UV LED

Discussion in 'Alternative Processes' started by Mainecoonmaniac, Feb 18, 2014.

  1. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    I'm a bit of a ghetto tinkerer. I saw some UV LEDS on Ebay and I thought I'd buy some to see if it's applicable for alt processes.

    http://www.ebay.com/itm/10x-1W-Ultra-Violet-UV-Purple-High-Power-LED-400-405nm-/110737016202

    After receiving them a month ago, I was able to play a bit on my work bench. So far, it seems to work with cyanotypes.

    Here are some pictures. http://www.flickr.com/photos/mainecoonmaniac/sets/72157641172069904/

    I got the heat sink from a scrap pile from work and borrowed a power supply. Optimal voltage is about 3.5v and draws .78 amps. My plan is to mount 10 on them on an aluminum heat sink that is 10x12 inches.
     
  2. Klainmeister

    Klainmeister Member

    Messages:
    1,492
    Joined:
    Jun 2, 2010
    Location:
    Santa Fe, NM
    Shooter:
    4x5 Format
    Interesting idea. So these guys put off enough heat to warrant a complete heat sink? Also, what was the exposure distance/time with that cyano?
     
  3. stormpetrel

    stormpetrel Member

    Messages:
    109
    Joined:
    Apr 24, 2010
    Location:
    Wellington,
    Shooter:
    Multi Format
    This is great!
    I have put ~30x 3W UV led in my enlarger head with the idea of using them for alternative process too.
    So far so good. The heat sink hardly change of temperature and the head light produces a lot of light!
     
  4. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    I'm still doing tests

    But so far, it's about an inch or less for 30 minutes for 1 LED. I expect the times to be lower when I have a bank of them running. They have a lens that concentrates the light and I plan to have the LEDs 5 inches to achieve even illumination.
     
  5. stormpetrel

    stormpetrel Member

    Messages:
    109
    Joined:
    Apr 24, 2010
    Location:
    Wellington,
    Shooter:
    Multi Format
  6. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    I don't know

    But those LEDs get hot. It's better to keep them cool or else they might cook. The distance is about an inch. Exposure time 30 mins. This is just a test.
     
  7. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    Yow! This can make you blind looking at it. Probably made for tanning beds.
     
  8. jakobb

    jakobb Member

    Messages:
    15
    Joined:
    Feb 4, 2013
    Shooter:
    8x10 Format
    I was also interested in the use of UV/far blue LEDs. There ones with high output at 360-380 nm which would be close to perfect for Pt/Pd printing. May be one of my next projects?
     
  9. Herzeleid

    Herzeleid Member

    Messages:
    191
    Joined:
    Oct 31, 2008
    Location:
    Ankara/Turke
    Shooter:
    Multi Format
    It is good to know such LEDs require serious cooling. I was intrigued by the idea of using LEDs for an exposure unit. But it seems they require good planning on cooling the unit. Thanks for sharing.
     
  10. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    The ones I'm using

    The LEDs I'm using are 410nm. I don''t know if it will work for Pt/Pd. I'd like to run a test for Ziatypes though. Still a lot of testing. Gotta find the time :sad:
     
  11. David A. Goldfarb

    David A. Goldfarb Moderator Staff Member Moderator

    Messages:
    17,980
    Joined:
    Sep 7, 2002
    Location:
    Honolulu, Ha
    Shooter:
    Large Format
  12. polyglot

    polyglot Member

    Messages:
    3,472
    Joined:
    Jun 12, 2009
    Location:
    South Austra
    Shooter:
    Medium Format
    Hey ghetto-tinkerer,

    You're grossly over-powering that LED and should expect it to have a (very) short life at 780mA continuous current as they're rated for 350mA. When driving LEDs, you need a current-controlled power supply, not a voltage-controlled supply.

    Best option is to put a whole bunch in series and control the current through the series string; you can buy dedicated devices off eBay that will do exactly that, for example, or 300mA, etc.

    What you do is:
    - pick a model of LED
    - pick a power supply current that is correct for the LED (300mA would be a good choice for these LEDs rated to 380mA)
    - decide how many you want in series (say, 10)
    - multiply the forward voltage (3.2-3.6V for these) by the number in series => 32V-36V
    - pick a power supply that supports the voltage range required at the current required, e.g. a 300mA, 30V-60V supply.

    With support for up to 60V, you could put up to about 15 of these 3.6V LEDs in series on the one power supply.

    Edit: easier than chaining a whole bunch in series is to buy 10W/30W chips. They typically have a bunch of diodes already in series and typically operate at about 31-35V and 300mA/900mA. 395nm LEDs are cheap and readily available but 360nm seems much more expensive. You probably want to know what wavelength your emulsion needs before buying up on LEDs...
     
    Last edited by a moderator: Dec 30, 2014
  13. rafaelrojas

    rafaelrojas Member

    Messages:
    8
    Joined:
    Nov 14, 2014
    Shooter:
    Medium Format
  14. ME Super

    ME Super Member

    Messages:
    1,234
    Joined:
    Apr 17, 2011
    Location:
    Central Illinois, USA
    Shooter:
    35mm
    Let's use a simple resistor to convert our constant voltage supply into a constant current supply. Suppose you have a 5v power supply and you want to run your LEDs at 350mA current. Further let's suppose that the voltage drop across the LED at 350mA is 1.4v. You can run 3 LEDs in series with a 5v supply, and the voltage drop across them will be 4.8v, so you only have to lose 0.2v with the resistor. Drag out ohm's law, which says E=IR, where E is voltage, I is current (in amps), and R is resistance (in ohms). 0.2=.350R. Solve for R, and you get 0.57 ohms. Power P=EI, so the power dissipated across the resistor is .07 watts, well within the max power dissipation of an inexpensive 1/4 watt resistor.

    The trick is to find a 0.5 ohm resistor. Unfortunately, I know of no such beast. Fortunately, Digikey has 1 ohm resistors for 10 cents each. Write two of these in parallel and you have your 0.5 ohm resistor, because 1/R=1/R1+1/R2 for resistors wired in parallel. Therefore 1/R=1+1 (you're using 1 ohm resistor, making the math very simple. Solve for R and you get R=0.5 ohms, which is what you want).

    BOOM! You've converted a constant voltage power supply into a constant current power supply suitable for driving 3 LEDs with a voltage drop of 1.4v across each diode at 350mA for the princely sum of 20 cents (plus shipping).
     
  15. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    An update on my UV LED box. I added more LEDs to my box and ended up with the same exposure time. Where I work, there's a electronics engineer saw my schematic and he said I got it completely wrong. I put all my LEDs in a parallel. The engineer said the LEDs are voltage driven, not current driven. He drew out his own schematic and he put 3 LED in a series with a resistor and since I had 30 LEDs, he put 10 of the LEDs with resistors in a parallel circuit. Also, the power adapter was inadequate and recommended that I used an old Dell power 19V adaptor laying around the junk pile.

    Now my print times has gone down from 30 minutes to 7 minutes. Correct my if I'm wrong, but the light is about a 2 stop intensity in brightness. The bulbs get hotter now, but the aluminum block absorbs most of the heat. I also had to glue the 3W resistors down to the aluminum with heat sink adhesive to keep the resistors cool.
     
  16. polyglot

    polyglot Member

    Messages:
    3,472
    Joined:
    Jun 12, 2009
    Location:
    South Austra
    Shooter:
    Medium Format
    I'm pretty sure you misunderstood your eleceng colleague, but his recommendation to put them in series is correct. The output power of a LED is dependent on current, but the current is extremely sensitive to voltage and there is chip-to-chip process variation between them. Therefore when you put a bunch of diodes directly in parallel you are forcing them to all run at exactly the same voltage, which means (because of their manufacturing variations) they will run at (possibly extremely) different currents; in the worst case you might get all the current through just one diode, which will die, then the current will go through a different diode until that dies, etc.

    Series is the best option (as per my post above) because when things are in series, they have exactly the same current through the whole chain and will run at very slightly different voltages. If you want to put several series-chains in parallel, you can, but you must control the current through each chain separately, e.g. using one of those controlled-current supplies I linked above per chain, or with a separate high power resistor for every single chain. You cannot just make 10 chains of 3 LEDs, put those chains in parallel and then one resistor on the whole lot, because that gets you back to the original problem of unequal current-sharing. You must have a separate resistor for every chain.

    Resistors are simple and fairly easy, but there's a bad tradeoff between the accuracy of your current control and the amount of power wasted in the resistor. If you can possibly get them, the current-controlled supplies that are designed to directly drive series chains of LEDs are a far better option. More voltage across the resistor means less uncertainty in the current due to manufacturing variations (voltage at chosen current) in the LED chain, but also means more power lost to the resistor.

    If you're going to insist on using a 19V supply, you should put 4 LEDs in series in each chain (~14V) and then use a 15 ohm 5W resistor per chain (resistor will dissipate ~1.7W). Heatsinkable resistors are easier to keep cool (they won't burn you), but non-heatsinked resistors should be used in cool, free air (half inch clearance) and not half-ass glued down to a heatsink. Air-cooled resistors are designed to get extremely hot at their rated power (100C+) but that's OK as long as their positioning is such that they can't burn people or other parts of your design.

    If you put 3 LEDs in series off 19V, you need a 27 ohm 5W or 10W resistor, which will dissipate 2.7W. Clearly that is much less efficient (0.9W lost in resistor per LED) than having 4 per chain (0.4W lost in resistor per LED). And of course you need more and larger resistors.

    Once you've assembled it, measure the voltage across the resistor while operating, and measure the resistor (out of the circuit), to determine the actual operating current. If the LEDs and the resistor are both at the extremes of their allowable tolerances (10-20% is not uncommon for power resistors), you might have a lot more or less current flowing than you expected from your design. In that case, you will need to replace the resistor with a different value.
     
  17. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    Wow thanks for the info

    I think I must have must quoted. I put one resistor in the beginning of each chain of 3. I have 10 resistors total.

    I really appreciate you taking the time in explaining. I'll send you his schematic.

    Best,
    Don
     
  18. polyglot

    polyglot Member

    Messages:
    3,472
    Joined:
    Jun 12, 2009
    Location:
    South Austra
    Shooter:
    Medium Format
    Yep, that'll work. The drawback is that it's pretty inefficient, but it will work fine.

    Have you measured the resistance of each resistor and the voltage across it while running?
     
  19. Mainecoonmaniac

    Mainecoonmaniac Subscriber

    Messages:
    4,001
    Joined:
    Dec 10, 2009
    Shooter:
    Multi Format
    Should I do that?

    I haven't done that yet. But if I do, I'll have to wear my sunglasses. It's sure bright.