36 LED reading lights -- HOT

[Josey]

I'm testing out some 36-LED 12V reading lights from TheLEDlight.com to replace the 12-LED bulbs that I have been using. The 36-LED bulbs are wonderously bright, but they are running hot. There is a row of resistors in the center with exposed wires that get too hot to keep a finger to. My smaller LED bulbs burn no more than warm. These are using about 350 milliamps. My question is whether these are overdriven and whether I can expect 100,000 hours from them.

Anyone have any ideas? Josey

[Nerd]

They are using around 32 ma per LED (including resistor losses) if my calculations is not wrong. As such, I would expect that they are mildly overdriven.

At the 100K hours mark, they should have their brightness dropped by half. At which, I don't think make a big difference which is because they gradually dim bit by bit over the years. [img]/ubbthreads/images/graemlins/smile.gif[/img]

[Josey]

Thanks, Nerd. I did a little more testing and now I'm really concerned, but I can see why the 36-LED light gets so hot. The light is pulling 415 milliamps or 11.5mA per bulb, including the 18 67 ohm resistors (one for each three bulbs)(blue/red/black/gold/brown).

My 12-LED light pulls 98mA, or 8mA per bulb(I don't know what kind of resistors they use because they are epoxied in)
My 24-LED light, which is about as bright as the 12, pulls 76mA, or 3.2mA per bulb.

So the super bright 36-LED is pulling a lot more amps per bulb; so I am concerned about it's life (especially since the reliability of the resistors is just 1 percent). I want to be able to give a good recommendation to people with off-grid solar systems (like me), but at $165 per bulb, I want to make sure it's good advice.

I've got a light meter coming, so I'll post the brightness of these bulbs when I get it.

Josey

[jtr1962]

Are the LEDs all in parallel, or are there 18 parallel series strings of 2, or perhaps 12 parallel series strings of 3? Your measure of the current is meaningless without knowing how the LEDs are connected. If they are all in parallel, then based on your measurement of 415 mA each diode is getting 11.5 mA. However, if they are in 18 series strings of 2, each diode is getting 23 mA. 12 series strings of 3 would mean a diode current of close to 35 mA. What is the voltage across the entire array (not including the voltage drop across any resistors? This should tell you how the diodes are connected. If it is about 3.5V then they are all in parallel. 7V is series strings of 2, 10.5V is strings of 3, etc.

As for lifetime, even driven at specs 5mm white LEDs degrade to 50% brightness after 6000 or so hours due to the epoxy in the package darkening from the short-wavelength blue light. If the diodes are overdriven and get very warm, the lifetime may be dramatically less. The 100,000 hour figure only applies to colored LEDs, except for shorter wavelength ones like blue and UV which have shorter lifetimes. I'm not aware that anything has been done yet to fix this problem for standard 5mm LEDs, although the white Luxeons use a clear gel instead of epoxy and do get at least 50,000 hours of life when driven to spec.

[Josey]

Thanks, JTR1962. There are actually 12 resistors, one feeding each set of 3 LED bulbs. I can't see how they are wired, but these are 12 volt lights, and I measured 12.5V from the top of the resistor to the far leg of the 3-LED string it feeds.

The heat really worries me. The 36-LED is heating the bulbs from a low of 145 degrees F to just over 200 degrees F. By comparison, the 12-LED light is running about 70 degrees F.

[Josey]

Now I'm starting to see what Nerd and jtr1962 are saying. With 12 62 ohm resistors and 36 bulbs, there are probably 12 parallel circuits, so the 415mA gets divided by 12, meaning each bulb is running at 34mA or so. Since they are not heat sinked, the heat is really building up and will probably shorten their lives.

I took the plastic diffusers off to cut down on heat build up and to release more light.

[jtr1962]

According to the info you provided and the fact that it is a 12V lamp, there are definitely 12 parallel strings of 3 LEDs in series, and each LED is getting 34 mA as you said. This means they are overdriven, but not dangerously so. However, you might very well be cutting the useful life to less than 1000 hours instead of the 6000 to 10000 5mm white LEDs usually get because of the heat buildup. There wouldn't be a problem with only a few LEDs, but when you put that many in a small space you do have heat buildup. 415 mA at 12.5V means you have to get rid of 5.2 watts of heat (actually less since some of the power comes out as light, and the rest as heat).

Removing the diffuser is a good idea. 145°F isn't too hot for LEDs, but since they make less light as the temperature rises, if it is at all possible reduce the lamp current (maybe substitute 82 or 100 ohm resistors for the 62 ohm ones, or just put a 2 or 3 ohm resistor in between the lamp and the 12V line). The LEDs will run cooler, and might not produce noticeably less light despite the lower current. You'll also get something closer to the rated life.

I have similar problems, but to a lesser degree, with the 20 LEDs in my bicycle headlight. On high (24 mA per LED), I'm dealing with about 1.5W of heat in a small (1.5"x2.5") area. They do get noticeably hot, and the lifetime will probably only be a few thousand hours. However, this is still worlds better than the few hundred hours of the halogen bulb build they replaced, and they give about twice the light with only 70% of the power.

[gwbaltzell]

There may be another issue if you use the light when the battery is being charged in that the voltage will more likely be 14.4V.

For this cost there might be better solutions. I think a top down approach might be better. How much light do you need for reading? The US Government Mil-STD 1472F (the Human Factors Guide) suggest 10-20 footlamberts. If the paper we were reading were perfect the footcandle reading of the light source would be the same. However in the real world be need at lot more. The Government's 2005 Solar Decathlon requires 30 footcandles for reading. We also need some general room lighting in addition. Remember the basic definition of glare is excessive contrast. The Decathlon wants the whole room at the 30 fc level but I doubt if this is yet practical for anything but a large solar installation. BTW, I am NO solar expert. My best guess, and only that, is 7 fc. If we had a fair amount of energy my likely choice for the reading light would be an infrared conserving halogen reflector lamp. This operates at a efficiency about the same as a Luxeon LED, cost about $12.50 US, 4000 hour life, requires no special equipment, has a CRI of 100 (though color temp. is a little lower than the most desirable). Soft starting equipment could add up to 50% to the life. Main drawback is the smallest size I can find is 20 watts, but might illuminate a whole newspaper to the right level. The most efficient system would be a well designed fluorescent which could operate around three times the efficiency of a Luxeon or IRC halogen. Cold cathode lamps, while now having a longer life, do not approach the effieciency of hot cathode or electrode-less. Bulbs are available from 5W up. Both a Luxeon or fluorescent system could be dimmed without a color shift. I'm not aware of any good off the shelf reading lights based on any of these, though the halogen would be very easy to make. Ideas anyone?

[jtr1962]

In the smaller sizes fluoroescent isn't that much more efficient than LEDs. It has to do with the arc length. A 4W tube emits about 180 lumens (45 lm/W), an 8W tube about 400 lumens (50 lm/W). I've already read about some higher binned Luxeons on this site that have a minimum efficiency of about 40 lm/W according to the bin number (those X-binned 5-watters in particular). At this power level, it really is almost a toss-up between LED, halogen, and fluorescent. LED and fluorescent both have a slight edge in efficiency and color temp, and halogen has the edge in ease of use (i.e. no driver circuit required). If I were doing this, I might use a pair of high-bin 3 watt Luxeons, or a high-bin 5-watter if lifetime wasn't that great a concern, and couple them with a highly efficienct step-down switching regulator.

Once you get to higher power levels (10W and over), fluorescent starts to become the clear choice. If you have the space, a 32W linear tube gets efficiences up to 100 lm/W depending upon make, and can be dimmed without shifts in color temp. Of course, at current rates of progress in a couple of years LED will blow everything away except for higher-wattage discharge lamps. Nichia is slated to release 60 lm/W white LEDs in 2005, and I imagine Luxeon will follow suit.

[Josey]

Even at similar efficiencies, the big advantage of the LED lights is that they are task lights. While most lights throw photons in all directions, the LED focues on whatever I'm reading. I don't need to light up the whole room to read. So the LEDs let me use this light more efficiently.

The 36-LED bulbs put out a huge amount of high-quality, high-contrast light for just 5 watts. No 5-watt fluorescent or incandescent can come close to putting out that much light in the task area.

For solar power in the rainforest, these LEDs cannot be beat.

[georges80]

[ QUOTE ]
Josey said:
Even at similar efficiencies, the big advantage of the LED lights is that they are task lights. While most lights throw photons in all directions, the LED focues on whatever I'm reading. I don't need to light up the whole room to read. So the LEDs let me use this light more efficiently.

The 36-LED bulbs put out a huge amount of high-quality, high-contrast light for just 5 watts. No 5-watt fluorescent or incandescent can come close to putting out that much light in the task area.

For solar power in the rainforest, these LEDs cannot be beat.

[/ QUOTE ]

For a tasklight - have you seen my flexiled coupled with a uFlex driver? That gives you a Luxeon that can be run from about 10mA (at the dim setting) all the way to 400mA at the bright. uFlex is a switcher - so feed it from 4-20VDC to run the Luxeon 1W in the flexiled head. At 12V it only draws about 100mA to drive 400mA into the Luxeon 1W. The flexiled head projects a smooth 60 degree flood beam from its sandblasted reflector surface.

george.

[gwbaltzell]

OK, I failed to notice the rainforest location. If I had, I might not have brought up fluorescent. However before we leave that entirely I was speaking of a well designed one not a US$5 battery operated POS. There are a few bulbs with better efficiencies: 6W 330lm, 8W 540lm, 11W 750lm and 13W 930lm. Off the shelf line voltage ballast however add another 3W. CRI is 85. Found this source of other fluorescent, LED and halogen fixtures designed for solar. Granted even the best designed fluorescent fixture can't be as directional as a halogen or Luxeon, this is not really a bad thing in a reading light. And that 20W halogen is available in 5 different angles from 10° to 60° and cd rating from 6,000 to 450 (2800cd for the 24° one). A Luxeon III and 12V power supply will run less than US$40 and produce 80lm (though it will need a heatsink and fixture). This should be about the same as the 36 LED array which is likely putting out 2 - 2.4 lm per LED (my best guess 20cd at 30°).

I am looking forward to seeing the light output of your array when you get your light meter.

George

edit: fixed misplaced decimal point.

[Josey]

Thanks George and George, those are good suggestions and great info. I did put the Meterman 631 on my LED reading lights (12-LED, 24-LED and the 36-LED bulbs).

The 12-LED measured 787 lux at 1 foot, and the 24-LED was about the same at 758 lux. (The 24-LED has more bulbs, but runs less current, 76mA vs. 98mA).

The 36-LED at one foot measured 883 lux, which doesn't seem like that much more but the meter measures only a small area (smaller then a dime) and is very sensitive to hot spots.

At two feet, the 24-LED and the 12-LED measured about 205 lux, which teh 36-LED measured 378 lux.

At two feet away and with a 6-inch offset (so the meter measured outside the hot spot), the 12- and 24-LED bulbs measured 120 lux, and the 36-LED measured 241 lux, indicating a much larger spill.

These 36-LED lights are awesome, with very high-quality light; I just hope that they have a good life span because they are so expensive. Josey

[gwbaltzell]

Since one lux = 0.0929 footcandles the suggested minimum of 30 fc would be 323 lux. You are getting a little over this at two feet. Is this a comfortable distance? Is the entire reading area covered or how large an area do you get until the intensity is half? Is it a fairly uniform amount of light? Sorry, these are a few of the questions I would ask myself if I were reviewing this.

GW

[Josey]

Hey gwbaltzell, I use two adjustable lamps about 3 feet apart and move them depending on the use. With the 12- and 24-LEDs I could easily read a book or magazine from about 18 inches, but newspapers required that I move the lamps a lot to illuminate whatever part of the paper I am reading. \\With the 36-LED lamps (two of them) I can read anything easily with both lights at 2 feet above the desk. In fact, two of these lights is almost too much light. But the light is uniform over my desk area, about 2 feet by four feet. So with two lamps, I guess I'm getting well over twice the light I need; sure seems like it, but wow can I see things clearly.

[gwbaltzell]

Josey

Thanks, I was thinking one was not enough to easily read a newspaper. Nice to know real world sometimes fits a SWAG (scientific wild a** guess).

[jtr1962]

[ QUOTE ]
Josey said:
Even at similar efficiencies, the big advantage of the LED lights is that they are task lights. While most lights throw photons in all directions, the LED focues on whatever I'm reading. I don't need to light up the whole room to read. So the LEDs let me use this light more efficiently.


[/ QUOTE ]
I was thinking of something similar. While a fluorescent might be twice as efficient as an LED, when the goal is simply to focus as much light as possible in a fairly narrow beam you will have losses, and the net efficiency might end up being no better than LEDs which throw nearly all their light into a fairly narrow beam. I noticed this once when comparing the light from a twin 8W tube 12V fluorescent (the tubes are somewhat underdriven at about 5W each) to 5 1 watt Q bin Luxeons (theoretical output ~150 to 200 lumens total). I figured the fluorescents should be putting out about 500 lumens total based on the drive level, yet the room seemed just as bright with the Luxeons putting out about a third as many lumens. I attributed the losses in part to the diffuser, and also to the poor optics of the fixture. The LEDs, on the other hand, put 100% of their output into the room. Ditto for the LED bike light I made versus the halogen bulb. There were undoubtably losses due to the reflector, so 100% of the bulb's output didn't go into the beam while most of the LED's output did. If you're interested in task lighting then today's best LEDs (30 to 40 lm/W) are probably as efficient as most low-wattage fluorescents once you account for the losses redirecting the fluorescent light's output.

[gwbaltzell]

Decided to try a fluorescent I had on hand. Its the same as one Backwoods Solar is selling for $22 and was $9.97 list (bought for $6.97). A Philips R-30 Flood 120V, 15W, 500lm, 8000hr. I removed the front diffuser. Readings taken with a Minolta Auto Meter IIIF with the standard hemisphere in place. ASA set to 100 and EV converted to lux using the formula (2^EV)*2.5
At one foot 1190 lux, 50% point at 12 inches from center (24 in circle).
At two feet 557 lux, 50% point at 24 in
Of course I don't think fluorescent is a good choice if typical humidity is more than 85% unless very carefully designed.

Added: BTW, I think you might get better overall performance from georges80's lamps. Less cost, less power, more and better light. I don't get a commission. [img]/ubbthreads/images/graemlins/frown.gif[/img]

From the current home page of TheLEDlight: Luxeon Star LED's "White models due sometime in 2003." [img]/ubbthreads/images/graemlins/smile.gif[/img]

[Negeltu]

[ QUOTE ]
jtr1962 said:


As for lifetime, even driven at specs 5mm white LEDs degrade to 50% brightness after 6000 or so hours due to the epoxy in the package darkening from the short-wavelength blue light.

[/ QUOTE ]

I thought only UV 5mm leds had went to half brightness before 100,000hrs. Not the white leds.

[gwbaltzell]

[ QUOTE ]
Negeltu said:
I thought only UV 5mm leds had went to half brightness before 100,000hrs. Not the white leds.

[/ QUOTE ]

You'll find this Luxeon lumen maintenance data widely quoted on CPF.

[Negeltu]

Well, that's interesting. I've never heard that before, but that is what the data sheets show.

[Josey]

This link suggests that the the Luxeon Star is the better choice for LED reading lights: more light, longer bulb life. Anyone know a source for LS reading lights or a design guide to build your own?

[gwbaltzell]

[img]/ubbthreads/images/graemlins/banghead.gif[/img]

[TORCH_BOY]

Do the 36-LED 12V reading lights throw pure white
or do they have artifacts in the emitted light.
I have been thinking of purchasing one purely
for reading.

[Josey]

Hey Torch Boy: No artifacts. The light is white, close to full spectrum I would guess, and high contrast. It's really nice light.

[TORCH_BOY]

Thanks Josey, I will have to get one.

[Josey]

I talked to the design engineer today and found out that the lights were overdriven, which is why they were running too hot. Some time ago, there was a factory mixup and some lights were built with 62 ohm resistors instead of 82 ohm resistors. My bulbs were from the 62-ohm batch that didn't get recalled. The company is replacing the lights for me. Thanks to all of you guys with your great information.

And OK GWB, I know you insulted my intelligence somehow, and you're probably right since I can't quite figure out how. Guess I should have stayed in school. Josey

[gwbaltzell]

I in no way was trying to insult your intelligence. It was after two attempts by me and one by georges80 to get you to look at Luxeon based lights, and with me providing a link to georges80 pre-made lights, you asked about Luxeons. This made me think I was either [img]/ubbthreads/images/graemlins/banghead.gif[/img] or [img]/ubbthreads/images/graemlins/dedhorse.gif[/img]

[Jarhead]

[ QUOTE ]
jtr1962 said:
In the smaller sizes fluoroescent isn't that much more efficient than LEDs. It has to do with the arc length. A 4W tube emits about 180 lumens (45 lm/W), an 8W tube about 400 lumens (50 lm/W). I've already read about some higher binned Luxeons on this site that have a minimum efficiency of about 40 lm/W according to the bin number (those X-binned 5-watters in particular). At this power level, it really is almost a toss-up between LED, halogen, and fluorescent. LED and fluorescent both have a slight edge in efficiency and color temp, and halogen has the edge in ease of use (i.e. no driver circuit required). If I were doing this, I might use a pair of high-bin 3 watt Luxeons, or a high-bin 5-watter if lifetime wasn't that great a concern, and couple them with a highly efficienct step-down switching regulator.

Once you get to higher power levels (10W and over), fluorescent starts to become the clear choice. If you have the space, a 32W linear tube gets efficiences up to 100 lm/W depending upon make, and can be dimmed without shifts in color temp. Of course, at current rates of progress in a couple of years LED will blow everything away except for higher-wattage discharge lamps. Nichia is slated to release 60 lm/W white LEDs in 2005, and I imagine Luxeon will follow suit.

[/ QUOTE ]

jtr1962,


JKL Lamps P/N BF26318 3.4 Watts puts out 200 lumens.

This works out to 58.82 lumens per watt. I've been able
to drop the power consumption on this bulb by reducing
the loss of power to metal around it, and have gotten even
higher efficiencies than that.

I use in our products some fluorescent bulbs from another manufacturer that have even higher efficiency ratings,
but I don't recall the exact numbers, but saw over 50% increase in efficiency over the JKL lamps.

I'd consider it safe to say they were +75 lumens a watt.

http://www.lcdl.com/ultrabrite4_1.html

[Josey]

Thanks GWB. I had gone to the Backwoods link and didn't see anything and thought that was what you meant. But I went back again and saw I missed the georges80 and have sent an email.

Don't they have a beginners slope on this forum?

And thanks Jarhead. That's good info; I'll be checking it out.