A PDP 11 By Any Other Name: Heathkit H11 Teardown And Repair

[Lee Adamson] is no stranger to classic computers. He recently picked up a Heathkit H11A which, as you might remember, is actually a PDP-11 from DEC. Well, technically, it is an LSI-11 but still. Like a proper LSI-11, the computer uses the DEC QBus. Unlike a lot of computers of its day, the H11 didn’t have a lot of switches and lights, but it did have an amazing software library for its day.

[Lee] takes us through a tour of all the different cards inside the thing. It is amazing when you think of today’s laptop motherboards that pack way more into a much smaller space. He also had to fix the power supply.

We are looking forward to seeing more videos on this computer. We miss the days that your computer broke down into multiple boards plugged into a backplane. Even though the computer is a Heathkit, the CPU board came from DEC assembled. However, Heathkit had its own boards that you did build along with things like power supplies.

The power supply needed some care, as you might expect. A diode wasn’t attached properly but it wasn’t clear if it had been damaged in transit or if it had never been installed correctly. Replacing it put the power supply right and now he’s ready to see if the thing will start up.

There are plenty of ways to emulate a PDP-11 on things like Arduinos. If you want to see what assembly language looked like on this machine, there’s a tutorial.

27 thoughts on “A PDP 11 By Any Other Name: Heathkit H11 Teardown And Repair

  1. I had to make enclosure for six DEC LSI-11 boards and their little tape drive back in the early 1980’s. I don’t recall what the OS was but I’m thinking RT11? Anyway, it took about 2 hours to boot from those tapes. Trouble shooting therefore took a very long time.

    We had some Apple IIs with floppy’s and they ran circles around the LSI-11 setups. The Apples ran DOS3.3 and BASIC and a good assembler, Orca/C, Kyan Pascal and Kyan’s Unixy OS KIX, and a Forth that did not need an OS and had a nice assembler built in. (You could get all that in an Apple II with a 7″ or 9″ monochrome (non-Apple) monitor for the price of a DEC VT100 terminal. And you got some OK graphics and lots of expansion as well.)

    However, I DID personally have the heath terminal that looks like the H8 computer without the floppy. Heathkit had very cool stuff and I can recall the surprised and excited people when this H-11 was announced. It was obsolete before it hit the shelves, but that didn’t matter much at the time since there were so many people who had learned on college PDP-8 and PDP-11 and VAX.

  2. There was an early article in Byte about the LSI-11. And the Southern California Comouting Society had a group buy or two, as reported in Byte.

    People were looking to 16bit almost from the start, maybe they’d experienced it at work already. In the fall of 1975, Godbout had a contest, I think it was to guess the CPU, for their 16 bit CPU. It was the National PACE, but somehow the computer didn’t happen.

    One thing that was in favor of the PDP-11 was DECUS, the user group, or specifically the program library. It was never clear to me how useful that actually was to home users, but going in it was seen as useful.

    Some people had the need, or maybe just the money, to go after the LSI-11. So when the H-11, arrived along with the H-8, it wssn’t unfamiliar, and not a surprise.

      1. Cromemco sold a 68020 on an S-100 card. It came out in the mid eighties and I think it was sold into the early nineties. Many, if not practically all, were bought by the Air Force. That’s what I think I remember, but I don’t trust my memory anymore.

        1. And I’m talking September 1975, first issue of Byte. Not the dawn of home computers, but before most people knew about them.

          Byte for Feb 1976, six issues in, a small bit that SCCS had already gotten fifty buyers into a group buy for the LSI-11. Sometime that year Byte had an article about the LSI-11.

          I know about the TI single board 9900 computer. And some small company had one too. The Alpha Microsystems using the Western Digital WD16. The Intersil 6100, CMOS that also happened to run PDP-8 code (well that was a 12 bit).

          These were before, or concurrent with the Heath H-8. A few years is nothing when you look back, but it’s a long time when looking forward. People bought into what was available, if they had the money and need. They didn’t know what was coming.

          By the time Cromemco and Godbout (and the Seattle Computer Products 8086 S-100 board) came after, things were already shaking out. The earky 16 bit turned out to be also-rans. Useful at the time for a few, but never making big inroads. (Yes, the TI 99/4, but I don’t think anyone used it for serious use.

          That’s the excitement for the H-11 when introduced in a multipage spread with the H8.

          1. The PDP11 instruction set and architecture was like a religious experience for me after all the quirky 8 bit instruction sets. The H8 was too expensive for my student budget though. In later years, I picked up a lot of DEC stuff on the surplus market and learned a lot from it as well as having a lot of fun.

  3. Ah, regrets. This is some serious minicomputer porn, right here. In those times, there were two great minicomputers. One was the PDP-11, and the other was the Data General Nova series. I managed to get a hand-me-down Supernova SC with assorted cards back in the early 80s, but unfortunately, this was way pre-Internet, and I couldn’t get all the documentation needed to seriously get this to work. The H11 was an amazing accomplishment, and I’m glad Lee got ahold of enough parts to make this one work. It’s close to incredible, to me, that a company like DEC would collaborate with Heathkit on a computer for the hacker/hobbyist, but here’s the proof.

    1. Re: Data General. I worked there as an engineer from 1978 through 1993. One of the annoyances was DG’s preoccupation with obscuring their IC part numbers by using company part numbers. We engineers finally collected our own cross reference files because schematics had to be done using DG part numbers instead of industry standard numbers. I still remember some of the common ones: a 100-799 was a 74LS00 and a 100-1253 was a 74LS244. This was all done so third party repair houses couldn’t repair broken boards or service DG computers. Service was a huge part of DG’s revenue. DEC wasn’t as paranoid…

    2. Check out Tracy Kidder’s book “Soul of a New Machine” for the story of Data General’s effort to develop a competitor to DEC’s VAX system. The result was the Eclipse MV/8000.

      1. …And that system running AOS/VS was a *dream* to work with. I read the book in one sitting in the weeks before our company began receiving our first MV/8000 systems, and the reality was better than the anticipation.

        The fact that AOS (and later AOS/VS) provided RDOS compatible “multitasking” was wonderful. It took decades before Linux supported native threads well enough to catch up to where we were in 1981/1982. (Maybe SunOS / Solaris had a competent thread implementation too, though I never worked on Sun equipment.)

        We were running a ton of IDEA data entry terminals and dabase lookup on small service code table entries was killing our performance. With less than 1000 lines of PL/I in the server side and simple modifications on the data entry side we were able to use interprocess communication to deliver cached data from those little tables and nearly doubled the number of data entry terminals we could support per system.

    3. The LSI-11 was developed as a DEC product for sale to anyone that wanted to buy it. They likely had app notes available and provided Heath with the support they would give any potential volume customer.

    4. DEC were very, very much open to people tinkering with their products. I have DEC’s ‘Digital Catalog Sales Fall 1978’ catalog of what they offered and there is loads and loads of items for (their words) “do-it-yourselfers”. Complete systems, peripherals, interfaces, Flip-chips, backplanes, mounting brackets, panels, Unibus/Qbus edge connector blocks, Unibus IC Drivers (ten to a pack: $26), blank EPROMS, DECkit, adapters, cables, clips, patch cords, tools, breadboards, extender boards, cabinets, power supplies, distribution panels, disk packs, LINCtape, paper tape, fan-fold, tables, stands, shelves, cupboards, racks, teleprinter paper for ASR33, ribbons, software, handbooks, logic books, reference manuals the list goes on, all available by mail order.
      And yes I am scanning this and will put it up when I get it finished. 179 pages of it.
      No wonder DEC was so popular with universities, it was one of the most open computer architectures ever made available.

  4. It’s always interesting to see inside these machines, also, this guys voice is oddly reminiscent of Leonard Nimoy, at least at times.
    The architectures of the period can often seem strange, but there was usually a good reason for the chosen design, I wonder if people will look back forty years from now and think of our current chip architectures as being as unconventional. Who knows what we’ll have by then.

    1. Google the MINC-11. It has A/D to D/A converters along with parallel I/O. It was the successor to the LINC-11. DEC had a niche in laboratory computing from the 1960s. I often miss my MINC-11…

  5. I worked on various DEC PDP-11 systems in the 80s. For one job, I did my work on my own PDP-11/23 and would deliver it to my boss who had an H-11 of his own for his work. He had an RL02 drive (10MB) on his, and I had an RL01 drive (5MB) on mine. At the time, the 22-bit single card RLV12 controller was well over $800, so we both had the older 18-bit 2-card RLV11 controller. Trick is, that’s a pair of cards that have to go into “CD interconnect slots” (mentioned in the video). Those slots are not Qbus at all, they are wired so that most of the pins on the bottom of one card are connected to the top of the next card, rather than running a cable from card to card.

    Some DEC backplanes only contain Qbus slots and others contain Qbus and CD interconnect slots. I got a BA11-N box for my PDP-11 that is 9 slots and is Qbus down the left and CD down the right. That’s how the slots got their name… Each of the 4 backplane segments has a name for identifying specific signals, and the convention going back to the 60s is it starts with “A”, so Qbus fingers are named A, B, C, and D. In my box, A and B were Qbus and C and D were the CD-interconnect slots.

    The H-11, as mentioned, has no CD-interconnect slots. There are slots named C and D, but they are wired as Qbus slots (same as for most DEC backplanes). I mentioned that my boss was using an RLV11 in his H-11. What he did was to unscrew the side panel from the H-11 box and extend it on threaded rod, then he installed the RLV11 in the middle two slots of the backplane, but half-out of the cage, with the C and D fingers hanging out in free space. He then took four Heath backplane connectors and wired up two 2-slot jumpers the same as the CD-interconnect and stuck them on the CD slots of the RLV11. It was ugly but it worked.

    Later, long after the company folded, my old boss gave me his H-11. I reversed his mods and returned it to a floppy-based system. I now have an RLV12 I could use to hook the RL02 back up to it. For a 1980s minicomputer, the RL02 was a fantastic drive. Not the fastest, but a lot less fiddly than the RK05s before it.

  6. I really wish the guy that owns whats left of Heathkit would just release all the schematics, parts lists, and circuit board artwork into the public domain. It would be great fun to reproduce a lot of these kits. They are not really doing much with the intellectual property at all which is a shame.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.