Single LED Ultraviolet Exposure Box

[Jacques Lebrac] built a UV exposure box for printed circuit boards using just one LED. He usually makes boards that are just a few square inches and didn’t think building a box that had upwards of 80 LEDs was worth his time. He passed by the low power LEDs for a single 5W unit. Pumping 1.5A through this LED makes for some quick exposures, but causes heat issues. To solve this, an aluminum arm was used to mount the LED, acting as mechanical support and heat sink at the same time. The voltage regulator was glued directly to the chassis, providing at least some heat dissipation.

[Jacques] came up with an eloquent solution for holding the transparency and copper clad in place. A piece of acrylic is hinged on the back using a piece of aluminum tubing. The front has a magnet glued to it, with another one in the base to hold the cover tight to the work surface during operation.

40 thoughts on “Single LED Ultraviolet Exposure Box

  1. Awesome! I’ve been really interested in laying out boards with those high powered LEDs lately (visible light for me).

    I finally looked around and found that has a really slick calculator that lets you specify some parameters, and it will help you design a whole circuit for your lighting system. Then, if you know EAGLE, it’s easy! (if you don’t know eagle, use the sparkfun tutorials and learn!)

    I designed a board this weekend that has 4 CREE XP-G leds and one little chip with just a couple of external components to drive it. It will draw about 10 watts and put out as much light as a 60 watt light bulb!

    I also thought this was interesting:

    There is a section in that paper that briefly explains why they don’t drive these things anywhere near their capacity normally (it’s not as efficient, so more LEDs at lower current is preferable).

    Digikey also sent me a little magazine that is all about LEDs. It explained why heat management is so important (heat drastically reduces the life of the LED). LEDs are efficient, but they still burn 75% of their energy as heat. Only its worse for them because the heat goes into the base, not radiation in the air like regular bulbs. So you HAVE to use a heat sink if you want things to be good.

    Lots of learning. I wish i knew a link to a PDF of that magazine. I don’t see it, but digikey has a lot of other info here:

  2. Oh, and I learned why constant current drivers are a must for high power LEDs. They have a steep voltage/current curve, so a slight change in voltage (like a resistor heating up) can equal a large change in current, and that will reduce the life of the thing. So constant current drivers are really important. Luckily they’re easy with the calculator I mentioned.

    My other comment had links and is awaiting moderation, so this comment may not make as much sense until that one is approved…

  3. Just use MG Chemicals pre-sensitized boards and use a regular fluorescent lamp. It’s quick and easy and cheap. The exposure is quick enough you don’t get undercutting, and you develop the same way, so nothing has to change really, except you’re using a simple light source.

  4. Why are we exposing circuit boards with UV again? My preferred method is painting on a mask, laser etching the negative away then ferric chloride to etch the copper away then acetone to remove the remaining paint to expose the traces and then water only (no abrasive) waterjet cutting the holes. Simple, really.

  5. eloquent
    el·o·quent (l-kwnt)
    1. Characterized by persuasive, powerful discourse: an eloquent speaker; an eloquent sermon.
    2. Vividly or movingly expressive:

    Lol. Wut?

  6. Ok mike, u got me.
    I see tons of these exposure boxes.
    I refuse to etch boards because my shop , is my bedroom. Not interested in storing chemicals.

    Can you find some good posts on how to make boards using this method.

  7. @Addidis
    “I refuse to etch boards because my shop , is my bedroom. Not interested in storing chemicals.”

    But your bedroom, your computer, your circuit boards, and you yourself are made of chemicals! Ferric chloride isn’t even particularly dangerous; you’ve got worse stuff sitting in your stomach right now.

    Don’t mean to make fun of you or anything, I’ve just never understood people being afraid of “chemicals.”

  8. OK I will ask a really dumb question. How does this work? Is there some chemical on the board that reacts to the UV light? Or does the UV light just eat away the copper?

    Are they just using a clear plastic sheet that was printed out on a laser jet printer?

  9. @devin

    No offense taken, So I have a cat who likes to eat things. I am also really sensitive to chemicals from working batching flavor chemicals. Worst job Ive ever had. After working it , im allergic to peppermint , cinnamon, and various other flavorings. Its just one thing I decided I wouldnt tolerate any more.

    @lobster clear plastic with traces on it that (if im not mistaken) blocks the uv light from the traces. The uv breaks down the exposed copper. The covered traces remain. I could have that bass ackwards though.

  10. There are health issues with any thing. If you eat the pcb , it probably wont do you any good lol.

    Here seems to be a good page on it

    Just add a W

  11. Wait… can you elaborate on “allergic to” peppermint, cinnamon, etc?

    I am genuinely curious. I have seen people become sensitized to things like the chemicals in latex rubber, the proteins in the latex, acrylates, adhesives but peppermint? Can you explain further what that means to you?

  12. @lobster & @Addidis

    In regards to the statement that “The uv breaks down the exposed copper”, that’s quite not correct :)

    You first put UV (or light) sensitive emulsion on the copper (make your board kind of an old fashioned photo paper, if you still remember how we used to make “permanent” memories before the “jpeg” and “ink jet” did what CD and the MP3 did to vinil records and record players), UV (or light) chnages chemical compisition of exposed emulsion, then you “develop” the board (basically remove exposed areas from the copper), then “etch” the exposed copper with FeriChlorid or some other chemicals.
    To make all this even more complicated, there are “positive” and “negative” emulsions – ones that get disolved where exposed or not exposed.
    Hope this helps a bit :)

  13. @George Johnson

    What I’ve never figured out is why people bother with this method at all. Ok I get that some people may want to do this at night, but it seems like an awful lot of work when the alternative is to go stand in the sun for 2 minutes which yields better results due to more coherent light hitting the board at a perpendicular angle (no undercutting even if overexposed).

    Though I get this may be a good idea if you’re stuck in England where the sun is never visible, or Norway where the sun doesn’t exist for half a year :)

    @Any who still persist with toner transfer:

    Try this photographic method. It will change your life, and blow your mind. You’ll probably just end up in a great depression when you realised how much of a bloody pain toner transfer was and that you’ve wasted your life perfecting that (in my humble opinion) CRAP method of making PCBs.

    Come join us enlightened :)

  14. Word of caution to anyone trying to do this. Some (read: most) acrylic responds to UV light, reducing it’s intensity. Prove this to yourself by half covering a sheet of standard white paper – you will see a “shadow”. I reduced my exposure times by 50% by switching to real glass…

  15. @CutThroughStuffGuy
    Please, provide access to a laser and water jet, and then I will do things your preffered way. However, with the limited resources most of us have, etching this way is far mor cost available.

  16. Access to a waterjet and laser cutter may be obtained by payment of $270,000.

    Or buy a surplus 40k psi waterjet pump off of ebay for $2000 and marry it to a CNC table then buy a chinese water cooled laser etcher for $3000 or so. Still not what most people would consider cheap.

  17. @jeditalian
    Thanks for wasting 40 femtoseconds of my life reading that ridiculous post.

    The toner transfer method works for me because I only use it for one-off boards (usually just for testing). Anything more than one gets sent to a board house.

  18. @Spork
    What difference is there doing one offs on toner transfer compared with this method? The way I see it the wastage is the same. 1 piece of inkjet paper vs 1 piece of overhead transparency. The chemicals for this process are reusable. I have used the same bottle of developer and riston stripper for close to 1 year now, though the developer now has taken a very purple colour so I expect to spend another $15 on a new one soon.

    Is your over-exposing problem due directly to time or due to lack of adequate light cut-off due to the screen you used. I used laser printed overhead transparencies. If I use just the one I pretty much have a very limited time band between nailing the exposed area and starting to eat away at the supposedly unexposed area.
    If you stack multiple transparencies and glue them together (carefully) then that band of error increases. With 3 transparencies stacked on top of each other my initial tests gave something like 1-2min for a correct exposure and 5+ min before the unexposed sections will start to deteriorate due to light leaking.

  19. Did anyone actually look at the schematic? The resistor is wrong, for 1.5A it has to be 1.25/1.5=0.83 Ohms. The 1 Ohm resistor will result in a 1.25A current. Also, the LMx17 needs a heatsink for *anything* over a 7V input voltage.

  20. @Garbz
    It seems like I had more materials lying around for the toner method. I have a printer, an iron, and etchant. I only had to purchase the photo paper. I also never get undercut/overexposure problems and can have traces as small as 8mil. Do you think there is much to gain by switching methods?

    Good luck finding that 0.83 ohm resistor. Oh, and did you even look at the article? The entire support arm is aluminum… as in an aluminum heat sink.

  21. @glanar,
    Yes it is less than 1.5 amp. I did not have a 0.83ohm resistor while doing this project (do you ?).
    1.25A is just fine, who said it needs exactly 1.5A ?
    And no you do not need a heatsink, i run it with 12Vdc, it is hot but much lower than the LM117 can stand, remember it will run for only 2 minutes a couple of times a day!

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.