MicroPDP-11/83 H960 Rackup and Such

[glitch]

Around the 4 JUL long weekend, I stopped by the old Sontec/ITI Audio building to check on a power outage situation for Burgess, and also pick up a few things. I decided to bring back the old DEC H960 rack that had lived next to my workbench since college (around 2007). Jack Rubin was right, they do fit perfectly in the back of a Volvo station wagon with just the back seats down:



Does anyone know what this power distribution block is for:



It's been in there since Burgess bought the rack as surplus from Virginia Tech, decades ago. He said the racks were empty except for the power distribution blocks and some top-of-rack fans.

The H960 was filthy, so I removed all of the hardware, spring nuts, etc. and pressure washed it:



That removed most of the accumulated grime, and knocked some loose paint off. There were numerous rust spots (the Sontec/ITI building has no air conditioning), so I decided to wire wheel the rack and repaint it. This is not some restoration job, just maintenance work to prevent further rusting and make the rack serviceable. It got primed with Rustoleum rusty metal primer (my favorite for work like this), and a first coat of Rustoleum professional semi-gloss black:



I wasn't super happy with the look of the semi-gloss, especially since it showed all of the paint flakes really bad, so I finished it with Rustoleum professional matte black. After drying for a few hours in the July sun, it got placed under the deck to finish curing and stop smelling like spraypaint:

 

[glitch]

Today, I finally had time to run out to the farm with the station wagon and bring the H960 back to the building. The goal is to rack the MicroPDP-11/83 I bought from Kipp Yeakel in 2016 (VCFed forums thread about Kipp having bought the PDP-11/83, thread from which I purchased it in 2016). I hadn't done much with the MicroPDP-11/83 until this year: when I bought it, I had *two* MicroPDP-11/73s, one of which I ran regularly, the other was sold to TangentDelta when I picked up the 11/83. I had no pressing need to run the 11/83, until I ended up selling my remaining 11/73 to techfury90 so he wouldn't have to start from zero on getting a PDP-11 capable of running 2.11BSD. I figured that would make me get the 11/83 up and going!

Well, it took until earlier this summer for me to actually get motivated to pull the MicroPDP-11/83 out and get it cleaned up. Like all BA23s, I knew it had three RIFA capacitors inside just waiting to fill my shop with smoke. I also knew if I was taking the PSU apart to replace the RIFAs, they were going to get a full recap. That made the project a lot more involved. When those Clearpoint QED1 RAM boards showed up, I decided I really did need the 11/83 operational, so that I could test them and see if I needed to return either for being defective. I didn't have any other PMI-capable systems, so it'd have to be the 11/83!

As usual, that turned out to be its own yak shave and I ended up having to repair the KDJ11-B due to a cache failure. But today, the machine was getting racked! I had the rails for double-mounting two BA23s from Kipp, and got them in the H960 first:



You start with this bracket, on the right side of the RFI shield mount. The RFI shield assembly swings out and lifts off this bracket, and the side rails unbolt. After I had that bracket installed, I went ahead and did the associated rails. Then I did the (looking from the back) left side:



Don't forget to put two spring nuts in the positions where the RFI shield's frame secures to keep it from swinging sideways -- they're the two middle holes on the left bracket in the above picture. Side rails mount the same. After that, the BA23s slide in from the front. I found it easier to close the RFI shield frame and secure it to the rack before sliding the BA23s in. Here they are, mounted:



Remember to remove the bottom rack mounting screws from the front of the side rail brackets before sliding the BA23 in -- that screw retains the BA23, and is hard to get out if it is already slid in.

For now, the H960 with the MicroPDP-11/83 is living in the front assembly/repair area, next to the small Steelcase desk it had been sitting on while I did repairs on the KDJ11-B and other QBus boards from that big lot:



The top of the desk is currently cluttered with paperwork and misc. QBus boards from having to move things to get the H960 rack in place!

This rack, and two other H960s, will eventually go in the new DEC lab we're setting up in one of the small offices at the building. I'm still working on getting the office in shape before I move hardware in: lights are on, convenience outlets are wired, but I still need to mount some Wiremold 4000 series raceway and run the 3/4 EMT to it. The raceway will carry 3x 120V @ 30A circuits terminated on Twistlok receptacles for the DEC PDUs I have for the three H960s, and a 120V @ 20A circuit for stuff like the corporate cab system I picked up from Colorado in December 2018.

I'll probably rack one of my spare (e.g. doesn't go with a specific system) RL02s under the MicroPDP-11/83 to make it easy to image and load packs.

 

[czunit]

Nice! The distribution panel there reminds me of what one sees on older pdp8's.

As for the BA23's, looks pretty good as well. I think I have a BA23 rack front panel that I'd rather have a BA23 pedestal cover in exchange.

 

[glitch]

Nice! The distribution panel there reminds me of what one sees on older pdp8's.

That's what I was thinking, but I don't have a lot of PDP-8 experience.

As for the BA23's, looks pretty good as well. I think I have a BA23 rack front panel that I'd rather have a BA23 pedestal cover in exchange.

I think I have the covers for these, but boxed them away so they wouldn't get damaged. I don't have any BA23 pedestal covers or I'd swap you for one I generally prefer rackmount.

 

[wa2flq]

Pretty. Just make sure you secure them against tip-over if you extend one or more chassis. See H952 or H9514 stabilizer feet as needed for your rack. https://www.1000bit.it/js/web/viewer.html?file=/ad/bro/digital/dec-cabinets.pdf

 

[m_thompson]

Nice! The distribution panel there reminds me of what one sees on older pdp8's.

The distribution panel has both fixed and variable +10 and -15V, so it goes with a transistor-only early PDP. The PDP-9 distributes both fixed and variable +10 and -15V to I/O cabinets through a Margin Cable, so maybe that is where it came from?

 

[glitch]

Pretty. Just make sure you secure them against tip-over if you extend one or more chassis. See H952 or H9514 stabilizer feet as needed for your rack. https://www.1000bit.it/js/web/viewer.html?file=/ad/bro/digital/dec-cabinets.pdf

I have the stabilizer feet for one of my H960s, and the pull-out stabilizer for the corporate cab, but I need more. Are these something that's available? I know the screw-in rubber tipped leveling feet are, but haven't seen the brackets. If they're not, I could have more made up.

 

[glitch]

This morning I thought I'd try and get more gear into the rack and out of storage, so I started with a RL02. I picked up these rails from Herb Johnson after VCF East:



They'd clearly seen moisture at some point, but cleaned up fine! They roll nicely now, after wire brushing I removed the old grease with WD-40, and oiled with some DTE Light. I will grease them at some point, once they've run in a little. They will of course get Lucas Red n Tacky Both rails need new rubber detent rollers, the originals turned to goo. This is the drive I want to rack:



Unfortunately, someone drilled out the pop rivets and removed the right-side inner rail. I do not know why -- this RL came with the DEC Datasystem (originally a PDP-11/23, upgraded to 11/53) I picked up from Colorado in December 2018. This RL had lived on a shelf above the corporate cab the PDP-11 was in, so I guess it didn't matter that a rail was gone...but the left rail is still there! There was also some yellow tape on that side (again, don't know why) that was shedding dried out adhesive really badly, so that was the first thing to come off.



It's dirty, but not filthy. I sprayed Windex on top and cleaned with an old toothbrush, which gets the grime out of textured plastic. I figure anything I clean off is garbage that won't end up in the pack. Here's the front, after the top got cleaned:



Notice the unit #2 plug in the front -- #0 and #1 were of course in the corporate cab. Next I opened up the drive for cleaning and inspection. Don't forget to clean the little groove between the pack well lid and the back:



It holds filth well! With the top removed, it's easy to clean the pack well and electronics:



I added a couple drops of DTE Light to the blower motor's rear bearing, too. I did of course lock the heads before the equipment made its journey from CO to VA:



Everything cleaned up, power applied, lamps all good:



I needed cables and didn't want to steal them from another system, so I decided to deal with this filthy bunch:



The above came with my PDP-11/10, from the dirty basement of an old mansion on a hill in Malden, MA. They were in a heap in the floor with other random cables. RetroHacker_ talked me into saving them, and I'm certainly glad he did! There's three RL cables, a bulkhead bracket, a transition connector, and the flat cable in there. I wondered why there were three, as the system only had two RL01 drives with it...turns out one is bad! I pulled a pack off the shelf that I believe I bought from Chris Zach at a System Source workshop. The label said it contained RSX11M 4 generated for a PDP-11/44, and it's a -EF error free pack (or at least was back then). I cleaned it as we did with RetroHacker_'s packs and mounted it, then discovered the bad cable After replacing the cable, and with the WRITE PROTECT switch in, it booted:



Currently it is being dumped at 38400 bps using vtserver:

 

[m_thompson]

So you did remember to unlock the heads...

 

[czunit]

*snort* Though I don't think I want to know what would happen if the heads rammed that plate....

 

[glitch]

So you did remember to unlock the heads...

Haha, yes -- you can see them unlocked in the lamp test picture! I check on any drive when I wipe out the pack well, which I do anytime they've sat.

*snort* Though I don't think I want to know what would happen if the heads rammed that plate....

Probably bend it right out of the way, damaging or jamming the heads in the process! Bad times, I'm sure.

 

[DDS]

This morning I thought I'd try and get more gear into the rack and out of storage, so I started with a RL02. I picked up these rails from Herb Johnson after VCF East:



They'd clearly seen moisture at some point, but cleaned up fine! They roll nicely now, after wire brushing I removed the old grease with WD-40, and oiled with some DTE Light. I will grease them at some point, once they've run in a little. They will of course get Lucas Red n Tacky Both rails need new rubber detent rollers, the originals turned to goo. This is the drive I want to rack:



Unfortunately, someone drilled out the pop rivets and removed the right-side inner rail. I do not know why -- this RL came with the DEC Datasystem (originally a PDP-11/23, upgraded to 11/53) I picked up from Colorado in December 2018. This RL had lived on a shelf above the corporate cab the PDP-11 was in, so I guess it didn't matter that a rail was gone...but the left rail is still there! There was also some yellow tape on that side (again, don't know why) that was shedding dried out adhesive really badly, so that was the first thing to come off.



It's dirty, but not filthy. I sprayed Windex on top and cleaned with an old toothbrush, which gets the grime out of textured plastic. I figure anything I clean off is garbage that won't end up in the pack. Here's the front, after the top got cleaned:



Notice the unit #2 plug in the front -- #0 and #1 were of course in the corporate cab. Next I opened up the drive for cleaning and inspection. Don't forget to clean the little groove between the pack well lid and the back:



It holds filth well! With the top removed, it's easy to clean the pack well and electronics:



I added a couple drops of DTE Light to the blower motor's rear bearing, too. I did of course lock the heads before the equipment made its journey from CO to VA:



Everything cleaned up, power applied, lamps all good:



I needed cables and didn't want to steal them from another system, so I decided to deal with this filthy bunch:



The above came with my PDP-11/10, from the dirty basement of an old mansion on a hill in Malden, MA. They were in a heap in the floor with other random cables. RetroHacker_ talked me into saving them, and I'm certainly glad he did! There's three RL cables, a bulkhead bracket, a transition connector, and the flat cable in there. I wondered why there were three, as the system only had two RL01 drives with it...turns out one is bad! I pulled a pack off the shelf that I believe I bought from Chris Zach at a System Source workshop. The label said it contained RSX11M 4 generated for a PDP-11/44, and it's a -EF error free pack (or at least was back then). I cleaned it as we did with RetroHacker_'s packs and mounted it, then discovered the bad cable After replacing the cable, and with the WRITE PROTECT switch in, it booted:



Currently it is being dumped at 38400 bps using vtserver:

In one of the above pictures one can see the right angle brackets I mentioned in a post that has since disappeared. You can see that the rail on the right has both brackets attached while the rail on the left has one attached and one laying more or less loose. Those brackets often don't come from a reseller if you buy a set of rails but they're not magic and replacements can be cobbled together rather easily.

That "missing" post was actually on another thread about adjusting for differences in rack depth.. My bad! I'm told that there are good things associated with being 75 years of age. If I find out what any of them are I'll be sure to let you all know.

 

[glitch]

In one of the above pictures one can see the right angle brackets I mentioned in a post that has since disappeared. You can see that the rail on the right has both brackets attached while the rail on the left has one attached and one laying more or less loose. Those brackets often don't come from a reseller if you buy a set of rails but they're not magic and replacements can be cobbled together rather easily.

Indeed, they would be the logical place to extend or shorten to a non-DEC rack post depth! The loose ones came that way, there's only one screw holding the bracket in on those, but they're just #8-32 so I figure once I set the proper depth I'll lock them down with something from the 1000 year drawer.

That "missing" post was actually on another thread about adjusting for differences in rack depth.. My bad! I'm told that there are good things associated with being 75 years of age. If I find out what any of them are I'll be sure to let you all know.

My dad says the senior coffee is the best thing he's found so far

 

[glitch]

The bad cable came back to life with some DeOxit -- I figured with its previous storage conditions, there might be something gross on there that was not letting a pin make contact. I imagine wiping force is really low on these connectors. Currently running the performance exerciser as I don't trust it yet

 

[glitch]

Three hours of testing with no issues -- must've just been a dirty connector!

 

[Terri Kennedy]

The bad cable came back to life with some DeOxit -- I figured with its previous storage conditions, there might be something gross on there that was not letting a pin make contact. I imagine wiping force is really low on these connectors. Currently running the performance exerciser as I don't trust it yet

I brought a VAX 4000-705A to the LSSM last week after months of testing*. Of course, the AUI/Thinwire switch decided to act up. A complete disassembly of the bulkhead to get to the PCB, along with DeOxit / board cleaner / flip switch multiple times / lather / rinse / repeat got it sorted out. I did bring a complete spare bulkhead, just in case.

* The system performed flawlessly the whole time I was testing it, except during one of the coronal mass ejections - an RF73 behaved as if it's ready switch was being pressed randomly by a user - mount verification but no device errors. Turned out that the PAL in the OCP was particularly sensitive to solar particles. A quick swap and all was well.

 

[Terri Kennedy]

I needed cables and didn't want to steal them from another system, so I decided to deal with this filthy bunch:

I have a mystery RL11 controller-to-drive cable that doesn't use a ribbon cable to a bulkhead connector - it is a direct drive to controller cable. Very handy when you need to stick an RL11 into a system so you can run XXDP from an RL02 on a rolling cart. I don't know what it was originally used for, but tomorrow I can look and see if it has a BCxx or 2-5-2 part number.

 

[glitch]

I brought a VAX 4000-705A to the LSSM last week after months of testing*. Of course, the AUI/Thinwire switch decided to act up. A complete disassembly of the bulkhead to get to the PCB, along with DeOxit / board cleaner / flip switch multiple times / lather / rinse / repeat got it sorted out. I did bring a complete spare bulkhead, just in case.

I've had a *lot* of DIP switches go like that, many of them Grayhill, oddly enough. Often they do come back with some DeOxit, but if it's an 8-position usually we just replace, since we stock them.

* The system performed flawlessly the whole time I was testing it, except during one of the coronal mass ejections - an RF73 behaved as if it's ready switch was being pressed randomly by a user - mount verification but no device errors. Turned out that the PAL in the OCP was particularly sensitive to solar particles. A quick swap and all was well.

Wow! I have not had any failures that I could directly relate to solar events. That's a pretty neat thing to get to observe, if not annoying. I wonder if my TEMPEST shielded VS3100 would've been affected

I have a mystery RL11 controller-to-drive cable that doesn't use a ribbon cable to a bulkhead connector - it is a direct drive to controller cable. Very handy when you need to stick an RL11 into a system so you can run XXDP from an RL02 on a rolling cart. I don't know what it was originally used for, but tomorrow I can look and see if it has a BCxx or 2-5-2 part number.

Probably a BC80J or BC80M -- Mini-PV Berg connector on one end, RL connector on the other. IIRC the BC80M has a ground braid connection, and the J suffix doesn't. I've seen them used on smaller machines, but don't have one myself. RetroHacker_ has one on his PDP-11/23+ corporate cab and they are indeed *super* handy for testing!

 

[glitch]

Hardware from McMaster came in yesterday, and in the order were some bits I needed for this rackup. I got the cleaned up RL rails mounted low in the rack:



Protip: use #10-32 square nuts behind the RL rail brackets, and you don't need to get a wrench on them! My corporate cab uses hex nuts, but the H960 I picked up from Andy Malloy in Syracuse years ago used square nuts, which I thought was pretty clever and convenient. Last time RetroHacker_ and I added RL rails to something, we used hex nuts because that's what the hardware store had, and it was a pain to get a wrench in there. I mounted the rails at the same height as the rails in my other H960, so the drives will sit at the same height if they end up next to each other in the DEC lab

Going to pick up a junker RL from czunit probably this weekend, which will provide this missing right inner rail. Then I can get the RL02 off the desktop!

 

[glitch]

Finally got out to pick up the parts unit RL from czunit over the weekend -- first we got sick, then he got sick, so there were delays. His drive appears to be an early RL02 or a RL01 conversion, the SN label actually has "RL01" modified to "RL02" on it with a marker! His unit had rivnuts to attach the inner rails, so removal was super easy. To determine the size of pop rivet I'd need, I pulled my RL02's faceplate and punched out the frontmost stub left from whoever drilled out the rivets.



I'd been wanting to pull the panel so I could clean the cooling air intake filter, but thought I'd wait since I didn't want to disturb the foam filter -- I figured it'd crumble when I did. I was right! This picture is for all of the "my RL drive looks clean on the outside...I'm just going to run it!" folks:



I replaced it with some filter foam from the hardware store that is intended to replace the filters in window air conditioners. It's about the right thickness and density, and was a little more than $3 for a huge sheet that would probably be enough for 20 RL drives! While I had it opened, I swabbed the air-to-air heat exchanger tubes:



Pretty dirty in there! I also noticed that the sealant which helps maintain the separation between HEPA filtered pack air was cracked in places:





I resealed it with 3M Polygun hot melt adhesive, which is the industrial version of hot snot. Finally, with the drive reassembled, I pop riveted the inner rail from czunit's drive on and racked it:



The correct pop rivet diameter is 5/32", medium length works well, but I also tried a 1/8" pop rivet to see how it would work, since they're more common at the hardware store. 1/8" does work, but 5/32" is a much better fit. Not going to drill out the 1/8" rivet to change it though!