Just off the CNC - Update: It's Alve!

Thank You! That makes two of us :)

The vast majority of the reason I 'built my own' light like this is that I found the thermal charachteristics of most of the commercial stuff.. well.. terrible.

All of the big brands out there build their lights on an extruded aluminum frame. The problem with this is that, even when they do extrude heat sheading 'fins' into their frame - they face the wrong way! I don't drive sideways!

After a lot of reading, I found that most of the available stuff has serious heat problems. ALL of the lights have a small-print disclaimer about their brightness claims and it's all because of heat. They have to back off on driving the LED's because the whole assembly gets too hot.

So - 6 months later - here we are, and I've started building my own. Before I went down this road, I did plenty of thermal analysis, and I thought you might enjoy seeing some of my results.

Here are two cut-plots of the steady-state temperature of the light. These temperatures assume that I am driving 24 XM-Ls at 3 Amps Each (240 Total Heat Watts), and they assume that its 90 degrees F (32 degrees C) in the ambient environment. They also take into account a light breeze (2 MPH) passing from front to back. The numbers would be far better with even a few more MPH of airflow. These are worst case scenario.

As you can see, I can easily drive all 24 emitters full-blast for as long as I want, even in 90 degree weather. The careful thermal design of this light is what allowed me to cram 240 watts of LEDs into a bar that is 15" long rather than 48" long.

Cut-Plot JUST behind where the actual LEDs are mounted. This is the hottest part of the entire light.
Also visible are the trails of cooler (but still quite warm) air as it passes through the fins at 2MPH.

Side view Cut-Plot through about the center of the enclosure. You can see one of the lenses and an entire fin.

Hope you're all enjoying my ramblings about this project. I really can't wait to get it lit up.

PPtk

Gonna need some beamshots of this baby! Looks fantastic!

-Garry

Interesting Pilot. You are so right about the heat sinking fins on lights. For bike lights, they usually face front to back so air is flowing over them. As a matter of fact, most bike lights get quite hot standing still because they were designed with air flow in mind. Here's an example you might find interesting. Troute started as a DIY guy.

wow

thermal analysis, and to an extent that very few, if any, manufactured lights get

that is one squared away flashaholic!

That is quite an impressive project!

Can you give us a list of software you used for this?

That's ok. It was just a picture of my old trail Jeep (XJ). But yeah, keep us posted if you do decide to make them for sale. I've been holding off installing any aux lights because nothing really struck my fancy. This looks like a winner!

Nice, thumbs up for figuring out the cooling!

Some serious driving, 24 XM-L´s at 3A :o

At least I would be satisfied even with a little less of current with that amount of emitters :D

Wow.

I wish there more threads like this. 24 XM-Ls . . . farking awesome.

farkFoy

Really great :)

Cant wait for the result!

Fritz

Sorry I've been gone for a while - Life got busy, as life tends to do once in a while..

dthrckt, Thanks very much for the 'wow' :) Thermal Analysis is quite amazing today - My suspicion is that the big guys don't do it because they don't want to know :)

The Solid Modeling was done with SolidWorks 2011

Thermal Analysis was done with Solidworks Flow Simulation 2011.

CNC CAM was done with Mastercam X4.

Circuit board Schematic Design was done with Orcad Capture CIS

Circuit board Layout/Route was done with Cadence Allegro 16.3

More to come!
PPtk

Had a few minutes the other day, so I threw together a schematic and layout for the circuit board that will carry the LED's and the drivers. I may consider sharing this once it's been debugged and proven to work.

Here's what went out to be fabricated. Should have it back in three or four days.

Overall, the board has 1,678 vias - the vast majority of which are strictly for helping move heat away from the LED emitters and into the enclosure.

The board was designed for 30 Amps of current (Max current with 24 LEDS at 3A each is about 20 Amps) with less than 0.05V of drop across the 14 inch length. Each group of 3 LEDs will be infinitely and independently variable in brightness. PWM Frequency of the constant current drivers will be 500 KHz, so there will be NO audible hum or visible blink. With the particular N-Channel MOSFETs and ultra-fast recovery Schottkys that I've picked, efficiency of the driver should be 92.. maybe 93 percent. Not bad..

PPtk

Circuit boards showed up yesterday. They look great. I'll be putting parts on and testing the circuits this week - should be making light by next weeekend.

The enclosure goes monday to be anodized.

2ish weeks and I should have a fully completed light!

Top Side PCB

Bottom PCB

Man, that's lookin' bad ass.

Pilot -

This project is so awesome . . . I can't wait to see this on the front of your Jeep. Regarding throw; would you say each triple approximates say, a Dry or Sky Ray in size/reflector depth/configuration. Will it be, basically 18 Drys and 6 triple XP-Gs driven to within an inch of their collective lives?

Sorry if I missd it . . .

Foy

Impressive. I'm really looking forward to this thing completed. :)

Absolutely brilliant project, looking forward tot his perhaps becoming available as a kit.

Awesome work! Did you learn or study something related to this?

wow you are a perfectionist also a dedicated board

but hey lazy ass you promised 2/3 weeks! ...kidding xD

Foy,

Thanks very much for the kind words. I'm happy that you are enjoying my project.

The optics are a LITTLE smaller than the SkyRay/TrustFire/DRY - but they're TIR, not reflectors. In my limited experimentation with them, the beam profile is quite similar to my trustfire 3-XML. The hot spot is beautiful and is hard to even define - it just rolls off into very useful spill.

I believe the Trustfire and the Skyray are actually driven less hard than I will be driving my emitters. I'm giving them TRUE current-controled 3A each. I'm not sure how hard the DRY is driven, but I think I recall it having a Direct-Drive mode where it's probably getting more than 3A per emitter.

For the moment, I'm going to populate the board with all XM-Ls (24 of them), as I don't believe I'm going to need the additional throw from a few XP-Gs. There will just be so much light coming out the front of this thing that I don't believe its going to matter, and I'd rather have the raw lumens of the XM-L's than the slight increase in throw from the XP-G's with the resulting fewer total lumens. I'm also more confident that I can get the tint to match using all the same emitter type.

PPtk