Printegrated Circuits: Merging 3D Printing and Electronics | Hacker News Reader

Printegrated Circuits: Merging 3D Printing and Electronics

spectrum.ieee.org

79 pointsrbanffy31 comments

Wanted to show off a similar project I did. I used tracks for dupoint wires in my model and used the GPIO pins to push through the wires to create connections.

For the LED eyes, I created THT connectors using the ends of the dupoint wire ends.

https://makerworld.com/en/models/672277-wabbt-wifi-bluetooth...

I have been thinking about adding conductive traces a-la PCB to 3D prints in a DIY setting for a while. Obviously, conductive filaments exist - but they are not remotely in the same category as copper.

My initial hope was that doping PLA, PETG or some other material with a conductor and then applying strong variable magnetic field near the print head to force creation of conductive domains while the filament is amorphous and hot. This turned out to be not feasible, as O3 explained to me repeatedly, over hours of chats.

A simpler and surprisingly workable solution appears to be adding a second printing head loaded with tin. Tin is not as good as copper - but it's still leagues ahead of conductive filaments. To offset the poor conductivity you can use thin, but very broad traces.

A speculative approach would work like this:

1. Print PETG layers using a regular filament, but leave "baths" for tin traces. A bath should be an opening at least 2-3 millimeters tall, to account for the surface tension.

2. After N layers, fill the baths from the tin head. Tin melting point is near PETG, but it would cool rapidly and, hopefully, weld to the plastic.

This way you could probably integrate a pcb into a print. I haven't tried that, but i recall people actually trying to print with tin - so that part is at least not a complete fantasy.

Apply PCB mask to copper sheet then laser etch it. UV printing has some more promise here, i just dont think FDM 3d printing is gonna be the solution.

This, I think non-planar PCBs are an incredibly niche thing, and very rarely necessary. you can just run wires/connectors between multiple PCBs, it's no biggie if your prototype needs some manual assembly.

Fiber laser etching is the way to go, the required equipment is already below $2000, and the process is quite simple. Additionally manufacturing copper/aluminium is next to impossible at home, considering their melting points.

You could even use a pre-masked copper plate (which is less reflective than copper), so your PCB would come with a mask.

People have been doing this for a long time, but it feels a little too purist to me.

The components will still be the same, so you'll still need some kind of pick-n-place functionality to make anything, so why not just have another print head for making the traces / doing the PnP?

The head could lay copper wire/foil tape for conductors and do standard PnP from trays / reels of components, which you'll need either way.

It would be a little more geometrically limited than what this post imagines, but it would have the upside that it would actually work today and with most real electronics applications, unlike the low performance conductors made via conductive polymers as the OP's process imagines.

> The head could lay copper wire/foil tape for conductors

Sounds awesome, but this is an extremely hard problem to solve. You can't simply lay down wire or foil into arbitrary shapes on 3D surfaces.

If you had a wire feeder with a actuated barrier just past the tip you can fairly easily bend wire into controlled shapes pretty well. If you printed channels for them to sit in, I think they could be placed.

When I worked at Markforged we had a printer that could put solid carbon fiber threads into the print using a second extruder (on the same print head), so it's certainly possible. It was $20k, though. Getting this down to something accessible to hobbyists is the challenge. I think it will happen one day, though.

CNC wire benders exist, but they're solving a different problem: They bend the wire in free space, not on to a surface. You would have to design the part and the wire such that the part never comes around into the space occupied by the head, which would limit it to only very basic and small shapes.

Maybe lay chains instead of wires. Apply tension to the chain, ensuring that it conducts current and use fast solidifying glue to fix it in place/make it adhere to the surface/insulate it.

I was thinking you'd print a channel where the wire would be placed, bend the wire to the shape of the channel, place it, and print over the top.

Some geometries wouldn't work, but it would be sufficient to make shapes like you see in breadboard connections (up, zigzag around, down)

Yea, agreed it's pretty hard, though with some tradeoffs it could be done pragmatically.

I would probably slightly overbuild the plastic and then use a heated tool to form the smoother surfaces the wire/foil would go in/on.

I've also seen laser sputtering of copper, etc. which could be the another approach, something similar is used for metalizing plastic already, though contamination would need to be controlled to maintain low resistivity.

A PnP placing the components upside down onto a surface printed by another head would be interesting. You could align the heights of the resting surfaces to optimise pads needing to be connected being on the same plane. I'd still want to lay copper but if you had the ability to squirt a little solder paste from (yet another) head, you could stack everything with wire connections into a very 3d circuit.

If the base material was thermally conductive you could have a heatsink block with the circuit embedded in it.

I wonder if you could just print a network of hollow tubes and fill it with mercury or something later. (Maybe with ABS you could use tin?)

There's probably a sweet spot for material melting points for some printable substances. Hotter would be better if the printer can manage it. If you really wanted to make a solid block circuit inside a heatsink you could print it not worrying about things like layer adhesion, then once printed, place it inside a frame to hold it in place and reflow the entire thing. Would mostly depend on the ability of the embedded components to endure high temperatures during reflow, but considering how modern batch soldering works, I'm guessing a lot of this problems have been addressed (or a the the very least, the bounds of capability well known)

I don't think there are any non-exotic 3D printing plastics that could survive a solder reflow

Well I think when you are going for something like this, we might already be in exotic territory.

I'm sure it's never wise to underestimate a materials scientist.

I think the "printegrated circuits" approach is roughly the right level of abstraction.

3D Printing the PCB itself is pretty much impossible for any non-trivial application. Doing multi-layer PCBs with 0.20mm wide traces, spaced 0.20mm apart? Forget it, not happening - and requirements like those are standard for hobbyist-level chips like the RP2040 these days.

And if you're not printing your own PCB, what's left is module-level assembly and connectivity. In other words, just printing a bunch of wires.

For this in 2D, see Sam's thesis, 6.3.3 (p. 86, CNC wire plotting). 3D would add a lot of challenges.

https://cba.mit.edu/docs/theses/19.09.calisch.pdf

This reminds me a bit of Multiwire, a somewhat unusual circuit manufacturing technique from the 1980s. A machine laid down wires and then encased them in resin. The best info I can find is this slide deck:

https://www.swtest.org/swtw_library/2013proc/PDF/SWTW13-22.p...

I believe this technique was used in the Three Rivers PERQ computers. Here is an image of one of their PCBs https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg...

That's pretty cool. Why is it not used for prototyping? What are the downsides?

There's a lot of potential for desktop rapid-prototyping with electronics. I think one of the things that is killing us is the tooling. One of the reasons I started building an autorouter was because I wanted to be able to have different "build targets"- e.g. a build target that is a PCB with no vias and only 0 ohm resistors (jumpers). If our EDA tooling supported different build outputs, then we could have earlier prototypes built with less-than-ideal equipment (e.g. conductive 3D printed filament, as the article suggests)

Has anyone used dark color 3D filament printed onto copper clad PCB as photo resist or etch resist?

It might be tricky printing PLA directly to copper clad PCB, but then you could expose the board to UV or etchant to make the PCB traces. Then remove the PLA plastic to expose the copper traces.

assumedly much much easier to buy a cheap vinyl cutter, but interesting idea

The toner transfer method uses a similar process and is probably easier. You don't need a 3D printer, but one of those laminating machines instead.

Yep, variations have popped up in the DIY community over the years.

Slightly modifying/abusing the cheap DLP resin printers is more effective because they're essentially a controllable UV mask anyway.

We are starting to see metal filaments and even this copper one[1]. Multi-filament fdm printers just might be able to make some rather large circuits. I doubt we'll get down to 0.2mm tracers, but if size isn't an issue, we can do better than the conductive carbon tpu(?) filaments which are common today.

[1] https://www.gsc-3d.com/3d_materials/copper-filament/

Anyone remember the Next Dynamics NexD1 Kickstarter?

It was pitched back in 2017 as a "Multimaterial & Electronics" printer. Got to half a million or so in pledges before some of the backers uncovered serious red flags and Kickstarter suspended the campaign.

Hope this effort fares better.

Printed Circuit Boards, as the name suggests, are already printed, with the extra step of submerging in a chemical solution.

I don’t see the advantage of this approach. You end up with a worse and less reliable circuit just to avoid the single chemical step.

On the one hand, I like the idea. On the other hand, I dread a future where you need an X-Ray and/or MRT machine to be able to inspect any kind of electronic device. And don't even think of disassembling or repairing...

Project Binky (YouTube) ep 39 or 40 did this with their CNC mill

Reply on news.ycombinator.com