Easy Driver Heatsink for Mag D Tube!

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
Hi Friends,

I was thinking of a heatsink for HighPower Drivers in a D Mag tube. So I have ordered 35mm dia. cooper tube with 1mm wall thickness. And made a heatsink for DX.26106

Here is the Results,

First The cooper Tube
IMG_1009.JPG


Heatsinked Driver
IMG_0999.JPG


IMG_1001.JPG


When Working
IMG_1004.JPG


LED's
IMG_1015.JPG


Thanks,
 
Last edited:

datiLED

Flashlight Enthusiast
Joined
May 9, 2006
Messages
2,023
Location
Atlanta, GA
That is a nice, inexpensive solution. :thumbsup:

How much of a drop in driver temperature do you notice? Is the inductor still getting hot?

What is the voltage into the driver, and how many LEDs are you driving?
 

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
For now driver is not mounted to a mag tube. Also at this point (as a only passive bulk heatsink) it works well. I can touch the inductor after heatsinked the driver, before it was very hot to touch. When mounted in mag tube, i wish it will work cooler.

- Power to the driver is 10.5V and 2.17A = 22.79W
- Power to the 4x Serial LED's is 14.3V and 1.40A = 20.02W

Wish Helps,
EpRoM

That is a nice, inexpensive solution. :thumbsup:

How much of a drop in driver temperature do you notice? Is the inductor still getting hot?

What is the voltage into the driver, and how many LEDs are you driving?
 
Last edited:

netprince

Enlightened
Joined
Sep 7, 2008
Messages
547
I like that, where did you get the copper tube? How snug does it fit in the m@g?

EDIT: Better yet, can I buy a few pieces from you? :)
 
Last edited:

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
I have used 3*1Ohm (1206 Package) resistor as sense resistor.
and make the mod's for this driver which seen on this thread


Today I make a 3 hour runtime test with 4S Cree and results were perfect. Input was 12.1V, Output current stabilizied at 1.43A. Efficiency was 92% after the mods.

How did you mod the driver to handle the increased input voltage and also increase the output current? 1.4A out is perfect for driving some XP-Gs.
Thanks! :thumbsup:
 

tinkerman

Newly Enlightened
Joined
Oct 15, 2007
Messages
16
EPROM, IMO you're connecting your heatsink to a non effective location on this driver.
The negative input (-) connection ring you connected your heatsink to isn't directly thermally coupled (i.e. vs metal to metal heat conduction) to any of the main heat generating components (and the following first two are also failure prone) - the power Mosfet, the Schottky Rectifier and the big inductor, so that any heat transfer to the heatsink is done via the glass epoxy material which is a lousy heat conductor, although there are several vias in the PCB below the power Mosfet that go to a small track on the other side, but they aren't connected to the (-) ring.
On the other hand, all three ARE thermally coupled (as well as electrically) on the components side of the PCB in close proximity - at the power Mosfet's cooling tab. Thus the best place to connect a heat conductor would be to the power Mosfet's cooling tab and nearby Schottky Rectifier pin (which is that part's specified cooling path). Of course it would be necessary to carefully isolate this power carrying heat conductor so it doesn't short to anywhere, especially not to the (-) connection, which would result in a direct short between the (+) and (-) inputs through the coil.
I believe that this method would give a much better cooling effect. Too bad this good driver board doesn't offer easier cooling options.
Another possible method would be to fill the gap between the coil, Mosfet and rectifier with thermal glue and gluing them all (at the coil side) to a heatsink, but I doubt this would be effective because such glues have relatively poor heat conduction at those thickness levels compared to the metal to metal heat transfer mentioned above, as well as eliminating any possibility of electronic repair if needed.
Even inserting a relatively thin (1-2mm) heat tab between these parts and directly gluing to their plastic packages and the coil won't be very effective IMO because these parts' plastic packages aren't very heat conductive either.
I'll try playing with these ideas and come back with results. In any case, applause for all the good work that you, HANSV and others have done about this great driver.
Cheers!
 
Last edited:
Top