An Entire PDP-11 On Your Bench

A PDP-11 at The National Museum Of Computing, Bletchley, UK.
A PDP-11 at The National Museum Of Computing, Bletchley, UK.

The DEC PDP series of minicomputers occupy a special place in computing history for us, because as the workhorses of commercial computing from the 1960s through to some time in the 1990s they provided the bedrock upon which so many of the computing technologies we take for granted today were built. If we think of any PDP, the chances are we’ll be imagining fridge-sized units with panels of blinkenlights that have become iconic in their own right. But that wasn’t the sum of PDP hardware, for at the end of the series of machines there were produced PDP-11s containing what had previously needed those fridge-sized units on a single chip-sized module. [Peter Schranz] had one of these modules, a DCJ11 that he’d salvaged in the 1990s, and he set to with it in making a modern desktop version of a PDP-11.

The PDP-11/hack is a PDP-11 as a set of daughter cards on a lightly modified Q-bus backplane. The DCJ11 and its memory sit on one, an emulated disc controller on another, and finally a multifunction board brings together clock and serial functions. Where the original would have had acres of 74 logic the PDP-11/hack uses more modern CPLDs and microcontrollers to provide glue logic and to emulate now-obsolete components. Given a serial terminal it will boot and run PDP operating systems and software, though it lacks a set of blinkenlights to display its status.

This isn’t the first PDP-11 using this chip we’ve shown you.

25 thoughts on “An Entire PDP-11 On Your Bench

  1. Big deal…. (rolls eyes)… I used to have a Heathkit H-11 minicomputer ‘back-in-the-day’, had the H11-5 serial I/O card going to a DEC VT05 CRT terminal (which I still have – and it’s for sale, taking offers), also had the Heathkit paper tape reader to input the boot loader, programs, blah, blah, blah. I should have kept it… eh, shoulda, woulda, coulda….The H11 was a desktop PDP-11 (ran the same programs. FORTRAN was pretty cool – but this was before the days of the *public* internet).

  2. Let’s not forget DEC’s own PDP-11 on your desktop, the DEC Professional series – I personally had a Pro-350 w/ HD, floppies, and extended bit graphics.

    DEC also had the VT-150 (?) which had a PDP-11 in a VT-100 chassis.

    As others noted there was also the Heathkut H-11.

    The title of “First Desktop PDP-11” was handed out long, long before this hack.

  3. I had a PDP-11 on my desk in 1979. I’m using that desk as I type this. Sadly the PDP-11 is long gone, along with the Altair that sat next to it. It was my dad’s. Technically, I think it was his employers but I think he kept it when they switched to IBM PCs a couple of years later.

    It wasn’t a Dec branded PDP-11, but some kind of clone or more likely a white labeled machine. I used it with RT-11 primarily to play Zork, but also to write some DEC Basic programs.

  4. What a bunch of whiny comments so far… The card used, was from the full size model, that’s the significant part. The hack, was to take an old card lying around, and put it back in use. Who care why, it was a learning experience, fun, gratifying. I have no doubts there are PDP-11 emulators, web based, for you PC, run on a smart phone, maybe a micro-controller, FPGA. Not sure of the use or value, other than a fun look at past computing, and the challenge to make it happen. We all hold more processing power in are hand, than was available back then.

    Why pick at details, it’s just a project for fun, and learn a few things along the way. Not everything has to be useful, better, or intended commercial project. Keep peeing everyone’s successes, and fewer are going to want to share. Think most the negative comments come from people who are all talk, no build, and feel a need to tear everyone else down to their own miserable, low-life, level.

  5. The picture used in the writeup is from entirely different project :) https://hackaday.com/2017/04/26/a-modern-day-pdp-11-front-end/ also by Jenny :)))

    Neat useless information corner: Very early Infocom history, maybe even a genesis of the company, can be traced to Robert Supnik (DEC engineering/research, SIMH https://en.wikipedia.org/wiki/SIMH) porting “Dungeon” to fortran/PDP-11 (LSI-11, pretty much a desktop PC with 2 floppies) in ~1978 to lighten the load of big iron PDP-10 time sharing system.
    Oral History of Robert Supnik
    https://youtu.be/lk7ygEZxV9Q?t=2h32m28s

  6. Harvey, I built one of those in the spring of 1978! I used it to launch a career in programming! It originally had paper-tape punch and reader, and had to assemble the monitor and keyboard, down to the springs under the keys. I later bought a dual 8″ floppy disk drives for it. In total, it cost about as much as a new car at the time. But still less expensive than quieting my job to go back to college to get a degree in computer science.

  7. That looks like a late model in the photo. No front Panel with Octal lights and toggles, like the second computer I used in school (after a Univac). I can’t get nostalgic about them. An Apple IIe with 4MHz accelerator was faster – in fact I think it is better than a VAX for single tasks.

    1. Looks like an 11/34 in the photo of the rack machine. (DEC badge #48818 here). TBH, we all liked the earlier ones, like 11/45 (esp the one with solid state ram!) and 11/70…and I was one of their specialists in the 4/7/915/10 series (there were no such things as bytes then….). The LSI-11 was…an object of humor among we DEC hardware guys, FWIW, the Q bus being a real nightmare compared to the Unibus and, well, there’s a long list of “the cheapening”. But hey, that instruction set – later copied almost verbatim in the 68000 (just add some registers and lose some orthogonality)…that was the good stuff.

        1. Yes, I did, PDP-15 was just about my favorite one. Lots of blinken lights (incandescent!) to do preventative maintenance on, especially in places with um, attractive secretaries. I kept a “scratched” ’15 front panel on my wall as a memento.

          I worked primarily around the beltway…most of the 15 users were government and pretty much didn’t like you talking about what they did. Most wrote their own environments, as for example, one was looking at realtime VELA data on vector graphics monitors (cold warriors) and the normal DEC opsys wouldn’t cut it.
          ARPA was cool. I didn’t get to work on their 11/40 that was the modem to the arpanet (well, it was part of that modem – the rest was a bunch of WW boards – but I’m not supposed to know, and wasn’t supposed to go in that room). I just did ARPA’s 15 and their line printer…it was on my way to work, and a great excuse to drink some coffee with the gals and then be late to the office “enhancing customer relations” – but nope, just the coffee…and a great excuse to not wear the customary suit – line printers are messy.
          Arpa was, I think, actually using that mainly for office work(!). The pre-internet foolery was in “that room” that had a serial connection to the ’15.

          Really, almost all customers – NRDC, FBI, NSA, CIA, State – you didn’t ask, they didn’t tell. They took all that stuff down before you got in the room at all, even if you had a clearance for other reasons. NASA was open about their stuff, but we didn’t usually ask (believe it or not, they didn’t teach hardware guys software, kind of discouraged that). Back then they had smart guys and most often just told us where to look and were often right about that, so no worries. We were more interested in the cool maps from space, the weird optical FFT processing they were trying to make, etc etc.

          So, I can’t really answer your question. Most customers didn’t run a sharing type system at all (maybe ARPA did). Most ran dedicated runs – run your science program and that’s it – they just used a loader and some drivers. FBI used a bunch to vectorize fingerprints, which then used PDP 11/40’s to interface with their huge quad 360 downstairs to match fingerprints with their database. Which cost a lot of women their jobs in the end. Sigh.

          Security at VISA was by far more intense than at the “alphabet agencies”.

          In general, that was a time with the only people who owned even DEC machines were essentially not resource limited at all. It was kind of an early version of something like linked in as far as the rest of your career might go – if you were good at repairing DEC machines, you never had to look for a job again, just stop saying no.

          A good memory from then – DEC did the most fantastic set of paper tape diagnostics in all of ever. Which started with the old PDP-8 (in diode-capacitor-diode logic, which was itself cool – an and gate didn’t need both inputs at precisely the same time, etc). They stepped up from checking the very most basic stuff – paper tape reader itself bit by bit, and could even tell you if one carry in the main adder was the issue (and in general, stuff like that got it’s own board then – per bit).

      1. We had an 11/70 in school. After I left they got an 11/780. I think a few years ago I calculated that an AT324 at 16 or 20MHz was a couple VAX MIPS. I was at Boeing for a while in 1978/79 and sort of managed one of several new installations of 11/70’s (Maybe 11/780’s? Did they have full toggle front panels?). It is hard to believe now, but they had problems with user’s code getting into loops and locking everyone out with no solution from DEC. I wrote a note to the groups saying how to toggle run/halt until you stopped in user space and entering a halt instruction (00000000?) and see who yells because their program “quit”.

        I do recall support was awful and those really difficult DEC documents cost the company a lot – they needed 40 or 50 of each. OTOH the systems were hooked to some nice Tek terminals and some monochrome and color Vector General terminals and gigantic lofting plotters. This was Boeing’s project leading to full CAD and simulation. There were systems for engineering groups on Aerodynamics, Structures, Sound, etc. They were developing the math to represent the curved surfaces and then simulate and all that.

  8. Wasn’t the DEC Pro 350 basically a PDP 11? I bought one when they came out, but it had essentially no software & I wasn’t a programmer. The operating system was a kludge, appropriately named – I kid you not – P/OS. It had an optional 10mb hard drive that cost $4,550. The cooling fan sounded as loud as a vacuum cleaner. I still managed to have quite a bit of fun with it.

  9. Back in the early 1970s I was tasked with designing some “peculiar” peripherals for a PDP-11 (an 11/35 IIRC). Having done interfaces to Honeywell (bad, and the fans howled) and Univac (far worse, and their data exchange protocol stunk) computers, I was still appalled by the wretched design of the DEC bus. It was simply impossible to do a synchronous logic design which would interface to it; it seems that the DEC bus depended on “crash” logic for the timing to work. It’s amazing that they were even somewhat reliable. And what idiot figured it was a good thing to have to go through the whole bus request/bus grant protocol just to assert an interrupt?

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.