I'm BLIND! (LS overkill)

[PsycoBob[Q2]]

I\'m BLIND! (LS overkill)

Well, I finally figured out why I keep seeing spots after looking in my latest creation.

A green 1w LS/o, mounted to the underside of a Golden Orb CPU heatsink. Running 4.05v @ 2350mA. The neighbors are probably wondering about those blinding flashes of green light.....

Creepy part is, I think it can go even farther- the aluminum PCB is fairly cool (less than 100F), and that monster heatsink can easily dissipate the heat, even with the fan off.

[TimAllen]More Power![/TimAllen]

 

[hotfoot]

Re: I\'m BLIND! (LS overkill)

Hmmm... 2350ma x 4.05V - CONGRATULATIONS! You now have the equivalent of a 9.5W Luxeon! (for reference, a 1W luxeon driven to spec is 3.6V x 350ma = 1.26W)

EDIT: Oh, and a 5W luxeon driven to spec is 7.2V x 700ma = 5.04W

 

[PsycoBob[Q2]]

Re: I\'m BLIND! (LS overkill)

Without the fan on, the sink gets nice and hot- after 15 minutes. Turn it on, and it's room-temp in 10 seconds.

 

[Willmore]

Re: I\'m BLIND! (LS overkill)

Your limit for this project will not be PCB to heatsink, but DIE to packaging. At this thermal rate, and with the junction resistance, you need to make sure that the die stays cool enough.

Have you stared at the datasheet and calculated any of this or are you going on karma?

 

[Slick]

Re: I\'m BLIND! (LS overkill)

Originally posted by PsycoBob[Q2]:
A green 1w LS/o, mounted to the underside of a Golden Orb CPU heatsink. Running 4.05v @ 2350mA...........

<font size="2" face="Verdana, Arial">That sounds awesome. I have always thought about using one of these (radial) heatsinks with a Luxeon, since they look much better than your "average" heatsink. Would like to see a photos if you could get one.

 

[Willmore]

Re: I\'m BLIND! (LS overkill)

Okay, looked up the data sheet. Max junction temp if 120degrees C. Thermal resistance between die and *case* is 15 degrees/watt. So, at 9.5W, you need to kee the *case* at 120-142.5 or -22.5 C. So, keep in mind that your case to heatsink thermal resistance will eat up some more. You're quickly getting to the point where you just need to set the heatsink on a block of dry ice.

If not, you're running the die way beyond spec and you can expect a rapid output loss or a complete failure before long.

 

[shipinretech]

Re: I\'m BLIND! (LS overkill)

Pretty soon the bar is going to be raised to the point where liquid nitrogen cooling will be required to impress people. I'm starting to worry how Craig is going to accept the 6D Maglites we're looking at adapting. How you are going to explain that your flashlight requires a handtruck to move the generator and cooling unit is your own business! I don't even want to think about the shipping costs!

 

[PsycoBob[Q2]]

Re: I\'m BLIND! (LS overkill)

Well, unfortunately, it died.... not from long-term overheat, either. I dropped it. Broke the optic-retainer right off, and it yanked the LS off with it.... mostly.

Such is the downside of having an enourmous heatsink- it has mass. I'm trying to get my webcam back from my brother, so I can sell a LS-mag mod I made, and I'll try to remember to get a pic of the remains.... maybe even put another LS on there, and fire it up.

Willmore, thanks for the data. I'd labled this greenie as expendable anyway... selling and shipping would too annoying for the $. I had been running this as the world's brightest hard-drive light, and occasionally plugging it in as other indicators, like network, power, sound-signal triggered, etc. DC solid-state relays kick ***.

 

[PsycoBob[Q2]]

Re: I\'m BLIND! (LS overkill)

Minor update- I have 2 cyan 5w LS's on order- should be here tomorrow, or friday. The PCB from the 1w green will have one of my bare-emmitters AS-epoxied to it, and it'll likely be stuck on that Monster heatsink.... almost. I'll be mounting the 5w LS/o on the INSIDE, protected by the aluminum fins, so it won't get killed like the greenie. Then it'll either be stuck on a flashlight, or plugged into my computer as a HDD LED.

I have the wielding mask all ready for my 5w experiments.

 

[BentHeadTX]

Re: I\'m BLIND! (LS overkill)

Psyco,
A Golden Orb heatsink on a LS overdriven to 9.5 watts? You da man! Watts next? Hmmmmm, a peltier on a Alpha 8045 copper/aluminum heatsink with a 75 CFM fan? Woohoo! Time to smoke some Luxeons

 

[PsycoBob[Q2]]

Re: I\'m BLIND! (LS overkill)

Got a 5w Cyan (Arctic-silver epoxied to a plate from a 1w LS/o... with the PCB under the emmitter removed) mounted to that monster Gorb again.

Only running it @ 620mA, due to a lack of resistors to make a truly oddball resistor pack (currently wired to a 12v source, with a 7.5ohm custom resistor pack). The heatsink doesn't warm up nearly as much as it did with the 1w greenie at the same wattage.

I think my attachment method was better than Lumiled's- the original 1w emmitter had several layers between it and the aluminum PCB. "Stock-epoxy (NOT designed for thermal conduction), copper-foil X on pcb, pcb (electrical) insulating layer, THEN the aluminum PCB core. My emmitter is epoxied direct to the core, with a better epoxy to boot.

If I'm lucky, the 5w die is running cooler than the 1w was at the same wattage-input. I'm not feeling brave enough to overdrive the hell out of this one, due to the cost.

Can I get donations to buy me some LN2? I'd love to try running a 5w @ 55w, in a pool of liquid nitrogen....

[Edit] I managed to mis-wire things after testing the "run a 5w off a 9v battery" idea, and It's currently running @ 1.25A, and sitting enxt to a fan. Don't have time for testing the max-temp with no fan.

 

[brightnorm]

Re: I\'m BLIND! (LS overkill)

Originally posted by shipinretech:
...Pretty soon the bar is going to be raised to the point where liquid nitrogen cooling will be required to impress people.

<font size="2" face="Verdana, Arial">Or the increasingly larger bezels will start overwhelming the flashlight bodies. The KL-3 is the beginning of this "tail wagging the dog" evolution.

Brightnorm

 

[EMPOWERTORCH]

Re: I\'m BLIND! (LS overkill)

What next... a nuclear powered supercooled superconducting torch?
You'll be saying "Here's my torch...it's battery powered..."and you then fail to point out the articulated lorry carrying the batteries!

Onn a more serious point though; I think that the super-high-powered torches of the future may have some kind od assisted cooling system.
Watercooling has done a serious amount to increase the power of motorbike engines without significantly increasing thier weight...the supertorches of the future need not be unweildy or heavy, and hopefully the efficiency of the LEd's will continue to increase so that less heat is produced in the first place.
The way technology is developing we will probably see the first handheld solid state floodlight!

 

[MicroE]

Re: I\'m BLIND! (LS overkill)

Wow. A torch that can fry two eggs simultaneously: one from the beam and the second from the heatsink.
Obviously the efficiency of the LED must be improved before we can expect a realistic (10+ Watt)handheld floodlight.
The power that is lost to heat (IxV) is too high. The forward voltage of these devices will come down a bit as the manufacturers figure out better techniques for creating an ohmic contact but there is a lower limit for Vf defined by the band-gap of the material (anybody know what that is?). The only option left is to reduce the current requirement (eliminating currents that fail to create useful photons) and that means improving the wafer-processing techniques.
It will take them time to create the LED that PsychoBob really wants: a low Vf, low current multi-watt die. In the mean time we will have to buy burn cream for Bob's fingers.---Marc

 

[MarcV]

Re: I\'m BLIND! (LS overkill)

I often think of the same thing you do, that we pay extra for cleverly designed heat sinks that efficiently dissipate our batteries.

It doesn't seem fair.

 

[Rothrandir]

Re: I\'m BLIND! (LS overkill)

Well MicroE, sounds like you've got quite a job ahead of you...Now get to work on those new sped's (super photon emmiting devices)!!!

 

[MicroE]

Re: I\'m BLIND! (LS overkill)

Right after I graduated from college (when dinosaurs roamed the earth) I went to work as a process engineer for Laser Diode, Inc. (wanna guess what we made?)
It took my team 2 years to figure out a good method for making electrical contact to InGaAsP/In LED's.
Each semiconductor material is funky in its own way and requires process optimization that can take years.---Marc

 

[Rothrandir]

Re: I\'m BLIND! (LS overkill)

well, then u'd better hurry, it's not a good idea to keep cpf'ers waiting...

seriously tho, that InGaAsP thing sounds interesting, what the heacl is it?

 

[PsycoBob[Q2]]

Re: I\'m BLIND! (LS overkill)

I want to get a Star-Trek battery- a battery that measures 3x5x1.5", is rechargable, and holds 4.2x10^8th megajoules.

Phaser powercell + Über-LED. All I'd need then is a job at a nuclear-power plant...

If nicha tried to get ultra-high Effeciency LED's while Lumileds continued for the ultra-high handwarmer catagory, maybe they could have a little 'Patent liscencing' to thrill their shareholders. Lumileds making LED's that are 98% efficient... and draw 25w.

 

[flash....]

Re: I\'m BLIND! (LS overkill)

Something I have noticed about pushing a LS beyond it's rated current...
After 650ma the star does not seem to get much brighter.
150 - 350ma big change in brightness. Obviously.
350 - 650ma another very noticeable change.
650 - 2amps .. not much increase in output. IMHO.

Can anyone confirm this with an actual light meter?
Based on this, I am assuming anything above 650 is getting real inefficient. 650ma is what I plan to use to over drive my LS's unless someone with the proper equipment can tell me otherwise...