• Please review our updated Terms and Rules here

What did I do to my PDP-8 today.

Well that was easier than I expected. Let me back up a bit.

My 8/e had a problem with the second (rear) backplane. When I put cards back there. the memory was not working correctly.

I had been working on an idea for making a diagnostic image for console serial disk and finally settled on how I wanted to do it. I made the image and put D0AB.PA on it and then used PAL8 to make a listing and binary for it. D0AB is also known as Instruction Test 1. It tests AND, MQ register functions (not the EAE), and the basic OPR instructions and combos. It is a paranoid program that tries to only use stuff it has already tested. It uses conditional skips over HLT instructions and I think there is only a single JMP in the whole program at the end to get back to the start in order to do the next pass. About the only instruction that gets tested only once is the HLT instruction. If that doesn't work then you would never get an indication of an issue. So now I have a test I can easily run and have a high level of confidence that the machine is working.

Previously, I moved the bus loads card (M8320) to the last slot. The second 8k core cards in the next 3 slots. The first 8k core cards in the three slots in front of that. The RF Shield card (M849) in the slot in front of that. The RK8E controller set goes in next with the M8655 console interface before that. With this configuration there are 7 open slots between the M935's which connect the front and rear backplanes. And one bit of memory could not be written. I think it was AC4 and that is probably mentioned in a previous posting.

This time I decided to move one function at a time to the rear backplane. I also had decided to leave space for two more 8k core sets (6 slots). Move the bus loads and test, no problem. Move the 2nd 8k of core and no problem. Move the 1st 8k of core and no problems found. Moved the RF shield card and no problems (that one would have been weird to cause a problem). Moved the RK8E controller three board set and again no problems. Finally I moved the M8655 and again no problems.

So it seems to work now. What did I do? I did pull the backplane out of the machine to inspect the backside but found no issues. This did have the side effect of the power cable getting reseated. Nothing else except put cards in slightly different places. And I would put it all back in the way I had it when it was failing but as anyone who has slotted cards in these machines, especially ones with lots of the top blocks knows, your hands will hurt, ,even if you don't suffer from arthritis. Which means that the problem is still there somewhere but that is an issue for future me to deal with when he gets more core memory to install. Or when the problem returns. Note I didn't say if it returns.

Next I am going to put the most useful tests on my 8/e specific CSD diagnostic image and make sure the machine is really healthy. INS test 1 is pretty good but it does not test the most complicated instruction in the machine which is JMS. It doesn't test ISZ or JMP and it does not test memory. I have two sets of EAE boards, one of which does work but I don't leave it in the machine because mostly it just adds additional load to the power supply and unless the program is written to use it does nothing. And I am not a writer for Byte magazine who inserted an 8087 into his S100 8088 CPU card and said that afterword the machine just felt faster, even though he had no software that used it.

Speaking of building software, D0AB.PA is a big program, approximately 64k of source code. Originally it was multiple paper tapes. If I assemble it with palbart you don't notice anything. It is just done. I like to build stuff on the real hardware when I distribute it so I wanted to do that. I put the source on device CSD2: and built it there with a listing file and cross reference and I noticed it was not done after 15 minutes. I walked over and looked at the blinkin lights and they were a blinkin! So it was still running. It took about half an hour to assemble over the Console Serial Disk. I moved the source over to the RK05 and the same assembly there took about a minute. Compared the binaries and they are the same.

I know VCF East is coming up this weekend. I wish I was going but I have another event to attend unrelated to all of our favorite Dusty Old Computers. Have fun and I hope to catch you there next year!
 
DougIngraham said:
So it seems to work now. What did I do? I did pull the backplane out of the machine to inspect the backside but found no issues. This did have the side effect of the power cable getting reseated. Nothing else except put cards in slightly different places. And I would put it all back in the way I had it when it was failing but as anyone who has slotted cards in these machines, especially ones with lots of the top blocks knows, your hands will hurt, ,even if you don't suffer from arthritis. Which means that the problem is still there somewhere but that is an issue for future me to deal with when he gets more core memory to install. Or when the problem returns. Note I didn't say if it returns.
Click to expand...
Hose down the backplane and power connectors with DeOxit.
 
m_thompson said:
Hose down the backplane and power connectors with DeOxit.
Click to expand...
That seems somewhat indiscriminate. I am thinking there might be one pin on one slot that is not making contact. I will go after it when it shows up again. Of course if you have a bad contact it is possible to work around it. The core plane board essentially has no bus connections except for ground. If that one went into a bad slot there would be no issues.
 
Hello all! This is a post about what I did to my PiDP-8, so I hope emulators are okay here.

I've written a "shut the box" game that just uses the front panel switches and lights for gameplay. I figured it would be a fun way to show off the kit to people and take advantage of actually having the physical switches and blinkenlights.

If anyone gets bored and wants to try it out on a real PDP-8, I would absolutely love to know how it goes. (Assuming it goes at all, that is.)

Here's the github repo: https://github.com/JeffJetton/pdp8-shut-the-box

Thanks!

- Jeff
 
JeffJetton said:
Hello all! This is a post about what I did to my PiDP-8, so I hope emulators are okay here.
Click to expand...
I have no issues with emulators. I take my PiDP-8 along with my 8/e to VCF shows just in case the big iron becomes reluctant from the travel.
JeffJetton said:
I've written a "shut the box" game that just uses the front panel switches and lights for gameplay. I figured it would be a fun way to show off the kit to people and take advantage of actually having the physical switches and blinkenlights.
Click to expand...
I am on the road but I will give it a try when I get home! Thanks for making it public.
 
Today, after a loooong story, I was able to assemble this rack.

1777758312940.jpeg
Well, thats the backside. Ther are very heavy transformers inside, around 150kg all powersupplys.
1777758503351.jpeg
So that lead to some problems with the wooden rolling platform.

1777758543700.jpeg
I had to repair some wooden parts and set two rolls more to balance the weight.

And here is the beauty:

1777758617906.jpeg

A very nice 8/I with Fast Papertape/Punch. The rest of the equipment will be fun to analyse, what it is and what it did. But lots of cables are cut and you see, there are 4 instruments missing.
 
I traveled to the Pacific North West to visit the VCFPNW show. Vince lives out here and asked me to assist with getting one of his 8/L's running so that is what we have been working on pretty much the last 3 days.

Vince's 8/L is serial #3661. It has 4k. It is probably like most 8/L CPU's out there. The first thing we did was reform the caps. We were going to use a Variac since these do not have ferro resonant transformers. But the variac did not work. And a quick inspection of its guts did not shed any light on why since everything seemed to be connected up properly and in good working order. We put the variac away and connected a bench supply to one cap at a time and did it the slow way. They all behaved themselves so it was time to put power to the supply and see if the 5V regulator would regulate. With the bench supply connected to the +8 regulator input we were able to see that the regulator was regulating with no load on the output. Put it all back together and flip the switch. We have fans and lights. And it looks like it can sort of examine and deposit. A bunch of the bulbs were not working so that became the next thing to fix. Pulled the front panel and discovered that there were quite a few bulbs missing, the wrong bulbs installed and foils and pads missing and damaged. This was fixable but going to be a lot of work. Vince does have a second 8/L but it is in quite a bit worse condition. I pulled the front panel off of it and found that at some point this panel had LEDs installed. There are other issues but it is a better front panel from an operational standpoint so we used that for the rest of the debugging and bulb and PCB repair on the front panel are a problem for future Vince. Next issue was that we could not deposit to any address ending in 7. We looked at the schematic and determined which board contained that decoder and swapped it with its neighbor but this changed nothing. went back and decided it could also be a diode in the memory. I decided to swap a couple more boards and see if we could move the problem but instead this cleared up the issue. Put all the cards back and it still worked fine. I did wiggle the core plane in its socket and I suspect this is what resolved the problem. It may have simply needed to be fully inserted. We then toggled in a few simple tests and couldn't find any issues. The next step was to get the console port working. The existing interface is 110baud current loop. Fortunately Vince has the RS-232 interface and high speed baud rate generator replacements he designed and he spent half a day and built two sets of these boards rather than steal a known working set from the PDP-12. After building those Vince built a 30ft BC01V cable with a DE9 connector. Time to debug. We had problems with the cable that were not helped by my plugging in the Berg connector in backwards. We had problems with a transistor installed backwards. We eventually got most of this sorted out and then discovered we had blown the 5V fuse. We are not sure how this happened but it did and after replacing the fuse we were able to have the 8/L echo characters back to the terminal program. Vince then toggled in the CSD boot code and PQS8 booted right up. With Vince's boards it is moving a jumper to switch to 115200 baud and almost as easy on the server side so we did that and that is where things stand.


IMG_20260505_161321594_HDR[1].jpg

LEDs simply do not look right on this machine!
 
This morning I wrote a tiny program to do the Battlestar Galactica Cylon eyes in the AC on the 8/L while Vince is working on repairing the lamps on the front panel. Thin of it as a screen saver for Vintage Computers. I took the opportunity to play with PQS8 and typed it in and ran it on the L using real hardware. I have to say that I would have been satisfied with setup as this in an educational setting back when I was learning to program. Here is the first cut although I was lazy and didn't rekey all the comments or formatting when I typed it into the L.

Code: 
/ PROGRAM TO WALK A BIT ACROSS THE AC IN A FASHION SIMILAR TO THE BATTLESTAR GALACTICA
/ TOASTER MODELS EYE DISPLAY.

TIMER=4000        / HIGHER IS FASTER

    *200

START,    CLA CLL CML    / INITIAL PATTERN FOR THE AC
LOOPR,    JMS COMMON    / DO ALL THE COMMON STUFF
    RAR        / MOVE BIT RIGHT
    SNL        / ARE WE OFF THE END OF HE WORLD?
    JMP LOOPR    / NO, CONTINUE
    RTL        / PUT BIT IN CORRECT PLACE

LOOPL,    JMS COMMON    / GO DO ALL THE COMMON STUFF
    RAL        / MOVE BIT LEFT
    SNL        / ARE WE OFF THE END OF THE WORLD?
    JMP LOOPL    / NO, CONTINUE
    RTR        / PUT BIT IN CORRECT PLACE
    JMP LOOPR

COMMON,    .-.
LOOP,    KSF        / WHEN A KEY IS PRESSED, RETURN TO OS/8
    SKP
    JMP EXIT
    ISZ CNTR    / IS IT TIME TO MOVE THE BIT?
    JMP LOOP    / NO
    DCA SAVEAC    / SAVE AC
    CLA OSR        / RESET THE TIMER COUNTER FROM SWITCHES
    DCA CNTR
    TAD SAVEAC    / RESTORE AC
    JMP I COMMON    / RETURN


EXIT,    KCC        / DISCARD THE KEYSTROKE
    JMP I K7600    / RETURN TO OS/8

/VARIABLES
CNTR,    7777        / LOOP COUNTER
SAVEAC,    0

/CONSTANTS
K7600,    7600        / OS ENTRY POINT.

    $
This runs slightly too fast so I kludged in a slowdown when I ran it. I will probably turn this into a read the console keyboard routine at some point just for fun. Feel free to use it! It runs until you hit a key and then returns to the OS. Compatible with OS/8, PS/8, and P?S8. The switch register controls the speed.

Vince has pulled all the dead bulbs on the L front panel and is soldering in flush sockets to make future replacements a non-issue. Suffice to say this is quite a lot of work but will be worth it.
 
Here is Vince's 8/L with the front panel with bulbs, not those ickyLED's.

IMG_20260508_175212529_HDR[1].jpg

The machine is running the program I posted above and the AC4 lamp is the only one fully lit up. The bit is moving right and you can see that AC3 and AC2 are still cooling off. To human eyes most of the other bulbs are more yellow than the red the camera shows.

We are close to buttoning it up. Here is a photo of the card side which is the bottom on the machine.
IMG_20260506_174223831_HDR[1].jpg

There does not seem to be any retainer mechanism so what kept the cards from eventually vibrating out of their slots? There was foam originally attached to the bottom cover. But it would have had to be over an inch thick to touch the cards.
 
Hi Doug,

It was foam holding the cards. The foam was still in my PDP8/L when I got it
Roland Huisman

Thread 'ITT Satellite navigation system 6000 (PDP8/L)'

So finally I got enough money to buy a second hand satellite navigation system. Somehow I'm
a bit worried about map updates... ;) But also the satellite is probably not working any more.

PDP8L - 01.jpg PDP8L - 02.jpg PDP8L - 03.jpg PDP8L - 04.jpg PDP8L - 05.jpg

I got this PDP8/L system from a man who installed and maintained these kind of systems.
When the system became obsolete he took one of these systems home. The idea was
to use it for his miniature trains. That never happened so it was waiting for 40 years in the attic.
So after...

The PDP11/20 was build with the same 'upside down' construction. And there are a few more common things between the 8/L and 11/20 like the wood around the bulbs.
Roland Huisman

Thread 'PDP11/20 restoring'

Hi,

I just started with tiny baby steps to restore a PDP11/20. It is the somewhat newer model with the black case. Once meant for parts, now this Dutch idiot is trying to make it complete and working again. Probably a long way to go... :neutral: For what I know now these things are missing.

- top and bottom covers
- lock switch and wiring
- power cable harness
- fan panel and wiring

I hope some people can help me with some pictures from or information about missing parts.

The top and bottom covers are not really a problem to make, but I'm missing details. I would like to have some...

In the demo with the PDP11/20 I wrote that I found an original core stack for it and I have to fix it. Now the demo is almost four years ago and I still have to test and fix that core memory module... Why does time go so fast...? :unsure:

Roland Huisman

Thread 'PDP11/20 demo RK05J, PC05 and TU56'

Okay, don't take this one too serious please :biggrin: I've put my PDP11/20 on YouTube for a demo. One can show programmers stuff, I thought I do something different. I hope you like it. And don't mention the way I close the lower RK05 at the end ;) This is not like DEC explains it in the manual... 😇


I did a topic about restoring the RK05 drives

And A story about the PDP11/20

I didn't do a story on the TC11 and...
  • Like

Regards, Roland
 
Roland Huisman said:
It was foam holding the cards. The foam was still in my PDP8/L when I got it
Click to expand...
We kind of knew that it had to be foam from looking at the residue on the cover. And from your photo it was thick.
Roland Huisman said:
The PDP11/20 was build with the same 'upside down' construction. And there are a few more common things between the 8/L and 11/20 like the wood around the bulbs.
Click to expand...
Vince has a collection of scrap remnants. These are boards where someone cut the gold edge connectors off the boards. One of these is a PDP-11 front panel. Vince managed to get one working bulb off of this board and reused it in the 8/L. I wondered what 11 it came out of since the connectors were on the top and it was a very similar arrangement to the 8/L front panel. Vince thought it was out of an 11/10 or 11/20 model. I could take a photo of it but it is pretty ugly.
Roland Huisman said:
In the demo with the PDP11/20 I wrote that I found an original core stack for it and I have to fix it. Now the demo is almost four years ago and I still have to test and fix that core memory module... Why does time go so fast...? :unsure:
Click to expand...
I mentioned to my Dad once that time seems to pass faster the older you get. He said yes, and it only gets worse. I turned 70 this year and he would have been 96. I am reasonably certain you will get around to fixing it when it becomes important enough to you.

So many projects, not enough time.
 
Another of Vincis's projects is nearly complete. Replacing the power cable in the 8/e that goes to the front 20 slots. Here is the new and old cables. He is installing it as I type this.
IMG_20260509_141139915_HDR[1].jpg

The old cable had excessive voltage drop on the 5V connection. The new wire is 14 gauge which is what the original was supposed to be but the new wire is more flexible due to higher strand count.

The 5V is 4.7 across a cap on the Space Wars board. Pre harness the 5V was 4.4V at the power supply connector. A huge improvement!

And the machine boots from the RK05 emulator.
 
Vince,

I need to replace the power cables in my 8/e. Well, not replace, because I don't actually have any at the moment!

Can you tell me what part number you have used for the connector shell and pins please?

I need cabkes for both the front and rear slots.

Cheers,

Dave
 
I found the old post which has the part numbers for the pins and housings for the power supply end. I'll try to post later with more detail about the wire and lugs, crimpers, etc. I used in the actual cable construction.

Vince
 
Yesterday I was able to test the RK05 emulator from George (gwiley). I managed to bring it to boot with the Plessey RK8E Conroller and two DEC ones on my Frankenstein 8:

1778869905181.jpeg

That made me happy, because now I can test and repair the controlles I have. And also have a nice litte "harddrive" for exhibitions. One of the next steps will be working on the real RK05 drives.
(But today I sortet FlipChips the whole day).
 
You must log in or register to reply here.
Back
Top