high power led reverse lamp build (Rev 1)

[soundman98]

this is my continuation of this thread. i figured making a new thread for my specific design would be better then crapping all over blurred talons thread...


to recap:
i had tossed the idea around in my head for a month or two before blurred talon started the thread, but after his thread, the ball started rolling...

in his thread, i posted some output shots of a cree xp-g r5 flashlight, good for about 350 lumens to compare to my current filament reverse lamps.

i wasn't happy with the overall light output, so i decided to order some cree xm-l's to be sure that i would have enough illumination.

to date, i have ordered 6 XM-L T6 high power led's, and 6 3000mA constant-current drivers.

i intend to use 4 total, or 2 per side.

for reference, a picture of the back of my eclipse:




measurements:
the reverse lamp portion of the lower housings is roughly 6" wide.

the area behind the reverse lamp has about 4" on the outside of the opening and 5" on the inside of the opening of clearance between the metal body-- measured off the front of the reverse lamp housing.

there is a crashbar about 2-3" below the reverse lamp housing, otherwise, the entire bumper area is pretty clear of obstacles.

and also for reference, the housings that i intend to use for the retro:





as i talked about in the other thread, these led's put out a lot of heat. with the led's mounted on a 6"x1"x1" square aluminum tube, they had a heat rise of 5 deg per min in a 5 min run time.

admittedly, a short length of raw square tubing is a poor heat sink, but has taught me a very good lesson that serious heat sinking is required for led's like this..


problem:

• average backup time is about 5-10 minutes-- about 200ft of alley, 30-40 seconds for garage door to open, and time to back into garage.

assembly issues to overcome:

• must run for 10 minutes, want to shoot for 20 min run-time.
• must stay below 100 deg F
• getting the 2 led boards+ heatsink assembly to look good under the clear lens.

assembly goals:

• cost should be below $50 per side for heatsink components(i'm not going to complain if it hits 55-60).
• doesn't matter if design is active or passive.
• passive is the preferred method due to the location getting exposed to some dirt/water.
• would prefer small-enough components to allow for a mostly-PNP design that would not require bumper removal.

that about sums up what i am trying to accomplish.

right now, i am looking at computer processor heatsinks, as many processors put out about the same amount of heat. though mounting the heatsink, and dust/water in a cooling fan is a little worrisome..

any ideas anyone has would be greatly appreciated.

[JaySilver]

I don't know if its possible but, would you try water cooling instead of air cooling? Like what they do in computers?

[soundman98]

i briefly considered it, but after recently talking about it in a unrelated forum, the minimal performance gains over standard heatsink cooling are usually not worthwhile in-car after considering the extra setup time and maintenance it takes. (link)

i will def. keep it in mind(why can't there be a easy solution to every problem!?)-- running extra tubing in/under the car might allow me to have enough fluid to get the run times i need without trying to actually cool everything--er, i mean, have enough fluid in the lines so by the time the fluid is hot, the car will be parked/finished backing up for a while so the lines can cool down again--so i wouldn't need a radiator to try to get rid of the heat.. the bonus is that everything would be sealed, so water/dirt isn't a problem, the minus is the extra maintenance, and driving the pump to circulate the fluid.. but the bragging rights of having water-cooled reverse lights...

whereas a heatsink setup would require me to displace any heat the led's create almost as fast as the led's can create it..


suffice it to say that i am still flip-flopping on the method, but would def prefer a heatsink design for the low-er maintenance...

[Raptor05121]

This is what I'm looking at:

http://www.luxeonstar.com/High-Effic...-LED-s/195.htm

They have 19mm square heat sinks and I'm going to use 25mm for more space. Just mount the LED to it, and cut a 19mm (or 25mm) square, slide it in, and epoxy in place.

[Mike S]

I didn't know you were planning to drive these at 3A each. That's a lot of light!

Could you cut out the back of the housing to accept a large computer heat sink? I haven't been able to drive my XM-L's harder than 1A, so I can't say for sure if it would be adequate. It might be worth a shot.

[upashi]

Are you trying to drive 2 emitters next each other, or some spacing between?

3A drive each will be about 20W of power you need to deal with if you use emitters next each other.

XM-L thermal resistance is 2.5C/W to PN junction to thermal patch contact.
Assuming avarage alluminium star board thermal resistance about 1.5C/W
Thermal paste about 0.1C/W

XM-L max junction temp is 150C, add 20% mergine to aim 120C target temp.
Ambient temp set as 25C

(120-25)/20=4.75C/W

4.75C/W-(2.5+1.5+0.1)C/W=0.65C/W

0.65C/W passive heatsink gonna be massive I assume.



When 2 emitters are seperate, can consider two different heat source,
each one will be about

(120C-25C)/10W=9.5C/W
9.5C/W-(2.5+1.5+0.1)C/W=5.4C/W

depend on heatsink design, location, angle you set, 5.4C/W can be rather compact.
when space is available, heatsink fin want to be not too dense so it won't get extra air flow resistance.
layout to enchance vertical air flow. Vertical longer each fin is better thermal capacity.

when air flow expected, heatsink capacity can be smaller.


So, now, if 2 emitter need to be sperate, good looking design in clear housing can be difficult I think.


Will one emitter won't be enough? or half of driving rate at 1500mA?

If driven at 1500mA, typ Vf is 3.1V, 4.65W each emitter and 9.3W total
This case, your target heatsink capacity is about 6.11C/W
and can expect max 520lm x 2 =1040lm

or one emitter at full drive
target heatsink capacityof 5.4C/W for about 10W power
expect 910 lm max

About 5C/W(ambient at 25C caluclation) passive heatsinking seems to be doable.

[corvettecrazy]

About 5C/W(ambient at 25C caluclation) passive heatsinking seems to be doable.

And since he probably doesn't live in an area that is 25C and less year round, what happens when it is 45C outside?


Personally, I think your run time estimates are kind of crazy. If you take 5 minutes back up time. Take out a minute for the garage door. Then you go 220 ft in 4 minutes. Which is .91 ft/second. (1mph =1.46 ft/s) Can you actually get a car to go that slow?

Regardless, I would get the largest thermal mass you can and run with it. If you back up your alley twice a day, everyday, for 4 years and it takes you 20 minutes each time then the total run time of the LED's is: 58,400 minutes= 978 hours. The XM-L has gone though significant LM-80 testing which states that it can maintain over 90% lumen maintenance at 1500mA over the period of 6000 hours. Your 1000 hours of run time at 3000mA is peanuts.

Get a chunk of aluminum with fins on it and move on with life. I'd be more worried about the plastic housing being able to handle the heat than the LED.

[soundman98]

This is what I'm looking at:

http://www.luxeonstar.com/High-Effic...-LED-s/195.htm

They have 19mm square heat sinks and I'm going to use 25mm for more space. Just mount the LED to it, and cut a 19mm (or 25mm) square, slide it in, and epoxy in place.

thanks for the link! that was exactly what i was looking for earlier. bookmarked.

I didn't know you were planning to drive these at 3A each. That's a lot of light!

Could you cut out the back of the housing to accept a large computer heat sink? I haven't been able to drive my XM-L's harder than 1A, so I can't say for sure if it would be adequate. It might be worth a shot.

yes, i was planning on cutting the rear of the housing to accommodate the heatsink.

Are you trying to drive 2 emitters next each other, or some spacing between?

they are about 4-5" apart, but in the tests the led's were thermally connected:

this is my earlier picture, before i screwed the led's down to the aluminum:

3A drive each will be about 20W of power you need to deal with if you use emitters next each other.

XM-L thermal resistance is 2.5C/W to PN junction to thermal patch contact.
Assuming avarage alluminium star board thermal resistance about 1.5C/W
Thermal paste about 0.1C/W

XM-L max junction temp is 150C, add 20% mergine to aim 120C target temp.
Ambient temp set as 25C

(120-25)/20=4.75C/W

4.75C/W-(2.5+1.5+0.1)C/W=0.65C/W

0.65C/W passive heatsink gonna be massive I assume.



When 2 emitters are seperate, can consider two different heat source,
each one will be about

(120C-25C)/10W=9.5C/W
9.5C/W-(2.5+1.5+0.1)C/W=5.4C/W

depend on heatsink design, location, angle you set, 5.4C/W can be rather compact.
when space is available, heatsink fin want to be not too dense so it won't get extra air flow resistance.
layout to enchance vertical air flow. Vertical longer each fin is better thermal capacity.

when air flow expected, heatsink capacity can be smaller.


So, now, if 2 emitter need to be sperate, good looking design in clear housing can be difficult I think.


Will one emitter won't be enough? or half of driving rate at 1500mA?

If driven at 1500mA, typ Vf is 3.1V, 4.65W each emitter and 9.3W total
This case, your target heatsink capacity is about 6.11C/W
and can expect max 520lm x 2 =1040lm

or one emitter at full drive
target heatsink capacityof 5.4C/W for about 10W power
expect 910 lm max

About 5C/W(ambient at 25C caluclation) passive heatsinking seems to be doable.

thanks for doing the calculations! that pretty much solves it.

i will use a single heatsink per led. for some reason, i was thinking it was better to thermally combine them....

i just remembered i have some old cpu heatsinks in my basement-- i am going to attach led to one, and see how well it works-- it is a little tougher when i don't know the exact c/w of the ones i have..

after thinking about it today, i am going to focus on using a heatsink that should hold up while being passively-cooled, but will add some fans to to aid in keeping the temp down--at least until the fans die from the tough area.

thanks for the advice!

[Raptor05121]

bumping for a good read

[azdave]

Are you sure you aren't building LED headlights?

[Bitter]

Just shove a small fan inside the aluminum tube and install some vented air intake tubes from the trunk, reverse lights come on, fans come on, lights go off, fans go off. waste air can be vented outside through some small holes or something, maybe some headlight style macaroni vent tubes.

[soundman98]

Are you sure you aren't building LED headlights?

the other option was retroing a set of blazers with the cutoff removed with a set of 4300k hids into those same housings.. this is more technical, and will have a little wider light dispersion without the projection lens...

[soundman98]

time for a little update.

i snagged some old computer heatsinks from work:



i first lined them up on the housing to see how easy they are to fit, and the fit was pretty close..


so i then drilled and tapped one, mounted the led to it, removed the fan, and tested for passive cooling:


results:
ambient:72
start: 74
5min: 81
10min:87
15min:91
20min:96
25min:98
32min:skewed--97--at about 30 mins, i was going crazy not doing anything, so i tried using my dremel to clean up some other stuff--that happened to be right next to the heatsink i was testing with. the airflow coming off the dremel was enough to drop the temps..

so with those tests, i am confident that these heatsinks are going to be enough, esp. when i use the included fan..

after drilling and tapping the other ones (and figuring out that you really do need thread cutting oil!!) i started lining them up on the housing to see what i need to cut:



you really can't see it in the pics, but with the fans installed, there is some overlap on both sides, so i think i will need to cut some of the fins down to accomodate them.

and then i remembered that i had to pay attention to the opening:


so because i can't access anything behind them with the bumper installed, that means that i really can't install them by removing the bumper, so they must fit through the opening. the height is going to be fine, but the overall width looks like the inside heatsink might 'just' catch the inside mounting tab...

i'm starting to overthink this, so that's it for today..

[Mike S]

Nice, I like the fan idea even if it's not necessary. It doesn't hurt to include it. I'm looking forward to the finished design.

[Raptor05121]

This is awesome.