• Please review our updated Terms and Rules here

I'M Back, PDP 11/40..

I can well understand there might be confusion surrounding the KM11 modules and their use. (now that I compare the docs to the actual thing)

For those not familiar, these links may be instructive:


Images from those links:

Original:
maint_panel.jpeg


DIY Alternative Approach:
km11x.jpg


Another DIY:
KM11s-in-11-40.jpg



With these in mind... I'd like to get some more details of your thoughts...

  1. The original KM11 has "NAND-Latches" using a 74LS00... how did you accomplish this function on yours?
  2. The original also has a transistor, connected open-collector to BUS SSYN L - how did you handle this?

Observations:

Your design uses numeric displays where individual LEDs were - for bus address bits. There are 9 bits of bus so values can range from 777 to 000. However, meaning of these is dependent on other states - you will need to interpret. (I'm sure you've considered this - I'm just catching up)

The original docs for the KM11 show "overlay cards" to assist in this interpretation. You'll need to translate as needed.

Since there has been some confusion about your display decoder's connections, I'd advise that you power your module out of the system and manually manipulate card fingers to prove bus connections will be displayed correctly before you proceed "in chassis".
 
Hi All;
RSX11M, I will try to answer Your Questions, First I have not looked at Your Links, I will do so after this posting is finished.. Thanks for the Links.. I had downloaded the information from ShireSoft as a guide.. But, Mostly, I used the Documentation that Came with my Machine, since it has the correct Overlay for my conversion..
In answer to question 1. I copied it Directly and used the same values, 74LS00 and 3K pull-ups..
2. I used a 7405 for that, You could use a 7406 as well.. And IF I remember correctly, the /40 doesn't use the BUS SSYN L .. But I'll re-look at my pages again.. I've just looked and the /40 has CLK Enable, and M CLK, which are the same as other configurations,
then "No Time Out" on other boards, but on the /40 it is M Stop.. And the Switch for BUS SSYN L is NOT used.. I have an empty IC socket for the 7405 If it would be needed and could be wired in..
""The original docs for the KM11 show "overlay cards" to assist in this interpretation. You'll need to translate as needed.""
Which I did, and is How I got the correct PUPP and BUPP Address connections, comparing the overlay, with the light connection and for a sanity check I also compared the one from the ShireSoft schematic..
""Since there has been some confusion about your display decoder's connections, I'd advise that you power your module out of the system and manually manipulate card fingers to prove bus connections will be displayed correctly before you proceed "in chassis".""
Since I have already put the Board in the Chassis multiple times, I can do as You suggest and OR I can "OHM" it to the right IC connection on the U Word Board, which is where the PUPP and BUPP originate..
""Your design uses numeric displays where individual LEDs were - for bus address bits. There are 9 bits of bus so values can range from 777 to 000. However, meaning of these is dependent on other states - you will need to interpret. (I'm sure you've considered this - I'm just catching up)""
The first version of the Board I built used the 10 space Bar Led display, (see earlier pictures), which was hard for me to interpret the addresses, which is why I went to the Numeric Display, which is in octal, as the original lights are also..
For Now I have wired the Switch portion on another Board, which I mounted on/next to the Front Panel and use that for the different switch settings.. Even though I have made provision on my Maintenance Module Board for such switches.. You will notice, four slide (on/off) position switches, with room for 3 more switches if needed, and 3 pushbutton switches.. None of these are wired in at the moment..
And remember that this Maintenance Board is Built for my /40 and not a Universal Maintenance Board, Later when or If I get to the 11/45 I will make the necessary changes to accommodate that machine as well..
I also plan on finishing building the UniBus Analyzer (UA) based on the one from ShireSoft, using their schematic, as a basis for my design.. that way I have all of my input/outputs covered, so to speak..
Here are some pictures..

01042014925.jpg 01042014926.jpg 01042014927.jpg

THANK YOU Marty
 
Last edited:
RSX11M+ said:
I can well understand there might be confusion surrounding the KM11
Click to expand...

Why anyone would slavishly copy the KM11 is a mystery to me. Just stick a microcontroller on there and prod and poke
the chassis remotely. Then you can have whatever display you care to implement. Back in the day, the KM11 was a pain
to use because of having to have them so close in to the chassis and you never had enough of them, so you'd be constantly
swapping the plexi's . The same goes for Unibus probing.

Marty, you really should save up your pennies and get a cheap logic analyzer. That initial uPC list you sent out makes no sense
according to the microcode flow charts. If you had a logic analyzer, you could trigger on the reset of the uPC and see exactly
what it is doing. You also need to dump the microcode proms. Something is very wrong and unfortunately, the next address
comes out of the proms, and they use wire or's to modify it. Have you checked to see if any of the inputs to the uPC are stuck
low?
 
Hi All;
Al,
"" the next address comes out of the proms, and they use wire or's to modify it. Have you checked to see if any of the inputs to the uPC are stuck low?""
That is exactly what I am working on.. And for now using my version of the KM Module, once I have narrowed down the problem area then I can use things like a scope and Logic Analyzer.. I have an HP 1661A, I just have to figure out How to make it work, I have the Manual Thanks to another Member, I just have to figure out "The How" of interrupting what it says.. The Setting up of the Logic analyzer intimidates me, why can't they make it simpler, I don't know, just let me look at the signals, without all the labels and such..
I am more of an LED person than trying to interrupt the levels on a logic Analyzer or sometimes even a scope.. I think both have their places and uses.. I Think once I learn to use a logic Analyzer, I would agree with You, I would say why did I ever use Leds or why did I not learn to use an Analyzer sooner.. I have to take Baby steps, learning and such..
THANK YOU Marty
 
Last edited:
Thanks Marty.

I think you've answered everything. Didn't know you found Shiresoft or were following anything other than the DEC docs.

Checking the display - I was actually thinking of asserting one bus signal at a time to see the appropriate reaction on the display before you trust it to interpret what's happening on the bus. Given that you're building a tool - the intent was to "verify" it before proceeding - and locate miswires before consequences happen.

Marty - apparently others are following your struggle too. (still)

I wish my support was more helpful. It seems all I'm accomplishing is a check function - much of it after the fact. Ok, you're not proceeding as I would - but that's not fair of me. You're deep into a path you've selected and my choices are limited to trying to follow and assist, or sit silent and watch.

Al's advice might very well end up being more expedient were you to follow that path. I don't know if that's tooling you would need to acquire, or have to learn to work. (or worse yet - have to fix) Whichever you choose - I'll keep trying as long as you do.

Frankly, you're attempting something that's practically mission impossible using stone knives and bear skins. It's been a bit frustrating to watch at times. I suppose what keeps motivating me is the thought that it might be possible to help you avoid more damage to your system. More than getting it working - I'll feel satisfaction if my contributions help achieve that.
 
Last edited:
Hi All;
RSX11M, and AL, You don't know HOW Much You have helped, even If its after the Fact, It provides a sanity check for me and it possibly help me re-check my thinking, since sometimes I have stinking thinking, and You and AL and Everyone else who have contributed help steer me on the correct path..
"" Checking the display - I was actually thinking of asserting one bus signal at a time to see the appropriate reaction on the display before you trust it to interpret what's happening on the bus. Given that you're building a tool - the intention to "verify" it before proceeding - and locate miswires before consequences happen.""
Yes, that is part of what I am in the middle of doing right Now.. I am making it So I can do just that.. and what I have said to AL below as well..
Al, Yes, I am going to be checking exactly what Address the proms are at by checking the proms themselves and comparing that to what my board says, as a wiring sanity check..
So, Yes, Everyone Please write what you think I need to do or should do, I have lots of issues to overcome on/in this machine.. Everything takes me awhile to do and to implement sometimes rechecking my findings, and trying to figure out If what I see is what actually is there..
THANK YOU Marty
 
Last edited:
Hi All;
I found "Why" I got such funny numbers on my Numerical Displays, I got the "finger Pins" swapped left for right and top for bottom.. Such as (looking at the Print Set) for PUPP0, I thought it was (finger pin) E1, which is on the right side of the page and it should be P1, and instead of F2 for PUPP1, its L1.. So I will check each one with a meter before I make a wire change.. The "why" of the problem is now solved.. I should have checked it in the First place, Oh well.. At least I found it, RSX11M and AL Thanks for the input and getting me started on this Path of actually Checking with a meter.. I can only Hope that this can Help SomeOne else in the future, to not make the same Mistakes as I have made.. I feel so foolish, especially for not checking in the first place..
THANK YOU Marty
 
It is cool that you ARE making progress on it, though!
Next step would be to see if it actually gets to the RESET block in the flow diagrams. There should be five microinstructions
as I read them, before the first branch possibility.
 
Hi All;
AL, Thank You for the encouragement and the Suggestion.. It will be awhile, while I re-wire the "fingers".. Yes, I would love to see a '030' or something similar..
When This is all Done, I am thinking of Writing a book, Can't tell If it will be a Comedy of ERRORS or a Detective Novel.. I even have a Title for it --
How NOT to Troubleshoot, Or (sub-title) How to make all kinds of Mistakes and still make Some progress..
THANK YOU Marty
 
Marty said:
Hi All;
AL, Thank You for the encouragement and the Suggestion.. It will be awhile, while I re-wire the "fingers".. Yes, I would love to see a '030' or something similar..
When This is all Done, I am thinking of Writing a book, Can't tell If it will be a Comedy of ERRORS or a Detective Novel.. I even have a Title for it -- How NOT to Troubleshoot, Or (sub-title) How to make all kinds of Mistakes and still make Some progress..
THANK YOU Marty
Click to expand...

Did you see the posting that Ed (quapla@xs4all.nl) is selling off his collection? He has a lot of 11/40 boards.
 
Hi All;
Well, I did a check on a Prom address that the system went to after starting up.. And I know its not the correct Address after startup..Which in this case is address 170 and I Verified it with my scope and the prom is Actually at Address 170.. The Data for E15 and E14 which is the next address to jump to, should be, coming out of the Prom 174 or after the 74H10's and it is 503 and it should be 204..
I even took out the 74H10's to check the output of the Prom, in case of the 74H10's were causing the Prom to read wrong.. But not the case..
SO it looks like at least some of the Prom(s), if not all of the Proms are Hosed..
WHICH means I most likely need another Prom (U WORD ) Board, or build my own or IF I can find enough Proms then program new ones..
I am trying to decide What to do next, I think I can address the Proms with a Counter circuit and then check other locations, just in case..
OtherWise not Only am I looking for Suggestions, but trying to decide what to do next.. I did a check of my bits in my Maintenance Module from the two 74174's for the PUPP and BUPP addresses and grounding one bit at a time, (after pulling all of the 74(H)S10's) and making sure it displayed the correct bit pattern, which it does.. So I have at least have a now working Display wise Maintenance Module..
I was Hoping against Hope that the Proms were OK..
THANK YOU Marty
 
Last edited:
Hi All;
Yes, my readings could very well be wrong or my trying to explain or inverting a number could be wrong.. Which is why I am Asking here..
012 345 67 (Order of bits)
101 000 11 (Out of Prom before inverting and shown in listing)
010 111 00 (What would be listed)
110 000 10 (What should be in Prom)
001 000 01 (inverted shown in listing
turning everything around to agree with listing
76 543 210
11 000 101 (coming out of prom) 305
00 111 010 (what would be listed) 072
01 111 011 (what should be in prom) 173
10 000 100 (what would be listed) 204
Hopefully, I did everything correctly.. Please check..
THANK YOU Marty
 
Hi Marty,

I have followed you thread for some. I am not entirely sure what to suggest. But here is a short version how I brought up my 11/04 (which of course is a much simpler machine). First I fixed the PSU. When the voltages were all right I put my scope on to the main clock oscillator circuit which didn't produce any clocks. It turned out that the DC OK, AC OK signals from the PSU were bad. Fixing those made the clock running (at least a few cycles).

Then I attached my logic analyzer (state mode) to the next micro address bus using the main clock as clock signal. Here I did the mistake Al mentions. The micro address bus is active low so that you can wire-OR together signal directly on to the bus creating micro branches. It took a while to understand this. I was in the same bad mood in considering replacing all PROMs until I took another look at the schematic and finally got it.

When I was confident that the micro code executed initially correct I attached the LA in state analysis mode to the address bus using MSYN high-to-low edge as a clock signal. This directed me to find a bad bus transceiver chip causing one bit to stick high.

Back to your machine:

I think it would be possible to find Bipolar PROM chips on EBAY to program if your PROMS are bad. I have an old DataIO programmer that can do quite a lot bipolar PROM chips. The chips themselves can be hard to find but Ebay might be a source. There is a company called hobbyroms that both sell and burn chips. Looking at the schematics the chips seems to be 256x4 OC chips which are not that uncommon. Signetics 82S126 or Harris 7610 should fit.

But I think it is unlikely that your chip has failed so much that 6 out of 9 bits has reverted polarity. Maybe Al is correct that it is in fact inverted?

You bought your self an excellent logic analyzer. Have you tried that one on your system? I found it invaluable when finding problems in the CPU. I have now continued fixing a DL11-W which had a bad 7408 chip and the TU60 drive where I tracked down a failing DEC8881 and a bad 75452 chip all entirely by using the LA.

Looking at the schematics: The outputs from the PROM definitely looks like they are inverted, active low OC types. There is these BUS Uxx signals which goes to J1-J3. Are J1 - J3 connected to anything pulling the signals low? The 74H10 are three input NAND gates which inverts the signal again. Thus the signal after the gates and 74H74 flip flop should be active high.

Is there any clock pulses on the clock input of the 7474s or do you single step it some how? If there are pulses here then attach the LA to the address inputs of the PROMs. Is executing any microsteps? Or halted? What are the signals at the CLR and PR signals of the 7474 chips? These signals seems to set the initial micro address. (instead of wire-ORing in a LOW in bit 0 of the uAddress as the 11/04 which thus starts at uAddress 1)

Ok. A lot of thoughts. Maybe something is useful. Good luck!

Marty said:
Hi All;
Well, I did a check on a Prom address that the system went to after starting up.. And I know its not the correct Address after startup..Which in this case is address 170 and I Verified it with my scope and the prom is Actually at Address 170.. The Data for E15 and E14 which is the next address to jump to, should be, coming out of the Prom 174 or after the 74H10's and it is 503 and it should be 204..
I even took out the 74H10's to check the output of the Prom, in case of the 74H10's were causing the Prom to read wrong.. But not the case..
SO it looks like at least some of the Prom(s), if not all of the Proms are Hosed..
WHICH means I most likely need another Prom (U WORD ) Board, or build my own or IF I can find enough Proms then program new ones..
I am trying to decide What to do next, I think I can address the Proms with a Counter circuit and then check other locations, just in case..
OtherWise not Only am I looking for Suggestions, but trying to decide what to do next.. I did a check of my bits in my Maintenance Module from the two 74174's for the PUPP and BUPP addresses and grounding one bit at a time, (after pulling all of the 74(H)S10's) and making sure it displayed the correct bit pattern, which it does.. So I have at least have a now working Display wise Maintenance Module..
I was Hoping against Hope that the Proms were OK..
THANK YOU Marty
Click to expand...
 
MattisLind said:
Hi Marty,

I have followed you thread for some. I am not entirely sure what to suggest. But here is a short version how I brought up my 11/04 (which of course is a much simpler machine). First I fixed the PSU. When the voltages were all right I put my scope on to the main clock oscillator circuit which didn't produce any clocks. It turned out that the DC OK, AC OK signals from the PSU were bad. Fixing those made the clock running (at least a few cycles).

Then I attached my logic analyzer (state mode) to the next micro address bus using the main clock as clock signal. Here I did the mistake Al mentions. The micro address bus is active low so that you can wire-OR together signal directly on to the bus creating micro branches. It took a while to understand this. I was in the same bad mood in considering replacing all PROMs until I took another look at the schematic and finally got it.

When I was confident that the micro code executed initially correct I attached the LA in state analysis mode to the address bus using MSYN high-to-low edge as a clock signal. This directed me to find a bad bus transceiver chip causing one bit to stick high.

Back to your machine:

I think it would be possible to find Bipolar PROM chips on EBAY to program if your PROMS are bad. I have an old DataIO programmer that can do quite a lot bipolar PROM chips. The chips themselves can be hard to find but Ebay might be a source. There is a company called hobbyroms that both sell and burn chips. Looking at the schematics the chips seems to be 256x4 OC chips which are not that uncommon. Signetics 82S126 or Harris 7610 should fit.

But I think it is unlikely that your chip has failed so much that 6 out of 9 bits has reverted polarity. Maybe Al is correct that it is in fact inverted?

You bought your self an excellent logic analyzer. Have you tried that one on your system? I found it invaluable when finding problems in the CPU. I have now continued fixing a DL11-W which had a bad 7408 chip and the TU60 drive where I tracked down a failing DEC8881 and a bad 75452 chip all entirely by using the LA.

Looking at the schematics: The outputs from the PROM definitely looks like they are inverted, active low OC types. There is these BUS Uxx signals which goes to J1-J3. Are J1 - J3 connected to anything pulling the signals low? The 74H10 are three input NAND gates which inverts the signal again. Thus the signal after the gates and 74H74 flip flop should be active high.

Is there any clock pulses on the clock input of the 7474s or do you single step it some how? If there are pulses here then attach the LA to the address inputs of the PROMs. Is executing any microsteps? Or halted? What are the signals at the CLR and PR signals of the 7474 chips? These signals seems to set the initial micro address. (instead of wire-ORing in a LOW in bit 0 of the uAddress as the 11/04 which thus starts at uAddress 1)

Ok. A lot of thoughts. Maybe something is useful. Good luck!
Click to expand...

Hi All;
Mattislind, Thank You for Your thoughts.. I will try and answer Your postings.. But, I will most likely do alot of back and forth, reading your entries..
I have an 11/04 with a Power Supply problem, I am deadly afraid of attacking the Power Supply.. I am not a Power Supply person..
I have some Proms, I just need to see what I have.. And I think I have a programmer that will do the Job, I know I have the programmer, I just need to see If it will do these proms..
yes, I know they are Wired OR.. and they Output low unless there is a high.. They also DO have Pull-ups, the J1 to J3 are for the EIS option and are all pulled high as well..
""But I think it is unlikely that your chip has failed so much that 6 out of 9 bits has reverted polarity. ""
You need to go all the way back to the Origional problem, where a "BAD" (unknown to me) Board shorted a possibly 12 Volt line to the +5 volt line, not only taking out what IC's were directly in its path, but any "weak" ones as well, and on ALL 5 CPU Boards, not all Ic's just some of them..
""Maybe Al is correct that it is in fact inverted?"" Yes, they are Inverted, from the Listings, Its even says So in the Listings files and in the DOcs.. The 74H10's Inverts them from the proms before they go out to the 74174's PUPP and BUPP..
""Is there any clock pulses on the clock input of the 7474s or do you single step it some how? If there are pulses here then attach the LA to the address inputs of the PROMs. Is executing any microsteps? Or halted? What are the signals at the CLR and PR signals of the 7474 chips? These signals seems to set the initial micro address. (instead of wire-ORing in a LOW in bit 0 of the uAddress as the 11/04 which thus starts at uAddress 1)""
The Clock Pulses come from the Single Stepping of the Maintenance Board and the Timing Board.. And I looked with my scope at the proms Both the Addressing and the Data lines, carefully recording what I found on paper.. I have also, before looking at the prom Data lines for the "Next Prom Address" made a separate list of each prom address and its "next Address" which is not what comes out of the proms, but it is the Inverted address coming out of the 74H10's.. The Listing even states that they are Inverted..
Also, all of the Ic's for "This section" have been removed and Socketed.. So, I can easierily look at the signals from the Proms, the 74H10's, the 74H74's and the 74174's..
Also, Remember, that "Normally" at Startup, after the AC LO and DC LO do their thing, then JAMUPP sets and Presets the 74H74's for Address '030' which is for Console start, and my System is NOT doing that as of Yet, and I know it.. Which is one of the reasons it starts at like address '170'.. So there are still alot of things to look at and correct, beside making the Proms do the "Correct" thing.. So, this is before the Proms even get "addressed"..
I am hoping that by going over all of this I can show that I have tried to cover all my bases.. I am also making a Counter so I can check more prom addresses, starting at '000' and going up from there..
THANK YOU Marty
 
Last edited:
Hi All;
After thinking about it since Yesterday, I have decided to put the PDP 11/40 away for now, Not forever, but for NOW.. It will give me some Time to Think about "This" machine and Time for some (Maybe) boards to show up.. Give me a Breather on this system..
But, I am going to Start a "New" Thread and Work on the PDP 11/45..
THANK YOU Marty
 
Hi All;
I think based on what I have Found in the PDP 11/45, that this type of Short, is what did the 11/40 in,, -12 shorted to +5.. It at least fits the scenario of what took out so much of the IC's on this machine.. So I think I know Now what did this machine in, a -12 Volt short to the +5 Volt Rail..
I don't plan on making the same Mistake with the PDP 11/45 !!
THANK YOU Marty
 
Hi All;
I think I have Solved the mystery of what went so terribly WRONG on the 11/40.. I plugged an un-modified Board into a modified slot.. The Communications Board, M7800.. After thinking about it, No the M7800 pluggs into an SPC slot.. But, I will still leave my warning below..
So the un-modified Board an M7800 plugged into a Modified Unibus slot shorts -12 to places it should not go.. My thinking is wrong since it is an SPC, but what happened somwhow (I think) is correct, somehow a + or - 12 got on the +5 voltage rail.. Thereby causing a Ripple effect and Blowing Many Ic's down the line from where the origional short was, or is..

So, Be WARE of Plugging a Board that is for UniBus into a modified UniBus slot !!

THANK YOU Marty
 
Last edited:
Marty said:
So, Be WARE of Plugging a Board that is for UniBus into a modified UniBus slot !!
Click to expand...

I'm going from memory here, but the modified part is on the E and F connectors. A quad board should work fine,
a hex board would not. Easy enough to check, look at a backplane designed for MUD devices, like the 11/34.
 
You must log in or register to reply here.
Back
Top