Show HN: KiDoom – Running DOOM on PCB Traces

mikeayles.com

318 points by mikeayles a day ago

I got DOOM running in KiCad by rendering it with PCB traces and footprints instead of pixels.

Walls are rendered as PCB_TRACK traces, and entities (enemies, items, player) are actual component footprints - SOT-23 for small items, SOIC-8 for decorations, QFP-64 for enemies and the player.

How I did it:

Started by patching DOOM's source code to extract vector data directly from the engine. Instead of trying to render 64,000 pixels (which would be impossibly slow), I grab the geometry DOOM already calculates internally - the drawsegs[] array for walls and vissprites[] for entities.

Added a field to the vissprite_t structure to capture entity types (MT_SHOTGUY, MT_PLAYER, etc.) during R_ProjectSprite(). This lets me map 150+ entity types to appropriate footprint categories.

The DOOM engine sends this vector data over a Unix socket to a Python plugin running in KiCad. The plugin pre-allocates pools of traces and footprints at startup, then just updates their positions each frame instead of creating/destroying objects. Calls pcbnew.Refresh() to update the display.

Runs at 10-25 FPS depending on hardware. The bottleneck is KiCad's refresh, not DOOM or the data transfer.

Also renders to an SDL window (for actual gameplay) and a Python wireframe window (for debugging), so you get three views running simultaneously.

Follow-up: ScopeDoom

After getting the wireframe renderer working, I wanted to push it somewhere more physical. Oscilloscopes in X-Y mode are vector displays - feed X coordinates to one channel, Y to the other. I didn't have a function generator, so I used my MacBook's headphone jack instead.

The sound card is just a dual-channel DAC at 44.1kHz. Wired 3.5mm jack → 1kΩ resistors → scope CH1 (X) and CH2 (Y). Reused the same vector extraction from KiDoom, but the Python script converts coordinates to ±1V range and streams them as audio samples.

Each wall becomes a wireframe box, the scope traces along each line. With ~7,000 points per frame at 44.1kHz, refresh rate is about 6 Hz - slow enough to be a slideshow, but level geometry is clearly recognizable. A 96kHz audio interface or analog scope would improve it significantly (digital scopes do sample-and-hold instead of continuous beam tracing).

Links:

KiDoom GitHub: https://github.com/MichaelAyles/KiDoom, writeup: https://www.mikeayles.com/#kidoom

ScopeDoom GitHub: https://github.com/MichaelAyles/ScopeDoom, writeup: https://www.mikeayles.com/#scopedoom

olelele 2 hours ago

Just looking quickly but it looks like I have to try to run this on my analog scope

dspillett 13 hours ago

> to extract vector data directly from the engine

That got me thinking “I wonder if anyone has done this on an oscilloscope” and oddly I can't fine anyone who quite has. That DOOM objects are sprites and not actual 3D objects would limit the fidelity, but the scenery could be rendered at least. There are several examples of managing to use a high-speed scope as a low-res monochrome raster device (scanning like a CRT monitor does and turning the beam on & off as needed).

I did find an example of Quake being done on a scope the way I was imagining: https://www.youtube.com/watch?v=aMli33ornEU - as all objects are actual 3D models in Quake that even manages to give them some presence & shape.

EDIT: then I read the second half of this post and saw ScopeDoom! I'm surprised there are no earlier examples that are easy to find.

  • mikeayles 12 hours ago

    Author here:

    I pulled inspiration from this port to a vectrex: https://www.youtube.com/watch?v=VuVnoqFF3II, there is a wayback machine link to the authors writeup credited in my writeup.

    I have seen a lot of ports of DOOM on overpowered Windows based test equipment, particularly the Keysight MXA's, but they're just using them as a computer. Spectrum DOOM though. Could it be done by taking snapshots of a waterfall plot?

oniony 14 hours ago

The obvious next step is to play Doom rendered as actual PCBs. Each frame is automatically ordered online and then, when it arrives two weeks later, slotted into a holder whilst the bored player contemplates their life choices.

  • Someone 7 hours ago

    You can speed things up with speculative execution. If you add one, you will want a good branch predictor to keep costs down.

  • junon 12 hours ago

    Assuming that's 15 FPS that's about 80-100 Euro (sans shipping) per second of gameplay and due to most PCB fabs having a MOQ of 5 you could play 5 at a time!

    • roygbiv2 11 hours ago

      Yeah but just liked an overclocked gpu, there's errors on those boards so you have to throw them away and wait for the for the next revision.

    • Artoooooor 8 hours ago

      But how motivating it would be to do the best speedruns!

dcuthbertson 5 hours ago

Could you print the designs to paper and make a repeatable flip-card stack? Oh no. How long before it becomes a PowerPoint presentation?!

actinium226 20 hours ago

Lol, I jsut started learning KiCAD last week, and I work in at a game dev coworking space, so this is a perfect combination of the two! Nice!

mlhpdx a day ago

I don’t care how this makes the world a better place, because it just does.

Lerc a day ago

One of my to-do-one-day projects is an audio jack display system out of a Microcontroller.

Was never quite sure if I should raw XY it or soft modem so I could decode on a web page on a handy device.

  • retrac a day ago

    > raw XY it or soft modem

    How about analog raster scan? a.k.a. slow-scan TV? [0] Like how they returned the live television images from the Apollo missions. (They only had 1 MHz of bandwidth for everything - voice, computer up and downlink, telemetry, and TV. Standard analog broadcast TV was 6 MHz. So they reduced the scan rate to 10 frames per second instead of 60, and halved the horizontal line resolution -- that could fit in 500 kHz.)

    Most modern SSTV standards are super-narrowband, designed to fit into just a few hundred Hertz for amateur radio. But what if you had the full 20 kHz of bandwidth of a nice audio channel? With 100 horizontal lines per frame, and 1 frame per second -- that is about 200 cycles per horizontal line, or enough to resolve, in theory, 100 vertical lines on each horizontal line. I.e., 100 x 100 pixels (ish) at 1 fps.

    [0] https://en.wikipedia.org/wiki/Slow-scan_television

    • Lerc 13 hours ago

      Now that does present some interesting possibilities.

      It looks like a web page with audio input permissions can be expected to sample at 48KHz I wonder what the quality is like from a cable bodged off a spare pin.

      A little webapp running on your phone could actually do some nifty on-the-fly display.

philipwhiuk 9 hours ago

As an extension, allow new maps to be added based on chip design CAD files - chips become rooms, soldering lines are corridors

junon 12 hours ago

Side note, this is using the new KiCad socket APIs in v9 right? What was your experience using those?

I've written my own s-expr library to inject footprints and symbols and it's a huge pain and flakey. I'd love to move to something a bit more fleshed out and official.

  • mikeayles 12 hours ago

    This isn't using the KiCad socket API, although it probably would have been cleaner. It's a bog standard Plugin that should work for v6+.

    I'd be keen to have a look at your s-expr library, it likely has some overlap and utility for one of my other projects, https://www.circuitsnips.com , which is like thingiverse for electrical circuits. I had to figure out how to feed the embedded kicanvas renderer a full sheet, as it can't handle subcircuits, but allow the user to export subcircuits to the clipboard.

    When I first shared CircuitSnips with the KiCad discord, the KiCad 9+ design block feature was brought up, which might be of interest to you as well?

thenthenthen 7 hours ago

Is there a video/demo of it somewhere?

  • maaarghk 7 hours ago

    Literally the first thing in the linked post after the titles.

    • p1mrx 7 hours ago

      That's a potato-quality gif, not a video.

jacquesm 18 hours ago

Inception version: make it so shooting components degrades the machine it runs on.

  • p0w3n3d 18 hours ago

    You remind me about Atari XL/XE game "Inside", which was fighting some dust balls or what? inside the Atari. If a circuit was broken you had to fly there and fix it, but when for example Pokey was broken the music was distorted, Antic - the screen was distorted etc

  • ollybee 14 hours ago

    A bit like psDooM

sho_hn 21 hours ago

Love ScopeDoom!

KiDoom I don't fully get. The website says "All components connected to a shared net; the PCB could be sent to a fab house (it just wouldn't do anything useful)" but I don't see any of the component pins hooked up in the demo video.

  • hbnjjgff 18 hours ago

    What don't you get? The pins are not hooked up, so the PCB wouldn't do anything useful

    • sho_hn 18 hours ago

      It means they're not actually "all connected to a shared net", no?

      Something that actually connects the components and routes the traces in a way that makes it somehow still recognizable as the 3D environment would've been cool, otherwise this is kind of just like piping draw commands into a <canvas> from a hook in the Doom renderer. KiCAD just happens to be a complicated line drawing app.

      Don't get me wrong, still a fun little hack. But some more PCB-ness would make it even cooler.

      It might be that the website undersells it and there's more PCB-ness than I can detect in the visuals. Is it using layers and vias between them for the z-sorting or so? Both the website and the commits have a distinct AI slop feel to them and are somehow not very detailed on this part.

      • mikeayles 16 hours ago

        So the reason everything is on one net is so kicad doesn't need to calculate a ratsnest/air wires.

        As for the drawing, we pulled the vectors as a list from the C, and used a painters algo and drew back to front using the distance from the player in the python code.

        We then treated them as polygons to allow us to work out occlusion to hide things behind walls, but the data pipes to kicad/the headphone jack is just the vector/wireframes/outlines, filtered by what's left after the occlusion test.

        So yep, using footprints as sprites was my (clunky) nod to electronics, as I didn't like the idea of drawing polygons. Kicad can definitely handle them, but they're less fun.

        Now, if I'm really bored over Christmas, I may port it to fusion360, which will have a 3d engine.

        I 100% abused Claude code to get here, and i tend to get it to write the bones of a write up, which I then populated with my own thoughts, else I can't get started. We are worryingly becoming more aligned.

robbru a day ago

Of course I love this. DOOM forever.

GaryBluto 19 hours ago

Awesome project, but why is the page a semi-transparent overlay on another page?

danishSuri1994 3 hours ago

This is one of those projects where the implementation is more interesting than the meme. Rendering DOOM isn’t the impressive part, hijacking a PCB editor’s rendering pipeline and making it behave like a real-time vector engine is.

The part I love most is how many unrelated systems had to cooperate:

extracting geometry directly from DOOM’s drawsegs/vissprite internals

mapping sprite classes to physical component footprints

running real-time updates through KiCad’s object model without triggering full recompute

and then running the same vector stream to an oscilloscope via audio DAC

That’s a really clever chain of “use the tool for something it was never designed to do.”

ScopeDoom might end up being the more interesting long-term direction, vector displays force you to think about rendering differently, and there’s something poetic about DOOM being rendered as literal analog voltage traces.

If you ever take it further, the combination of:

faster DAC (or multi-kHz arbitrary waveform generator)

true analog persistence phosphor scope

and dynamic sprite simplification

…could get you surprisingly close to a smooth vector-shooter aesthetic.

Either way: great hack. The world needs more playful abuse of serious tools.

  • mikeayles 3 hours ago

    Thanks very much. ScopeDOOM definitely has potential, I need to work on the occlusion logic to clean up the display, whilst it's not particularly visible on the demo, I did actually use distance to 'camera' to control pcb trace thickness, so close walls look thicker.

    In lieu of an intensity channel for the scope implementation, slowing down the vector drawing for segments would make them 'brighter' on a proper scope, but I don't know if my relatively cheap Siglent would be able to distinguish between them.

    I've got some NI DAC's here, or could use a mcu, the practical limit of the DAC on a teensy 4.1 is reportedly around 1Mhz, however for me I don't think theres a practical reason to do so.

    Although, a native implementation on a teensy, with the wad on an sd card and direct input, no computer at all is very tempting...

    Back on track, i've got to spend a bit more time focussing on work and other projects, but I have my next ludicrous port planned already.

TOR22 16 hours ago

[dead]

djmips a day ago

This is just a meme now. Doom running on X. I don't get it but congratulations on your very whimsical accomplishment!

  • sethaurus 21 hours ago

    This is a little different from most "Doom on X" projects, because the accomplishment is less about the hardware (it's just a normal computer) and more about turning a circuit-board designer into a real-time game display.

  • georgefrowny 15 hours ago

    Doom on X (or rather "Can it run Doom?") is one of the original internet memes, dating way back to the 90s.

aqid1 16 hours ago

[dead]

aqid1 16 hours ago

[dead]