• Please review our updated Terms and Rules here

PDP11/20 restoring

I have photos of the G8000A (lacks the diode, I think)

Hi Vince,

Thanks a lot! Google didn't show that card on your site. And I didn't expect 16 bit related stuff on your site ;) :D Your pcb has also the diode. I have updated my PCB and ordered a few. I refused to add the 'FLIP CHIP' text mirrored on the back of the PCB...

G8000 render final front.jpg

And tonight I have finished my brand new power harness. I used an old 9 slot backplane to keep the right distance between the boards. The GND and +5 are made double so the full current does not have to go trough one single pin any more... So soon I can check my power and also if I have the right voltages on the boards before I power it up with them in the backplane off course.

IMG_20210408_191936.jpg IMG_20210408_215904.jpg Nieuwe kabelboom.jpg

Regards, Roland
 
Tested the power supply and power harness. I made one mistake, I swapped the -15 and -25 on the H720 connector. So I corrected that and the band new cable was ready to use in the machine itself.

The front panel had a rough time once... A resistor broken, and four bulbs were replaced by LED's. When they removed these four bulbs they damaged the pcb vias at this place. You need to be very careful to remove these bulbs. And to use LEDs they removed the pre-heat resistors. Since 25% of the bulbs were broken I decided to replace all of the bulbs with new bulbs. Removed the LEDs and replaced the pre heat resistors.
01 front.jpg 02 front.jpg 03 front .jpg 04 front.jpg

There were several modifications on the backplane. One of them was even a relay to switch a ssyn signal. There were some other modifications done to the backplane. Most of them were for the two SPC slots. I've brought everything back to original, added the wiring for the G8000 boards.
05 backplane.jpg 06 backplane.jpg

I've made a drawing for the fan panel but in the meanwhile I've decided to test with a temporary solution :D
08 fans.jpg

Top cover is ready, when all covers are done I will paint them black as the rest of the cabinet.
07 top cover.jpg

First test learned that bit 6,7 and 8 were always on and bit 9 always off. I fixed that with replacing two 7474, one 7420 and one 7430 on the M725 board.
09 PDP11-20 first power on.jpg

Loading an address works, examine addresses is behaving weird while depositing is counting up the addresses in a good way. Still a lot to do... A little demo from what it is actually doing now...


To be continued...
 
Some people are missing the XX2247 key, but I have the opposite problem.
I'm missing the XX2247 lock itself and all parts with it. I've been searching on the
internet for pictures of the lock construction. It seems that it is the same construction
as from the PDP8/L. So I can copy that construction.

The biggest problem is getting the lock itself. It would be great if someone could actually
sell me such a lock. But I'm afraid that most people would not let go of their parts...

So I've been looking for a tubular lock on Ebay. Most locks have one or two slots to pull back
the key in a certain position. But DEC had three positions off-power-lock. I think I have to live
with that problem. But I was wondering, are these locks also available on the XX2247 spec?
Or am I forced to use just a different key?

Lock.jpg

Regards, Roland
 
I don't have the original KA11 overlay but here's a picture of the overlay from Guy Sotomayor's replica. His overlays match the originals that I have (11/35, 11/45, RK11-D) so I'm confident this will also match the original. Guy's boards use LEDs and machined light baffles, so the physical dimensions of the display board are slightly different from the original but functionality and signals are the same.
1120_overlay.jpg
 
Hi Jack,

Thanks a lot! It seems he still has them on his website. If these fit on the
original KM11 than it might be a good idea to order a set of these overlays...
I only have a set for the 11/05 and the RK11D. The paper protection foil is still on the
RK11D and I have no idea how to get it off without damaging it... The glue is hardened...

The RICM has most of a PDP-11/15 system. I believe that we have all of the boards to make a processor. We don't have the backplanes for the KA11. Any leads would be appreciated.

Well, I wonder if there is any wiring list for the 11/20 backplane like there is for the PDP8/L. I haven't found in yet.
When you have a wiring list you should be able to make your own backplane from three unibus backplanes.

If anyone knows where to find the wiring list then please let me know. Sometimes it is
hard to find a signal on the backplane... So I could use it too...

Regards, Roland
 
Last edited:
Unfortunately, Guy's overlays don't quite line up with the lamps on the original KM11 but you could certainly use one as a basis for a replica to fit the original. Given your talents, it should take you about 5 minutes to scale one up and print it on a transparent sheet. :>)

Not sure why the uploader chose to invert my image, but so be it.
KM11_twice.jpg - Click image for larger version  Name:	KM11_twice.jpg Views:	0 Size:	143.6 KB ID:	1204268
 
I've been given an original lock for the machine. But it had some issues. Two of the rollers were cracked at multiple places.

12 lock broken.jpg

So I made a negative silicone mould in vacuum to prevent air bubbles in the copy.
Then I filled that with a two component casting resin.

13 Making a mould.jpg

Next is to drill the hole for the screw and you are ready to go.

14 copies.jpg

Now the lock is complete and working again. I made a few spares as well.

14 lock fixed.jpg

In the meanwhile I painted the covers. The fan panel is made by a watercutter at the moment.


10 covers black.jpg

All baby steps to get the least original PDP11/20 in the world :wink:

Regards, Roland
 

Attachments

  • 12 lock broken.jpg
    12 lock broken.jpg
    192.2 KB · Views: 3
On the M727 board I found this modification and the backplane is also modified for this modification.
A timer IC is added. When Start is pressed there is an INIT_L generated from the original START_L signal.
START_L had now become a delayed START_L. So now there is always an exact timing between INIT_L and START_L.
But I wonder, is this an official modification or was it also for application specific needs?

Does anyone have documentation about all modifications for these PDP11/20 machines and boards?



start init circuit modificatie.jpg start delay.jpg
 
Most of the early 11/20 Modules were replaced by -YA versions in the KH11-A, including the M727-YA.

Options_and_Modules_List_197312.pdf has useful information regarding module swaps from the KA11 (released to production 1/72) to the KH-11A (released to production 6/73).
 
Most of the early 11/20 Modules were replaced by -YA versions in the KH11-A, including the M727-YA. Options_and_Modules_List_197312.pdf has useful information regarding module swaps from the KA11 (released to production 1/72) to the KH-11A (released to production 6/73).

Interesting. There are no YA board in this machine. And besides the init/start circuit the backplane looks original to me for the CPU part. Any idea where I can find the M727YA schematics? Then I could at least figure out if the modification is DEC or just user specific.
 
I've made a drawing from the right side cooling openings and used that as a reference for the fan panel. The mounting holes for fans are standardised and the size of the panel could not be too difficult. My colleagues had a job for a water cutting company so they could order my panel too.

Click image for larger version  Name:	01 water cutted metal.jpg Views:	0 Size:	92.3 KB ID:	1209097

When I got the panel I've bought a piece of stretch metal and cut the holes for the fan screws.
Click image for larger version  Name:	02 panel bended.jpg Views:	0 Size:	196.1 KB ID:	1209098

The moment to draw the holes on the piano hinge to drill holes at the right places.
Click image for larger version  Name:	03 panel in machine to drill.jpg Views:	0 Size:	186.1 KB ID:	1209099

A bit of painting...
Click image for larger version  Name:	04 painting.jpg Views:	0 Size:	118.5 KB ID:	1209100

And then the fun part, putting everything together.
Click image for larger version  Name:	05 fans mounted.jpg Views:	0 Size:	183.5 KB ID:	1209101

Looks good enough for me...
Click image for larger version  Name:	06 in machine.jpg Views:	0 Size:	197.7 KB ID:	1209102

So now my machine is at least complete again with cloned parts.

Cloned so far:
- top cover
- bottom cover
- switch cams
- fan panel
- power harness including G772B +5 -15 etc...
- mains power harness 115/230
- G8000 filter boards

Regards, Roland
 
@ Doug and Tom, thanks for the compliments!

I don't see anything that looks like this schematic on the M727 documentation. Where did you extract it from?

Oh sorry, I should have clarified my drawing. The part with the switches and 7400 is the front panel. On top is the original front panel schematic. On the lower part is in my machine. The INIT_L and START_L are disconnected from the backplane under front panel, and rerouted trough a few spare pins on the M727 where the 74122 is build on. I assume this will work fine, but I'm curious if this is a DEC designed modification or a customer specific add on.
 
Ah, OK. If this were an incorporated feature in a later -YA module, where would we be looking for it?

Where is the schematic for the front panel in KA11_ProcessorMan_Sep1970.pdf, anyway?

And, no, I don't know where to find the -YA schematics; I'd like to see them as well. All of my key modules (eight of them) are -YA.
 
As written earlier, the first tests learned that address bit 6,7 and 8 were always on and bit 9 always off on the front panel lamps. I fixed that with replacing two 7474, one 7420 and one 7430 on the M725 board. But that wasn't enough to get it back online. To see what is happening on the bus I wanted to have a Unibus probe and ordered one from Joerg Hoppe. Since I don't have a memory module in this machine I was already emulating that with the Unibone. This is all on an external DD11CK panel together with a serial card.

IMG_20210516_130642.jpg
With the 11/20 connected and powered on, it immediately showed that not only the address flip flops were broken but also the bus drives for address bit 6...9. So I had to replace two 8881 bus drivers as well. Now all the bits on the bus behaved normally.
IMG_20210503_194444.jpg

But it was still impossible to let the machine do anything. So I borrowed a board set with the idea to fix both sets. When you have the opportunity to swap boards you can narrow problems down to specific boards. So I removed the user added stuff from my boards and wrapped the backplane back to original. Great these wire lists in the documentation!

Swapping boards learned that my M820 and M725 boards were broken. On the M725 I already replaced 6 chips, but there was another 7402 defective. And on the M820 was a bad DEC8815.

M725 7 IC's defect.jpg M820 1 IC defect.jpg

M725 broken chips:
2x 7474
1x 7420
1x 7430
2x 8881
1x 7402

M820 broken chips:
1x DEC8815

Especially the chips on the M725 were broken in a strange way. Inputs were just short circuited, and outputs just blown out. Another thing that got my attention is that the broken boards were next to each other in the machine. And the broken chips were all connected to finger D. So I think that someone had the machine running in the past and dropped something ugly on the backplane when it was running...

Or is it possible to damage a machine in this way when someone just reversed the unibus cable?
 
Back
Top