Results 1 to 18 of 18
  1. #1
    mtbr member
    Reputation: peterh88's Avatar
    Join Date
    Apr 2007
    Posts
    41

    LED intensity has decreased

    Hey Guys,
    Hope somebody can help here.
    I made my own triple leds's a about 5 months ago, the first one I used Cree P4's.
    The second light I used Cree Q2's.
    check out my original thread here Another homemade triple Cree
    OK now to my problem, you will see from the beam shots in the original thread the Cree Q2's are a lot brighter than the P4's.
    I have been using the Q2 light exclusively for the last 4 months and the light output seems to be decreasing.
    I just bought a new battery pack and wired it up to my lights, so armed with two battery packs I am able to run both lights side by side. The Q2 light is definitely not even as bright as my P4 light.
    Do led's degrade over time?
    Why would the output of the light be decreasing.
    Maybe the leds are degrading due to heat?
    I know you guys will have some answers for me

    Regards

    Peter

  2. #2
    Master of None
    Reputation: presslab's Avatar
    Join Date
    Jan 2007
    Posts
    1,462
    As LEDs age the phosphor and the epoxy lens degrade. Overheating the LED will cause this to occur MUCH faster. This will decrease the light output, maybe that's what happened to you.

  3. #3
    mtbr member
    Reputation: bikeny's Avatar
    Join Date
    Feb 2004
    Posts
    1,602
    I think you have a heat problem. I read through your initial post that you linked and checked out all of the pictures. I sounds like you made the heatsink only 10mm thick for the second Q2 light, and you do not have any kind of thermal pathway from the heatsink to the housing. Also, you say the housing takes a while to get warm during use. All of these things tell me that the heatsink inside and the LEDs are getting very hot, and the heat is not being disipated efficiently through the housing.

    I would try the following. Turn on the light for like 10 to 20 minutes. Then turn the light off and quickly take off the back cap so you can see how hot the heatsink is. I have a hunch that it will very hot in there!

    Good luck,
    Mark

  4. #4
    Spanish biker
    Reputation: msxtr's Avatar
    Join Date
    Dec 2006
    Posts
    734
    Quote Originally Posted by bikeny
    I think you have a heat problem. I read through your initial post that you linked and checked out all of the pictures. I sounds like you made the heatsink only 10mm thick for the second Q2 light, and you do not have any kind of thermal pathway from the heatsink to the housing. Also, you say the housing takes a while to get warm during use. All of these things tell me that the heatsink inside and the LEDs are getting very hot, and the heat is not being disipated efficiently through the housing.

    I would try the following. Turn on the light for like 10 to 20 minutes. Then turn the light off and quickly take off the back cap so you can see how hot the heatsink is. I have a hunch that it will very hot in there!

    Good luck,
    Mark
    Yes, I think that this are the problem, surely don't have enought heat sink, then the leds begin to damage and do less bright light.

    Greetings - Saludos

    msxtr
    Warning!!! my english is very very bad, sorry.

    Easy DIY led light1
    Easy DIY led light2

    The Beast!!!

  5. #5
    mtbr member
    Reputation:
    Join Date
    Oct 2004
    Posts
    3,051
    Take a look at the following links:

    http://www.colorkinetics.com/support...EDLifetime.pdf

    http://www.philipslumileds.com/pdfs/WP12.pdf

    As others have said, and you yourself suspect, you most likely have a heat problem.

  6. #6
    mtbr member
    Reputation:
    Join Date
    Nov 2005
    Posts
    500
    Well, it could be heat, that's one possibility. Running LEDs hot will definitely decrease their lifespan. I assume you've checked the voltage and current at the LEDs to see if it is within spec.

  7. #7
    My Rusty Knee
    Reputation: Stuart B's Avatar
    Join Date
    Mar 2005
    Posts
    1,118
    have they turned blue too? If so this is another indication that they have overheated.

    Stu
    What exactly is a rigid hard tail?

  8. #8
    mtbr member
    Reputation: peterh88's Avatar
    Join Date
    Apr 2007
    Posts
    41
    I think you have a heat problem. I read through your initial post that you linked and checked out all of the pictures. I sounds like you made the heatsink only 10mm thick for the second Q2 light, and you do not have any kind of thermal pathway from the heatsink to the housing. Also, you say the housing takes a while to get warm during use. All of these things tell me that the heatsink inside and the LEDs are getting very hot, and the heat is not being disipated efficiently through the housing.

    I would try the following. Turn on the light for like 10 to 20 minutes. Then turn the light off and quickly take off the back cap so you can see how hot the heatsink is. I have a hunch that it will very hot in there!
    Thanks for the info Mark, I will do that with the lights.

    Well, it could be heat, that's one possibility. Running LEDs hot will definitely decrease their lifespan. I assume you've checked the voltage and current at the LEDs to see if it is within spec.
    ,
    At the risk of sounding like an idiot, how do I check voltage and current at the leds?

  9. #9
    mtbr member
    Reputation: California L33's Avatar
    Join Date
    Jul 2005
    Posts
    699
    Quote Originally Posted by peterh88

    At the risk of sounding like an idiot, how do I check voltage and current at the leds?
    You'll need a multi-meter. Even an inexpensive model should be able to do this.
    To the troll mobile, away...

  10. #10
    mtbr member
    Reputation: peterh88's Avatar
    Join Date
    Apr 2007
    Posts
    41
    I think you have a heat problem. I read through your initial post that you linked and checked out all of the pictures. I sounds like you made the heatsink only 10mm thick for the second Q2 light, and you do not have any kind of thermal pathway from the heatsink to the housing. Also, you say the housing takes a while to get warm during use. All of these things tell me that the heatsink inside and the LEDs are getting very hot, and the heat is not being disipated efficiently through the housing.

    I would try the following. Turn on the light for like 10 to 20 minutes. Then turn the light off and quickly take off the back cap so you can see how hot the heatsink is. I have a hunch that it will very hot in there!
    Ok I measured the temp of both the heatsink and the housing and got the following results after 15 minutes of operation with the light sitting on a bench.

    Housing temp 45 deg C - 114 deg F
    Heatsink temp 56 deg C - 134 deg F.

    Are these temps to hot for the leds.
    Please keep in mind this is the first time the light has sat still with no air flow for so long.
    I dont know if this would make a difference but I use 14.4 voly Li-ion and a 3021 buckpuck with the dimmer pot.
    Thanks

    Peter

  11. #11
    mtbr member
    Reputation:
    Join Date
    Nov 2005
    Posts
    500
    Quote Originally Posted by peterh88
    At the risk of sounding like an idiot, how do I check voltage and current at the leds?
    You can measure voltage across each LED's + and - tab. There will be some slight variation, but all should be fairly close (approx 3.7v). The only way to measure current is by placing your meter's leads in series with the LEDs. Of course, that means you would have to break the existing series wiring of your LEDs and put the meter in line just as if it were an additional LED wired in series. You should see about an amp, assuming that's what you were intending to drive them at. (Make sure you put your meter on a setting that can handle one amp).

    I believe most LEDs are spec'd out at 25 degrees C. at the die. (about room temp) Anything hotter than this results in lost lumens (immediately) and will accelerate the degredation of the LED's performance and useful life. This is why you need to keep them as cool as possible. Moving air works wonders, which is why I don't bother with heatsinks. I never run my lights stationary for more than a minute inside while bench testing (if longer time is needed, I point a small fan at it) at a full one amp drive current. I usually tell folks that if you're not gonna keep 10mph of air flowing across the light to add additional heatsinks. But remember, moving air only works to cool your LEDs if you are conducting the heat away from your LEDs to the outside of the housing that is receiving the air flow.

    This chart from Cree shows how output decreases with increase junction temp: (remember, this is junction temp, not heatsink or housing temp.)


    It's possible that one of your LEDs might be bad and is pulling them all down. So do the voltage check. If one has an unusually high or low voltage, it is bad.

    -Good luck.
    Last edited by achesalot; 07-14-2007 at 07:09 PM. Reason: Added chart

  12. #12
    mtbr member
    Reputation: peterh88's Avatar
    Join Date
    Apr 2007
    Posts
    41
    Many thanks to all who replied.

    I have performed the tests as suggested and found current and current and amps to be within specs.
    I am not convinced that I have a heat problem as the test shows 56 degs standing still.
    Anyway I have ordered some new led's and I am making my heatsink slightly larger and I am going to use some thermal compound between the heatsink and the body of the light.

    I will see how this goes after a couple of months.
    Has anybody else had the same or simular problem?

    Regards

    Peter

  13. #13
    Gone riding
    Reputation:
    Join Date
    Jan 2004
    Posts
    3,369
    Sorry for the delay and I know some of the points below have already been covered, but I think it’s worth mentioning again.


    There are many factors that lead to a decay in lumen maintenance, but the underlying reason for most cases is an excessive emitter junction temperature. Please note that the junction temperature is the internal temperature between the die and the substrate inside the emitter itself, and is quite likely to be another 10, 20, 30 or more degrees hotter then the heat sink in your case.

    Having a large heat sink with a big thermal mass is fine, but if you don’t have a way of dissipating that heat somehow you’re not going to be able to keep temperatures nice and low. I still believe that you don’t have a suitable thermal connection between your housing which is exposed to the cool air, and the hot heat sink. Assuming you measured both the heat sink and housing close to the emitters, the fact that there is such a large difference in temperature between your housing and heat sink with no moving air shows that you certainly haven’t got an ideal thermal connection between the two.


    You also mentioned in your original thread that your Q2 system was smaller and lighter with a thinner heat sink, which may also be a contributing factor, although I still believe your issues are with the thermal path. You need to be thinking about minimising the thermal resistance between each of the elements in your thermal path, which in your case is emitter / star / heat sink / housing. Obviously you can’t do anything about the emitter and the star, but you need to think about the connections between the star and your heat sink, and your heat sink and your housing.

    To start with, did you use any thermal interface material between the stars and the heat sink? Was the heat sink reasonably smooth? The connection between your heat sink and housing is even more important, and I suggest looking into some thermal tape or thermal adhesive at the very least. Ideally a housing design in which the heat sink is exposed to the air, or indeed is the heat sink would be well worth considering.


    I recommend having a read of the "Thermal Management" and "Reliability" documents on the Cree website here for more information and a better explanation.


    Good luck with your next light!

    Dave.

  14. #14
    mtbr member
    Reputation: California L33's Avatar
    Join Date
    Jul 2005
    Posts
    699
    Quote Originally Posted by Low_Rider
    Sorry for the delay and I know some of the points below have already been covered, but I think itís worth mentioning again.


    There are many factors that lead to a decay in lumen maintenance, but the underlying reason for most cases is an excessive emitter junction temperature. Please note that the junction temperature is the internal temperature between the die and the substrate inside the emitter itself, and is quite likely to be another 10, 20, 30 or more degrees hotter then the heat sink in your case.

    Having a large heat sink with a big thermal mass is fine, but if you donít have a way of dissipating that heat somehow youíre not going to be able to keep temperatures nice and low. I still believe that you donít have a suitable thermal connection between your housing which is exposed to the cool air, and the hot heat sink. Assuming you measured both the heat sink and housing close to the emitters, the fact that there is such a large difference in temperature between your housing and heat sink with no moving air shows that you certainly havenít got an ideal thermal connection between the two.


    You also mentioned in your original thread that your Q2 system was smaller and lighter with a thinner heat sink, which may also be a contributing factor, although I still believe your issues are with the thermal path. You need to be thinking about minimising the thermal resistance between each of the elements in your thermal path, which in your case is emitter / star / heat sink / housing. Obviously you canít do anything about the emitter and the star, but you need to think about the connections between the star and your heat sink, and your heat sink and your housing.

    To start with, did you use any thermal interface material between the stars and the heat sink? Was the heat sink reasonably smooth? The connection between your heat sink and housing is even more important, and I suggest looking into some thermal tape or thermal adhesive at the very least. Ideally a housing design in which the heat sink is exposed to the air, or indeed is the heat sink would be well worth considering.


    I recommend having a read of the "Thermal Management" and "Reliability" documents on the Cree website here for more information and a better explanation.


    Good luck with your next light!

    Dave.
    This is what everyone here is wishing they wrote- great post. Maybe we should have a DIY LED sticky, with this as message one. Heatsink size doesn't matter unless the heat goes into it through some sort of thermal coupling. A hot heatsink is a good thing, not bad- then get that heat into the air to combat global cooling
    To the troll mobile, away...

  15. #15
    _I_
    _I_ is offline
    mtbr member
    Reputation:
    Join Date
    Jun 2007
    Posts
    206
    I wouldn't go so far as to say a hot heatsink is a good thing, while it is indicative of heat transfer from the die it is also indicative of a higher thermal density, lower thermal gradient to the die. In other words, it is usually a sign the heatsink is a bit on the small side or the driving current a little overambitious for the design.

    If an (star mounted) LED is ran for a fair period of time to reach an equilibrium in device case temp, say 30 minutes, and the enclosure is relatively sealed, you can end up with a similarly hot heatsink with a great thermal conduction between the star -0R- a poorer thermal conduction to the star, because the amount of heat radiated from this closed system is the same (actually with the hotter star that system is generating more heat because it lowers the forward drop voltage->more current at higher temp but let's ignore that for this example), and with the poorer conduction scenario the star simply gets hotter.

    The central idea is correct, that of utmost importance is the best star (or die with a naked LED) - next transmission medium possible (whether that medium be directly to a typical heatsink or the metal case) interface efficiency possible. Towards this end it is a bit puzzling that arctic epoxy is the preferred choice as it does conduct worse than plain old silicone grease, but even so there will come a point where these bike lights are still more limited by the same heatsink surface area to keep weight and size down.
    Last edited by _I_; 07-17-2007 at 07:43 PM.

  16. #16
    Gone riding
    Reputation:
    Join Date
    Jan 2004
    Posts
    3,369
    Quote Originally Posted by California L33
    Maybe we should have a DIY LED sticky.
    It has been discussed by a few of us in the past. Just about everything one needs to know is actually hidden away in various posts here on the forum, but it would be nice to combine it all in to one document that covers everything from batteries, drivers, emitters, thermal design and lumen maintenance and optical components.

    Iím actually slowly working on a few bitís and pieces that Iíll eventually post in a FAQ / Lighting Information style sticky once I have a little more time and a few people have looked over and checked things.

    There are a number of regulars here who have some excellent guides.


    Quote Originally Posted by California L33
    A hot heat sink is a good thing, not bad - then get that heat into the air to combat global cooling
    While I do know what you mean, it could be interpreted the wrong way by those who might get confused by my ramblings. You could also say a cold heat sink is a good thing, but itís not if itís cold because thereís such a large thermal resistance between the emitters and the heat sink that no heat is being transferred between the two.

    Ultimately, the aim of the game is to connect the emitters to something that can dissipate the heat (draw it away from the emitter) with the lowest thermal resistance between all connections as possible.

    In Peterís case where he has a heat sink (perhaps better described as a base plate as really itís just a thermal mass) connected to an outer housing (a true heat sink as it dissipates / radiates itís heat to ambient air), itís that housing that should be dissipating the heat to keep the base plate the emitters are connected to cool. With a high thermal resistance between the two, the housing is probably having a minimal effect on dissipating the heat of the base plate.

    Dave.

  17. #17
    mtbr member
    Reputation: peterh88's Avatar
    Join Date
    Apr 2007
    Posts
    41
    Thanks for all the info Dave,
    I have started to re-build the light today.
    I have made the heat sink larger and this time the the surface is a lot smoother. The last heat sink surface was not very smooth at all.
    Last light I used some thermal paste from the local Jaycar shop.(no brand name or specs available) But only under the stars, nothing between the heat sink and housing.
    I have used some artic silver thermal paste this time around and will be using the thermal paste between the housing and heatsink.
    The last build was not water proof in any way at all and this may have contributed to the failure of the leds. As the light got saturated on a few occasions.
    This time I am going to spend some time to water proof with silicone sealant designed to seal automotive gearboxes with no gaskets.
    An added bonus for this light, the new Cree Q5 binned led's have been used.
    I will add some photos as time permits

    Cheers

    Peter

  18. #18
    Gone riding
    Reputation:
    Join Date
    Jan 2004
    Posts
    3,369
    Were you able to make the heat sink a press / friction fit in the housing? Getting the heat away from the emitters is critical in any light, but it’s especially important between the heat sink and housing with your design.

    Also remember to try and make your thermal paste as thin as possible, a thick layer would probably be more detrimental then none at all with a nice and smooth base.

    Looking forward to some photos of those Q5’s in blazing action!

    Dave.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •