Modding An Inkjet For PCB Production

Like all of us, [Ryan] is tired of waiting for board production houses. To reduce some of that turnaround time, he modded an Epson inkjet into a PCB printer. The Instructable of his build is extremely thorough and it looks like he’s getting some quality boards out of his project

The build started off by disassembling an Epson C86 printer he had lying around the house. Going with an Epson printer is important – Epsons have a piezo print head accepts ink that would clog other printers. After tearing all the plastic off his printer, [Ryan] set to work raising the printer (or lowering the bed, whatever) and was off to the races.

The cartridges were filled with etch-resistant yellow ink and a piece of copper clad put onto the printer. After printing, [Ryan] etched his board in ferric chloride. Sadly, he’s getting small pinholes in his traces where a bit of the ink was eaten during etching. He’s tried HCl and Peroxide, but those turn his boards into green junk.

If you’ve got any tips to help [Ryan] out, leave them in the comments. Before that, check out the printing demo [Ryan] put up.

49 thoughts on “Modding An Inkjet For PCB Production

  1. I dont know if is already doing this. but when i exprimented with printing directly to pcb i got the ink to stick better by pre-etching the board for a short while. The board turns really pink but it loks kind of pink in the video so maybe he is already doing this

    1. Exactly. I don’t know why anyone would bother with ferric chloride these days. Ammonium persulfate was available and cheap when *I* was making my own boards (which is damn near 20 years ago). Ferric chloride is slow, nasty stuff to work with.

  2. I’m not sure if printing the etch mask on is the best way to go about using a printer to fashion pcbs.

    There are a couple things I’d like to see tested:

    First is using the printer to print etch fluid in a reverse mask. Might be worth testing if you’ve got an accurate print head, although it would probably kill the printer and be really slow to etch (if at all).

    The other option I’d like to see explored would be directly printing copper to a substrate. I’d imagine it to be quite hard as not only would you need an incredibly fine mix of copper, you’d also need some way of affixing it to the board, either the solvent solution the copper is in reacts and turns to glue, or something that reacts to copper.

    Still, I’m impressed by the work that [Ryan] has put in, rather detailed instructions.

    1. ive read a paper a while ago where they used copper nitrate and ascorbic acid (vitamin c) to produce copper traces.
      in a 2 part process the traces would be printed with copper nitrate first then with vit c. the vit c would then reduce the copper nitrate to copper. ill try to find the article in question and post it here..

  3. There has been a lot of research into printable electronics, including the components. And as I understand it, flexible circuits, like you find connecting to an LCD through the hinge of a laptop, are printed. I’m not sure exactly the process though.

  4. if we assume ryan tested the etch-resistant ink and the etching fluid before embarking on hacking the printer, then surely the fault lies with the way the printer lays down the ink? multiple passes? thinner/thicker layers?

  5. What i would like to se is using a uv-laser in combinaton with a prinetr / scanner / cnc.
    and use that with a pcb with etch-resistant.

    and so print the parts that we are going to etch-away. would need some cover to not expose the rest to ligth

  6. There are some groups on yahoogroups that experiment with all of this kind of stuff, printing traces directly to a substrate, using inkjets etc.

    I believe the preferred method of stopping the pitting is to double print the board. also heating the board once the print is on there helps, it cures the ink ready for etching. perhaps he should try FE CL as his etchant, either way I would use the sponge method (its on instructables) to etch the board. Uses very little fluid and you control the etch as you can see it happening, you’ll know if the etch is good within 5-10 minutes max and you don’t need to warm the fluid.

  7. The shack hackerspace in Stuttgart, Germany successfully prints and etches pcbs with an inkjet printer (of course an Epson). We even did smd boards instead of through-hole with that thing.

    The difference is, that we did not have to modify the printer. It is an Epson photo R800 and as the name says, it is build for photo printing and is able to accept the usual copper clad board without modifications. Oh and our etch-resist ink is black instead of yellow.

    Funny thing though is, that the inks original purpose was to print labels for front panels (metal ones) and someone discovered that this stuff is etch-resist. Oh and we’ve got all the stuff from a fairly local board production house for testing purposes.
    for more Info visit
    (i hope all the necessary pages are in english)

    1. Oh almost forgot: The ink we use rocks with ferric chloride but sucks hard with sodium persulfate.
      On the wiki page is a link to the manufacturer of the ink, but the page is definetly german.

      We had huge trouble producing smd pcbs until we started using ferric chloride.

    1. try searching it without “Triton” but just with “N-GEN XN 00070 Black”

      I’ve found a supplier over here in Germany, but the 250ml bottle runs for 220 EUR. Don’t know if there are any suppliers in the US. We got ours from

    2. normal ink will not be a good etch resist since it’ll wash off (it’s usually water based).

      as for ordering the ink i honestly have not checked yet since we have another source for it.
      the full name of the ink we’re using at shackspace is: Tritron N-GEN XN 00070 Black

  8. I used ferric chloride in the past, with hand drawn traces, protected by using permanent marker. Sometimes acid would eat through a little, but not if you cover the trace area with a marker twice.

    Therefore he could try to use a cnc setup to simply draw his traces on a PCB with a marker. And then etch that.

  9. It might be worth trying a salt water etch with some modifications. Use only a partially saturated solution rather than saturated, and use low voltage as 1 1/2 VDC. It would be less aggressive. Also, as aeiah suggested, a double pass might help if registration is not a problem.

  10. At shackspace we’re using a similar approach to print our PCBs.
    We’re using a heat drying ink and Fe-III-Cl to etch the board. We’ve found that Sodium Persulfate is no good (at least with our approach) and had similar results (the goo) using peroxide which just seems to oxidize the copper but not in any way that yields results.
    More details can be found in our wiki linked above.

  11. The boards should be well-cleaned and degreased before inking. The etchant tank should be constantly rocked (or agitated, or bubbled) throughout the etching process, that shortens the etching time greatly. [Heating the solution also helps, but that’s a pro app – the benefit out is < the effort in.]

  12. Hey everyone. Thanks for the suggestions and ideas up and down the comments. I’m still working on the project and I’m going to try a few different things and see if I can increase the quality of print and remove the pinholes. My first plan is to try and use a more controlled or faster etching method so I’ll let everyone know how that goes in the next few days. Otherwise keep the suggestions coming as that was my goal in posting everything to get as many suggestions as possible and get this process working as perfectly as I can.

    Also to randy: you can buy or build a chip resetting device that can by-pass that problem.

  13. By using an open source printer driver like gutenprint you can probably adjust the dot pattern to be more regular. Normal printing uses a randomized dot pattern that looks better but might levae holes. Also coverage could be increased when slowing down the printing by a lot, so the ink can dry between passes.

  14. Some ideas that I learned from building a DIYDTG:
    – cure the ink by heating it (curing temperature and time depends on the ink)
    – print at as high resolution/quality as possible, this slows down the printer and puts down more ink
    – make sure your print surface is ~2mm from the printhead, too far will make the print “fuzzy”, too close risks a collision with the printhead
    – printing while the surface is hot should help the ink stick better
    – printing more times rather than 1 really wet layer ends up having a sharper image (if you can get stacking prints)

    Sorry for repeating some the ideas in the comments. I haven’t tried doing pcbs yet, maybe I’ll document the second printer I build.

  15. I’ve recently seen something similar on this PDF pages 16-18:
    (warning it’s about 100 MB)

    Don’t know who actually did this, or if there is some more info about it. They used some sort of wax instead of ink. The results look much better than Ryans. I got the link from Metalab (a hackerspace in vienna) so maybe someone on their mailing list ( knows more.

  16. Anybody heard of Tec-200 sheets? They are clear plastic sheets that you laser print a PCB image onto. Then you use an iron to transfer the toner to your PCB. Place in acid and presto!

  17. Try a different container of ink? An improper mix of pigment and base will result in an uneven print, and with printer ink quality standards are more for looks than for being used to protect a circuit during etching – I had a similar problem using an ink transfer to protect a metal blade while etching, I’d suggest trying another brand or even from a different supplier, some batches are thicker than others.

  18. I tried the same thing with an Epson Artisan Photo Printer. However, I got the same thing, lots of pin-holes in my circuit traces. I finally tracked it down. The Epson printers fire tiny amounts of all colors when you print black, even if you force all colors to black. I was never able to stop this and had to get it working by replacing ALL colors with the yellow ink. That was expensive but it would work. You should check to see if it was the same problem.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.