Connect with us

laser ablation of PCB resist

Discussion in 'LEDs and Optoelectronics' started by (*steve*), Feb 19, 2014.

Scroll to continue with content
  1. Fish4Fun

    Fish4Fun So long, and Thanks for all the Fish!

    462
    105
    Aug 27, 2013
    *steve* Thanks for the Link! I "joined" their forum and posted a newbie question about the suitability of a Laser Diode for paint ablation....we will see what kind of response a 1 post newbie receives ;-)

    While I know almost zero about lasers, I do have a pretty good handle on driving LEDs...what I don't know is how much power is required to efficiently ablate paint...If it can be done with a <5W Laser diode then I am willing to give it a go! Even if the laser has to be pulsed, developing the driver will be faster than attempting to modify a laser//Ink Jet printer for direct to PCB printing...though direct printing has always been a fantasy...well that and a process that directly prints "conductive traces".....lol

    Fish
     
  2. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
  3. Arouse1973

    Arouse1973 Adam

    5,164
    1,080
    Dec 18, 2013
    Wow cool Steve :)
    Adam
     
    Last edited by a moderator: Apr 6, 2014
  4. Fish4Fun

    Fish4Fun So long, and Thanks for all the Fish!

    462
    105
    Aug 27, 2013
    Yes, if those boys and girls in England could just catch up with Jennifer Lewis in Chicago my life would be a lot easier....http://www.matse.illinois.edu/faculty/Lewis.html

    I have been following her work for a while, "conductive inks" are all the rage on the unlimited budget side of macro-economics, bit frustrating on the DIY side....I follow the chem, get all the notes, help etc, but at the end of the day, direct print PCBs for the DIYer are still in the future......somewhere behind the 3d printer revolution coming to your house in the next few years...

    My 1st post @ Laserpointer <dot> com is going to be a bust :) Too much too fast on my part....good folks by all appearances, but I utterly failed to "follow the learning curve"....I have little interest in Lasers aside from the PCB photo-resist ablation perspective....

    Steve,do you have any thoughts on the viability of using something like this: http://www.ebay.com/itm/2W-445nm-M1...909?pt=LH_DefaultDomain_0&hash=item2a2158fe9d for PCB paint ablation? I have a CNC stage I could mount it to, and building a pulsed driver shouldn't be hugely complicated.....just don't know if 2.8W is even in the ball park....

    Fish
     
  5. jpanhalt

    jpanhalt

    426
    4
    Nov 12, 2013
  6. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    That should be fine.

    The microslice can do it with a 250mW laser diode! https://www.kickstarter.com/project...ini-arduino-laser-cutter-and-eng/posts/740663
     
  7. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    Here are some very quick results from a recent test.

    I've gone back to the yellow paint I used initially and tried at various power settings for a single pass.

    Here is the result for 35%, 30%, and 25% on the particular unit I have access to.

    IMG_5955b (Medium).JPG

    Clearly 30% (the middle one) is the winner.

    This gives me a baseline. Multiple passes at lower power seems to be the way to go, and it is what I will be testing next time.

    The laser focus seems critical. I had the laser focused before I did this, so this is as good as it gets focus-wise.

    OK, so what did they look like before etching?

    Before etching (the 3 that I showed above are the top left three -- only partially visible). I thought these were done at fat too high a power.

    IMG_5943b (Medium).JPG

    And here are the ones done at very low power. You can see the cuts are a lot thinner, but you can also see that the laser wasn't working correctly on the one to the right. There were 14% and 13% respectively. The laser is not guaranteed to work under about 20%.

    IMG_5945b (Medium).JPG

    After etching the three I showed above looked like this:

    IMG_5951 (Medium).JPG

    Not exactly impressive is it?

    So, I've got more work to so, but I have a baseline to work from.
     
  8. BobK

    BobK

    7,680
    1,685
    Jan 5, 2010
    What size are those traces?

    Bob
     
  9. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    The tapered traces go from 1mm wide to 0mm wide over 10mm. The bars on each ens are 1mm wide by 4mm in length at one and and 5mm in length at the other.

    The serpentine path was a 0.2mm trace, but in the examples which etched nothing remained of it.

    The two pads are 2mm wide.

    I told the software that the cutting path was 0.05mm wide. Clearly it's more than that. Part of this test was to determine exactly how wide. Measurements of the tapered tracks will tell me that exactly. And I'll do that later...
     
  10. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    Last night this process was used for real when a colleague needed a board.

    I have some pictures to post later.

    Drawing the board took the longest (about an hour -- there was no schematic!)

    Converting the drawing to vectors took about 20 minutes (it probably should have taken 5 minutes)

    "Cutting" the pre-painted board took about 30 seconds (after a few minutes of set-up time)

    Etching in warmed Ferric Chloride took 8 to 10 minutes. Nobody had remembered to bring a toothbrush to clean up the tracks, so the etching took a while to start.

    Cleaning of the board took another 5 minutes or so (after finding a suitable solvent for the paint)

    Drilling took about 15 minutes (we had to search around for the right size drills)

    Soldering took about 30 minutes, mostly because we had to rewire stuff.

    All up, less than 2 hours of "work" from idea to final construction.

    This was a single sided, through-hole board, about 6cm x 4cm. Sure, we could have done it on veroboard, or even dead-bug style, but there are several more of these to make up and the PCB will allow less skilled people to contribute at various stages.

    W00t
     
  11. Arouse1973

    Arouse1973 Adam

    5,164
    1,080
    Dec 18, 2013
    Look forward to seeing the photo's
     
  12. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    I'm a bit like a one-armed paper hanger at the moment. I have some "live" photos of it on my phone, and I'll grab them as soon as I can have time to scratch.
     
    Arouse1973 likes this.
  13. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    *finally* some pictures of a real life board produced using this purpose.

    This is the second real-life board, I have photos on my phone of the first, but this one was actually planned rather than being drawn up on the spur of the moment :)

    The board itself is a combination of a buck SMPS and a flash trigger.

    Here's the Circuit

    FlashPSUTrig.png

    I don't have the PCB layout or the dxf file produced from that to drive the laser printer at this stage. Let me know if you're interested.

    The board was painted with two light coats of water based matt black acrylic. (I wanted an even coverage without runs). Due to time constraints the paint was dried using a hairdrier.

    The laser cutter is an LG500. The settings are 100mm/s at 90% power with 45% corner power. These settings are significantly above the minimum required, but will ensure a good cut even if there is a blob of paint.

    The width of the cut was assumed to be 0.4mm and the design rules were such that the closest permitted spacing was 0.45mm.

    Three sets of milling files were produced, for cut widths of 0.4, 0.7, and 0.95mm. These were later combined so most paths were cut three times producing a wide clearance. (Incidentally, you might ask why those figures? the answer is that 1mm produced a weird cut that missed a lot of the pads, and 0.4/0.95 seemed to risk a small gap between the cuts (in retrospect it wouldn't)).

    I don't have a picture of the board prior to etching, but here is the etched and mostly drilled board with the paint still on it.

    IMG_6292b.JPG

    The board here is untrimmed. the extent of the board was marked by a very thin line around the edge. Part of it survived the etching.

    Holding it up to the light it looks like this:

    IMG_6294b.JPG

    I'm pretty happy with that. Spot the hole I didn't drill!

    After trimming I made sure that everything actually fit.

    IMG_6295b.JPG

    And as an added bonus the box fitted like a glove (I was worried I had not allowed sufficient for the angled sides of the box)

    Here's another view of the other side of the board prior to removing the paint, and with the holes drilled to their final sizes.

    IMG_6298b.JPG

    And then after cleaning

    IMG_6299b.JPG

    And then after the surface mount components were soldered in place:

    IMG_6300b.JPG

    Yeah, I know, it's a very poor job. It's a very good idea not to leave your solder paste out with the lid removed. It was, shall we say, "non-optimal" in consistency.

    Also note that the resistor at the bottom doesn't fit the pads :-( And also that there are some solder bridges that I later removed.

    And if you're wondering at any choices of components, they were simply what I had on hand. Yeah, I realise that an air core inductor is less than typical for a SMPS, but it's the only one I had that I didn't have to fear the core saturating.

    For those interested in the PCB layout, notice the star ground and the Kelvin measuring of output voltage.
     
    chopnhack and Arouse1973 like this.
  14. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    Yet another test!

    The laser cutter I use has had an upgrade, now having an air line added so it can blast compressed air at the work while it is being cut.

    I made up a quick test target, cut it, and I have just etched it.

    Here is the board prior to being etched.


    IMG_6514b (Custom).JPG

    The white flecks you see are dust and/or reflective stuff in the paint, they're not pinholes!

    I have a series of 8 squares at different power settings and three SO-8 patterns (the top right one was a mistake) also at different power settings.

    The test targets have no expansion to allow for laser kerf.

    The powers used are as shown below:

    IMG_6516b (Custom).JPG

    These are percent power and the laser speed was 100mm/s in all cases. The corner power was *not* reduced.

    You may note some interesting issues. Firstly the top and left edges of the squares seem to have "cut" differently to the bottom and right edges. The difference is small, but if you note that all cuts appear to be done in an anticlockwise direction, you may notice some interesting issues later...

    After etching, the board looked like this:

    IMG_6517b (Custom).JPG

    You will note that I didn't etch until all copper was removed in the SO-8 test targets. I was already losing the tracks!

    What you may notice is the slight increase in line width with power. The broadening of the lines near the corners (due to the slowing of the laser head) is visible, but not huge. It was noted that the corners of the squares started etching first.

    The lowest power square was not etched well on the top and left sides, despite attention with a brush.

    A closeup of the SO-8 targets (rotated 90 degreed anticlockwise:

    IMG_6519b (Custom).JPG

    The bottom two show a failure to completely etch on the bottom of several traces. Due to the rotation, this was the left edge of the lines in the previous images. It seems that there is something in the direction of the compressed air flow (perhaps) which affects the cuts.

    Whilst none of these cuts was successful, it seems likely that only a slight expansion of the cutting path would allow the traces to remain. In this case (passing traces between the pins of an SO-8 package) it may not be possible due to the width of the cut.

    It also suggests that pin spacing of 0.65mm pin spacing would be the minimum that could possibly be hoped for (and these would have very thin pads).

    Now I'm off to remove the paint and see what the traces look like (has undercutting removed even more of the traces?)
     
  15. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    Amazing what a bit of acetone will do to paint...

    IMG_6520b (Custom).JPG

    It's painfully clear where the copper didn't etch. However powers of 30% and above are pretty good. It will be worth looking at powers between 20% and 30% with multiple passes since the "failure" (top right SO-8) which was 3 passes at 30% (from memory) had better results than the single pass at 30%, and a single pass at 35%. Sometimes these "failures" can be successes!

    And some closeups, for no really good reason...

    IMG_6522b (Custom).JPG

    These are the 30% and 35% SO-8 patterns

    IMG_6523b (Custom).JPG

    And this is a slightly interesting part of the squares

    IMG_6526b (Custom).JPG

    And of course, proof that an SO-8 package fits :)

    It wouldn't be very pleasant to solder though.

    Finally, using a rough estimate, here are the cut width of the laser I am using. These will depend a lot on the focus of the laser (which was pretty good when I did this).

    20% 0.17mm
    30% 0.25mm
    40% 0.27mm
    50% 0.3mm
    60% 0.31mm
    70% 0.33mm
    80% 0.35mm
    90% 0.38mm

    If multiple passes at 20% work, I would be very happy.
     
    Last edited: Oct 19, 2014
    Arouse1973 likes this.
  16. BobK

    BobK

    7,680
    1,685
    Jan 5, 2010
    The traces in the SOIC breakouts are definitely smaller than I can handle with photoresist. I am only really confident in traces and spaces of 0.012, and will use 0.010 in a pinch. I would need a maximum of 0.008 to get a trace between the pads of and SOIC. These look like 0.006, which I definitely cannot do. So your process seems to be about equal to the photoresist process that I use.

    Bob
     
  17. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    The traces there are probably as thin as 0.004 in some places, but they fail (get etched away) when they go at 45 degrees. I'm pretty sure this is due to the stepping of the laser head. Some traces of stepping were visible before etching.

    In any case, the size of the trace here is pretty much determined by the planned width less the kerf of the laser (and undercutting). I did not draw 0.1mm traces :)

    Having said that, if I can get 0.1mm traces (0.004) with a 0.3mm (0.012) spacing I'd be very happy.

    The limiting factor becomes the spacing, but it's a 2 edged sword. If I can get smaller spacing I can be more confident with thinner traces, but I also leave islands of copper closer to the pads. Having to cope with (say) 0.1mm gaps between my pads and the surrounding "unmilled" copper, especially in the absence of a solder mask, would be inviting trouble.

    I'm considering how I can best combine a low power cutting step (to get fine edges) and an etching step (laser etching) to remove more of the "waste" copper areas.
     
  18. Mongrel Shark

    Mongrel Shark

    260
    18
    Jun 6, 2012
    Nice project steve. I have been thinking about making a cnc mill to engrave pcb's for a while now. Because I don’t like the etching chemicals.

    But if you must etch. might as well do it with lasers. Lasers are cool. :D

    What kind are you using? Home made or commercial? Would love some details on the cnc set up too.
     
  19. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,363
    2,758
    Jan 21, 2010
    The process is probably not what you are thinking.

    I get a piece of PCB, paint it, use a laser cutter to remove the paint, then chemically etch the exposed copper.

    I did refer to "etching" with the laser cutter but this is simply a process of using the laser cutter in raster mode (as opposed to is normal vector mode) to remove large areas of paint.

    The laser can't remove the copper because the energy required to cut through the copper is large compared to that required to go straight through the fibreglass.

    Another possibility is to use a UV laser to expose photosensitive boards directly. It's a great idea but the laser I have access to is IR.
     
  20. Mongrel Shark

    Mongrel Shark

    260
    18
    Jun 6, 2012
    I got the etch idea. I did read the thread... Thats why I mentioned my dislike of the etching chemicals.

    I made a UV(nitrogen) T.E.A. laser a while ago.



    Looking at making Co2 lasers in the not too distant future. With 3d printing applications in mind.


    Lmao at the thought of using co2 laser to remove the copper. Would likely melt the pcb a bit. Somehow I dont think it would go too well... Copper is a bit shiny too, could lead to problems...

    Is it a diode laser? or gas type? What wattage? do you have a photo of the CNC rig? I'd like too see how the beam is directed? If its diode laser, the diode likely moves around. If its Co2 it may have a mirror bassed beam aiming system, which is what I am particularity interested in. Although diode movement systems are of interest to me too. Always on the lookout for good ideas to pinch. :D

    More info on the laser rig please. Do you know what processor it uses?
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-