• Please review our updated Terms and Rules here

PDP-11/23 Bootstrap Failure?

NeXT

NeXT

Veteran Member
Joined
Oct 22, 2008
Messages
8,455
Location
Kamloops, BC, Canada
Okay lets try looking at this again since the last thread a while back was a spectacular disaster.

I have a KDF11-B M8189 CPU board and a memory board M8067 that Lou gave me a number of years back. It was sent to me with two EPROMS in the bootstrap sockets that would start the machine up to a series of nice menus you could navigate from the VT100 and added support for local RX, RL and RD booting support as well as primitive diagnostics. About a year and a half ago I attempted to add a disk controller and somewhere in the process something happened. the status LED's on the CPU are now all lit which apparently means there has been a bootstrap ROM checksum error and the system halts immediately.
As the board stands right now everything is reset back to the configuration it was at when it still worked. Here is where I am:

-All but the CPU and ram are pulled form the BA23 chassis to aid in troubleshooting. The absolute minimum I can do.
-I can verify that the BA23 power supply and backplane are not faulty. I have a spare KDJ11-B 11/73 CPU and M7551 memory board that when plugged in and powered I can reach the ODT.
-I can verify the M7551 memory board is healthy enough to work with the test CPU.
-I cannot verify if the M8067 is faulty as I do not know if the M8067 and M7551 can be interchanged.
-I cannot verify the integrity of the bootstrap EPROMS as I do not have any image to compare with and all three of my EPROM programmers either cannot support the MCM68766C or are currently not functioning.
-I cannot jumper anything on the KDF11-B that requires a wire wrap tool.
-I cannot FORCE the system to drop to the ODT, even with the halt button.
 
If you haven't already, I would check the PowerUp Mode Selection. It should be set to Mode 2, J18 to J19 installed and J18 to J17 removed - to boot from EPROM. Its also a good idea to check the other jumpers and switches for the correct setting.

Then I would change it to Mode 1 - Console ODT and pull all the other boards including memory. This will bypass the EPROMs. The configuration is J18 to J19 removed, J18 to 17 Inserted. The small removable configuration jumpers will fit between two adjacent pins. That's what used on my board.

If it won't go directly to ODT when power is applied, then the CPU board is the problem. If ODT works, check to read the bootstrap location (EPROM) @ 7773000 for valid data. Only then would I add memory and check for read/write, etc..


Hope this works for you.
Jerry
 
EPROMS do occasionally suffer from bit rot. If one of the 68766 prove to be the root of your current trouble, I offer the suggestion shown in this photo : http://www.vcfed.org/forum/album.php?albumid=49&attachmentid=2719 . You should be able to at least program 2764s with your eprom programmer. Ok, it's really crude, but it does work. The rightmost socket in the picture had to have another socket plugged into it. The IC's overlap each other in a stairstep fashion. Hey I'm a mechanical engineer, we rig stuff up all the time!

Lou
 
Hella bump time.

I was able to snag someone in Seattle who let me borrow his KDF11-B CPU and I came down with my entire BA23 chassis and verified again my cabinet/PSU/ram was not at fault by putting his board in and getting to KDF11B-BE ROM V0.9 on a terminal. We then swapped EPROMs and found my CPU board would still halt with a checksum error and no ODT could be forced and my EPROMs in his CPU would also halt his CPU but ODT could be forced. We then swapped EPROMs back, set my boards jumpers/dip switches to match his (which were correct) and found it still did not work. We then swapped the MMU, CPU and FPU and still could not get my CPU to even reach ODT.
 
So 2 faults?
Eprom fault in your system and in his working sytem? That will make a Eprom fault.

Then CPU board defect?
Did you swap your MMU, CPU and FPU on his board and that did not work or
Dis you swap your MMU, CPU and FPU on his working KDF11?

Be bit more specific of the stapes you made.
That will point the next problem.

With fault analysing it is important to a secure step by step to get to the broken part.

Could be multiple faults also so make a list of steps.
Then with multiple faults at start up after long time not been in use beware of your supply!
Could be that a startup power peak has destroyed couple of logic.

Best is to start up first (after long time out of use) only the supply with all boards out.

NEVER put a plug in what ever electronic that long time been out of use.
ElektroCaps have to be reformed other wise you end op with lots of defect equipment.

Its like a totally dead battery those capacitors.
And need to be reformed by time, say half an hour at least.
Reformation is done by power-up and a new oxide layer is formed chemically inside the cap.

Correctly it is done by separately voltage on the big capacitors with a series resistor.
Read about it.

So long time out of use...start up with only the supply and check voltage and ripple-AC.
Simple test 5Vdc with multimeter at position AC and read the AC-voltage
Only some minor mV to be read.
 
So 2 faults?
Eprom fault in your system and in his working sytem? That will make a Eprom fault.

Then CPU board defect?
Did you swap your MMU, CPU and FPU on his board and that did not work or
Dis you swap your MMU, CPU and FPU on his working KDF11?

Be bit more specific of the stapes you made.
That will point the next problem.

With fault analysing it is important to a secure step by step to get to the broken part.

Could be multiple faults also so make a list of steps.
Then with multiple faults at start up after long time not been in use beware of your supply!
Could be that a startup power peak has destroyed couple of logic.

Best is to start up first (after long time out of use) only the supply with all boards out.

NEVER put a plug in what ever electronic that long time been out of use.
ElektroCaps have to be reformed other wise you end op with lots of defect equipment.

Its like a totally dead battery those capacitors.
And need to be reformed by time, say half an hour at least.
Reformation is done by power-up and a new oxide layer is formed chemically inside the cap.

Correctly it is done by separately voltage on the big capacitors with a series resistor.
Read about it.

So long time out of use...start up with only the supply and check voltage and ripple-AC.
Simple test 5Vdc with multimeter at position AC and read the AC-voltage
Only some minor mV to be read.
 
MauriceH said:
So 2 faults?
Eprom fault in your system and in his working sytem? That will make a Eprom fault.
Click to expand...
I cannot confirm this as his EPROM set did not work in my board, even though they were spec and configuration identical.
Then CPU board defect?
Did you swap your MMU, CPU and FPU on his board and that did not work or
Dis you swap your MMU, CPU and FPU on his working KDF11?
Click to expand...
The MMU, CPU and FPU were all swapped and tested. They all tested okay in his healthy board.

Be bit more specific of the stapes you made.
That will point the next problem.

With fault analysing it is important to a secure step by step to get to the broken part.
Click to expand...
I'm trying to keep my cool on this, but I keep having to repeat myself, sorry.
Could be multiple faults also so make a list of steps.
Then with multiple faults at start up after long time not been in use beware of your supply!
Could be that a startup power peak has destroyed couple of logic.
Best is to start up first (after long time out of use) only the supply with all boards out.

NEVER put a plug in what ever electronic that long time been out of use.
ElektroCaps have to be reformed other wise you end op with lots of defect equipment.

Its like a totally dead battery those capacitors.
And need to be reformed by time, say half an hour at least.
Reformation is done by power-up and a new oxide layer is formed chemically inside the cap.

Correctly it is done by separately voltage on the big capacitors with a series resistor.
Read about it.

So long time out of use...start up with only the supply and check voltage and ripple-AC.
Simple test 5Vdc with multimeter at position AC and read the AC-voltage
Only some minor mV to be read.
Click to expand...

Please read more carefully. I have stated twice now that the chassis, backplane, power supply and the ram board are good, known good and tested good with an alternate CPU board. It's a fault exclusive to my CPU board, not an external component or peripheral.
 
I mean , you had a working station way back and is now after long time broken.
That's what I mean by the first start up seconds after long time off power, a power peak at that first state could make a lot of trouble.
Especially with old caps.
Then after minutes I can believe the power is again ok.
But damage could be done.
There for after long out off use, first take boards out.
But Ok, to late now. Next time try the method startup sequence .

"..and still could not get my CPU to even reach ODT"
ODT is that basic, if then already halt your KDF11 then I think even the Clock is not running.

Rule outs are
At ODT it does not go look for an Eprom or the contens in it., so that complete circuit is to left alone.
Should start in ODT with even none Eproms present
At ODT serial communication is not the issue because what ever that circuit the LEDs still will go normal

Check with a scope the clock pin on the CPU.
CPU cristal it self is on the far edge of the board, right of the LEDS.
(The one close to the 2 dip40 Serial communications IC is pure for the baudrate those serial lines)
So circuit left CPU,MMu not an issue at this moment.

Could be all kind of trouble with logic.
F.i. It tries to communicate with the RAM board and can not find it.
Due to sent or read issue's.
Time to get a schematic I think.
 
EK-1T23B-OP-001 PDP11-23-PLUS System Manual.pdf

check page 106 -PDF =page 94 book "Microprocessor reconfiguration.
Table 8-1 Power-up mode.

Setup - Console ODT

That way you force at startup to enter ODT without
a eprom check.

Tip:
check Those Dip switches also for really continuity, I had once a problem with
a dipswitch block.
 
So, obviously something broken on your CPU card. Find the schematics. Check power through the board. Check if there is some microfuse on it. Then start chasing various signals. I would probably start with the power OK signals, since getting the CPU into ODT is very basic. If that don't work, the CPU is never starting to run.
 
Hello mondo BuMP.

Since the pre-COVID days I located the drawings for the board and between me and some other folks at the office we suspected a bad bus driver was fouling up the bus, but we stopped short of pulling drivers and checking further, so the entire system sat on a shelf while the world lived through COVID.

Glitch to the rescue. I handed the board to him this past VCF MidWest to look into and he quickly found what the issue was: Bad EPROM.
Why couldn't we drop it to ODT? Halt button wasn't pushed in. Didn't I strap out the EPROM and force ODT? Yes, but somehow I still did something wrong.

I'm both embarrassed and a bit at a loss here. I did an EPROM swap with someone else and the issue remained with my board, so we quickly ruled out bad ROM (An EPROM dump would of verified this but I neither had an image to compare against or a Data I/O 29B that actually wanted to work, so that was a fault on me not having working tools). As for the Halt not working, no idea. I know I tried both with and without the halt button pushed in so I gotta look. Perhaps there's an open line somewhere?
Well he reports it is working now and is currently exercising it to validate the repair.
 
Ran it all day yesterday when trying to figure out The Weird Noise from my RL02 in the MicroPDP-11/83 rack, and then for testing, exercising, and qualifying its replacement, so I think it's good to go!

One EPROM was all 0x00 -- I don't know offhand if these erase to 0x00 or 0xFF but they are late enough I'd assume 0xFF.

You may have swapped with someone who only had the 2K EPROMs, which requires jumper changes. There could also still be a fault in the system unit, either the backplane or power supply. Seriously, it seems like 50% of PDP-11 problems are backplane faults nowadays!

As stated above, you could have a DCOK/ACLO issue from your PSU. If you haven't gone through it, you ought to, as there are three RIFAs in there (one very large and very full of smoke!). Mine also had some leaking capacitors, both low ESR and general purpose, so I did a complete recap (cap list here).
 
Would a PSU fault potentially hit the CE and VPP/OE pins and cause an accidental write to a UV erase EPROM? The only person I know I swapped EPROMs with marked their before we did it as to prevent a mixup. Before that the board came from another member here who had it tested and working as-is before it went into my BA23.
 
I'm not sure, I haven't looked at how those EPROMs work, but I'd suspect there's a requirement for a supervoltage on a regular I/O pin, being as how it's a 24-pin device but holds 8K. If that happened, it'd probably have gotten blown up, or at least damaged something else on the board.
 
Yeah, PSU or backplane -- definitely check voltages first, and give it a cleaning if they're OK. Electrosolve doesn't hurt the backplane.
 
NeXT said:
Would a PSU fault potentially hit the CE and VPP/OE pins and cause an accidental write to a UV erase EPROM? The only person I know I swapped EPROMs with marked their before we did it as to prevent a mixup. Before that the board came from another member here who had it tested and working as-is before it went into my BA23.
Click to expand...
The only way to clear an EPROM is to UV erase it, or blow it up. I don't know if @glitch reprogrammed the part or just put a new one in. I'd expect that anything that would take out one EPROM would get the other as well, since the only difference is high/low byte.

You probably want the KDF11-B3 EPROMs for the most console functionality. I don't remember what the part numbers are offhand (my memory not only never had ECC, but the parity checking apparently failed a decade or two ago). I'm not looking forward to the age where I get a stuck bit...
 
Terri Kennedy said:
You probably want the KDF11-B3 EPROMs for the most console functionality. I don't remember what the part numbers are offhand (my memory not only never had ECC, but the parity checking apparently failed a decade or two ago). I'm not looking forward to the age where I get a stuck bit...
Click to expand...
Some relevant pages:
- - - - -
 
You must log in or register to reply here.
Back
Top