Prototype off-road/boating spotlight - King COB!

Can’t machine the head to fit?

I think I can make it work if I can remove the plastic bits of the Vero29 from the actual COB assembly… or just use a CXB3590 and machine my own adapter for the Lenina.

I don’t think I actually need the mounting lugs… just a tight press fit boss similar to what the LEDiL adapter plates actually have in the Lena series.

I cant work out to post photos but i used the L6 reflector with cmt 2850 and taskled HBflex in a scratch build. But it fried the driver, i think the shelf was too thin and ambient heat got it. Happened twice and now im waiting for more drivers. I will redo it with copper inserts for the led and driver i will also put them back to back with a gap in between so the copper protects the driver from the led.

Would like to see it. Easiest way is to upload your photos to imgur and copy the url to the box when you click on the add photo box above the comments box.

As I said earlier, my evil plan involves a tapered copper “shelf” for the COB and the HyperBoost will be mounted to the back of the shelf with thermal tape.

Bezel got used for something else. The switch is an automotive push button with led.


Pretty cool!

Some “fog lights” im working on, not sure how to do the heat sinks, the drivers will be mounted outside the lights.

It is a big light with 5 x 26650. I dont do drawings trial and error works out more costly but good fun.

Well, I can confirm that with some minor modifications, a LEDiL C13867 will fit over a CXB3590. So, I think I’m probably going to design a custom spacer for the L6 and put the CXB3590 in there and then get a Vero29 for the Battle Lantern. Both will get Lenina-S reflectors.

So, I took the test rig out with the reflector double stick taped to the heatsink (lol) and the front lens light pipes like CRAZY. It is actually almost painfully bright.

Screen grab with the Lenina-S, a Vero29se and the new and improved front plate.

I did some tests with the CMA3090’s I have. Couldn’t figure out why I couldn’t get them powered up… until I figured out I was sent the 72V versions instead of the 48V versions! Does almost 12.8klm @1.2A, and up to 175-180% at max current, 2.4A. Not as bright as the Vero29 at 68V (18.5klm), but the LES is a lot smaller too, about 6mm in diameter smaller. Gets hotter faster. Have not tested the CMT2890 yet, but the output is theoretically approximately the same as the 3090, just at a lower Vf (51.2V) and higher current (1.6A). The LES is about the same size as well. What’s nice about the 2890 is that I can run it at max and draw less than 6A from an 8s config…

Have not tested the Vero29, CMT2890, or CMA3090 yet with the Stratus LED reflectors. I would have to make a custom interface to bolt everything together, and possibly shave down the bottom of the reflector at least for the Vero29. It does focus with the CXB3590 however.

How much current were you trying to push into the CMT2850?
.

The L6 head should be able to handle about 40w or so if you had it clamped down well. Before I disassembled my L6, I ran it on turbo outside for at least 10min and it didn’t really seem to get all that warm. allegedly the L6 on turbo, should be about 30w.
.

The “manual” for the HBFlex leaves a lot to be desired. It sounds like it natively supports an e-switch interface but there is no documentation that I can find. I may e-mail TaskLED about this.
.

— I actually typed this post yesterday but it seems I never hit “post”. Also, now I realize that you were just using the L6 reflector and your own version of the L6 head. —

Half the fun for me is drawing some of the parts. But in the case of the L6, I knew that a drawing of the head would be needed since I don’t have x-ray vision to see what the wall thicknesses are and where there is room inside the head.

I lost the contact name at TaskLED but they were pretty helpful with whatever questions I had. Not sure either about the e-switch other than to change modes?

The HBflex peaks at 3 amp output and the voltage can go to 45v. The cmt 2850 at 3amp was 35.8v about 107.4 watts it got hot quick but i didnt get a chance to test it before it fried itself. I have another 3 hbflex on order but i will work out the bugs and test before i ramp to full. I think the heatsink is most important thing and thinking a copper core is the go. Even on low it is a good work/flood light
I think the ideal combination is the hyperboost 80v 3.2 amp max with cmt2870 54v 2.8 amp this combination will give that bit extra above max current. But i havent really looked at the cma 72v

Possibly because you had the driver and COB mounted on the same piece of metal? I do plan installing one of these COBs in a flashlight host but haven’t decided yet on which one. I agree that heat sinking is paramount as well. I run all my tests on active heatsinking - even then the 3090 heats it up relatively fast. So any plans for using a smaller host for me will involve active cooling of some sort. I may just mill a head out of a solid billet but still looking for a faster, simpler retrofit solution (maybe someone has an MS12 with busted LEDs they want to dispose of…)

I don’t think copper is required as long as the aluminum is THICK. The base of that massive pin sink I have on my testing rig is at least 3/8” thick. It might even be 1/2”. Thermal modeling also confirms that for something passively cooled, the plate needs to be pretty thick to not “bottle neck” the heat. I want to use copper in my L6 but I have a LOT of aluminum at my disposal.

I would think though that would only be for a short time - you’d still need quite a bit of surface area to dissipate all that heat - i.e. fins?

Oh, of course you need surface area to actually get rid of the heat. I actually have an idea for solving that problem that I may use on the Lumen Bazooka. At some point, I need to machine some parts and do some testing. One key to passive cooling is fairly widely spaced pins/fins to minimize the resistance to airflow. On flashlights, I think these are usually designed and machined much thicker than optimal simply for durability reasons. If you look at Alpha Novatech passive heatsinks, they always have some crazy thin fins. On the order of 0.7mm to 1mm thick.