View Full Version : UV Light Box

Doug K
12-14-2010, 11:53 PM
I've got parts ordered to build a light box for exposing alternative process prints. I'll be using LEDs for a light source. I found some pretty high power wide angle LEDs, and I'll be putting them on a 2 inch grid pattern, and at the distance I have calculated (conservatively) they will overlap each other 100%. The only thing I haven't worked out is whether or not I'll need a diffuser for the light. I'm sure I can find some low iron glass for transmitting UV, but whether I should make it like viewing glass on a large format camera is the question. Also, what would be a good way of testing the evenness of the exposure?

In theory I think this will work pretty well, we'll see how it works in practice...

12-15-2010, 08:37 AM
I would coat a sheet of paper the same size as the contact frame with cyanotype (or whatever your first alt process sensitizer will be) and just expose the sheet with no negative. If you get areas of darker and lighter exposure that match the LED pattern, then you either need closer spacing or some type of diffuser. I suggest cyanotype because it is inexpensive and easy to use.

How expensive are the LED's? Wouldn't it be better to put more of them at a closer spacing (and thus decrease exposure along with increasing the uniformity) than to diffuse the light and make your exposures that much longer?

I am intrigued by UV-emitting LED's. This seems to be a great way to go - I would think that they should provide the most efficient UV light source that you can build yourself. What is the wavelength of the light emitted by the LED's? I have heard of a few people trying this type of exposure unit, but I have not heard how successful they are. Show us some pictures of the exposure unit and of the images that you print using it.

Doug K
12-15-2010, 10:01 PM
Thanks Earl. This is a big experiment for me so we'll see how it goes. As I said, these are pretty high powered LEDs, 1/2 watt, and they are wide angle, supposedly 140 degrees. My plan was to try and make a 16 x 20 light box with 99 LEDs. That would put them on a 2 inch grid. The distance to the paper will be about 3 1/2 inches which will give me 100% overlap. So the edge of one LED's circle will be on the center of it's neighbors. The peak wavelength of these LEDs is at 400nm. Perhaps a little high, but after reading Sandy King's article, I think it will be OK. Once I get the LEDs in hand I can experiment a little before committing to the final design. I think there will be plenty of light however.

The LEDs weren't horribly expensive, about 70 cents each. The price will continue to come down as LEDs become more utilized as general lighting. There are some high powered ones already being made, I think the highest I found in my search was 100 watts, though those were pretty expensive. I love the idea of them for their instant on, low heat, and low power consumption. I've seen a few things made for PC boards, but never anything for Alternative process. I can't imagine I'm breaking too new ground. I'll definitely keep everyone posted as I experiment.

12-16-2010, 04:58 PM
Yes this is an interesting project. Where did you find the units? I'll be anxious to see your results also, keep us posted.


12-16-2010, 06:52 PM
99 LED's at a half watt each. That sounds pretty good compared to my 3000 watt plate burner with the 220 wiring. Have you looked into controlling exposure with an integrator? Come to think of it, with LED's, you might not need one. If the light output is consistent for the life of the LED, then a timer will be just as effective as an integrator. I think that you might be on to something here, Doug.

Doug K
12-16-2010, 07:51 PM
Steve, I got them at the world's largest garage sale: ebay. I've been doing a lot of reading to get up to speed on LED nomenclature and characteristics. It helps having an electrical/controls background. Parts have just started showing up today in the mail, and I imagine in a couple of weeks I should have something working. I'll absolutely share my learnings on this. I haven't found much to go on, so even if things don't work out so good, at least people can learn from my mistakes.

Earl, I absolutely haven't thought of using an integrator. LEDs have a very long lifespan, so I think with the amount I'll be using this, the exposure should be consistent. It's worth thinking about from a testing perspective though. I'll use your earlier suggestion to test with cyanotype. That's a great idea.

Tom Persinger
01-13-2011, 10:39 AM
this sounds great.... i'm very interested to see some pictures of the apparatus as well as the tests made with it! I've been thinking of getting a plate burner like Earl, but may wait to see how this project transpires... Thanks for pioneering the way Doug!

Doug K
01-14-2011, 07:50 PM
Well, Tom, the first try with the LED to just see what it can do doesn't look so promising yet. I am going to try a few other things though, and see if I can't get something workable.

The good news is I made my first 3 cyanotypes today using sunlight. I just winged it with the negatives, and the results, while not great, are not bad. Now to calibrate the system...

Doug K
11-30-2013, 09:58 PM
This project just gained some new life. I was able to source some 3 watt 365 nm UV LEDs. I am tired of having to hope for sunny weekends, which actually is pretty normal here, but also I'm tired of variable results because of clouds etc. I've made some nice prints with sunlight, but now its time to move on and have a measurable, predictable source of light available at night and snowy weekends. My original batch of UV LEDs were relatively low power compared to the ones I've just ordered, but more importantly, their peak UV light was at 405 nm, which with my tests proved ineffective.

More to come! I'm glad this project is back on track.

11-30-2013, 10:18 PM
365 nm should put you in the sweet spot.

12-01-2013, 05:16 PM
A few weeks ago someone asked me to make them a UV light source.
I finished soldering it this afternoon………and it works:D
Big enough to cover 5x4.
I have used a matrix of 84 LEDs, 7x12 rows on Veroboard.
Spacing is 10mm one way and 13mm the other.
5mm round, 3000mcd, clear, rated at 20mA. 395nm typical wave length.
Projection angle is 20 degrees so they need to be about 45mm above the surface to give a reasonable overlap and even illumination.
Just need to make a box to put it in.
This is experimental; if he finds it works well there will be a larger unit, that could be a huge number of LEDs but they are dirt cheap from ebay. I bought 100 for 7 pounds and every single one works :)
I have kept this unit cheap and simple, I am running it from 12vdc with the LEDs in groups of three with a 100 ohm resistor. This is not optimised but it will work reliably, if I build a larger unit I will probably use 24vdc with properly regulated current probably via LM317s. That will still be a reasonable price, LM317s are less than 40 pence each, maybe a lot less if I shop around.
I will post some pics when I have finished the box.

Doug K
12-01-2013, 09:53 PM
Sounds interesting DaCh. Your LEDs sound similar to the ones I had, though I payed more 2 years ago. I make prints larger than 4 x 5, and my experiments with them gave pretty uneven results. I was inspired by a few people who were making PC boards, but for photography, I just couldn't make it work to my satisfaction. I was going to put 3 LEDs in series with a resistor like you did. Now I'm on to UV LED Light box version 2. I was able to get a good price on 3 Watt 365 nm LEDs with what they say is a 120 degree angle. We'll see what my initial exposure experiments show up. I test them by exposing a cyanotype with a test strip to get an exposure pattern. I've not worked with LEDs this big, so I'm not sure yet how I'll power up my grid of 9. I do have heat sinks for them already, and a way to mount them. I've been salvaging parts for a while now. I'll have more information once I receive and test the new LEDs.

But really, I can see that this is going to be the way to go in the future. It's just a matter of time before we start seeing commercially made UV LED light sources. Smarter people than me will figure the whole thing out.

12-07-2013, 08:29 AM
Here is a picture of the board, not much point in showing you a picture of a box I put it in, not very exciting.
The good news is that it works really well only tested with cyanotype so far but it gives exposures shorter than the huge lamp that is normally used, that is two feet from the vacuum table glass and during a 20 minute exposure the glass reaches more than 40 degrees, and the room warms up quite a lot :mad:
The LED box needs about 10 minutes and is barely warm :cool:
This little box is used for testing different coloured and different density digital negatives. Different colours give different tonal ranges for highlight and shadow so the tonal range can be expanded further by using two different coloured negs rather than just a grey neg for highlight and another for shadow, very clever stuff. The guy who is doing it has a huge amount of experience of printing most alternative processes and seems to be advancing the techniques a long way, he is doing a lot of well planned analysis with step wedges and densitometry.
Should be some interesting stuff in the future. :)
I am going to make another small light source using 345nm LEDs so the difference of UV light frequency can be plotted. Then probably a large one to be used for the actual printing.

12-07-2013, 01:00 PM
That's a neat project.

You are probably aware that some of the higher end LED video lighting accessories now offer adjustable color temperature, using dual-color LEDs where one side is bluish and the other yellowish, with the capabilty of adjusting the temperature of the light to suit the conditions. These might be interesting as a lighting source for a future project, with the possibility of adjusting the color of the panel for the desired contrast grade of the print.

Hope to see some results, thanks for posting.


Doug K
12-15-2013, 10:57 PM
DaCh, For my small array of 3 Watt LEDs I picked up a constant current power supply. It was only 23 dollars including shipping from a US supplier. Yes, not as cheap as a LM317, but then I also don't need any other power supply. I can just wire up my 9 3 watt LEDs in series, and the power supply does the rest. No resistors, and no custom circuits. I guess it takes the intelligence out of making this kind of thing, but I'm getting tired of intelligence. Give me something dumb that works....

Interesting to note the price jump when going from 395nm LEDs to 365...