How to build an Ultra Violet Exposure box using LED's.

Your last Veroboard project!

A UV exposure box is an extremely useful piece of kit. It can be used to make proper PCB's. It can also be used to make other things such as intricate photo etched parts (a subject for another instructable). The trouble is they can be a little pricey for the hobby enthusiast especially if you want the double side type.
This instructable outlines the construction of a double sided UV exposure box using the recent generation of high brightness UV LEDs.

Why use LEDs?
LEDs are far more energy efficient than either incandescent or fluorescent lamps offering between 5-10 times more efficiency making them cheaper to run and kinder to the environment. They also (unlike fluorescent tubes) do not contain mercury. LEDs have a far greater life span than the other types of lamp measured in decades rather than months. The frequencies being emitted are also in a tighter band making UV LEDs safer than the traditional UV tubes. There's also just something cool about LEDs, I can't put my finger on it, but ever since I was a kid I've found them to be one of the more fascinating electronic components.

Is there a disadvantage to using LEDs?
Not really, however the UV exposure box I have detailed here is a little less powerful than the commercially available ones. This means that your exposure times will be around 2 ~ 3 minutes as opposed to 30 ~ 40 seconds, but come on, do you really need your PCB's to be produced that quickly? Anyway sometimes having a slightly slower exposure time can be an advantage allowing you a little more control.

This UV Exposure box will consist of 2 UV panels; each having 84 LEDs a total of 168 LEDs. Each panel will draw about 700mA at 12v. This makes each panel 8.4watts a total of 16.8 watts for the whole thing.

Step 1: Materials

The most critical parts of this project are the UV LEDs, you are looking for 5mm Ultra Violet LED 2000mcd 395nm, 3.4V 20~25mA.
I bought two 100psc packs from eBay.
If you find something better then ensure that they are;
- At least 2000mcd in brightness
- Have a peak wavelength of less than 400nm.
- A viewing angle of at least 20 degrees.

You will also need 2x 160mm x 100mm pieces of Veroboard and 56x 75R resistors.
Another important choice is the PSU. I used a plug in, 12 volt 24 watt switch mode power supply. Switch mode power supply's are far more energy efficient than most other types and are also very stable.

All the other parts and materials are easy to find. Some I bought, some I salvaged. This is where you own judgement and taste comes in. In the end it's up to you how closely you follow my design. I've included all the CAD drawings and schematics as metafiles so they're easier to read when you print them out.
hello, I made the same box ! But i don't have results after 30 min . I don't know what is the problem . There is no traces on the pcb When i put it in the development solution.<br>What do you think !?
<p>Maybe the problem is the glass. Glass is know to attenuate uv wavelengths. Try a thinner glass so it doesn't attenuate so much. Also don't use pexyglass. It attenuates even more. If you are doing the right thing now and want better results you can: -&gt;Use Quartz Glass(Expensive) or upper the led power(increase the number of leds, or change the ones you are using for ones with higger luminous flux, so that the amount of uv lights that isn't blocked is bigger). </p>
<p>Not sure what's wrong, but if you were to point your LEDs at a bottle of tonic water with quinine, does it cause the tonic water to glow? Maybe a highlighter marker would be an easier test, but I'm wondering if this would confirm that you're emitting at the short-enough wavelength to be in ultraviolet, or if you just have purple LEDs.</p>
<p>Hello, Maybe it was my mistake ... i tested the box for 25 Seconds and it works perfectly with a great results !! thank you</p>
<p>Very clean Instructable, and beautiful finished product! Question I have is based on optimal distance between LEDs and between LED panels. I'm resource constrained, and want to be sure I can make the best product possible. Did you give any consideration to optimal LED density and resultant luminous intensity at a given focal length? Not suggesting that you'd have to, the math is certainly beyond me, but I'm wondering if you did, if you could share how you arrived at your final dimensions?</p>
<p>*NB Make sure you get the less the right way round. They will still work backwards but burn out and smoke in 60secs!!! ouch!!</p>
<p>This is a great project. The instructions are pretty good. However I found that 2 leds on each 3 led blocks failed to power up. Seems like it needed more soldering than directed. It's perfect to get started making a more refined UV lightbox. Bravo!</p>
<p>Why use so many resistors?</p><p>In step 7, you have 4 sections of led's on the board/print.</p><p>You could have 1 resistor per section, instead of 7 resistors. </p><p>And if I am right that resistor need to be 10.71 ohms.</p><p>So instead using 28 resistors, you would be using 4 resistors </p><p>.</p>
<p>Yes, if the LEDs were perfectly identical, that would work. (Also, you could merge the 4 section resistors into one 2.7R resistor for the whole thing.) The thing is, they're not perfectly identical, and LEDs do not behave like a resistor would. Check out the voltage-current graph on any LED datasheet. They have a &quot;voltage drop&quot;, most of the time the voltage across their terminals is this much, regardless of how much current flows across them. If you connect more voltage, they'll attempt to conduct infinite current, which results in the Magic Smoke(TM) leaving the LED (burning out). If you connect less, they will not conduct any current. This voltage drop for these LEDs is approximately 3.4V, but it varies a little, even within the same batch.<br><br>The good thing is, we can use current limiting resistors to set the current of one or more LEDs connected in a row. Their cumulative voltage drop is approximately 10.2V, the rest of the 12V is across the resistor, so the whole row gets approximately 25mA. If the cumulative voltage drop happens to be 0.1V less or more, the resulting current doesn't change significantly, the brightness stays the same.<br><br>The problem comes if you use a common resistor for two or more rows of LEDs. If one of the rows has a lower voltage drop than the others (and since they're inequal, there will always be a row with the lowest drop), the voltage across the whole circuit will be equal to the drop of that row. Then that row will conduct all the current that was meant for all the rows, because the others don't conduct anything at all. Burnination! When the row with the lowest drop burned out, comes the next and so forth. Of course, in reality, this doesn't happen this dramatically, because the LEDs don't keep their voltage drops perfectly (look at a datasheet again), but the brightness of the different rows will be vastly different. Some won't even light up, some will be so bright they will cure your PCB on the opposite side too, and that's not what we want.</p>
<p>I don't know if this has been posted before but, I found this site that seems to have everything a decent prices. http://www.flexfireleds.com/shop-by-project/uv-led-lights/</p>
<p>Can i use the plexiglass instead of normal glass ? Or the result will be worse ?</p>
<p>Plexiglass will filter out the UV light. </p>
<p>Can you recommend a seller on ebay? Seems all I can find are sellers who can't provide me with a datasheet and it would be nice to have the information handy so I know exactly what I'm buying.</p><p>Great build by the way. I was initially going to use an old scanner, but I think I like this project better, plus it keeps me away from mains power and I like that. I'm still not that comfortable working with mains. This project makes me want to build my own design, but I like yours quite a bit. It's definitely got my mind reeling right now.</p>
<p>What is that attached to your iron? Is that a hot air rework or something like that?</p>
you could also add a PWM controller for dimming.
I just finished soldering up one board. I did spoil one board. It is a good idea to lay out one row of resistors and LEDs to make sure it fits on your strip board. Mine have a wide trace down each side which had be accounted for. I cut the traces with a small burr in my dremel. I placed the resistors on the copper side but fed the leads thru the holes, soldered at the end of each resistor and then cut the leads flush on the other side. I put my power rails on the component side. There is just enough room with the layout that is provided. In my case when looking from the copper side it would be best to move everything up and to the left at least one row.<br><br>Now on to building the box.
<p>Which sellers on that auction site are you guys getting your LEDs from? I've been reading the reviews for alot of sellers and they mostly turn up negative reviews. I've been coming back to this particular listing (link below). Has anyone done any business with this individual?</p><p>http://www.ebay.com/itm/50-1000pcs-5mm-Flat-Top-Ultra-Violet-LED-UV-Light-Purple-Wide-Angle-Lamp-2000mcd-/191132829063?pt=US_Light_Bulbs&amp;var=490302606065&amp;hash=item2c80678587</p>
<p>I built one of these (1 sided) and it works brilliantly. I compared it with a commercial UV PCB exposure unit and I find the results are much more reliable with this one.</p><p>The only problem is that I got a bit of falloff of light at the edges on a large board so I I rebuilt the board using a single sided PCB which was a bit larger and fitted right to the edges of my existing box. It was much easier to solder up and works great. This one has 117 LEDs</p><p>I attach a PDF of the track layout in case anyone is interested</p>
<p>Decided to go thro' hole with the resistors instead of a facsimile of smd. I had no problem with solder track shorts with this method, mainly because I used a 2mm <strong>BEVEL</strong> tip with .015&quot; (.381mm) 62/36/2 silver solder</p>
<p>I've come to the conclusion, despite the foregoing, that the only way to make sure that each led is flat is to put something flat directly below the led being soldered. Even so you need to check there is no slack.</p><p>Sometimes it might not quite work out that way so it's a good idea to solder only one leg &amp; check for slack. If there is, just apply more heat to that leg. I don't like doing this as there is the risk of lifting the copper track, which has happened to me, so I'm very wary (leary) now on that score. A lifted copper track can cause big problems &amp; if you are near completion of the board it is a real bummer to have to correct with bridges.</p><p>Hope this helps.</p>
<p>Just remembered-I found the the recommended spot tool not nearly sharp enough &amp; very very laborious work when you consider the number of spots you have to do. Save your money on this excuse for a cutter.</p><p>On the other hand rolling a new drill bit back &amp; forth between the thumb &amp; forefinger was surprisingly quick &amp; effective. No comparison really.</p><p>I originally used a Proxxon hobby drill press with a Dremel drill using an adaptor &amp; a burring drill bit (the Dremel drill press is well known not to be an accurate design). This method is quick but requires a degree of control, and its OK if your doing them all at once, but the real downside is that the resulting dust from strip/Vero board, albeit not huge, is apparently very toxic to the respitory system. If doing it this way make sure it is done in the open air.</p><p>Hope this helps.</p>
<p>A few more tips.</p><p>Before doing any soldering I found it obligatory planning to do a dummy placement of leds on the top horizontal row &amp; the left vertical column to make sure that the design is workable for the board size, also making sure that there is room at the bottom of the board for the two power rails. I then marked component placements with texta around the side of the board. When you start soldering use your continuity meter for every solder joint to ensure there are no short circuits between adjoining rails. Its easy to slip up otherwise. </p><p>Don't be tempted to follow the instructable blindly, particularly the pics. Make sure you understand how the first of all the blocks works. If necessary do a practice circuit of 2 vertical blocks (of 3 leds each) on scrap board to make sure you have continuity working between blocks. </p><p>The author says it tricky to get the leds to sit flat with only two hands. Yes it is, but that can be overcome by using the following simple method. Place temporary unsoldered leds in the 3 corners, ie., other than the corner where you are soldering the first led. Bend their leads over flat to the board. That's a reasonable start. Then place the board on an even flat surface with the temporary leds facing down, whereupon they act as stabilizing legs. There will still be some small slack between the board &amp; leds but if you place a reasonably heavy weight on the top of the board ( I used a cheap steel sticky tape dispenser) the slack will be taken up completely &amp; voila the leds then sit perfectly flat. </p><p>I hope this helps for those of us that are challenged with circuit design.</p>
<p>Not immediately obvious what the diagonal spot cuts are for? They are necessary to electrically isolate each vertical bank of 3 leds.</p><p>Saw elsewhere a tip recommending cutting the leads before soldering to avoid mechanical strain on the solder joint and the potential for later breakdown of the joint. Seems a worthwhile precaution, although too long with soldering iron heat may destroy leds with such shorter leads.</p><p>Probably wise to test each LED before soldering because if you have to wrestle with removing a dud LED, depending on technique, you could destroy the copper strip, which too easily lifts with excessive heat on a standard board, &amp; if you don't have room to bridge the broken continuity you've wasted a hell of a lot of time &amp; energy building the array.</p><p>I found it preferable to install a vertical column of 3 led sets, rather than horizontal banks in order to make sure from the outset that the there are no errors in the design. I found it easy thereafter to duplicate succeeding columns use the first as a guiding template. </p>
<p>Got it to work at last. This is very, very detailed work. Easy to make mistakes along the way. Needs some serious time, effort, continuity testing &amp; unlapsing concentration to get it to work first time. Don't do it when you're tired. Screw up with the spot cutter too many times &amp; you'll need to toss the board.</p><p>Saves heap on buying a commercial unit, but hey you need to be a masochist to do it. </p>
Great project mate, followed as best I could but am having problems getting all of the LEDs to light up despite checking the circuit and replacing components. I am using .5w 75r with 3.2-3.8 V, 75ma peak, 20ma forward LEDs. And a 12v 7.5a power supply to test. Could it be a problem with this combination? Any advice appreciated!
<p>I just completed my own UV exposure box. I used 96 UV LEDs in a 12 x 8 matrix (good for 4&quot; x 6&quot; boards). LEDs were approx. 3&quot; from exposure window. The LEDs were 1200 mcd with 160 degree viewing angle. Exposure time was 2 minutes.</p>
hey! very useful post! <br>I was wondering instead making a new circuit board can we use the LED light strip? <br>I am a graphic student and want to use it to expose my screen printing screens will it work on them? <br>Many thanks.
Parts express has UV led kits of 100 for $9.99
This is a very nice instructible. It is the most complete UV box I've seen. I do have one question. Is there a reason you put the jumpers and resistors on the copper side of the board and not on the component side? Was it just an aesthetic decision? Or will the UV cause component degradation of some type?
Hi, please, were did you bought your leds? I did looked in ebay but I can't say for sure that those leds there really have 20&ordm; view angle. All leds that I found there are kinda suspicious, chinese cheap fake stuff, even tough they say they are 20&ordm; I can't trust. I'm in Brazil and at here ebay is the more easy way to get electronics by a decent price but not in this case.
I found these in Germany. They are more expensive, but because of the angle of 120&Acirc;&deg; they are more suitable for illuminating flat surfaces. This is exactly what we need. http://www.ebay.nl/itm/50-UV-Leds-4-8mm-120-200mcd-Led-STRAWHEAD-schwarzlicht-ultraviolett-/370543226494?pt=LH_DefaultDomain_77&amp;hash=item564618ee7e#ht_1926wt_1139
Isnt 200mcd WAAAY too little?
I have a question, would the output from this UV box be enough for it to be used as a sterilizer box? I do my own beer/wine brewing and I'd love to be able to fit a rig I could use to sterilize equipment in between batches..
Sterilization also takes place at 365nm, however, to get the type of power needed would require a much more robust, i.e. $$$, set of LEDs that can output some real power. You'd also have to do some extensive testing to make sure sterilization is taking place, the research articles I've seen were using 15 to 30 minutes exposures to achieve complete sterilization of bacteria. <br> <br>I think using a non-toxic spray method would be a lot quicker and cheaper in the end.
No, the wavelength for sterilization is much shorter than this type of LED can emit (&lt;300 nm where these are likely closer to 400 nm). UV LEDs for sterilization are expensive and difficult to find in small batches.
I built a double sided UV exposure box with vacuum pulldown around 10 years ago using tubes and it is still going strong without replacing the tubes. Also my box will do boards up to A3 size, How many LED's and what would they cost for an A3 sized double sided unit? Yes they're more energy efficient but no one is going to have a unit running for more than a few minutes at a time so cost is negligible. You've made a nice looking unit but I just can't see any real benefit of LED's over tubes both in cost and construction time. TBH. Hopefully I'll post an indestructible of my unit soon but as I built it so long ago I'll need to dismantle it for photos.
I'm a bit lazy to research this so I guess I'll just ask, <br>Which will give better and faster results; total watts, total mcd or both?
I managed to finish this project, but with some mods of my own :p <br> <br>Instead of painting the wood white, I covered it with metallic duck tape which gives me extra brightness. <br> <br>I also tested some pcb that usually take 120s in a retail exposure box with UV lamps and they took the same 120s in my box with the LED. <br> <br>Thanks for the idea and the design, it is excellent!
So i found some LEDs but they peak at 400-405nm, everything else checks out. Will they still work?
Awesome project! Are those Metal Film Resisters? <br>
Looks very nice, I might copy your project and make a scaled-down version! ;)
can i just use a normal LED.....?
no because a normal coloured LED does not operate at the right frequency to be used for this purpose due to the wavelengths being different this diagram i found on google images shows the uses of the wavelength. :)
can i just use normal LED's ??? this parts are hard to find here back in Pilippines.<br><br>
UV LED is available at e-gizmo (taft ave, along DLSU). 3.50 PHP a piece.
this glass too thick for LED UV going pass thur glass..
I built the box, but I seem to keep blowing LED's. I'm using the supplier that you recommeded and everything appears to be okay. However, I noticed that when I turn off the wall-wart type 12v power supply, I have some latent voltage residing on the board. I can see some residual glowing of some of the LED's immediately after turning the unit off. It does eventually fade out, but if I turn the unit back on too quickly, I have a bright flash of the LED's. It's as if there is a capacitor somewhere that's holding current in the 12v transformer. Any ideas or suggestions to bleed off the current? I'm a newbie at these sort of things.
Okay. After surfing around the web I've found out what a &quot;Bleed Resistor&quot; is, but I haven't been able to find out definitively how to calculate the amount of resistance needed or exactly where to put it in the circuitry. Any suggestions would be welcomed.
Why guess resistors? Calculate: <br> <br>Most efficient way is to use make strings of few LEDs in series with resistor (and you can connect many such strings in parallel) <br> <br>R=(V-n*Vf)/If <br> <br>P=R*(If)^2 <br> <br>where <br> <br>R= resistance (Ohms) <br>V= power supply voltage (Volts) <br>Vf= LED forward voltage drop (usually about 2.2-3.8V for most LED, read specs) <br>If= LED forward current (usually 20mA, read datasheet) <br>n=number of LEDs in series (in one string) <br>P = resistor power (Watts, minimum rating) <br> <br>note: you should aim to get the voltage difference to be about 2V: <br>(V-n*Vd) ~ 2V <br> <br>reason: <br>if the difference is greater, the more energy is wasted into heat by resistor (you need more powerful resistor and power supply). <br>if the difference is too small you loose current regulation because LEDs are non-linear device. <br> <br>calculation example, assuming you are using <br> <br>V=12V <br>Vf=3.2V <br>If=0.020A (that is 20mA) <br>n=3 <br> <br>R=(12-3*3.2)/0.020=120 Ohm <br>P=120*0.02^2=0.048W (you can use 1/8W or 1/4W without problem) <br> <br>note how P is small. that is because we have (12-3*3.2)=2.4V <br> <br>using large resistors makes no sense because to get current to match spec, you would need high voltage (this is not just dangerous but also means that most of the energy is wasted in heat so those resistors would have to be for high power). <br> <br>for example using same scenario <br>Vf=3.2V <br>If=0.020A (that is 20mA) <br>n=3 <br>R=47k <br> <br>we would need resistor to have power rating of at least <br>P=47000*0.020^2=18.8W <br> <br>(and that was just power three LEDs) <br> <br>Also DC power supply would need to provide much higher voltage which is <br>V=n*Vf + R*If <br>V=3*3.2 + 47000*0.020 <br>V=9.6 + 940 <br>V=949.6V <br> <br>That is way too much just for poweering few LEDs. <br> <br>So that was introduction for current limiting resistors. <br>For bleeding resistors we calculate it from time constant. <br> <br>t=RC <br> <br>where <br>t = time (seconds) <br>C=capacitor in power supply <br>R= resistor (note this is combined resistance of load and bleed resistor) <br> <br>R=t/C=Rb*Rc/(Rb+Rc) <br> <br>Rc=V/(m*If) this is load resistance (for m LED strings drawing current If) <br>Rb=bleed resistor <br> <br>Say you have simple 12V PSU with bridge rectifier and 10000uF capacitor and 6 LED strings drawing 20mA <br> <br>C=0.010 F <br>m=6 <br>Rc=12/(6*0.020)=100 Ohm <br> <br>Without bleeding resistor t=RC=0.010 * 100= 1 seconds <br>which is shourt enough <br> <br>adding bleed resstor in parallel, we can bleed the capacitor faster (total resistance is smaller and t=RC is smaller). <br> <br>Say we use Rb=18 Ohm <br> <br>R=18*100(18+100)=15.25 Ohm <br> <br>t=RC=15.25 * 0.010 = 0.1525 sec <br> <br>so turn off time was reduced by about 6x. <br> <br>note that value of bleed resistor was quite small. usinglarge value such as 10k or 47k would not make any difference. <br> <br>also keep in mind power disipated by bleed resistor P=12^2/18=8W <br>(which is more than used power for LED circuit in this example). <br> <br>good luck

About This Instructable


297 favorites


More by djhamer: Soldering an SMT MOSFET Driver with a hotplate UV LED Exposure Box
Add instructable to: