CDP's 11/04-ish? | Replicating and Modernizing a Dinosaur

Back to "Computers" | Home


I like PDP-11s. I've seen 'em at shows, I've seen 'em in museums. I've played with 'em at home, I've played with 'em at work. I've moved a BA123, I've moved an 11/70. I have an 11/23+ that gets used to fire up RT-11, run some TETRI2, control various GPIB test equipment while working on stuff, and just have fun making silly little programs. It's pretty cool.

I like TTL stuff. I remember hearing someone was doing a Pong replica PCB and I thought it would be cool to build. The original PC-9801 is my favorite PC-9801 because you can at least source the darn chips to repair the thing. My first soldering project was an all-TTL alarm clock. I've always loved the earlier PDP-11s because, even though some of that 74-series logic is pretty darn hard to get a hold of these days, it's neat how relatively repairable they've proven.

I got talking to myself at work like a week ago (April 14th or 15th, 2023?) and spat out the idea of "what if I built my own little PDP-8?". I looked on VCF and saw that some people are cloning PDP-8/E PCBs, some people are working on making a new Omnibus backplane with modern connectors, and those projects could be mixed to maybe some day make a new PDP-8/E. Neat! But I haven't seen anyone try to make a new PDP-11...

As stupid as it's going to be, I want to try to build my own PDP-11. I want to start small, a simpler CPU in the line which could be made into say a silly Elektronika-esque desktop form factor. This page is all about that. Will it work out? Will it crash and burn? Only one way to find out.

Entry #1 - CPU Selection - April 22nd, 2023

One major goal of this project is going to be using discrete logic, not CPLDs or FPGAs, to do what I want. It's a deisgn aspect that I find really cool, I'd love to be hands-on with 74181s to learn how they operate, and this would be a lot easier if I could juse run a Pi Pico!

The first thing I need to do if I wanna make "my own PDP-11" is figuring out which CPU I'd like to target.

As much as I'd love to claim I'll start off learning the PDP-11 by replicating the original, the 11/20 is a lot being that I don't actually have a ton of experience with designing around TTL. Yeah I can just copy the engineering drawings, they're all out there, but that's nine whole boards! Maybe for the future, not for right now.

As much as I'd also love to do a top-of-the-line processor, the 11/70 is both massive in scale and way over my head. I won't be touching that until I have a good understanding of how PDPs actually work internally. Yes I've three years of experience diagnosing an 11/70 but that's mostly just looking for messy lines and finding the source, I can't claim to have an honest understanding of the architecture.

As much as I'd also love to claim I'll make an 11/05, I don't think I could do the front switch panel justice just yet.

I think I'm going to settle on the PDP-11/04. It's a weak model, sure. It can't support a MMU, a FPU, the EIS, or any of the added instructions after the initial generation. It won't be running Unix. That all being said, it does everything I like in a PDP-11; It's 16-bit, it runs RT-11, and it can even run RSX-11M. It has Unibus, or whatever basterdization I'm going to make of it, so it can connect all sorts of perephrials. It even has the option for a light front panel!

If I were to buy the ICs to remake the board, it would cost me about $125 (with extras on all the DEC-marked parts in case some are DOA) via UTSource. That's not horrifying. Why not give it a shot?

I'm working today and tomorrow. I'm going to print out the processor user's manual, start giving it a read, and start analyzing the schematics. I'd also like to take the 145 pages of microcode listing and translate them into a modern copy-pasteable format. Sounds like a good weekend project, no?