LED Safety/Eye Safety

Discussion in 'General Electronics Chat' started by Metalmann, Jan 2, 2013.

  1. Metalmann

    Thread Starter Active Member

    Dec 8, 2012
    700
    223
    Last night, I was playing around with a few of these:
    http://www.ebay.com/itm/290818283891?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

    and I got flashburned, similar to welding flash...maybe just as bright. I didn't think about them flipping around...and Bamm!

    Had two of them hanging on test leads, and fired them up at 9 volts, not so bad...then at 13 volts, they slipped in my grip, and naturally I glanced down....Damn!:eek: Sure as hell lit up the whole room.

    After the flash, I had trouble seeing/focusing, for a good 1.5-3 hrs. I figured they would be bright, but not that bright. Never thought that 10 Watts could be so damn bright.:eek::rolleyes:

    You can probably tell that I don't know much about new electronics, especially digital crap. Used to repair TVs and stereos in the 40s to 70s. Sometimes.:D Simple repairs....

    Anchor these babies down, when you make a test run. Only takes a few seconds.

    "Don't go into the light...";)
     
  2. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    How about this one?

    30 Watts.
     
    Metalmann likes this.
  3. Metalmann

    Thread Starter Active Member

    Dec 8, 2012
    700
    223
    Sweet.:cool:
     
  4. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    Here is the 50W hybrid (10K white and 450nm blues) I've been playing with for my reef tank.. (Will be using 2 x 50W hybrids and 2 x 50W 450nm blue LED's without optics)
    https://dl.dropbox.com/u/9726802/Reef/IMG_20121215_164729.jpg
    Only a 12 deg rise over ambient when running at 54W on the 60W Nuventix synjet cooler setup (heatsink and pwm diaphragm cooler (not a fan)) .

    and some information on "eye safety" with LED's (its VERY important) NEVER look directly at the LED's
    http://www.cree.com/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/XLamp%20Application%20Notes/XLamp_EyeSafety.pdf

    and PLEASE make sure you run these (and any led) with proper current limiting.. ie a constant current driver or other suitable circuit) DO NOT run them directly from a constant voltage power supply. I use a meanwell 60W driver for each chip
     
    Metalmann likes this.
  5. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    how many chips? would be helpful to see the actual LED
     
    Metalmann likes this.
  6. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    Specs for the hybrid LED I'm using
    50W
    Color Temperature/Wavelength: 24 chips 10000K + 24 chips 445-460nm Royal Blue
    45 mil chips
    DC Forward Voltage (VF): 24.0V-28.8V DC
    DC Forward Current (IF): 1800-1900mA

    Specs on the blue
    50W
    Color Temperature/Wavelength: 40 x 452-455nm chips
    36 mil chips
    Vf= 32-36
    If = 1750ma

    I just have the hybrid running now (great "cool white" color IMO).. Will have the blue running in 2 days when the synjet gets here.
     
    Metalmann likes this.
  7. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    I want to install one more 30W white lamp,
    and 2x 2x 10W red 660nm lamps.

    The silicon glue I use is great, bonding well enough on plastic and aluminium, but takes 8 hours to harden properly.

    I reduced the amount more and more.

    The 660nm LEDs are actually UV LEDs with a red phosphor, I removed the covering from one LED. They develope a lot more heat, so I can only use 2 chips on one sink. I had 3 on one sink, but had to increase fan voltage to 28 volts!
     
    Metalmann likes this.
  8. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    Whats the point of the silicone adhesive glue on your build? If anything it will just hurt the thermal performance.
    I simply drilled/tapped the 4 mounting holes (For #4-40 since I'm American) and used thermal grease.

    I'm gonna do a 660nm build too shortly for my fuge/macro algae tank too. Probably 2 x 10W or something like that.
     
    Metalmann likes this.
  9. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    To fix the LED to the sink.

    Yes there is thermal grease on the bottom.

    I tried carrier PCB and screws, but takes far too long to assemble.

    Actually I never heat tested the silicone glue, but that test is going now :)

    Maybe after some months it will break off- I hope not!

    Also I do use constant voltage from a SMPS.

    The idea is to keep costs for each Watt as low as possible.

    If you buy all the regular parts for 150 Watts, 5 lamps, how much does this add up?

    I paid about $60 for all the LEDs. All together my cost is less than $100, if you don't count the 500VA transformer.

    I have 4x RED 10W, 6x 10W blue, and 3x 30W pure white.

    OK actual real cost is higher, the silicone glue counts, the cables, the timer, the additional voltage regulators, and who knows what.

    But I still can keep it less than $1/Watt.
     
  10. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    Surely drilling/tapping/screws is quicker than an 8 hour dry time..
    I hope outgassing isn't a problem either for you.. No to mention it just looks like crap.

    Well good luck with that. Hopefully you know its a very bad idea as we ALL know (day 1 of any beginner electronics class) LED's require proper current limiting (via resistors/constant current circuits).. Right now you are just pushing your luck and just plain doing it wrong at the cost of a few dollars.
    Granted its obviously a DIY project but doing that in the real world is how you destroy a company (even the Chinese know better..ouch). A CC driver of that size is like $20 USD.
     
  11. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    It's for plant growing. They don't care about that.
    Also I have a good gap for the thermal grease.
    As I said I made some lamps with screws. Was a lot of effort.
    So I was thinking, how can I build such lamps with less efforts.

    20 dollar? more than two times the cost of the lamp. Why would I want to spend so much money on that.

    I carefully increase voltage until I get 1 Amps.
    Then I wait until the lamp has heated up, and regulate back a little.
    It turns out this is less than 0.1 Amps increase.

    I don't understand why people are so scared about constant voltage.

    OK 1 single LED is a bit difficult to control.
    But a string of 10 LEDs will have some internal resistance already.

    As I use a chip controlled SMPS, the voltage will not have variation. Actually I use a 78l05 as reference voltage.

    If the AUX voltage goes off, so will the output voltage. I can leave it on with the toroid powered. Indeed I only use the AUX voltage switch as a main switch.

    I don't really get why I should invest 100s of dollars, when I can have it for much less money.

    I don't need to use these LEDs for 10 years. I had CFL boxes, and many of them lived about 6 months to 8 months, then died. The heat developement if you have 15 CFLs together is horrible. As well dangerous high voltage all the time, and big boxes required, which also steal some light, and become brittle.

    I changed over completely to LED!

    And by the way, it is a pineapple plant.

    I also plan to use LM2576HV instead of the big TL494 board. So I can care for different voltages.
    Right now, I want to try hook up 4 red LEDs in series to 49.5V
    I hope it will decrease a little further, since nominal voltage is 36V AC.
    When I connect one to 19.5V (laptop adapter), they only get some 300mA.
    One blue + one red = 500mA.

    And yes the blue LEDs are a pain to look at even at 500mA!

    I have now made a lamp with 3 of them!
    6x the light volume! But have not yet powered up that lamp.

    The white LEDs are possible to look at but I would not really want to.
    The 10W blue LED radiates heat to some cm distance, even at only 500mA.

    A 30W blue lamp at full current might be safety issue. So I must place it in a direction that the LED array is not visible directly.
     
    Last edited: Jan 2, 2013
  12. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    Hey if you are happy with it and can ensure there is NO voltage fluctuation then go for it.. But a small change in voltage can mean the difference between a longlife LED and a pile of junk. And you are just replacing up front cost with assembly time.. I'd bet your solution is NO cheaper when you account for time spent. It takes me seconds to pull an LED out of the box and solder up a CC driver to it.. and I don't need to worry about measuring/stabilizing voltage,etc.. Guess you don't value your time like I do.
    Takes a minute to drill/tap 4 holes
    Takes a minute to do all the wiring with a CC driver and done..
    In under 5 minutes I can go from a pile of parts to a finished fixture that will last very long.
     
  13. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Here is a test bed for 3W LED...

    [​IMG]

    SMPS regulators have another advantage, they input less current than they output, due to it being a conversion process.

    High Power LED Flasher

    original thread
     
  14. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    You may find it a bit impractical to screw the LEDs, because there are fins where the holes should be.

    I made a PCB, pushing the LEDs on the sink, but it bends, so I need extra work to even it out, and to twist a piece of wire around it.

    Together with the screws, nuts and washers + carving the square holes + twisting the wire + soldering the LEDs, it takes nearly an hour. And it is inferior to the glueing solution. This silicon stuff is so good, the LEDs have full contact through the grease layer.

    It is not so hard you could not remove it. But it has enough bond so it won't come off from dropping the assembly, or from vibration.

    If it turns out to be capable to withstand the heat for a year or so, it will be a great material to work with.

    Even recycleable easily. A big screwdriver is enough to twist off the LED after end of life. The silicone is not a critical resource, can be burned off the sink or whatever.

    Actually I have LED driver chips here. They need a special inductor + a printed PCB + heatsink on the other side. Otherwise, they can't be used for rated current. They are rather meant to be for mass production, where they purchase a reel of matching inductors, so the price is low. Otherwise 1MHz inductors are not so cheap, for instance you can not use all kind of toroids. you need special materials for 1 MHz.
     
  15. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    The TL494 and MC34063 both have provision for current limiting. If I wanted to, using them in current mode is just a matter of a minor modification.
     
  16. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    I'd be greatly concerned with the silicones ability to keep consistent contact pressure over time (I'd bet the thermal resistance from LED to sink will change greatly over time).
    Why not just use a 2 part thermal adhesive? http://www.arcticsilver.com/arctic_silver_thermal_adhesive.htm I've been using that for securing 3W Cree leds for years.. Just took apart a 4 year old fixture this weekend and the thermal adhesive was still holding strong. Took a flat screwdriver and some muscle to pop them off for reuse.
     
  17. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    I'm not using it between the LED and the sink. I use it to secure the LED in place so it can not move.

    Under the LED, there is a layer of ordinary thermal grease.

    http://www.makingdiyeasier.co.uk/

    Unfortunately the website says nothing about temp. range.

    I have concerns too but I guess at least 80C should be sustainable.

    I don't really think I can go higher temp than boiling water.

    Most plastic materials can withstand boiling water. Some will deform. 80C should be sustainable for almost any plastics material.

    I guess the stuff I am using is silicone based.

    The price is good too, 10 Euro or something for a big cartouche. The dried out material can easily be removed with a big screwdriver. So I can keep using the cartouche for many months.
     
  18. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    What I'm saying is SKIP the silicone and regular thermal grease and JUST use thermal adhesive under the LED.
    The thermal adhesive alone will secure the LED and act as a thermal grease all in one.

    Any movement in that silicone after application could simply lift/raise the LED off the heatsink surface and just ruin the thermal resistance. But heck it will probably be dead from thermal runaway long before that :) ha ha ha
     
  19. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    I have run one of the lamps for a day already.
    The grease also has some adhesion to the sink.
    It all works out fine so far.

    What I am afraid off is the silicone could become brittle after some months, and then the LED falls off.

    I will use the lamps freely suspended with flying wires.
     
  20. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,770
    970
    That would be the least of my worries doing something like that..

    If its brittle after a few months or even few years its NOT 100% silicone.
    The best thing about silicone caulk is how it retains its flexibility year after year.
    But its NOT dimensionally stable enough which is needed for an application like this.

    The loss or inability to ensure a proper bond/contact pressure would be my #1 concern..clearly.
    Contact pressure (up to a point) is directly related to thermal resistance.
     
    Metalmann likes this.
Loading...