger a DC or AC low?

Thinking of it. The microprogram counter has to be in the backplane since it gets to the KM11 board. I just don't know what pins. It has to be in the schematic. Ahh. The NXT field is in the backplane. But that is good enough.



Screenshot 2022-08-17 at 13.02.14.png

If I parse the the MPC value you gave above I think it ends up in 277 octal. Unfortunately it is not a valid microprogramstep. If you look at the microprogram listing in the schematic you will see that it starts of at 000 octal (the inital condition when the MPC is reset) and then progress through 241, 347, 074, 351 and then 305 which dispatches what to do next, execute instruction or handle a key or whatever.

One thing to keep in mind is that the MPC bus is an open collector bus. Several outputs can wire-or in here to modify the NXT field, creating conditional micro-jumps.

A broken PROM or OC-driver could make that you ended up in an invalid micro-step. If we have the entire flow from startup it would be easier to understand what is going on. I see one difficulty though when examining the schematic. The CLR signal to the MPC is directly from VCC. So a reset of MPC can only be caused when the system is powered up. Unless you can force it low for a short period, which could be done using a toggle switch and a debouncing-circuit and wire it to the RESET line of E101 and E111. Trigger the logic analyzer on reset going high and see what is coming on the MPC bus in the backplane clocked by the processor clock on pin 9.
 
Thinking of it. The microprogram counter has to be in the backplane since it gets to the KM11 board. I just don't know what pins. It has to be in the schematic. Ahh. The NXT field is in the backplane. But that is good enough.



View attachment 1244977

If I parse the the MPC value you gave above I think it ends up in 277 octal. Unfortunately it is not a valid microprogramstep. If you look at the microprogram listing in the schematic you will see that it starts of at 000 octal (the inital condition when the MPC is reset) and then progress through 241, 347, 074, 351 and then 305 which dispatches what to do next, execute instruction or handle a key or whatever.

One thing to keep in mind is that the MPC bus is an open collector bus. Several outputs can wire-or in here to modify the NXT field, creating conditional micro-jumps.

A broken PROM or OC-driver could make that you ended up in an invalid micro-step. If we have the entire flow from startup it would be easier to understand what is going on. I see one difficulty though when examining the schematic. The CLR signal to the MPC is directly from VCC. So a reset of MPC can only be caused when the system is powered up. Unless you can force it low for a short period, which could be done using a toggle switch and a debouncing-circuit and wire it to the RESET line of E101 and E111. Trigger the logic analyzer on reset going high and see what is coming on the MPC bus in the backplane clocked by the processor clock on pin 9.

Just to clarify, are connections to the backplane represented by a two-letter designation followed by a 1 or a 2? For pinouts of the NXT field, do you know which schematic I should be looking at?
 
The one I took a screenshot from above. The full schematics is here unless you already found it. The manual is also very good to read to understand the machine. It goes into great detail on explaining the operations of the all aspects.

The full microcode listing is in the schematics as well.

For example MPC0 is at FP2. F is the connector closest to the front panel. 2 designates the solder side. DEC ”number” their connectors with letters. A B C D E F H J K L M N P R S T U V. 18 contacts per connector. They skipped a number of letters that easily got mixed up. So contact V is closest to the front panel.
 
Hello everyone-

It’s been a little while but now I have assembled a KM-11 maintenance module from Guy at Shireshoft. (Excuse the large toggle switch- there weren’t any spdt momentary mini toggles in stock so I had to use a dpdt)

I’m going to do a little reading on how to use the KM-11 to troubleshoot the 11, but if anyone wouldn’t mind giving a quick run down on what to expect, I’d appreciate it.

65B98BCB-0B53-404D-8540-847F6B5EB57F.jpeg
 
Power up the machine with manual clocking enabled. Check what microaddress it reports and compare with the microcode listing. Then toggle the manual clock and see how it compares with microcode list.
 
Power up the machine with manual clocking enabled. Check what microaddress it reports and compare with the microcode listing. Then toggle the manual clock and see how it compares with microcode list.
Bit of a stupid question;
The manual says to plug the KM-11 into position A2. If I’m reading the backplane layout correctly, the M7260 module is in slot 2, rows A-F, and there is nowhere to plug in the KM-11. Is this correct? The label on the chassis lists KM-11 in slot 1, along with M9970 and DF11.
 
Bit of a stupid question;
The manual says to plug the KM-11 into position A2. If I’m reading the backplane layout correctly, the M7260 module is in slot 2, rows A-F, and there is nowhere to plug in the KM-11. Is this correct? The label on the chassis lists KM-11 in slot 1, along with M9970 and DF11.
Yes. Exactly right. It goes into slot 1.
 
Yes. Exactly right. It goes into slot 1.
Alright then. The KM 11 is plugged into the designated slot in the row closest to the front of the machine. I started it up and followed the manual on using the KM11 with the KM 1 overlay. Depressed the halt switch and held down the start switch. Toggled the m clock twice. Released the start switch. Toggled twice again. MPC 06 indicator is dim along with AMUX 08. All the other LEDs except for BBSY are illuminated. Toggling the M clock did not do anything. Also, there is no change in lights from startup.
 
It would be more interesting (i think) to see what happens directly after power up. Can you set it to manual and THEN power it up. What microstep us it at then? The MPC should be cleared at power up but what you see as microstep is the next microaddress not the current.

You should have the microcode listing nearby. At page 42 of the engineering drawing on bitsavers is the restart from power-fail. I.e. When the computer starts up.

The first address is obviously 000 since MPC is reset. But on the LEDs you should see the next step, which is 241 or 01100001. If you then do a double toggle on the manual clock you should have 347 or 11100111.

Check this flow against your machine.
 
It would be more interesting (i think) to see what happens directly after power up. Can you set it to manual and THEN power it up. What microstep us it at then? The MPC should be cleared at power up but what you see as microstep is the next microaddress not the current.

You should have the microcode listing nearby. At page 42 of the engineering drawing on bitsavers is the restart from power-fail. I.e. When the computer starts up.

The first address is obviously 000 since MPC is reset. But on the LEDs you should see the next step, which is 241 or 01100001. If you then do a double toggle on the manual clock you should have 347 or 11100111.

Check this flow against your machine.

I tried again, this time with clock set to manual before turning the system on. MPC 00-05,07 are lit and MPC 06 is dim.

When reading the micro codes from the KM-11 do you read them from lamps 00-07 or 07-00? They would be in octal too, correct?
 
Bit seven is MSB of MPC and bit zero is LSB of MPC. Your result would correspond to octal 100, which is wrong.

With manual set can you check the inputs and outputs of the MPC register with a multimeter. The non-inverted outputs should be 0.
 
Bit seven is MSB of MPC and bit zero is LSB of MPC. Your result would correspond to octal 100, which is wrong.

With manual set can you check the inputs and outputs of the MPC register with a multimeter. The non-inverted outputs should be 0.
Do you think it could be an issue with the ROM that’s giving invalid MPCs? Just to rule it out, I also checked every LED on my KM-11. (I had replaced two that were dead and one that was always dim to avoid confusion in the future)
 
Yes, it would be interesting to know if the MPC ROM is ok.

Just to be sure it would be good to verify with a multimeter that the outputs of the ROM is what you see on the KM11.
 
Yes, it would be interesting to know if the MPC ROM is ok.

Just to be sure it would be good to verify with a multimeter that the outputs of the ROM is what you see on the KM11.
I’ll take a look at those. Do you know which prints would be the most useful?
 
I have attached a screen capture above. It is the same signals as the two 74175. Inputs to the ROM is the output of the MPC. The outputs of the ROMs are the inputs of the MPC.
 
Back
Top