Ultimate LED Thrower Challenge

[RecycledElectron]

http://www.dealextreme.com/details.dx/sku.5971
With 3 Q5's/R2's might be interesting.

I have a MR16 12 volt bulb with 3 3watt emitters in it. Actually, two of them. I'm going to put them in housings for motorcycle running lights. They're a drop-in in the Kurkalyn motorcycle teardrop housings. It's easily as bright as a 55 watt xenon headlight, maybe brighter. I don't know what it would do in a big reflector with an aspherical lens in front of it. They can be had in both spot and flood configurations, mine are 45 degree angle medium floods. I believe that they can be had in a 15 degree spot configuration. I got them almost a year ago on Ebay from a Canadian who imported them from China. I believe that they are much more easily found now.

At 13.8 volts they draw about 400 ma. That's about 6 watts. I don't know if that's going to be brighter than the 3 led light from DX or not. On the site, they claim the 4 CR123's are driving the led's at 700ma/yielding 400 lumen. I remember reading of someone modding it with Q5's and claiming 600 lumen.

You might have to mount a little water pump and tubing to cool the flashlight's heatsink. :devil:

 

[J_C]

With a good heatsink a XR-E Q5 delivers "up to" 210 lumens.
http://www.cree.com/press/press_detail.asp?i=1169819309344

Suppose you had a really, really good heatsink and put 2A through each and managed to get 350 lumens (which is doubtful) per each LED. We already have the rating of the halogen spotlight, 1650/350= 4.7, rounded up to 5 LED even with extreme overcurrent attempts and using a very good heatsink.

 

[saabluster]

With a good heatsink a XR-E Q5 delivers "up to" 210 lumens.
http://www.cree.com/press/press_detail.asp?i=1169819309344

Suppose you had a really, really good heatsink and put 2A through each and managed to get 350 lumens (which is doubtful) per each LED. We already have the rating of the halogen spotlight, 1650/350= 4.7, rounded up to 5 LED even with extreme overcurrent attempts and using a very good heatsink.

At its highest spec(1000mA) the Q5 puts out 214-228 lumens. What you linked to is old news. The Q5 came after that release. 350 lumens is indeed possible. Its already been done.

 

[J_C]

At its highest spec(1000mA) the Q5 puts out 214-228 lumens. What you linked to is old news. The Q5 came after that release. 350 lumens is indeed possible. Its already been done.

Barely, and it's not practical.

And no, the Q5 did not come after that release, that is referring specifically to it @ 1A. That is the whole point of the release to announce the gain with Q5.

My main point is, there is no possible way a single LED die like the Q5 can reach even 50% of the light output of the linked spotlight. It will necessarily take more than two of them no matter how heroic the measures taken are.

I LOVE these newer high efficiency LEDs, but claiming they are something they are not does nothing useful.

If we only contrasted a spot measurement of some extremely focused beam versus a less focused halogen beam, it is not a fair comparison. Given unlimited current halogens easily beat LEDs but that's not the point, rather that halogens burn out more quickly and are far less efficient.

 

[BBL]

Forget that idea... outthrowing a spotlight with its large reflector and high output is not possible.

 

[Jay T]

A LED vs a 1300 lumen spotlight? Just to give some scale to the battle here are some beamshots.

Here is a Tiablo A8 Q5, a decent thrower.

I don't have a 1300 lumen spotlight so here is a 1000 or so lumen Mag85. Now the spotlight's big reflector will take all that spill and add it to the already powerful center section. Just double the center brightness to get an idea of what a spotlight will do.

 

[LukeA]

I've got a project in the works that might be applicable. It's a Vector/B&D 1MCP spotlight (the model down from your friend's) with 7 DX cree dropins running in it in parallel. I'm probably going to replace the stock lead-acid battery with 3 Ni-MH D-cells. Should end up about 1400 emitter lumens (2.5x stock output). I don't know if it'll outthrow your friend's Vector, because the Vector has a much larger reflector (I dont know if this is true or not because my Pelican 7060 throws on par with the stock spotlight) It should be a nice wall of light anyhow.

Parts cost ~ $90.

PS: My aspheric mag, when it was working, could outthrow anything else, and by a wide margin.

 

[saabluster]

Barely, and it's not practical.

If I remember right it was the Q4 that someone got to 350 lumens. We now have R2. The R2 will easily do 350.

And no, the Q5 did not come after that release, that is referring specifically to it @ 1A. That is the whole point of the release to announce the gain with Q5.

Guess again. The proper press release is dated JUNE 27, 2007 . Your link is dated JANUARY 26, 2007. The Q5 did not exist then(not that we knew of anyway).

My main point is, there is no possible way a single LED die like the Q5 can reach even 50% of the light output of the linked spotlight.

If we only contrasted a spot measurement of some extremely focused beam versus a less focused halogen beam, it is not a fair comparison. Given unlimited current halogens easily beat LEDs but that's not the point, rather that halogens burn out more quickly and are far less efficient.

Keep in mind the challenge is max throw not max lumens. It doesn't matter that the led cannot reach the same lumens. Its whether or not he can focus what he does have farther. This is possible.

 

[J_C]

If I remember right it was the Q4 that someone got to 350 lumens. We now have R2. The R2 will easily do 350.

One will, now think about putting enough in a small bodied flashlight to reach the output of a 55W+ spotlight. The implementation scenario is always significant, in practice you cannot use a single XR-E die at 350 lumens output without unreasonable measures.

Then again, you make no reference to who this "someone" is or their measurement technique which could easily be flawed. I stand by the industry standard manufacturer ratings including peak current. 1A is already a lot, it would be silly to build a spotlight that fries out in a few minutes. If all one wants is the see smoke or a flash of light there are cheaper ways to do it.

Guess again. The proper press release is dated JUNE 27, 2007 . Your link is dated JANUARY 26, 2007. The Q5 did not exist then(not that we knew of anyway).

They announced it when they had achieved it, then release again when their production is soon capable of meeting demands. This is fairly standard in the semiconductor industry. The spec sheets also clarified this.

So, guess again.

Keep in mind the challenge is max throw not max lumens. It doesn't matter that the led cannot reach the same lumens. Its whether or not he can focus what he does have farther. This is possible.

Wrong. He's building something competitive to that spotlight. If all he wanted was an infinitely tight beam at the furthest distance he could just buy a $5 laser pen and be done already.

Pay attention to what was written. The competition is a 1650 lumen spotlight, not some other random thing.

 

[saabluster]

The implementation scenario is always significant,

:thinking:

in practice you cannot use a single XR-E die at 350 lumens output without unreasonable measures.

I'm doing it right now. And I don't have to do anything unreasonable.

Then again, you make no reference to who this "someone" is or their measurement technique which could easily be flawed.

Try this thread. I don't know of anyone who has a problem with the way he tests.

They announced it when they had achieved it, then release again when their production is soon capable of meeting demands. This is fairly standard in the semiconductor industry. The spec sheets also clarified this.

You are WRONG. Absolutely wrong. It is mind numbing how you just cannot get it when you have crees own press releases to go by. Can someone else PLEASE help him see the light? I'm tired of this.

Wrong. He's building something competitive to that spotlight.

Pay attention to what was written. The competition is a 1650 lumen spotlight, not some other random thing.

Here is what was written. " The Challenge:
Buy, Build, or mod an LED flashlight for under $100 that can out throw his light." It says out throw not out flood. Pay attention to what was written.

 

[J_C]

:thinking:


I'm doing it right now. And I don't have to do anything unreasonable.

You conveniently forgot to mention any details of this miracle. Care to ask Cree if these are capable of that output without "anything unreasonable"? Want to bet money on what they'll say?

Try this thread. I don't know of anyone who has a problem with the way he tests.

Well now you do.

You can't just run > 1A through one without doing exactly what the spec sheets indicate, which is derating the output per the die temp. If there were a self-done measurement that contradicts it, I can assure you that those who made the LED (Cree) do have the correct measurements and do know what they're talking about, not some 3rd party. Believe me, if they could easily (or even not so easily) claim no extraordinary measures to get their die up to 350 lumens, that is exactly what they would do (but note, they don't).

You are WRONG. Absolutely wrong. It is mind numbing how you just cannot get it when you have crees own press releases to go by. Can someone else PLEASE help him see the light? I'm tired of this.

Fine, be tired. Next time before you're tired, take some temp readings and account for the star-die thermal junction so you know how much to derate the light output. You can't just keep pumping more current through the die, there is a finite limit unless as I stated earlier, extraordinary measures are taken to cool it. For example 2A will not yield 350 lumens. Did you note the curve on the currentutput curve instead of just trying to continue it in a gentle curve? The curve is an equation and it steepens at overcurrent levels (over about 750mA actually, those running at even 1A are beyond the suggested current).

Here is what was written. " The Challenge:
Buy, Build, or mod an LED flashlight for under $100 that can out throw his light." It says out throw not out flood. Pay attention to what was written.

Yes, out throw doesn't mean some tiny spot, it means it's comparing to THAT specific floodlight. Pay attention to what was written (and so sorry you are sore when wrong).

 

[saabluster]

JC just a little reminder. If you look under your screen name you will notice it says you are "Unenlightened". Please keep that in mind.

 

[djblank87]

Wow it is time for a group :grouphug:.......everyone feel better .


Ok, now I will not get into details of light and temp readings and cree this or that, I will stay with the first thought that came to my mind when the OP stated he wanted to out throw a B&D spotlight.

Throw and flood are two different things. I'm postive my Tiablo A9 Q5 could smoke a B&D Spotlight but the spotlight will flood the area better. But the Tiablo is over $100.00 so that wont work.

You could make a modded Maglite that will out THROW not out FLOOD a spotlight.

So when stating that one wants to out throw another light, that is not to hard to do for under $100.00.

To out flood a light is a whole nother story. The OP stated out throw which in that case does mean a TINY spot reaching out farther that the flood of a light. That is throw!

Once again Flood and Throw are two different things.

 

[J_C]

JC just a little reminder. If you look under your screen name you will notice it says you are "Unenlightened". Please keep that in mind.

LOL, nice try but no cigar.

 

[AzN1337c0d3r]

Fine, be tired. Next time before you're tired, take some temp readings and account for the star-die thermal junction so you know how much to derate the light output. You can't just keep pumping more current through the die, there is a finite limit unless as I stated earlier, extraordinary measures are taken to cool it. For example 2A will not yield 350 lumens. Did you note the curve on the currentutput curve instead of just trying to continue it in a gentle curve? The curve is an equation and it steepens at overcurrent levels (over about 750mA actually, those running at even 1A are beyond the suggested current).

If you actually read the technical evaluation thread, evan DROVE the LED at ~2A and obtained a ~350 lumen output EXTERNALLY. He didn't try to extrapolate from a curve, he actually took a Q4, and ran 2A through it, and measured the output. The measured output is what your eyes would see. It makes no sense to derate it.

The reason why Cree don't bother to spec their LEDs to 2A is that of practicality.

Here's several reasons:

1) Driven at 2A, the LED won't last 50,000 hours (to 80% brightness)
2) The runtime on a set of batteries with such an emitter would be dismal.
3) The cost of designing a heatsink to cool such an LED would cost a lot more than what a typical consumer would pay.
4) To reach 350 lumens, it is cheaper to put two or more lower-powered LEDs in there.

 

[J_C]

If you actually read the technical evaluation thread, evan DROVE the LED at ~2A and obtained a ~350 lumen output EXTERNALLY. He didn't try to extrapolate from a curve, he actually took a Q4, and ran 2A through it, and measured the output. The measured output is what your eyes would see. It makes no sense to derate it.

The reason why Cree don't bother to spec their LEDs to 2A is that of practicality.

Here's several reasons:

1) Driven at 2A, the LED won't last 50,000 hours (to 80% brightness)
2) The runtime on a set of batteries with such an emitter would be dismal.
3) The cost of designing a heatsink to cool such an LED would cost a lot more than what a typical consumer would pay.
4) To reach 350 lumens, it is cheaper to put two or more lower-powered LEDs in there.

If we assume the measured output was accurate, the question is then how long it runs like this. The problem is what I've already stated, and that you have as well, the heatsinking. It's not just getting it to 350 lumens it's keeping it running. It isn't likely such a heatsink can even be put into a typical cylindrical flashlight body.

Since there's no practical way to heatsink it, it won't just decrease from 50K hours at that output like it would at more managable heat levels, instead it will soon damage the chip (package) which is not designed for the extreme temps. So it goes for many power electronics, they can be ran pretty hot if you accept lifespan reduction but there is a threshold past which the package itself is destroyed.

 

[AzN1337c0d3r]

I'm pretty sure evan's setup is accurate, since he has calibrated with other CPF members, and they generally agree to within a couple of percent I believe.

As for catastrophic failure due to overcurrent conditions, this guy ran his XR-E at 2.5A for over 30 minutes without issue: https://www.candlepowerforums.com/posts/2089898&postcount=10

In a throw contest, i'd doubt you'd leave your light on that long.

 

[Gunner12]

My thought on this project:

Aspheric lens or multi lens setup(like this) with a Cree Q5 driven at around 2amps on a good heatsink.

 

[saabluster]

My thought on this project:

Aspheric lens or multi lens setup(like this) with a Cree Q5 driven at around 2amps on a good heatsink.

I agree completely. With the right aspheric this is very doable. I just started work on just such a flashlight. I took the lens from a subaru svx's headlight. The way it can collimate is outstanding. Far better than any of the others I've tried. And that includes the one from surpluss shed that everyone raves about. Its 75mm so it will not fit in a mag though. No problem. I found one of those cheap shower head flashlights that this will fit into nicely. I have no doubt this will out throw that BD spotlight.

 

[J_C]

As for catastrophic failure due to overcurrent conditions, this guy ran his XR-E at 2.5A for over 30 minutes without issue: https://www.candlepowerforums.com/posts/2089898&postcount=10

In a throw contest, i'd doubt you'd leave your light on that long.

I'm suspecting he'll want to keep and use the flashlight after he wins the contest, so it could be important. If you just want to impress someone with $100 you can take them out to dinner instead.

As for the guy who ran one for over 30 minutes, there are a few problems that need addressed.

1) Lack of specifics. Heatsink, actual current measurements, etc. It would not be surprising at all if 4 AA cells he used, cannot push 2.5A or even 2A through an emitter for over 30 minutes. A typical NiMH cell fully charged and draining at this rate would have pretty bad voltage depression from this high drain rate.

The capacity of an AA cell and high rate vs capacity just don't make AA cells suitable for a controlled current test. If he had a current-controlled supply we could be more sure it was 2.5A. Instead with the direct drive he used, it is most likely is it started out at 2.5A and when down in current quite a lot over time. Actually it is certain that either the cells started out putting far more than 2.5A through the LED, or that it was nowhere near 2.5A through the LED shortly after the measurement was taken.

2) Lack of runtime. 30 minutes isn't very long, and does not account for wear on the PCB due to thermal cycling. Repetitive heating and cooling (just testing this design before there is any contest with a halogen spotlight) can be a problem separate from whether it survives one run for 30 minutes (which as mentioned in #1, we can be sure didn't happen at continuous 2.5A).

It's just not an 8W LED. Other manufacturers and parts are rated for more current when they are designed capable of withstanding the heat, and with these other higher-wattage rated LED they don't think to themselves that they shouldn't rate it that high because of an end-use that would require a large heatsink or bigger batteries. They could just include an extension of a graph for lifespan vs temp or current if the design were sound very far beyond 1A.