Compact 4AA host for 5W Luxeon

Compact 4AA host for 5W Luxeon

Ya know what I think the modder world needs is a compact 5W luxeon light powered by four AA NiMH cells.

I'm imagining something like a metal replacement head for a UKE 4AA or PT 40. The head would have a BB750 and 5W LS mounted on a nice thick copper slug, possibly with a radial finned heat sink, or else a pin-and-fin Pentium-type heat sink turned to a circular shape with the fins cut out of the middle to mount the Luxeon Star. There could be an NX05 nestled in the heatsink fins and the fins would protect the NX05 and emitter from getting poked. The copper slug and radiator fins would keep the LED reasonably cool, and the plastic flashlight body would be nice and lightweight, and far enough away from the action to not have heat problems. Finally, those forward-pointing (or radial) fins would result in a very radical looking light.

The resulting light would weigh less than most other 5W light that's come up here so far, have reasonable runtime (probably 1.5 hours with good NiMH's), and have very low operating costs (unlike the 123-powered lights), while using ubiquitous AA NiMH cells instead of exotic ones like 2/3 A's or whatever. Main prob might be the heat sink fins snagging on things. But I guess it could be a cylinder with turned channels like a Surefire LED head.

Does this sound feasible? Anyone with machining capabilities want to try?

Re: Compact 4AA host for 5W Luxeon

I like it! It may not be able to run continuous at 5W without overheating the LED but it would not be hard to add some closed loop temperature control to reduce power once the heatsink gets too warm for the LED. I think that I would stick with aluminum since copper adds a lot of weight for only a modest increase in thermal conductivity.

Re: Compact 4AA host for 5W Luxeon

The idea is to make the heat sink big enough for continuous operation. Would that be really excessive? I also thought copper turns out to be worth the extra weight, per tvodrd's calculations.

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
paulr said:
The idea is to make the heat sink big enough for continuous operation. Would that be really excessive? I also thought copper turns out to be worth the extra weight, per tvodrd's calculations.

[/ QUOTE ]

I would be inclined to use increased thickness of Aluminum rather than copper to offset the reduced conductivity of the aluminum. Equal thermal conduction performance can be had with less weight of aluminum vs copper. The principal design challenge is the heat transfer from the sink to the air. Copper has no better performance than Aluminum at this interface. I think that 3 to 4W continuous would be attainable with a reasonable size head. I tend to be conservative on the question of running at high die temperatures. Others here will cheerfully tell you to go for the full 5W and more.

Re: Compact 4AA host for 5W Luxeon

Maybe we could really go crazy and mount the LS on the back side of one of those CPU heat sinks with an integrated fan. I think some of those fans might run on 5 volts.

Re: Compact 4AA host for 5W Luxeon

Actually here's another nutty idea that may make more sense:

What do you think of a PT40 or UK 4AA to Surefire head adapter? That would allow putting on a KL4 or KL5 or whatever. Do you think the body is an important source of heat sinking in a light like the E2e?

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
paulr said:
Actually here's another nutty idea that may make more sense:

What do you think of a PT40 or UK 4AA to Surefire head adapter? That would allow putting on a KL4 or KL5 or whatever. Do you think the body is an important source of heat sinking in a light like the E2e?

[/ QUOTE ]

I've never seen/disassembled an E2e/KL4/KL5/whatever so I cannot comment. It would depend on the nature of the thermal paths from the LED to the head and body.

Re: Compact 4AA host for 5W Luxeon

Back to your original proposal, I suspect that those with machining knowledge will tell us that matching the threads on the PT or Uke will be a PITA.

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
paulr said:
Maybe we could really go crazy and mount the LS on the back side of one of those CPU heat sinks with an integrated fan. I think some of those fans might run on 5 volts.

[/ QUOTE ]
Taking this idea one step further, has anyone seen those "fanless" CPU heatsinks utilizing the Peltier principle for cooling. The only downside that I see is that some of the battery power would go to the Peltier unit itself thus cutting into the total runtime. How much? I haven't yet worked the numbers, but I suspect that it could be signifigant. The Peltier principle utilizes the principles of thermo-electric cooling. Pretty neat invention. Here is an example of a Peltier CPU Cooling

Ed.

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
paulr said:
The head would have a BB750 and 5W LS mounted on a nice thick copper slug, possibly with a radial finned heat sink, or else a pin-and-fin Pentium-type heat sink turned to a circular shape with the fins cut out of the middle to mount the Luxeon Star.

[/ QUOTE ]


Radial finned heat sinks get my vote! /ubbthreads/images/graemlins/grin.gif Overall, for a 4AA light I've been picturing something like a McLux with the width of a D-cell Mag.

Only thing is with 4 NiMH cells, you only get 4.8V, 5 cells would give you a full 6V which would be a better choice, IMHO. But the design wouldn't be as neat. For a 4 cell design, it would have to stick to regular alkalines or lithiums.

As for Peltier cooling, sadly I think the Peltier element might suck more power than the Luxeon itself /ubbthreads/images/graemlins/frown.gif

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
e=mc² said:
Taking this idea one step further, has anyone seen those "fanless" CPU heatsinks utilizing the Peltier principle for cooling. The only downside that I see is that some of the battery power would go to the Peltier unit itself thus cutting into the total runtime. How much? I haven't yet worked the numbers, but I suspect that it could be signifigant.
Ed.

[/ QUOTE ]

Hi Ed,

The cooler would take about 20 watts to cool a 5W Luxeon and would need a heatsink 5 times larger than the heatsink to cool the Luxeon alone.

Ron Schroeder
Brookhaven National Lab

Re: Compact 4AA host for 5W Luxeon

Peltier CPU headsinks I've seen have all had fans, I thought.

I still like this KL4 idea, if the KL4 doesn't depend on the flashlight body to transfer heat.

If threading a metal adapter to the UKE plastic threads is too hard, another approach might be to just epoxy the adapter over the UKE threads, so it would be a permanent mod. It would still have outer threads to attach the LED head, so you could swap heads as you wanted.

Re: Compact 4AA host for 5W Luxeon

If you want it to be compact and also run off of 4AA NiMH cells, it's hard to get it much smaller than my 5W "mod" I posted here a few weeks ago. Although the shape may be awkward to some, it does fit easily in a coat pocket and puts out a lot of light despite being underdriven. On the higher settings it can become quite a hand warmer after 20 minutes. Adding fins will help keep it cooler but I wanted to keep it simple and avoid sharp edges sticking out. It will be hard to get around the heat issue in a small light.

Re: Compact 4AA host for 5W Luxeon

That light is pretty nice, any reason to not use a BB750 instead of that ZLT circuit? I think I'd want a light of more conventional shape, though. And the side-by-side arrangement of the cells (basically any arrangement except 4 cells end-to-end) is important to keep the total length reasonable.

Re: Compact 4AA host for 5W Luxeon

I didn't use a BB750 mainly because I already had left over ZLT parts from earlier projects and I wanted to see if I could easily modify the circuit to work for the 5W Luxeon Star. I also wanted ,at minimum, a 2 hour runtime on my brightest setting. Considering how hot it gets running at 540mA, I don't think I would be able to run it for more than 5 minutes at a time with a current of 750mA. On the charts, perfectly flat regulation looks nice. However, for my uses, a non regulated circuit like the ZLT works fine because I get longer runtimes and I can't visually detect much change in light output until it's time to change the batteries. This is probably due to the fact that the color doesn't change much with varying current levels.

Re: Compact 4AA host for 5W Luxeon

paulr,

What are your constraints (measurements) for a compact 4aa host? In other words, what would you call too big? /ubbthreads/images/graemlins/crazy.gif

Would you stick to NX0x optics or would you want 30mm? /ubbthreads/images/graemlins/cool.gif

Re: Compact 4AA host for 5W Luxeon

In a 4aa host I'd say four aa's end to end would be too big.

Given the space needed for four aa's, 30mm is probably better
than nx05. Or a mini-flood could also be interesting.

Really though, I'd probably rather have a 4AA battery holder with a threaded fitting to take a Surefire or McLux head, instead of a complete 5W flashlight.

Re: Compact 4AA host for 5W Luxeon

Well, that rules my idea out! /ubbthreads/images/graemlins/rolleyes.gif

Anyway, I'm waiting on parts from ELEKTROLUMENS for a homebrew LS light. Maybe when I post pictures you might find it interesting! /ubbthreads/images/graemlins/grin.gif

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
Doug S said:
[ QUOTE ]
paulr said:
The idea is to make the heat sink big enough for continuous operation. Would that be really excessive? I also thought copper turns out to be worth the extra weight, per tvodrd's calculations.

[/ QUOTE ]

I would be inclined to use increased thickness of Aluminum rather than copper to offset the reduced conductivity of the aluminum. Equal thermal conduction performance can be had with less weight of aluminum vs copper. The principal design challenge is the heat transfer from the sink to the air. Copper has no better performance than Aluminum at this interface. I think that 3 to 4W continuous would be attainable with a reasonable size head. I tend to be conservative on the question of running at high die temperatures. Others here will cheerfully tell you to go for the full 5W and more.

[/ QUOTE ]

Doug,

Wouldn't it be better to use a copper slab near the emitter? This would reduce the temperature gradient across the slab, and keep the junction temperature a little lower, although with a bit more weight. I can see using aluminum for the rest of it, since the convection at the surface is a function of the surface area, not the material.

Re: Compact 4AA host for 5W Luxeon

[ QUOTE ]
Lux Luthor said:
[ QUOTE ]
Doug S said:
[ QUOTE ]
paulr said:
The idea is to make the heat sink big enough for continuous operation. Would that be really excessive? I also thought copper turns out to be worth the extra weight, per tvodrd's calculations.

[/ QUOTE ]

I would be inclined to use increased thickness of Aluminum rather than copper to offset the reduced conductivity of the aluminum. Equal thermal conduction performance can be had with less weight of aluminum vs copper. The principal design challenge is the heat transfer from the sink to the air. Copper has no better performance than Aluminum at this interface. I think that 3 to 4W continuous would be attainable with a reasonable size head. I tend to be conservative on the question of running at high die temperatures. Others here will cheerfully tell you to go for the full 5W and more.

[/ QUOTE ]

Doug,

Wouldn't it be better to use a copper slab near the emitter? This would reduce the temperature gradient across the slab, and keep the junction temperature a little lower, although with a bit more weight. I can see using aluminum for the rest of it, since the convection at the surface is a function of the surface area, not the material.

[/ QUOTE ]

Lux, Yes I agree that for a given slab volume the copper will perform better. If you go with the copper/aluminum mix you have to consider the resistance of the interface between the copper and aluminum parts of the light. It can probably be done well but I think that if not done well the resistance of this added thermal junction will more than offset the gains of using the copper under the LED.