Active cooling, would take care of any heating issues. Remember that modern CPUs like the i7s draw more power than our measly XM-Ls do (8a minimum on i7s), so your XM-L star on one of those would keep it cool with no issues.
Just a little status update… I wired everything up with the east-092 driver and I think it cooked the OEM XML. It lit up just fine for 2-3 seconds then flickered and died. It was never even on long enough to build up any heat. Now it only emits 1-2 lumens. So I think that east-092 driver doesn’t like solder+wire point to point designs like this. I think it relies on spring and contact resistances in the HD2010 (and other lights) to tame its output enough to keep the LED alive. Bummers I got a little lucky on the OEM emitter tint lotto.
Looks like I’m headed to Illumination Supply tomorrow to do a little shopping.
WHAT? You live how far from me? C’mon, it’s a simple matter of coming up with a plan and seeing if we can pull it off. BTW, bigC gave me a 20mm copper star that someone had CNC’d some time ago that I’ve never felt the need to use until now. I’ll let you have it for the same cost to me: gratis! Then I make a copper core aluminum heatsink, drill/tap some holes into it and hang a pair of square CPU heatsinks off the side. You’ll need to cut some holes for them to stick out of the case then bolt it all together. Done! How hard is that? Hehe!
wow keep up the good work!
Well, the east-092 is DD on high and with 4 cells in parallel you probably gave it way too much current since voltage sag would be very low thus causing a lot more current to flow.
Yeah, shitty DD drivers will burn emitters. For example, when xm-l’s first came out, folks here would sometimes boast of how much light the “drivers” in their new toy puts out up until it fried. 7135 drivers are slightly less shitty in the sense that they (over)heat themselves with the excess power from DD.
This thread turned out to inform the “why some chinese lights are cheap” one, in this case the HD2010.
Yeah I tossed the east-092 driver. I liked it a lot in the HD2010 and had no problems in that light. I used it in a MCE P60 build (dies running 4-P) and it cooked the MCE…. almost in very similar fashion. It lit up for ~1 second and turned angry blue, way before any heat could be generated. This particular build went great, I used brass rod stock for the +B spring and it worked great for 1-2 weeks. Then I got the urge to add copper braid wire to the -B spring on the module and on the -B tailcap spring… Oops, I think thats what killed it.
The 7135 drivers get really hot too. I have been using these for 3-4 years in various DIY builds (including a stanley fatmax 3.4A mod) and this is the first time I payed much attention to it. The 10x1735 modded board almost instantly gets almost painfully hot to the touch. So when I get the cooling fan thing all figured out I am going to have to make sure the driver also sits in the cool air flow.
Sadly the bountiful space and room is quickly being eaten up inside the light. So I don’t think my copper heatsink idea will be as BIG as I originally had planned. I think its going to be a smaller sized sink with some air flowing around it. I really want to do the 4-eneloop + fan thing too. Theres just “something” appealing about idea of two independent dedicated circuits in the same housing. For some strange reason (flashaholic OCD I guess) I want the fans to run full throttle at a nominal 4.8V. So that kind of mandates a separate eneloop power section.
I pretty much have the light engine part finished… just now gotta think about forced air cooling.
Cheers BLFers!!
Looking good!
-Garry
thanks Garry… this has been a really fun build.
Enjoying the reading guys. Interesting. It has made me think in a passive cooled situation such as this light, what would cool better? A solid chunk of copper or a heat sink such as the intel one without the fan? I realise size makes a difference but in this case?
Mass absorbs/contains heat but surface area radiates it (or via convection when a fan is present). Imagine a bucket with holes to drain water. More massive heatsink means bigger bucket, more fins equals bigger holes.
if this is going to be used on high for any length of time I think some way (active or passive) of getting the heat out of the case is essential. Not just for driver/LED life but also to avoid murdering the batteries. Don’t forget that it’s probably being driven at 1A or less, so as stock that heatsink and case only have to deal with 3W, whereas 10W is a whole different game thermally 
Great progress! Love the way you take your time to document and share details about it. And great picture of your son. He looks like a sweet kid.
I am excited about this flashlight although I cannot get my hands on one because it is stateside and I am not. But the way you build this up I sure see great things coming and coming fast by the frequent updates. Thank you for sharing this with us. Your rock!
That looks like the inside of a van I had in college…
Very nicely done, and well-documented to boot !!!
I would try small 5 volt CPU/Video fans and run them directly off the 4P 18650 pack. All of mine run fine on 3V. It only takes a small amount of moving air to do a lot of cooling. I bet a single well placed fan will keep everything under control.
Please see post #39.
I wanted to run the fans at a higher voltage than the 3.6. One of the things I was doing when bench testing the fans was evaluating their performance with different battery configurations. First I ran it with a partially depleted 18650 ~3.7V, then I wired up the 5S-AA pack, and ran the fans with 4S and then the 5S cells. In each scenario I could both hear and feel different airflow output levels from the 4p fans. The little suckers really do move a lot of air. I can easily feel the exhaust air blowing on my hand through the Chicago X disco-van holes. Also if you look at the pics most of the shell-casing is tab-in-slot, so the case has a decent air-seal around the perimeter. The trigger is also a pretty air-tight design. I can smell the shoe goo fumes venting a little when the fans are running.
I was thinking about it some more and realized I am making a trade-off. I don’t have any room for a huge heatsink mod, but I am running the fans at a much higher voltage and output level with the big eneloop pack. I still have a little room in there to piggy back another heatsink, it just can’t be very big. Although now I am wondering would it be better to just leave the fans blowing directly onto the small / thin OEM sink? Would piggy backing another sink on top of that, do more harm than good?
Thanks for the comments fellas!! I just realized this is the first DIY mod where I purchased everything local… no mail order!!
Oh yeah… my groovy son said Dad it looks like a poorly designed hair dryer.
if you have airflow then you don’t need a large heatsink - the stock one should be more than enough. I think the dinky little fans are a really cool touch, they give it a very futuristic look! One thing to consider would be to add an airflow guide to each of the heatsink fans so that the air is directed onto the heatsink as it looks like it’s flowing past it at present. Also, if you can make sure the exit holes are unobstructed as possible, that’ll also help a great deal as the fans wont be labouring under such a high positive pressure.
so, when are the beamshots?!
Yeah airflow is one of the things I was wrestling with. I wasn’t sure if I should try and aim the fans bluntly / directly at the side fin area of the heatsink, or if I should try and skim air off the flat surface. I chose the latter, being that the flat area is the heat source and gets warm noticeably faster than the perimeter around the fins. I was also thinking that hitting the fin area bluntly with a stream of air will introduce more turbulence in the “box”… so I would in theory create swirling pockets of trapped hot air inside.
Good point about the air duct and baffle… I hadn’t thought about that. I might also knock some more holes into it. As far as I can tell, placing it right next to my face I get pretty uniform air flow through all the holes. Just did some number crunching, and the 20 exhaust vent holes add up to 351mm-sq area, and the fan blade surface area (not including the hub of each fan) adds up to 84mm-sq. So I am giving the fans a lot more surface area to vent than they have coming in.
thanks all … beamshots tonight.
Hahahaha, Kramer, hands down 8) One of the most interesting/impressive mods I’ve seen so far 
Waiting those beamshots 
(try to make the same comparison you’ve made initially)