Hacker News

I made my own PCBs with a 3D printer(stavros.io)

285 pointsStavrosK posted 8 months ago144 Comments
tomstokes said 8 months ago:

I love seeing DIY PCB manufacturing projects like this. Using the 3D printer as a plotter is a creative take on the traditional DIY process.

For any aspiring EE hobbyists: Manufacturing your own PCBs is almost never worth the effort. You'll spend days or weeks getting the process and tooling right, and you still have to manually wire any vias that connect the front and back sides of the PCB. These DIY PCB manufacturing projects are fun if you're in it for the experience, but very impractical for getting work done. It's quick and easy to order small quantity PCBs online.

OSH Park is a popular option: https://oshpark.com/#services You can get 3 boards in 9-12 days for $5/square inch. If you need the boards sooner, $10/square inch will get you a 4-5 day turnaround time. You won't save any money by buying all of the gear to DIY etch your boards, and you certainly won't save any time.

franga2000 said 8 months ago:

> Manufacturing your own PCBs is almost never worth the effort

I don't know, for me, since I already had an entry-level 3D printer and Dremel, I was able to mod it to engrave the negative of a PCB onto these copper plates in a few days. Now I can get from design to finished PCB in an hour.

I suspect many people getting into EE might have some of these tools already, and being able to iterate quickly with locally-available supplies is a huge advantage, especially for people new to PCB design.

Edit: I would like to add that my method was rather poorly though out and the method in article is much better. Putting a Dremel on a 3D printer is not a good idea, it was just the first thing I came up with and it happened to work ok-ish for my very basic needs.

archi42 said 7 months ago:

Actually, PCB milling[1] is a thing. I suppose a MPCNC can be used for that. But it still has the huge drawback that you're limited to one layer if you're lazy (like me) or two if you fell like spending the time to add manual vias - though I was told there are tools for that, which work essentially like... Err... riveting pliers? (sorry, non native speaker)

rm445 said 7 months ago:

Given that multi-layer PCBs are effectively laminations of individual boards, hobbyists have already made their own multilayer boards [1]. I don't see anything particular about milling as a means of copper removal that would stop you layering up boards?

[1] https://www.instructables.com/id/Flexible-Double-Layer-OR-Mu...

archi42 said 7 months ago:

Thanks for the link, nice to know it can be done at the hobyyist-level. But it's still a lot of work I (personally) do not enjoy spending time with. If I had the space I'd probably setup my (hypothetical) MPCNC for milling PCBs, so I can do quick single layer prototypes (maybe some simple stacking using header rows), but anything more complex would still go to a fab.

archi42 said 7 months ago:

[1] https://en.m.wikipedia.org/wiki/Printed_circuit_board_millin... (derp, I forgot the link and can't edit anymore)

shaftway said 7 months ago:

It's not a terrible idea, and I've wanted to play around with the concept myself, but I've never really had time or an acute need.

The Snapmaker (1) is a 3d printer with changeable heads. For mine I mostly use the 3d printing head or the low-power laser engraving head, but it also came with a milling head and they sell a higher-power laser head that can cut some materials.

I looked around and people have had trouble trying this with the milling head. The bit usually breaks and the X axis isn't stable enough. Maybe v2.0 will be more sable since it has a second Z axis mount.

That being said, their mounting platform is modular. There's nothing to stop you from fabricating a dremel mount and ripping apart an existing milling head to interface it with the dremel.

[1]: https://snapmaker.com/product

thedance said 8 months ago:

Those prices and lead times don't seem very competitive. At Sunstone I pay $5/sq in for 1-day service on a 2-layer board. If I was willing to wait a week it's $1/sq in.

I think people who are impressed by 3d printers often aren't very experienced with how easy it is to make or get made the things they are printing. I can draw a board in my EDA program, hit a button at 6pm and come back to work with the boards on my desk at 8am the next morning, if I want to pay for that. Same deal for parts: I can order a box full of crap from Digi-Key at 10pm Pacific Time and the Fedex guy will drop it on my desk 10 hours later. Electronics prototyping is extremely well developed, fast, and dirt cheap.

kickopotomus said 8 months ago:

Do you have some sort of special deal on pricing? What sort of quantities are you ordering? For oshpark, it's $5/sq in for a batch of 3 boards. I just checked the prices at sunstone and they quoted ~$35 per board for a 1 square inch 2-layer board for 1-day service or $14 per board for 1-week.

StavrosK said 8 months ago:

JLC PCB is less than $10 for 10 5x5 cm PCBs, including shipping. They just take way too long to deliver where I live.

cozzyd said 7 months ago:

Not to mention many of their orders are stuck in limbo right now.

sschueller said 7 months ago:

All my JLCPCB orders are moving and arriving without an issue but that is not the case with their parts supplier (LCSC) or others.

thedance said 7 months ago:

I just looked at my last order, it was a much larger board. I suppose there are different optimal choices for very small boards or larger ones.

quadstick said 8 months ago:

You are right, we live in the golden age of electronics prototyping. I'll throw in plugs for the Octopart search engine for finding parts, and PCBEX for inexpensive prototype boards up to ten layers.

StavrosK said 8 months ago:

Let me guess, you live in the US?

weaksauce said 7 months ago:

that's actually really cool... sunstone is also US based if you are into keeping it in the US... win-win.

though, i just did a quick quote and they are quoting some seriously high prices so i am not sure where you are getting that number from?

beeforpork said 8 months ago:

> For any aspiring EE hobbyists: Manufacturing your own PCBs is almost never worth the effort.

Worthiness is something that every hobbyist needs to evaluate themself. You cannot do that for them.

For a hobbyist, time, money, efficiency, quality may all be totally irrelevant.

nwallin said 7 months ago:

Yeah I had kind of a weird reaction to his comment.

I was inspired by Ben Eater's YouTube series about making a breadboard computer. Rather than copying his schematics, I designed my own. Took a very long time. Went on digikey and Amazon and dropped probably around $1k in parts and tools. Now I'm building the damn thing and it's very time consuming.

I also recently bought a new laptop for $200 which is many orders of magnitude more capable than my breadboard computer eventually will be. Does that fact mean my time, effort, and money spent on the breadboard computer is a waste? Hell no.

stephen_g said 7 months ago:

I think that just proves his point. You're spending time doing the parts of the project that interests you.

Manufacturing PCBs isn't electrical engineering, it's process engineering. I'm sure there are some process-engineering hobbyists, so if people want to spend their time doing that, then cool. But his point is that if it's the electronic engineering part you're interested in, spending a lot of time tweaking and improving your DIY PCB manufacturing when you can order dirt cheap boards from China that arrive in less than two weeks, that are far and away better than what you could do at home, that generally isn't worth it.

dwild said 7 months ago:

> Does that fact mean my time, effort, and money spent on the breadboard computer is a waste? Hell no.

Well you did spent $1k in parts and tools instead of building theses parts and tools yourself, didn't you? His argument wasn't that it's not worth the experience, in fact he even mention how the experience can be worth it if that's what you want to experience. If your goal is just to get a PCB, then paying for a professionally made PCB will worth it (just like in your case spending $1k in parts and tools will worth it).

StavrosK said 8 months ago:

Yep, this project was basically worth it just for the fun of it.

lostgame said 7 months ago:

I love to hear stuff like this. I’m glad you can find your own joy through these kinds of technical arts.

Sometimes it really is just passion that drives us.

StavrosK said 7 months ago:

Oh yeah, I don't know what I'd do if I stopped finding making things like this fun. It's like being on drugs, except productive.

lostgame said 7 months ago:

Hey, you can be productive on drugs... >.>

StavrosK said 7 months ago:

I guess it depends on the drug.

StavrosK said 8 months ago:

I disagree that it won't save time, if just for the fact that 30 minutes is less than 9-12 days. I usually need neither the quality nor the quantity that these fabs will give me, so being able to quickly make a PCB is definitely worth it for me.

wpietri said 8 months ago:

That makes a lot of sense to me. One of the reasons I stick mostly with software is that it allows for much quicker iteration. That increases both my ability to explore a space (when working divergently) and polish a product (when I'm in convergence mode). Hardware is, well, hard.

Iteration speed can make a huge difference in outcome. I read that the average consumer electronics company goes through 3-7 physical product iterations. Apple, on the other hand, went through more than 100 generations of prototype for the first iPod.

Obviously what you've done here is just getting started, but it points in a really interesting direction. A 3D printer with swappable heads for plotting and an add-on PCB etching kit should definitely have a market. That would mean recurring revenue for PCB supplies at a good margin, too.

tomstokes said 8 months ago:

> That makes a lot of sense to me. One of the reasons I stick mostly with software is that it allows for much quicker iteration.

4-5 day turnaround from OSH park is quite fast. Keep in mind that you usually need to order parts from Digi-Key or another supplier when you finish the design anyway. If you have a mistake on a PCB, it's faster to use a tiny stitch wire than to make a whole new PCB.

On a professional level: You can always buy fast turn PCBs if you need them ASAP, but you'll pay for it. Usually cheaper than paying engineers to fiddle with finicky PCB milling or etching machines, though.

> Obviously what you've done here is just getting started, but it points in a really interesting direction. A 3D printer with swappable heads for plotting and an add-on PCB etching kit should definitely have a market.

Chemical etching is more of a novelty these days. A cheap CNC mill produces much better results without the mess and uncertainty of chemical etching: https://hackaday.com/2018/01/04/guide-why-etch-when-you-can-...

On the professional level, several manufacturers make dedicated PCB mills: https://www.lpkfusa.com/products/pcb_prototyping/machines/ These machines have helpful features to index tools to the surface and align the panels as you flip them over for two-layer designs.

marcosdumay said 7 months ago:

I have extra parts at home, exactly because of this. But I can not order extra PCBs with my future mistakes corrected. So it's either make them myself or try a single design, spot my failure, correct, order a new one, wait a week or two until I have some more free time to mess with it.

Or avoid PCBs completely, what is clearly the best option whenever possible (what is becoming less and less common).

StavrosK said 7 months ago:

To be fair, modern EDA software helps a lot with not making mistakes. If your schematic diagram and tolerance settings are correct, it is pretty hard to get the PCB wrong. Every time I had a problem with one, I had made a mistake in the schematic (and once the fab house mistakenly bridged pads because my tolerances were too tight).

I wholeheartedly recommend KiCAD if you aren't already using it.

HeWhoLurksLate said 7 months ago:

Seconding KiCAD. Their software is fairly intuitive- especially if you're used to Linux- and it's super helpful for making sure that your designs are logical and stuff, and you can also use it to figure out how to best breadboard your designs as well.

StavrosK said 7 months ago:

I disagree and think saying it's intuitive is doing people a disservice. It's not intuitive at all, but it's more intuitive than the alternatives and it won't take you very long to learn how to use it.

If you don't get it immediately, don't worry, that's normal. Just watch some tutorials and you'll be good to go in a few hours.

HeWhoLurksLate said 7 months ago:

Okay, you're right. I'd forgotten that I'd A) had a failed start with it and B) only gotten past that once I had to use it for a school project. That being said, once you figure it out, it's really powerful. It just takes a bit to get there, and I'd forgotten the beginning of my learning curve.

Sorry for the confusion, and thank you for reminding me of how hard it was for me personally to start.

StavrosK said 7 months ago:

No problem, and I agree, it's very powerful and it's fun to use. It doesn't take very long to get a feel for what's going on, and it's very rewarding afterwards, as you know.

sokoloff said 8 months ago:

Digikey, Mouser, and Arrow, for the most part, give me next business day delivery for cheap (often free). From Arrow, I think the minimum is only $50 for "free" overnight shipping.

squarefoot said 7 months ago:

"Manufacturing your own PCBs is almost never worth the effort."

It depends on the target device. For analog or simpler digital circuits prototypes where single side boards can be used it's still very convenient, with the only annoyance being the etchant treatment and disposal. In the old days I made things easier by adding one intermediate step in which I redrew the schematic as the ICs and keyed parts were seen bottom up, that is, ICs had the 1 pin up right, etc. Also the schematic wasn't arranged functionally, like an opamp drawing for example, but it rather showed how the pcb was going to look from solder side. From then on, copying it on the actual pcb became a breeze. As for drawing on the pcb, I did all by hand, using both transfer sheets with pads where necessary and thin (1mm or less) normal water resistant markers for connections, buses etc. Faber Castell and Staedtler were my favorites, as I found those specific for pcb drawing were a lot worse quality wise, not to mention like 4x more expensive. This method would of course not scale with circuits complexity, more than one layer and or series production, still I enjoyed a lot being able to design, etch and solder even moderately complex circuits in one evening or a week end in "mom's basement" (actually dad's attic...:). With time I also trained myself to draw my schematics directly that way, so I could get rid of one step between drawing on paper and the final drawing on the pcb. I miss those days...

OJFord said 7 months ago:

> You'll spend days or weeks getting the process and tooling right, and you still have to manually wire any vias that connect the front and back sides of the PCB.

Fancy! At school we had a UV box & bath, and made single-sided PCBs. I assume you can use the same process just with boards that have copper both sides. It was great for demystifying the process, and I suspect economical in a school setting.

(Aside: I then studied EE (+CS mix) at university, and there was none if that. Many of my peers without the opportunity to take Electronics at GCSE & A Level would be nonethewiser. Maybe that's fine, lines have to be drawn, but seems to me like an easy extension of existing lab assignments.)

As a hobbyist, if I had the room I'd certainly want the kit. Nothing beats the immediacy for prototyping , or turning around fixes/modifications. The basic stuff isn't prohibitively expensive, we're not talking even hundreds of units before break-even.

madengr said 7 months ago:

I do it at home (PCB mill) for RF circuits on Rogers 5880 and similar materials. At work though (more RF), I just order it. Seems to always be $3k-$5k per run no matter how many boards and what material. You are buying the PCB lines time yourself rather than sharing the cost.

I had twelve boards recently from OSH Park for a whopping $3.37 total, which includes shipping. They were small boards.

panpanna said 7 months ago:

PCB Mills create a lot of fiber glass powder. That thing is not healthy to inhale.

You can buy PCB of other materials such as compressed papper (F2?) but they are not easy to get hold off.

madengr said 7 months ago:

My mill has the cutter fully surrounded by a vacuum manifold, and the vacuum itself is bag and HEPA filtered.

The compressed paper PCBs have horrible properties. The RF laminated all have specific properties for electrical performance; dielectric constant, low loss tangent, low surface roughness, homogeneity, controller thermal expansion, thermal conductivity.

fest said 7 months ago:

Professionally- agreed. However I look at it as a learning exercise, which is a wonderful way of calibrating one's sense of time/money quadrants.

If you don't have time, but have money- send it to overnight PCB service. If you have time, don't have money- do it yourself. If you have both- send it to regular PCB service.

Now the last one is the least desirable quadrant to be in- if you don't have neither time nor money and this is the first time you're doing this- expect to find either (or even both) of those, at least until you have gotten the equipment and gone through the process a couple of times.

bochoh said 8 months ago:

I've seen hobbyists use JLC PCB in the past for low quantity runs and the quality always seems astounding with relatively quick turnaround times (several days). https://jlcpcb.com/

Looks like they do SMT mount now too which wasn't the case last time I looked at their offering.

StavrosK said 8 months ago:

I've used them many times and have always been extremely satisfied. The downside is the three week wait times.

gmurphy said 7 months ago:

FWIW PCBWay generally gets my weekend orders to me within the week, and costs about the same as JLC ($5 for 5 boards plus $15 shipping). In some cases I've put an order in on Friday night and gotten them on Tuesday.

StavrosK said 7 months ago:

Where do you live?

tdiggity said 7 months ago:

Have had several orders with this timeline. I’m in California.

StavrosK said 7 months ago:

Yeah, shipping in the US is much faster. Here in Greece you have to pay 3 euros minimum on shipping charges and shipping from anywhere in the EU will take at least a week.

kazinator said 8 months ago:

I've achieved good results with just a laser printer.

(Print mirror image onto glossy magazine paper, transfer toner to copper with clothing iron, wash paper away with water, etch, drill.)

https://i.imgur.com/JtMQemK.jpg https://i.imgur.com/2olwewD.png

I have used markers for touching up toner-transfer boards (not that one). I found that markers do not resist the etching solution as well as toner. Toner provides a higher build of plastic resin material on the surface than the trail left by a marker. (Perhaps doing multiple coats with the marker could build it up.)

I would also highly recommend doing "copper pours" as I did with the above board. Etch away as little copper as possible. The etch will be efficient and fast.

The circuit board in my above images was etched inside a tiny container, in minutes! In fact, the "etching tank" I used was a plastic bottle cap from a 1.9 liter V-8 juice.

If you etch away most of the copper and insist on getting those large, etched areas very clean, that may take so much etching that it will eat into the traces that you want to keep, leaving them in rough shape, and maybe even nonworking.

The resist only protects the copper from the top. The sides of every trace are exposed. Leave it in the solution long enough and the etchant will work its way under the resist and etch them away.

Of course, copper pours require large black areas, which probably doesn't translate to good economics if you're using markers. A mixed approach might be possible: do smaller copper pours in the the nooks and corners around the traces and pads, and cover large areas with rectangular pieces of electrical tape.

StavrosK said 8 months ago:

You're basically spot on, I had all these problems that you mention. I didn't think of adding planes (I actually had them but removed them) because I didn't want the printer to spend an eternity drawing, but after you mentioned it, it makes sense that it's definitely worth it.

That's what I'll be doing from now on, thank you!

Zardoz84 said 7 months ago:

My father (and my self did a few) made a lot of prototype PCB with the toner technique. A trick that I learned from him, was giving a coat of tin alloy to fix any micro cut made by defects on the copper or the toner transfer.

throwanem said 8 months ago:

You asked about safety. Sodium persulfate isn't that risky [1]; your main concern is respiratory, so keeping the temperature low is a good intuition. If you decide to work with higher heat to accelerate the etch, wear breathing protection.

Sodium persulfate is a strong oxidizer, which is what makes it an effective etchant, but that also makes it a fire risk. In storage, keep it cool and dry; silica gel moisture scavenger packets probably aren't a bad idea. And, of course, keep it away from anything flammable.

Finally, it's a good idea not to dispose of the used etchant solution down a drain. Check with your municipality to find a safe disposal method, rather than add it to the ecosystem where it will poison wildlife.

[1] https://www.fishersci.com/store/msds?partNumber=O61141&produ...

StavrosK said 8 months ago:

Thanks for this! Would it be okay to etch with the container lid on? That would take care of fumes even at slightly higher temperatures, but I don't know whether it'll cause some other problem.

throwanem said 8 months ago:

I'd hesitate to try that. You'd be making a pressure vessel, and while this doesn't seem a tremendously energetic process, it will become more so with increased temperature. Having a small explosion in your flat probably isn't a good time, either!

If it were me, instead of a sealed container I'd just run the process under an open window with a fan propped in it to exhaust whatever fumes develop, and maybe under an improvised fume hood made of a cut-open garbage bag and masking tape. (If you do that, make sure to leave space at the bottom for intake air. The idea is to make a funnel that will pull room air in, mix it with process gas, and exhaust the lot outside instead of into your living space.)

The window-and-fan method worked well when using vinegar-based iron acetate stain, and later polyurethane sealant, on a quite large table I made back when I lived in a tiny apartment; for a low-volume process like this, it'll probably do just fine, I'd think.

StavrosK said 8 months ago:

That would be a good option in the summer, but with sub-zero temperatures in the winter, that would probably not be ideal :/

I might try a sealed container with some kind of valve or loose seal and see, thank you!

pjbk said 8 months ago:

Don't use a marker. I use 2-3 layers of TPU filament or Ninjaflex over copper treated with washable adhesive. Works wonderfully. As always, clean the copper very well with isopropyl before anything. You can also do flex PCBs this way:

- https://www.instructables.com/id/Make-Flexible-Circuit-Board...

xenomachina said 8 months ago:

Nice! My first thought when I saw this post was that extruded filament of some sort would probably be more precise, and easier to set up.

Why TPU or Ninjaflex? Does that make it easier to eventually remove the mask? Or is it that other materials would break down in the etchant?

pjbk said 8 months ago:

Adherence. Normal PLA/ABS or other plastic filaments will shrink when they cool down, and cause the thin layer of material to pull and separate from the copper. TPU is rubber-like and will have less inner stress to pull the material off. Also PLA absorbs humidity and will tend to change volume in a solution, with the same propensity to separate from the copper during the etching. TPU has more impermeability.

I have also tried using adhesion glues and hair sprays. Sometimes a thin application of hair spray (removed with a quick spray of isopropyl before the etching) helps a bit to define the edge traces. However I found that cleaning the copper very well, having a clean chamber in the printer and controlling well the etching temperature to reduce the etching time has the same effect and is less burdensome.

penagwin said 8 months ago:

Have you tried PETG? I know loves to adhere to glass (if rather aggressively) - maybe it'd work here too?

It'd be cool if it does because PETG is a lot easier to print because of its rigidity (especially in bowden drives)

pjbk said 7 months ago:

Good question. Not yet. When I carried out the tests (3+ years ago) I didn't have PETG filament around. Then TPU worked well for me and I have been sticking with it. I have a couple of PETG spools now so I will certainly try it next time!

beeforpork said 8 months ago:

OK, I am puzzled: 2D printing is a dark art, but 3D printing is not?

And comparing this to Chinese PCB manufacturing services is weird, too. If you want double sided, with vias, down to tiny traces, with plating and solder-stop and text, then you may be willing to pay the shipment and wait a few days.

Yes, I admit that this is a fun method. But what is the advantage over laser printing and then, say, using UV light to transfer to a photo PCB?

Honestly, I do not get why laser printers are a dark art, please explain!

bob1029 said 8 months ago:

I came here with the same question. I can see how the laser printer approach might have its own difficulties, but from a layman's perspective this seems like the most accessible & reproducible path.

hwillis said 8 months ago:

It's very much an art. Gotta have the right paper, printer, and temperature application, and there's no good way to know until you have a feel for it. Plus finding a laser printer is a hassle now.

jacquesm said 7 months ago:

> Plus finding a laser printer is a hassle now.

Huh? Aren't they a complete commodity? $250 will get you a half decent one.

StavrosK said 7 months ago:

I don't know about everyone else, because what you're saying seems to be a common sentiment and multiple people here mention "just" getting a $200+ piece of equipment, but neither the $200 nor finding the space to put the printer on is worth it to me for the ~1 PCB per year I want to make (and ~0 pages I want to print).

At most, I might go to a printing place and ask them to print a page, but my 3D printer is right there and this saves me a trip.

jacquesm said 7 months ago:

Right, but a 3D printer is :)! Bet that cost a lot more than $200.

StavrosK said 7 months ago:

I think it was actually right around $200, but I'm not saying people should get a 3D printer instead of a 2D printer, I'm saying that making PCBs with what you have is better than making them with something you have to buy. If you already have a laser printer, that will probably be much better for you (unless you have a 3D printer too :p).

This post is basically saying "if you don't have a laser printer but have a 3D one, you can still make PCBs!" :)

jacquesm said 7 months ago:

Fair enough. Someone should figure out how to print photoresist directly with an inkjet printer. That would be a neat trick.

StavrosK said 7 months ago:

Or even better, conductive ink! We probably aren't very close to that, but maybe a better etchant isn't too far off, and I'd settle for that.

kragen said 7 months ago:

You can buy conductive ink pens from Adafruit and many hobby stores.

kragen said 7 months ago:

There are pages about how to use inkjet ink as a PCB resist. Apparently it's a bit trickier than the laser process.

StavrosK said 8 months ago:

It's not the printing that's a dark art, it's the making a PCB by printing.

lstodd said 8 months ago:

It's fun to turn a 3d-printer into a plotter, but alas only for simple stuff.

A given-away laser printer (that is, $0) and $25 laminator deliver 15 mil / 0.4 mm traces easily and consistently.

I very much doubt any pen can do that, even with proper mount - the rubber band will just ruin any accuracy and consistency, and even if the pen is hard fixed, the tip will still wobble, plus always imperfect levelling will mess with the track width.

StavrosK said 8 months ago:

This is true, that's why v2 of my mount uses a zip tie! That one works much better, and I don't really need 0.4mm traces for the quick PCBs I need. If I need tighter tolerances, I can make the PCB this way, see that it works, change the tolerance and send it to a fab.

How do you use the laminator with the printer?

lstodd said 7 months ago:

1. Preheat the PCB with common iron. 2. Apply the paper to the heated PCB 3. Pass the sandwich through the laminator several times.

This ensures that the temperature and pressure are uniform across the entire PCB and toner is never smeared.

djmips said 7 months ago:

A laser printer isn't exactly free. You might be able to find one cheaply or free but you still have to pay for toner.

lstodd said 7 months ago:

Literally free on local equivalents of Craiglist.

If the toner gets low or photodrum dies, I just get another one.

armagon said 8 months ago:

Hmm... I thought there were special pens you could use to draw your own circuits.

Something like this, perhaps: https://shop.circuitscribe.com/

I think that'd be even cooler, as it would save a step and not involve etching chemicals.

teraflop said 8 months ago:

Yeah, I'd be really interested in hearing how well those conductive inks work when used for a complete project.

Based on what Circuitscribe's website says, their ink has about 100x the resistance of a standard 1oz copper PCB (for the same length and width), so I can imagine that you would have to be extra careful about how you route your power traces and position your decoupling capacitors.

StavrosK said 8 months ago:

Oooh, if conductive ink works well, that could be amazing! I need to look into these some more, thanks for the info!

45ure said 8 months ago:

I have used Circuit Writer and Caikote for fixing lifted traces and remotes etc. I don't know if it will be either feasible or cost-effective, for small to medium projects. Good writeup, by the way!


StavrosK said 8 months ago:

Thanks, and thank you! These are a bit expensive, but I usually only need one or two PCBs, so it might work. I'll give them a shot, thanks again!

alphabetter said 8 months ago:

It's not unknown for people to hand-draw PCBs with permanent markers in much the way you describe. It's not a new idea, but with 3D printers being common these days, it's nice to remind people of this approach to PCB manufacture.

Having made a few PCBs at home, my experience is the etching if the weak link. It's hard to get consistent, high quality traces - particularly if there is any kind of high density layout.

With some effort you can to double sided PCBs using this kind of technique, but any kind of through-plating has to be done manually.

pathartl said 8 months ago:

I have a very wonder Precision Fidelity C7A tube preamp and it has some very wonderful hand drawn traces: https://www.audioasylumtrader.com/images/y2016/06/156317/IMG...

Being from the 90's and only ever seeing CAD-designed PCBs, this blew my mind the first time I took it apart.

hinkley said 8 months ago:

A while back there was a discussion here about how crosstalk in rectilinear traces (I think there’s a term for this but this is how I think of them) are becoming a bit of a problem and that more organic looking lines (not unlike these) can help.

Audio people are notorious for detecting noise. That hand drawn circuit might be less artistic and more practical than you’d expect.

pathartl said 8 months ago:

I do remember that article, but what I took away from it is it only really happens when you're working on the small, extremely dense scale.

DanBC said 8 months ago:

I designed a couple of small PCBs using crepe tape and etch-resist transfers. Sometimes this was directly onto the copper, but usually it was onto acetate to be used with UV sensitive etch-resist.

It was fun, but kind of a pain.

Here's a very old (1966, before I was born) article giving instructions for how to do this: http://www.rfcafe.com/references/popular-electronics/etch-pr...

krallja said 8 months ago:

The 8-Bit Guy repaired a Commodore calculator that had what appeared to be a hand-drawn circuit board as well: https://youtu.be/Q0rSv8NfQXQ?t=65

StavrosK said 8 months ago:

> It's not a new idea, but with 3D printers being common these days, it's nice to remind people of this approach to PCB manufacture.

Sure, but I had neither a 3D printer nor magazines nor an iron, and I didn't really want to get into that whole process, as it sounded like a hassle. Exporting gerbers directly to my 3D printer is trivial enough that I can do it in one minute.

> It's hard to get consistent, high quality traces - particularly if there is any kind of high density layout.

That is very possible, I still need to refine my etching technique. It's certainly been the hardest part in the ones I made, but that doesn't say much because the rest of the process was trivial.

> With some effort you can to double sided PCBs using this kind of technique, but any kind of through-plating has to be done manually.

I really want to experiment with that as well, but drilling will be harder. I wonder if it would just be easier to route some traces to the edge and solder them there instead of drilling a via.

jgamman said 8 months ago:

Note to chemists: PCB is a printed circuit board not a polychlorinated biphenyl. i know i came for the wonder of atomic scale printers but still, printing circuits is cool too.

fest said 7 months ago:

Another way to step up the resolution a notch is using photoresist coated PCB and a small UV laser. It is typically a negative process (exposed photoresist dissolves in developer solution), so you'd have to invert your plot/gcode, there are two extra steps (developing the exposed photoresist so it dissolves and dissolving the remaining photoresist after etching in acetone bath).

However, that brings you closer to being able to prototype a PCB for something with 0.5mm pitch.

madengr said 8 months ago:

Back in the 20th century, we did this with flat-bed pen plotters. Ink makes lousy resist anyway. Much better results using old fashioned transparency and photo resist.

StavrosK said 8 months ago:

It wasn't too bad, as you can see from the photos. I think the problem was that I left it way too long in the etchant because I didn't have a copper plane, but I will try with one next time.

cronix said 8 months ago:

Why not attach a Dremel with some routing bits instead of the pen and route it?

throwanem said 8 months ago:

Two problems. First, it's going to be hard to get any kind of rigidity without heavy clamping that may damage the printer head. Second, the mechanism isn't designed to drive that kind of load. So you're unlikely to get usable results, and you are likely to destroy the 3D printer while you're trying.

If you really want to do that kind of work, you'd be better off with something like a CNC-equipped Proxxon micromill. But that's several thousand dollars' worth of investment for a tool that's honestly not good for much beyond what you're already doing with it - those things are tiny, so unless you're making intricate model parts or something like that, you're going to have a hard time finding uses for it that don't run up against the physical limits of what it can handle.

And, after all, we're only talking here about removing a thou or two of material that can be oxidized away just fine. Not that machining isn't an absolute joy, but a chemical process makes a lot more sense here.

cronix said 8 months ago:

I use a lulzbot Taz and the toolheads are easy to swap out. I've seen a dremel tool holder for it on thingaverse, so that would take care of the weight and rigidity since it would be on its own mount instead of mounted to the actual extruder.

What do you mean by "drive that kind of load"? Are you talking about the weight, or pulling the dremel through the material to cut it (resistance)? If the latter, it seems that just going slow with the steppers and a high speed on the dremel would eliminate that?

Other people have converted their 3d printers to CNC machines. I'm not sure why this wouldn't be possible with a little effort. It seems cleaner/safer than dealing with chemicals?


crankylinuxuser said 8 months ago:

That may be true, but 3d printer gantries are not built for the stresses of a dremel or cutting tool.

First, you'll get loads of wobble. Normally cnc machines use a 2 smoothrod with a leadscrew in the center for strong stability and smooth motion. Once you turn on a source that spins in X000 rpm, it will send motion through the whole system.

Secondly, once you start cutting anything stronger than foam, cutting will introduce shear forces on the gantry head itself. All 3d printers don't need to worry about XY rotational motion. That's usually a non-constrained degree of freedom.

Third, you need to mount the piece you're cutting or it'll move (duh!). But tabletop vises are HEAVY. Your bed probably can't support it. The only bed I know that could would be a 3 point Z leveling bed. Those're rare.

Feel free to try, and not take my word for it. But this has been attempted, and failed on many accounts.

However, this would be a good place to try. And this tool is good up to alu cutting. https://all3dp.com/2/mostly-printed-cnc-mpcnc-all-you-need-t...

ur-whale said 7 months ago:

Just get one of those:


They're really cheap and work perfectly for PCB routing.

StavrosK said 7 months ago:

$270 for routing PCBs isn't very cheap though, considering...

kazinator said 8 months ago:

Plus, if you fill all the empty areas with copper pours (see my other comment), there isn't that much copper to route away!

StavrosK said 8 months ago:

The printer doesn't have enough torque to dremel things, and fiberglass dust would make a toxic mess in the flat.

mighty_atomic_c said 7 months ago:

I've had some success converting an old printer to do pen plotting, but I haven't tried PCB manufacture yet because Sharpie marker does not do a sufficient, reliable job at working as a mask. I haven't been able to even use it to touch up my own projects!

My experience is mostly with warmed Ferric Chloride as the etchant, so perhaps ymmv with other etchants. However, in my case, glossy photo paper + laserjet transferred with clothes or hair iron (a tedious and awful process) works well, but Sharpie has not been able to help me touching up any areas damaged by the transfer.

If anyone else has a better mask suggestion than Sharpie, I could try out that sort of thing instead.

StavrosK said 7 months ago:

I used an edding marker (I mention the model in another comment) that worked excellently.

ur-whale said 7 months ago:

This does not solve what (to me) is the main problem how DYI PCB making: the chemical step.

For me, that is the number one things that needs to be solved: using nasty chemicals to etch copper away.

Not drawing on the copper. That's easy (and there's a million ways to do it).

djmips said 7 months ago:

If you're patient you can use vinegar, hydrogen peroxide, and salt. You still create a copper salt that you need to dispose of properly. http://quinndunki.com/blondihacks/?p=835

StavrosK said 7 months ago:

Etching the small PCBs I created took a minimal amount of chemicals in the cut bottom of a water bottle. It's not nothing, but it wasn't a big problem.

ur-whale said 7 months ago:

How did you get rid of the byproducts ?

osrec said 8 months ago:

Rather than using an etching solution, could you just use a Dremel style attachment to shave off the copper from the board to "reveal" your PCB design?

ur-whale said 7 months ago:

There's tons of videos and web sites on the net describing this, usually with tiny and cheap CNC machines.

It works very well, and there's no need to use nasty chemicals anywhere in the process.

JshWright said 7 months ago:

> and there's no need to use nasty chemicals anywhere in the process.

Yeah, instead you get lots of microscopic airborne fiberglass strands (a known carcinogen). Yay....?

waiseristy said 7 months ago:

That's another possibility. I don't think you can get as good of a resolution though. Also dremels are loud as hell

prashnts said 8 months ago:

Nice! I tried doing this a few years ago with a RepRap build. It didn't go well! Bad marker, pigment that dissolved, not enough ink layers, but the most interesting one was with Staedtler markers which sorta dissolve previously laid down ink when I made multiple passes. Ended up doing toner transfer after giving up.

I wonder how would a thin plastic film cut with one of those vinyl sticker cutters would fare for masking.

StavrosK said 8 months ago:

Oh huh, I'm using an Edding 140S, in case it matters. It was what they gave me when I asked for a permanent marker, and it worked beautifully. Didn't dissolve at all in the etchant, yet wipes right off with some alcohol.

prashnts said 7 months ago:

No doubt, I had no idea which marker to use. Thanks for the suggestion, I want to give it another go since toner transfer is often a hit/miss.

StavrosK said 7 months ago:

No problem, give it a shot, it worked really really well in my tests, and very easily.

armitron said 7 months ago:

I wouldn't describe the results you show in the photos as "worked really really well". The traces are uneven and full of holes and discontinuities.

To me that's completely unacceptable. I'd throw them in the garbage but I guess I'm too OCD and obsessed about quality to work with hand-etched PCBs.

prashnts said 7 months ago:

As long as it works or can be fixed with some bodge, I'm okay with that for prototypes. I do tend to tin the traces wherever they seem 'porous'. For broken traces, I use wrapping wire core soldered on tracks.

But yeah, that isn't pretty!

StavrosK said 7 months ago:

To me, the circuit did its job, so it worked really well.

itronitron said 7 months ago:

You can also use a Circuit Scribe pen [0] to draw the circuits directly without the messy step of acid etching the plate. Related [1].

0. https://www.circuitscribe.com/ 1. https://chibitronics.com/sailboat-card/

jstanley said 7 months ago:

I couldn't tell from their website: can you solder to this stuff? If so, it sounds incredibly promising, but if not, then I'm struggling to see how you use this to connect standard components together.

itronitron said 7 months ago:

Good point, the 'ink' is silver but there is no mention of whether it can be successfully soldered. I wonder if the ink lines can at least aid the flow of solder.

salamanderman said 7 months ago:

Taking a guess at the issue the pcb2gcode author is dealing with, they probably are trying to solve the same problems as this patent: https://patents.google.com/patent/US10061301B2

kragen said 7 months ago:

Homo Faciens made a page about the mechanical side of this process last year:


ur-whale said 7 months ago:

Just get yourself a tiny CNC machine from banggood and then do this:


No chemicals involved, no hacking a machine that wasn't designed for the task.

JshWright said 7 months ago:

No _liquid_ chemicals... Airborne fiberglass threads are a lot more dangerous.

Jemm said 7 months ago:

I used my 3D printer with a sharpie rubber banded to the carriage to draw labels on CDs for a friends band. Worked very well.

Can’t remember the exact software in used but it was a script that converted a png to gcode.

jpindar said 8 months ago:

This is not what I was hoping for (literally printing the traces).

falcolas said 8 months ago:

Given that there are wood and metal filaments, it may be possible (the flip side being the filaments are still only around 70% wood/metal, and are hard on extruders).

StavrosK said 8 months ago:

It shouldn't be too hard to print the traces, I just didn't try it. I don't know how well the plastic would adhere.

pjbk said 8 months ago:

As I stated above, use TPU or Ninjaflex filament. It adheres very well to a clean copper surface. Start etching immediately after printing.

StavrosK said 8 months ago:

You also mentioned an adhesive, what's that for? Also, how do you find it better than the marker method?

pjbk said 7 months ago:

It will prevent the plastic layers from detaching too easily. Specially if you do the etching in a warm bath. If you do it right, the finish is far superior than using a marker, in particular edge traces that turn out very well defined. That means it's easier to do fine pitch SMD pads, for example.

Through the years I have done many PCBs using different methods: Sharpie markers, Posca pens, inkjet and laser printing, transfers of different materials, etc. Except perhaps for pro laser etching or routing, nothing beats photomasks. In my experience the quality I get with 3D filament compares to a DIY routing table, so if you have a 3D printer and don't mind chemical etching, you don't need one.

StavrosK said 7 months ago:

Hmm, that's interesting. I'll give it a shot, thank you.

convolvatron said 8 months ago:

another idea that's on my list...apply a ground (as in classic copper plate etching). use a scribe to remove the negative and etch. my hope is that it will produce higher precision than a felt-tip for small pitch components

but I still haven't gotten over the activation energy of getting kicad output into gcode for the dirt cheap grbl machine I got

StavrosK said 8 months ago:

Can't you just use my code? It should work well.

krallja said 8 months ago:

How does copper cladding work? That is, how do the manufacturers get copper onto the FR-4?

lux_scintilla said 8 months ago:

It is glued down. The fiberglass (FR4) is filled with glue, the back of copper foil is lightly roughed up with an oxidizer and then the materials are laminated together.

skykooler said 8 months ago:

I believe it's actually done the other way around - the "substrate" is deposited on copper foil.

qaute said 8 months ago:

There are other ways to make PCBs with a 3D printer:

Instead of etching copper, some people directly print the circuitry with a solder extruder [1]. The idea's been around for a while [2] but circuit complexity is usually very limited. Here's a guide [3] (to a similar method) that uses a hobbyist 3D printer.

To improve on the article's pen-masking method, you can mount a laser instead of a pen to the 3D printer and expose a specially coated PCB [4]. Or expose UV light through an LCD [5].

One can also just mount a milling tool to the printer and cut away the copper directly [6]. A 3D printer's not designed to take the forces from milling well, so similar but specially designed machines are made [7].

The most impressive methods are geared toward industry:

A cutting-edge industry-grade electronics 3D printer looks like [8]: an inkjet-style printer with conductive and insulating (dielectric) inks.

Somewhat related, you can use a laser etching and electroplating process ("3D Moulded Interconnect Devices" "3D MID") to make PCBs on 3D printed weirdly shaped (i.e., definitely not flat) surfaces. [9] is an impressive example, and definitely check out a search engine's image results [10].

[1] http://diy3dprinting.blogspot.com/2015/01/voxel8-conductive-... [2] http://blog.reprap.org/2009/04/first-reprapped-circuit.html [3] https://www.instructables.com/id/3D-Printing-3D-Print-A-Sold... [4] https://dangerouspayload.com/2017/12/17/pcb-uv-exposure-scan... [5] https://www.youtube.com/watch?v=vxl7glJMKOQ [6] https://www.instructables.com/id/PCB-Milling-Using-a-3D-Prin... [7] https://www.bantamtools.com/pcb-milling-machine [8] https://www.nano-di.com/dragonfly-pro-3d-printer [9] https://www.festo.com/group/en/cms/10157.htm [10] https://duckduckgo.com/?q=3d+moulded+interconnect+devices&t=...

yezr said 7 months ago:

lol and how can you not enjoy the surprise ending!

poseid said 8 months ago:

simple and beautiful!

gswdh said 8 months ago:

Is there any real value in PCBs like these these days? 2 layers from China can be fast and is very cheap. You can’t do anything complicated on these single layer DIY boards, they also carry the risk of defects potentially causing a lot of frustration.

NikkiA said 7 months ago:

There is mild value to photo-resist processing DIY, 2-sided is easy and it's a lot quicker than waiting for a shipment from china.

But 3d printing, CNC routing, all those methods are an order of magnitude coarser production and not suitable for anything except very simple projects, and veroboard/stripboard is about on-par signal integrity wise.

pixelface said 7 months ago:

one niche where i'd argue it's not only valuable but also period correct is restoring and modifying vintage electronics. Frequently they are single sided through hole boards that were not born from a CAD program, and are not available for purchase. You can certainly make the new board with CAD and a fab house, but for things like vintage synthesizers a hand etched board may actually be more appropriate and original.

Also the importance of latency in some cases is significant. When the goal is to iterate a board quickly several times and a diy board is good enough, being able to have a board in an hour or less could mean multiple board iterations in one day where sending out to fab might draw out each board iteration to a week long wait.

krallja said 7 months ago:

Do people still do wire-wrap prototyping? Strong enough for (tiny-scale) production, yet no chemicals, soldering, or drilling required.