07/07/15 Finished Handmade CXA3590 - Lum5-90 light build - Beam shots are in Post #1

The voltage will be adjusted and set. There will only be access to trim the amperage. There will be no moonlight mode. Bright as hell and then a little less bright, but still bright as hell. Moonlight is for wusses who only have 5000 lumen or less lights.

Just playin. I don't really intend on modes at all, just a way to back off of full amperage. The on/off switch in line with the battery pack output, will completely shut off the light.

I am taking that advice. A fan is on the way and I will put it in. I don't think it will do a lot, since I really can't cut fins in the bottom of the heat sink, but it will be there 92mmx92mmx38mm deep and 48v.

Do you mean replacing or moving the pot with some wires? Would it have to do with the wire length, or was it the remote pot itself? I was assuming a mechanical connection to the original part.

Or do you mean instability in general when trying to adjust during operation? I had a similar problem in my COB floodlight(still not quite finished but prototype stage published) build occurring even at about 50-70 watts, but then again I was running at more than 100% voltage boost and the converter was a smaller underpowered one. It's probably not bad at all(the version that has two trim pots, most have only one), but it's just too small. The one I have now(identical to what OL is using) seems stable at least up to 100 watts even when stepping up from 12 volts.

Is there any way to add a block with some fins? On my build, even a small fan makes a huge difference. I'm cheating by using a CPU heatpipe tower cooler, but I think any finned solid heatsink part would help quite a bit. Of course your build has a huge heat capacity, so thermal-wise it will tolerate short runtimes at almost any output. Given the huge amount of metal everywhere, fins might also help even if not directly mounted on the... ... pill.

That's what I was assuming. Probably I'm just being paranoid. That professional grade array should be strong enough to take a lot of any possible momentary overload during power up.

Hey! I'm sure the generic "100 watt" China COB with 11k chinalumens will reach at least 5001 lumens when slightly overdriven!

I replaced the original (current) trim pot with an external one connected with some small shielded coaxial cable. When turing the pot there we places where the output went unstable and the LED started flickering and the inductors started squeeling. I think the circuit may not like being adjusted “on the fly”.

OL: I'm sorry if we're adding semi-OT to a project thread but this seems relevant. The external pot thing is something I would at least consider in a build like this and I'm assuming you might. Unless you're set on mechanically extending the original current pot.

My converter was happy being adjusted on the fly, powered up. To me, that sounds like a faulty trim pot. Like when an audio amplifier makes static when you're adjusting volume or gain, and most often it's at the same spot(s) every time. In other words, the pot resistance itself is non-linear at that point because the resistive material is worn out or the contact is bad to begin with. Inductor squealing was noticable when the little DX converter went unstable, but has not occurred with the "600 watt" unit.

Nope, it was a multi-hundred-dollar mil-spec 10-turn pot… plus I tried a couple of different pots.

I think this explains what happened if anyone cares to read it. In a previous life I studied this and I have a vague recollection of actually understanding it (at that time) :slight_smile:

For a few interesting ideas, here’s a build with 100 watt COB, boost converter, external pot mod + hot components relocate.

Edit: not sure why the vid isnt working here. Here’s the direct link : https://youtu.be/c—5c3Egv4E

I am typing this on a computer with a 4 core CPU that needs to dissipate 125W of heat when running full tilt. It runs fairly cool with a relatively small heat sink and cpu fan. Moving air has tremendous cooling capabilities. I remember the progression in PC’s as the cpu’s got faster and faster.

1. No heat sink on the cpu chip
2. Heat sink with passive cooling
3. Heat sink with mounted fan,
4. Heat sink with heat pipes etc.

Most of our lights are still stage 2

I’ll say it again.

liquid cooling.

Liquid Nitrogen. It worked for AMD!

:bigsmile:

I am following OL’s lead on this build. I have ordered one of the LED’s and 2 of the drivers. One of the things I am going to try is running this on 2 - 18V drill packs in series, 42V fully charged!. First with the booster board and if the LED can handle 42V and from OL’s video I am pretty sure it can, DD of the 2 packs with an FET switch. Making and FET switch is almost trivial, of course there would be no modes and decreasing output right from startup.
The only problem I can forsee with 42V input to the booster board is perhaps there might not be enough voltage differential between the input and output at the voltages the LED booster board can handle

I will definitely be using a fan, but check out the name of the representative that OL and I have been dealing with. :smiley:

David xxxxx
2015-06-17 04:11:26
Shipping address is
xxxx xxxxxx Avenue
xxxx xxxxxxxx, CT 06xxx
My email is
xxxxxxxxxxx@gmail.com

Chuck Fan

2015-06-16 03:57:48
OK,i know,thank you!can you tell me your mail?
David xxxxx
2015-06-16 03:30:00
5000K please
Dave

Chuck Fan

2015-06-15 21:23:25
Hi,what color temperature you need?

Chuck Fan

2015-06-15 08:12:19
Hi,David:
do you need 5000K or 5700K,or 6500K?

dchomak - Great! Maybe Chuck can also sell fans? I kind of doubt that it's his real name, just his "for the americans" name. He seems like a good guy though. I enjoyed dealing with him.

Looking forward to your build.

06/17/15 - My build has come to a halt. I think it will be at least 2-3 weeks before I progress. I ordered a 48v fan, but it has not shipped, so now I have to play the ebay wait game, to get a refund and start all over again. I also had to send out the heat sink to have someone machine fins in it, so the project is just shut down for now.

Sux to wait…

definition:

chuck

informal

verb

throw (something) away.
“they make a living out of stuff people chuck out”
synonyms: throw away/out, discard, dispose of, get rid of, dump, scrap, jettison

:smiley:

A fan would draw very little current, it would be easy to drop the voltage down for a fan, but then again you have to wait for the heat sink anyway.

I’m quite new to the game, but every time I’m slowly recovering from a dropped jaw, you manage to surprise me/us with something even bigger.
What’s you next project gonna be? A portable worm-hole that projects the light of a super-nova in the far end of the galaxy?
Kidding: this takes me back to the time I pressed my nose against the windows of a Ford V8, while the average car in my street was a VW beetle.
Just go on and build the stuff that dreams are made off.

So now we have to wait.

I just wanted to mention that I ordered those boost boards that OL is using. Ordered Monday and got them today! only 2 days later. they ship from New York. For the money, they are quite impressive looking.
Now, while we all wait, I have something to play around with. What fun!

Back at it again, more photos in the OP.

Thanks for the inspiration OL.

I just ordered a CXB3590 to use. I wonder if I can get a little more out of it than the CXA3590. Instead of just using AA to hold my LED down, I am considering ordering one of the COB holders and D&T’ing my base plate so that I can screw it down tight. Since you have some experience with these LED’s, do you think that I will get adequate heat dissipation if I mount the LED to a 6” copper disc and have a heatsink with fan mounted on the other side of the disc directly behind/under the LED?