Last chance to enter The Hackaday Prize.

Testing The Limits Of Home PCB Etching

PCB

[Quinn Dunki]‘s Veronica, a homebrew computer based on the 6502 CPU, is coming along quite nicely. She’s just finished the input board that gives Veronica inputs for a keyboard and two old Nintendo gamepads. [Quinn] is building this computer all by her lonesome, including etching all the PCBs. She’s gotten very, very good at etching her own boards, but this input board did inspire a few facepalming moments.

In an earlier post, [Quinn] went over her PCB etching capabilities. As demonstrated by the pic above, she’s able to print 16 mil traces with 5 mil separation. This is just about as good as you can get with homebrew PCBs, but it’s not without its problems.

[Quinn] is using a photographic process for her boards where two copies of a mask is printed on an acetate sheet, doubled up, and laid down on a pre-sensitized copper board. The requirement for two layers of toner was found by experience – with only one layer of toner blocking UV light, [Quinn] got some terrible pitting on her traces and ground planes.

Two photographic masks means the masks must be precisely aligned. This example shows what happens when the acetate sheets are ever so slightly misaligned. With a 5 mil gap between traces, [Quinn] needs to align the masks to within ±2.5 mils; difficult to do by eye, and very hard once you factor in flexing and clamping them down to the copper board.

Even when this process goes perfectly, [Quinn] is pushing the limits of a laser printer. When printing at 600 dpi, the pixels of the print are about 1.5 mils. While GIMP, printer drivers, and the printer itself have some fancy software to help with the interpolation, [Quinn] is still seeing ‘bumps’ on the edges of perfectly aligned parts. This is one of those things that really makes you step back and realize how amazing fabbing PCBs at home actually is.

With most of the hardware for Veronica out of the way, it’s just about time for [Quinn] to start programming her baby. We’re not expecting a full-blown operating system and compiler, but those NES gamepads are probably crying out for some use.

Comments

  1. James says:

    5 mil is seriously good even for pre-sensitised boards. Especially with doubled artworks, Quinn’s hands must be a lot steadier than mine to align those two artworks so precisely.

    I use dry film and single tracing paper artwork but I seldom bother with 10 mil separations let alone 5. Better things to do than try and find and fix hairline shorts.

    • fartface says:

      Actually get yourself a good stereo scope for SMD work. you can align it quite easily that way when you have magnification. The hard part is you need to find a light source that you can use that will NOT start exposing the boards. I discovered that the LED light source on my surplus unit was putting out enough UV to actually start exposing boards. I added two panes of standard window glass to solve that issue. (Glass blocks UV significantly) But I need to build a better more board safe light source for the bench anyways.

  2. These “bumps” on the edges might not have anything to do with the printing, but rather be and artifact of the etching process.

    • I suspect so. In some of the boards I’ve etched, even though the transparency has a pretty dead-straight edge, I still see little bumps. I wonder if it’s because the copper is maybe a slightly different thickness over the strands of the fiberglass weave in the board material.

      Branded toner, although more expensive, is definitely more opaque (at least HP is). I used to be able to expose boards with only a single transparency when I was using HP toner, but now I definitely need two transparencies with cheapo toner.

      I don’t know about track spacing, but I was able to etch down to 4mil _wide_ tracks when I did a test ages ago –

      http://imajeenyus.com/electronics/20130322_making_pcbs/photos/smd_and_tracks.jpg

      Note, the transparency for this was printed on a big office machine capable of at least 1200dpi. No way would I actually trust a 4mil track though! Generally I go for 24mil for simple boards, 16mil if I need to get between pads, and 10mil if I’m really stuck.

      • Andy says:

        I used to have this problem alot when screen printing. Having to use two sheet per print with my laser, to get a dark enough mask for screening. I switched not long after to using velum or semi transparent tracing paper, It holds ink/toner so much better. Not only that but it has a more even tone all over when you hold it up to the light.

        The stuff i use is Daler Rowney tracing paper, it’s 90gsm and work well in laser printers ( just remember to set the printing option for velum/tracing paper or plane paper if there isn’t one. For the exposure time it’ll take slightly longer but that’s generally because it’s opake, I always use a glass pane over the paper to make sure that the mask is super flat against the surface to give tighter/sharper lines.

    • S says:

      That certainly is a problem with immersion etching. Can be avoided using bubble agitation in a vertical tank or better with spray etching that allows finer track spacing.

  3. Harry M says:

    Surely the pixels at 600 DPI are 600/2.5/1000 = 0.24mil

  4. Harry M says:

    My bad ignore that…..

  5. 0xfred says:

    Inkjet printers are so much better than lasers for photo exposure. I use a Canon MG5250 set to T-shirt printing mode and get a lovely thick black print that’s perfect for UV exposure without doubling up. I’ve been getting TSSOP happily with my first etching attempts and going to try LQFP soon.

    The uneven edge to the traces is probably not due to the exposure but needs tweaks with the etching process. Slow etches can cause this.

    • Ace says:

      Agreed. I always had trouble with the way the fuser of a laser printer distorts the transparency, especially when doing double sided boards. I use an HP Deskjet at it’s highest DPI and turn the black way up and I can make 8 mil traces. Also I don’t need to double up. I also agree slow etch causes problems, that’s why is so important to use warm etchant and agitate constantly, using a foam brush to wipe down that board now and then.

  6. ioch says:

    Using GIMP might be the cause of “pixeling” as it raster graphic program. Exporting as PostScript and feeding directly to printer might increase precision. Even if printer is said to be 600dpi normally it can be better than that.

  7. Somun says:

    Toner transfer method is easier and once you get the hang of it, more capable. You need a laminator for that but still beats the uv exposure equipment, size and cost-wise.

    • jpa says:
    • James says:

      There is a very highly collimated and extremely powerful UV source available completely free of charge… most people call it The Sun.

      • And totally uncontrollable as well ;-)

        To be honest, I really don’t understand why people fart around with all the different techniques of doing circuit boards. If your aim is to produce good boards, which I assume it is, then spend a little bit of money to get a UV exposure setup, learn the developing/etching chemistry, and you’re set.

        It’s repeatable, accurate, and lets you do double-sided boards relatively easily once you make an alignment jig. From start to finished board, maybe 30min if you’re lucky.

        The presensitised board lasts forever – I’ve got 4yr old stock and it still works. Maybe a little bit longer in the developer, but it’s still fine.

        Toner transfer – some people seem to have success with this, but it’s just too variable. Paper quality, toner quality, how long you heat it, board prep blah blah. Too much trouble. It’s OK for a quick rough board, if you don’t mind spots and doing a bit of touchup afterwards.

        • jpa says:

          Getting a good printout for exposure is not really that simple either. My current laser printer is spraying around the toner too much.. And different brands of transparencies also act differently.

          So no DIY method is really 100% controllable.

        • fartface says:

          Dude for prototypes it’s fine. For single boards “controllability” is worthless, I can get a solid exposure every single time for a single prototype with no special setup. For more than 1 board, I dont even bother, it’s far FAR cheaper to simply have sparkfun do a run of 10-20 for me, plus I get a better board with silkscreen and soldermask.

        • sneakypoo says:

          Well the same goes for toner transfer, spend a bit of money on proper paper and a cheap laminator and you’ve removed a lot of variables. And I don’t have to worry about boards getting old, basic copper clad lasts forever too :)

          That said, I very very rarely etch at home these days, it’s so damn cheap to get real boards.

      • fajensen says:

        That so? Round these parts we have almost forgotten that celestial phenomena even exists. Solid cloud cover at 150 meters and mist for 3++ weeks.

    • 0xfred says:

      Toner transfer is the Arduino of the PCB production world.

      • George Johnson says:

        I have tried and tried, and never could get that stuff to work. I could only get a partial board at best. I finally gave up after about a year of trying.
        I have been using the regular florescent light method and MG chemicals for some time without a problem. But don’t generally do really tight spec boards.

        • 0xfred says:

          In case my comment was ambiguous I was implying that toner transfer is what people who don’t know what they’re doing are drawn to but frankly long of amateur. UV is definitely the way to go.

          • Totally true. Anyone who knows what they’re doing knows full-well that the board fab houses use photoresist! Lucky me, saved me a bit of scratch when I wanted to get into screen-printing stuff; I already had the photo exposure gear. :D

            I’ll tell you this though: I sure wish my Silhouette SD (CraftRobo) could cut Rubylith fine enough to make film negatives for doing PCBs, it would sure beat the hell out of laser transparencies!

          • tekkieneet says:

            You may say breadboard or perf boards are the “Arduino” of prototyping.
            Toner transfer is the exact opposite as “Arduino” as you have to have
            skills and understand the limitations of what you can do consistently.

            I have been getting good enough results with toner transfer. I
            regularly do double side board at 10/10 and even sneak in some 7/7
            tracks. My routing limit is on getting small vias for breakouts and not
            on track width and spacing. What’s your PCB geometry?

            I have the MG Chemical kit with their “exposure light”. There was a
            limited an improvement to use photoresists. That was just a metal stand
            added to a standard kitchen light fixture sold at the hardware store for
            $10 more.

            I have been shown a more professional setup in my University. I had my
            layout printed on film and they use vacuum to clamp down for exposure.
            They use a spraying etching tank. Now that’s a system I would switch to.
            Anything else is half-ass and not worth the effort.

            My rework skill is more than enough to make up for any short fall of the
            toner transfer process. I have taken off and reused 0.5mm TQFP.

  8. Mikes electric stuff says:

    Clear transparency films are not ideal as toner density is often inadequate – tracing paper works much better – use thick >90gsm stuff to avoid crinkling. 10/10mil easily achievable with no special care.

  9. strider_mt2k says:

    I stand in awe of folks who can do this kind of stuff.
    Very impressive!

  10. Nik B says:

    canon IP4200 inkjet + overhead projection foils + UV light => 0.10mm trace/ 0.15mm gap. The aligning however is a whole different question, I wonder how she got such good results! http://www.flickr.com/photos/67917568@N00/2221198596/

  11. Quinn Dunki says:

    One small correction- I believe the visible pixel features are around 0.25mils.

  12. Dax says:

    How much time before someone figures to use an old overhead projector to optically shrink down the transparency projection.

  13. Joseph says:

    What about using TRF instead of a second toner layer? Since it sticks directly to the toner it seems like that would solve the alignment problem but I don’t know if it would work with toner on acetate. I only have experience with the toner-transfer method using PCBs.

  14. Dave says:

    Had to comment about that facepalm picture. Don’t be too hard on yourself. This happens even at multi-national design and build OEMs. I can tell you from experience!

  15. 16/5 thou is not pushing the homebrew limit yet guys.

    8/8 is routine for a few of the guys on homebrew-pcb.

    There is a standard test dj delorie has for anyone that wants to test themselves

    http://www.delorie.com/pcb/spirals/

    Using a multimeter will let you know if your 8/8 really ended up 6/10 even :)

  16. George Johnson says:

    There’s a couple things to look at here. One is undercutting while you’re etching. That’s where the etchant actually starts eating away the copper UNDER the resist. I wouldn’t expect that on such small dimensions though. You normally see that when you’re etching large areas, and have small areas exposed to the etchant for a long period.

    Another is pretty much the same thing, but it’s light exposure during the PCB exposing. Light getting under the transparency.

    That’s what I would suspect. If you have a flat piece of glass, and you clamp the corners and edges down, you’ll actually bow that glass out. So that there’s actually less pressure in the middle, than at the edges.
    Perhaps if you examine the boards close, and compare the outer areas, vs/ the inner areas, and see how much difference there is, that may give you a clue.

    I used to do this a while back, and used a lithograph film. So you only needed the one transparency, exposed the film, and then used that to expose the board. Because the film was more of a “binary” exposure, it either came out black, or clear (opaque really). That took away all the printing difficulties. But I can’t remember the type of film I used. Pretty easy to work with once you get the exposure down right. Didn’t have to be controlled down to the micro second or anything.
    You could try some print shops in your area, to see if they can come up with a name. But I just can’t for the life of me remember who it was. But that solved a lot of my problems like that.

    Without the film, the only other suggestion I could offer, is exposing the PCB in different ways, in different areas. Less light, or shorter exposure, in tighter areas. So you get less “leakage” under the transparency. But that would be a big pain.

  17. George Johnson says:

    Oh yeah, also matters which side of your transparency is next to the PCB. If the “ink side” is up, then you have a layer of clear film, in which light can “leak”. You need the “ink layer” next to the PCB so the light can’t leak under it.

  18. 0xfred says:

    I’ve been doing a bit of experimenting and research on PCB production myself. A new kid has prevented me getting round to documenting it properly. I hope it’s OK to summarise here.

    * Inkjets can create great transparencies, especially in T-shirt print mode.
    * Zip lock vacuum food bags can be used to press the transparency down very evenly on the board.
    * Standard fluorescent lighting has worked better for me than UV tubes. Not sure why yet.
    * Don’t bother using an ultrasonic cleaner to agitate your etchant. It has almost no effect.
    * Holding your PCB with a cheap electric toothbrush helps a bit.
    * Dynamask 5000 dry film solder mask will greatly improve how professional your board looks.
    * Cheap eBay USB microscopes are surprisingly good for inspecting your etching or SMT soldering.
    * SMT really isn’t as hard as it might look. Try it and you won’t go back.

  19. tekkieneet says:

    From the signal integrity point of view, those “fill plane” that are
    only connected on one side serves no purposes when you are driving it
    with a digital signal (i.e. low impedance). Capacitive coupling or
    conduction by contaminants is not significant when you are dealing
    with a signal that is a few volts and tens of ohms of impedance.

    If you are worrying about crosstalk, either connect *both* sides of the
    “fill planes” to power/ground so that there is a path for the return
    current to flow.

    You can also leave them out entirely and increase gaps between signals.
    This way you can get higher tracks density without pushing track to
    track spacing limits on your board process.

    I would thicken power/ground track that are actually connected on both
    sides since now there are more space on the board.

  20. Alien Spy says:

    and what about CNC for milling boards as a comparison? Anyone found the limit using that yet? and what setup have you tried: CNC machine, mill bit, copper board, etc.

  21. Anonymous says:

    Would MICR toner (which contains metallic dust) be dense and consistent enough so you only need to make one mask?

  22. Cazna52 says:

    I wonder if you could rig up something like a photographic enlarger except with a lens reduce the projection instead of enlarge it and with a UV light source? Optically reducing the pattern from a 600 dpi print could result in much sharper and more precise etches.

    The other problem you get though which I have gotten before is if the depth you are etching is relatively large compared to the width you end up getting etching under the masked area. Like this in cross section “`/_\“`. My experience was on a much larger scale doing deep etching in brass for a boilerplate type effect but the same thing may hold true on a tiny scale if the thickness of the copper layer becomes significant proportional to the width of the gap. Perhaps a board with a thinner copper layer could reduce this effect but then you would be increasing the resistance of the traces.

    • Mike Dodd says:

      I’m kind of looking at this, but with a variation – print out the PCB on whatever printer at largest supported scale (i.e. scale to page) then photograph this onto fine grain B/W film, process the negative, then use an enlarger to project this onto either lithographic film, or possibly even direct onto the presensitized board. I think your post was what brought me to this site with a google-hit.

      Anyone any experience? if not, it should be a week or so until I can report any success

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 91,149 other followers