I'm surprised I've never heard of this before. Toner transfer was always kind of a pain. Now I just prefer milling because it avoids all the nasty acid and lets you accurately drill holes and cut the PCB to size/shape at the same time. Small CNC machines are pretty cheap these days, the one I use cost $130.
PCB houses are amazingly cheap and fast compared to just a few years ago but if you don't need multiple layers nothing beats having it done in a few minutes for ~$0.10 per square inch for prototyping.
I have the Linksprite engraving machine[1], there are a bunch of companies selling the same design if you search for "CNC 1610" on any of the major e-tailer sites. It's called 1610 because the build area is 16x10cm, there are larger versions called 2418 and 3018.
At the price point it's obviously the bargain basement model of the CNC world and the quality is not great. But it works fine for PCBs as long as you don't need super narrow traces, the backlash is surprisingly not terrible thanks to the preloaded nuts.
I use FlatCAM to generate gcode from the PCB gerber files. Then just load up a copper clad board on the machine and let it cut away the copper until only your traces are left.
Not OP, but I've had luck making PCBs on my Taig mill. Might want to choose something with ballscrews or at least preloaded nuts or potentially a belt drive. PCBs hate backlash, but don't constitute a high load.
Yes, safety is always important. That's why I put the machine in my garage where there's a dust collection system.
Breathing in fine dust in general is not very good for you so dust collection and/or wearing a respirator is generally a good idea when doing any major wood/plastic/fiberglass cutting.
Genuine question: with all the great quality, super fast, and incredibly cheap PCB fab options available now - what motivates making PCBs at home?
I totally get that it's a fun process (and back in the day made plenty of boards with toner transfer + sharpie cleanup + ferric chloride), but if there's an interest in better quality product, the commercial solutions are going to be lightyears better and designs meant for them are far easier for others to replicate.
Nowadays, The final result is always going to be manufactured by a PCB house. The primary motivation is (still) hobbyists' home experiments, but I guess it's not too uncommon for rapid prototyping and verification in the workplace, especially for RF stuff below 1 GHz. Often, you find yourself prototyping a component, and you've never used or designed around it before. To begin with, you need to give your design a try and see whether it works or not.
* The process needs to be fast, you want to see the feedback immediately, even waiting for 48-hours (many PCB houses offer this option) is a waste of time.
* It's highly possible that your design contains one or more logical and layout problems, you need the ability to do multiple iterations in a weekend.
* More and more state-of-art components are surface-mount, it's simply impossible to prototype many things without using a PCB. RF is another interesting example, a double-sided PCB is simply irreplaceable for even the most simple RF circuits. Above 1 GHz, it becomes tricky and it's better use a factory-made PCB.
I guess it's down to what qualifies as "fast" and "cheap" - I tend to use dirtypcbs, mainly out of habit, and they're usually well within 2 weeks to my door in New Zealand via DHL shipping. OSH Park is in the US and claims under 12 days to ship with their standard option, their 5 business day option doesn't look too expensive. There are probably other competitive options available.
For my hobby stuff, I feel like even if the board fabrication were instantaneous it wouldn't affect my project start-to-finish time very much. The time designing boards, making mechanical bits, and developing software are dominant, and my hobby time comes in little bursts, often several days apart. It would also take me longer to design a board for home manufacture and to do the assembly, since my hypothetical home fab won't produce reference designators, solder mask, nor vias.
Anyway, I'm not meaning to talk down at folks making their own PCBs, but more wondering if there's some missing information or niche there.
I’ve never used JLCPCB for anything too intense, but turnaround has always been 5 days on the high end in my experience. Is it different if you require more layers or more sophistication?
JLCPCB always requires atleast two weeks for shipping in my experience. And I'd rather test out 2-3 designs on a single day if possible, especially iterative designs. That isn't possible without DIY PCB etching.
while still not "in a weekend" I've had cheap service PCB's from AllPCB in under a week order to door. it does help to have the "ready for shipping" point not fall on a weekend though. I think my record was 5 days from submitting my gerbers to them hitting my doorstep but i suppose everything aligned exactly though (and needless to say, that was using fast shipping like DHL. Allpcb aren't crazy cheap like they were though - for a while it was $5 for 10x prototype boards (under 100mm^2 area) with free DHL. But they've been gradually been increasing their prices and they don't work out as competitice as the other main prototyping offerings)
If I need double-sided boards I'll usually farm them out because it's an absolute faff. But if I need an really thin adapter to go from Wide SOP to Thin SOP (typically for replacing blown Flash chips), I'll make a board. When you ask a large PCB fab for anything that isn't 1.6mm thick FR4 white-on-green and standard stackup, the prices rocket. (because they have to stop the line, set up for your custom run, then switch back)
If I just need one board to fix one radio, I'll coat a piece of Pyralux laminate or paper-thin FR4 in photoresist, expose it, etch it and cut it out with scissors.
For quick-run -- I was involved with a theming/prop building group a few years ago. Last-minute unforeseen issues weren't uncommon, and you just had to roll with it. Being able to spin a bodge-board in a couple of hours can be all that's needed to snatch success from the jaws of failure.
>When you ask a large PCB fab for anything that isn't 1.6mm thick FR4 white-on-green and standard stackup
actually most of the china fabs (PCBWAY, JLC, elecrow, allpcb etc) include most of these options in their "prototyping" offers now.
some even have MATTE BLACK and controlled impedance at no extra cost. competition is fierce right now so it always pays to shop around (and not just using PCB shopper)
Low latency. I still sometimes fall back to at home (well, hackerspace) PCB etching when I need a breakout for a connector/SMD package _right now_ in order to progress with a project. With practice, you can go from design to etched and drilled single-sided PCB in 30 minutes.
3D printing services have been quite helpful to me in designing and building things... I do some printing with our local makerspace's FDM machine, but also really like being able to order high quality parts made on a big commercial SLA machine.
I do think the situation with 3D printing is a little more nuanced than the PCB discussion. Currently, not much gets manufactured using 3D printing. So, time spent on designing a 3D printed part will be time spent prototyping, whether you print it or pay someone else to. With PCBs; I imagine that there will usually be design compromises to enable home vs commercial manufacture. So, if there's a possibility of scaling up production, then the home route is going to involve some combination of a suboptimal design for commercial manufacture, and burned design time. I'd rather just aim my PCB design effort at being able to buy 10^x boards, then only order as many as I need.
To try an analogy: home PCB making seems to me like mechanical hobbyists making their own nuts and bolts. Sure, it's fun to know that you can do it, and on occasion it does make sense. But, I don't understand why there is so much interest in actually doing it, and refining the tooling/techniques to do so.
Basically, there's interest because there's no alternative for me. I either have to use perfboards and solder a bunch of wires, or wait a month for my PCBs to arrive through the cheap post, or pay $50 to get 10 PBCs fast when I really only need one or two and will throw the rest away.
None of these options is really viable, and if I could make a PCB myself in half an hour for $1, I would jump on that.
> None of these options is really viable, and if I could make a PCB myself in half an hour for $1, I would jump on that.
You kind of can, depending how you amortize equipment costs. This is how I've been doing this and it seems to be a good tradeoff between low initial cost, low effort and hhigh efficiency.
Up front you'll need a) a laser printer (any old laserjet will do) b) a hacked paper laminator (to increase its temperature) c) an etching tank with an aquarium bubble pump for water circulation and an aquarium heater. You'll find most of these in any hackerspace.
Then, for each PCB you'll need wax paper (even glossy magazine paper works perfectly) and copper clad FR4. Print your mask (mirrored) on the paper, clean your copper board (with steel wool, then with acetone), transfer the mask using the laminator. Then etch it.
This gives you a one-sided, undrilled PCB with 6 mil traces [1]. Good enough for the occasional SMD breakout.
That's almost good enough for me but the drilling for vias/through-holes kind of ruins it for me :/ There's no perfect home solution, so I make do with $10 JLCPCBs for now (which are cheap and fantastic quality, if you don't mind waiting).
Agreed, DIYing vias is also a big blocker for me now, which is why I only really use this method for single-layer prototypes and experiments. There's been some recent improvement in at-home via electroplating but it's still not there yet.
I used a method somewhat similar to this as an undergraduate researcher to cheaply develop patterns on copper-clad kapton for flexible circuits.
I used a solid ink printer, which I found by crawling the university network for printers and querying them for their model number until I found one. Then a very nice sysadmin let me use it on a thick sheet of copper clad kapton and it luckily did not jam! One of my few research success stories.
I used to make my own PCBs at home, and I still really enjoy the technologies that allow making high quality boards from home. I've found for my personal stuff that going through a low-cost fab house is just so economical, and the quality is so good that I usually can't be bothered to do it myself anymore. Still great if you want the learning experience, or if you really need it now.
This site is ancient (the printer used is from mid 2000s,) which makes this all the more impressive. PCBs with 1 mil traces are still quite expensive to prototype. OTOH, if you need such thin traces, you also probably need multi-layer PCBs.
Volkan Sahin, as far as I can tell, is one of the few people to truly master printing from a high end inkjet cartridge as a hobbyist. Really really impressive work.
PCB houses are amazingly cheap and fast compared to just a few years ago but if you don't need multiple layers nothing beats having it done in a few minutes for ~$0.10 per square inch for prototyping.