MT-G2 direct drive mag mod

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
I've done a number of mods before, but this is the first one I'm posting a thread about. I've been working with using the MT-G2 in an EDCable light for awhile, and was wondering how hard the emitter could be reliably pushed. I decided to use a 2D maglite as a host and go from there. Next I had an idea for a heatsink, and happened to stumble onto Bucket's heatsink thread: http://www.candlepowerforums.com/vb...g-Mod-Screw-In-Aluminum-Heat-Sinks&highlight=

Bucket is FANTASTIC to work with! I had a lot of ideas banging around my head, but had some trouble organizing them. Bucket was able to ask the right questions and gather the necessary information to create exactly what I wanted. After many PM's, I sent bucket some measurements for the heatsink and he produced this:
sinktopview_zps02ba2595.jpg
[/URL][/IMG]
It is an aluminum heatsink with a press-fit copper rod core. I figured the copper core would help pull the heat away from the copper mounted MT-G2 better than the aluminum. The two set screws are used to secure the star to the sink. I used some IC Diamond 7 carat diamond thermal compound between the star and the sink. The sink has a lip at the top that keeps it from being pushed too far into the mag tube.

wiring_zpsa74cd0db.jpg
[/URL][/IMG]
Sorry for the blurry image; my only camera is on my cell phone. My plan was to use 18ga. OFC speaker wire to deliver the current to the led.

switchandsinktopview_zps9f788708.jpg
[/URL][/IMG]
The switch is prepped and the heatsink holes fit the bigger wiring.

soldering_zps68c5ccbe.jpg
[/URL][/IMG]

Soldering has begun, and the heatsink inserted partway into the tube. It was a very tight fit, which was perfect for creating the thermal pathway.

mounted_zps06d1cf3a.jpg
[/URL][/IMG]

Soldering done, and the sink pressed fully into the mag tube. I used a rubber mallet to get it all the way seated.

spacerandspringmods_zps516e4c37.jpg
[/URL][/IMG]
Battery spacer tube made from some PVC piping, holding two AW IMR 18650s. Next to it is the shortened tail spring. I was going to use the 18ga. speaker wire and some electrical tape to do a resistance mod, but decided to wait until after I had done some current readings.

headview_zps861fa74d.jpg
[/URL][/IMG]
Here is the assembled head view. Yes, the stock mag lens is scratched and nasty. I'll have to order a borofloat one soon.

So after everything was assembled, I threw the batteries in and turned it on... Holy Cow this thing is bright! It has a huge hotspot, and a very minor artifact in the center of the beam. I'm going to need to file down the screws and soldering pads so that the reflector can sit a half millimeter lower. I'm using a sanded down maglite rebel reflector.

I first tried the light with a set of protected panasonic 3400s. Their protection circuit is rated at 7 amps, and the light shut off in under one second. I tried a couple more times with the same results; so I know that the emitter was pulling more than 7 amps.

I switched over the the IMRs, and lit it up. It seemed brighter than the split second of light from the protected cells. Next I tried to get a current reading using my harbour freight multimeter (I know I know)... and I'm pretty sure I blew the 10 amp fuse. So I got my other harbour freight multimeter...and the same thing happened. It just reads a "0" for the detected current. I have used them before to measure current in a maglite on various incan mods, and they worked fine. I guess this is just too much current for them to handle.

I decided to run the light for about three minutes straight with a tight grip. It got pretty hot, but not too hot to hold. There was no tint shifting or dimming that I could see. I let the light cool down for about fifteen minutes, then ran the light for another four minutes straight. I decided to shut it off at that point, because it was getting uncomfortable to hold. But again, I noticed no tint shift or dimming. I pulled the cells out of the light; they read 3.68 and 3.69 volts.

I tried to get a beamshot, but the camera on my phone couldn't focus at all. I'll try again tonight in an outdoor setting with a couple other lights for reference.

Let me know what you think! And a HUGE shout out to Bucket for helping me along with this mod :)
 

Bucket

Newly Enlightened
Joined
Jul 17, 2003
Messages
77
Hey Zach! I'm glad to finally see this beast. I'm pleased to see that the LED looks well centered. I agonized over drilling and tapping those holes in the right place since I didn't have the star in hand.

I can't wait for the beam shots.

Most multi meters only measure up to 10 amps, so don't feel too bad about the HF meter. If you really want to measure over 10 amps, you'll need a clamp on type. Be very careful buying one though. Some say they will do DC amps, but actually only do AC.
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
Buck, the LED is perfectly centered! You did fantastic with properly locating the holes. Thanks for the tip regarding the multimeter. I'll have to find what is out there. i know the Fluke is what everyone recommends; are there any others out there that are worth looking into?

Truth be told, I'm not sure that the LED is actually drawing ten amps. Based on HKJ's discharge graphs of the AW IMR 2000mah cells, they look like they have a hard time holding their voltage at a load that high; I could be wrong though. I think what I may end up doing is buying a couple of the MNKE IMR 26650 cells to see how they do.

EDIT: Using HKJ's graphs of the battery's capacity at a ten amp load, I did the math for a basic idea of runtime: 10 to 11 minutes. So I'll give that a try tonight and see what kind of results I get.

Hey Zach! I'm glad to finally see this beast. I'm pleased to see that the LED looks well centered. I agonized over drilling and tapping those holes in the right place since I didn't have the star in hand.

I can't wait for the beam shots.

Most multi meters only measure up to 10 amps, so don't feel too bad about the HF meter. If you really want to measure over 10 amps, you'll need a clamp on type. Be very careful buying one though. Some say they will do DC amps, but actually only do AC.
 
Last edited:

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
Good call! I'm concerned that I wouldn't be able to keep the star centered though. How would you go about the direct soldering? Given that there is not apparent tint shift or lumen loss, I will likely wait to alter the setup until I have a stronger power source.

My only suggestion would be since you went through the trouble to insert the copper into the heatsink, direct solder the copper star to it and forego the thermal paste. And... Beamshots!
 

Bucket

Newly Enlightened
Joined
Jul 17, 2003
Messages
77
Amprobe makes a decent product. I saw one I liked on Amazon for under $70.00 that did both AC and DC amps.

I'm thinking about how to solder the star directly...

Temperature 60/40 solder melts is 190C...about 375F. If you were to heat the sink up in the oven to 385F just to pick a number and somehow preheat the star/LED to some lower temperature. Flux both, put some solder on the sink. Then apply the star and gently cool with a fan it could work. I worry that the bare LED might move around on top of the star though.

I could make a jig that fits on the heat sink so that you could just plop the star in it and it would be centered.
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
The bare led moving around is my concern too. I know while cutting down the copper stars with a dremel, i have to stop every thirty seconds otherwise it gets hot enough that the led slides.

And thanks for the recommendation on amprobe!

Sent from my DROID RAZR using Tapatalk now Free
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
Well, I definitely need to invest in a semi decent camera; the camera on my phone just doesn't cut it. I post the following photos hesitantly. They are taken with no flash, but also no exposure "lock." They are only indicative of what I actually see in the sense that one light is a thrower, one a close range flooder, and one semi-insane :) The reality is that my camera could not capture how far any of the light was going. I could clearly see the baseball backstop and logo with the thrower and the DD MTG2.

Here goes:
I was standing on the basketball court (red dot) in the upper middle of the picture. The target was the backstop of the baseball field.
distance_zps7ac5e7ad.jpg
[/URL][/IMG]

control_zps14031a13.jpg
[/URL][/IMG]
Control shot. The light in the center of the photo is about 45 feet beyond the backstop.

3Amtg2M2_zps22fd8f8b.jpg
[/URL][/IMG]
3000k 90 CRI MTG2 at 3 amps (estimated 1200 led lumen)

1AVoBxpe2_zpsad3fa57f.jpg
[/URL][/IMG]
VoB cutdown Maglite with an XPE2 at 1 amp (266 lumen per cree's datsheet). I had to move the camera position to even get it to focus on the light. This shot is VERY deceiving; in person I could see the huge backstop logo very clearly.

DDMTG2_zps9792788b.jpg
[/URL][/IMG]
DD MTG2... who knows how many lumens :) This shot is also deceiving. It was considerably brighter in person; I am pretty sure the camera auto adjusted the exposure.

I'll see if any of my friends have a decent camera they will let me borrow. Can anyone recommend a cheapish ($100) camera that will take better shots? I know nothing about photography.
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
Just finished the runtime test. I had to do it between last night and this morning, so it is not as accurate as I would like given that the cells got to rest for several hours.
Breakdown:
90 seconds for the beamshot and just playing around.
fifteen minutes later, I ran it for 5 minutes continuously (it got really hot!) I had forgotten my multimeter at home, so about 40 minutes after this 5 minute run the cells read 3.89 volts.
This morning I ran the light for an additional 4 minutes continuously. Directly from the light, the cells read 3.69 volts.
I put the cells back in the light and ran it again for 2 minutes 27 seconds. At the 45 second point of this run, I noticed that the light seemed dimmer. The last 60 seconds of the run I could see the light dimming continuously. Direct from the light the cells read 3.60 volts. After resting for 10 minutes they bounced back to 3.65 volts.
I put the cells back in the light, and it dimmed after 5 seconds; so I'm just going to say that those 5 seconds don't really count. The batteries upon removal read 3.60 volts again.

So total runtime to dead batteries: 12 minutes 57 seconds. Given that at the 11 minute point I started noticing that the light had gotten dimmer and that in a direct drive setup as the battery voltage drops so does supplied current, I'll go out on a limb and say that the initial current draw on full batteries was at or above 10 amps.
 
Joined
Jul 2, 2008
Messages
553
Location
Gainesville, Florida
Amprobe makes a decent product. I saw one I liked on Amazon for under $70.00 that did both AC and DC amps.

I'm thinking about how to solder the star directly...

Temperature 60/40 solder melts is 190C...about 375F. If you were to heat the sink up in the oven to 385F just to pick a number and somehow preheat the star/LED to some lower temperature. Flux both, put some solder on the sink. Then apply the star and gently cool with a fan it could work. I worry that the bare LED might move around on top of the star though.

I could make a jig that fits on the heat sink so that you could just plop the star in it and it would be centered.

It is probably academic at this point since it is working so great, but my thought was basically this ^^^ (heat up the sink first), then after applying solder to the sink and flux to the star, screw down the star as it heats up using those perfectly centered screw holes. The LED should self-center on the star from surface tension.

If you machined a circle in the heatsink as Bucket is offering to do :D then you would not need the screw holes at all. I am direct soldering the heatsink -> star -> LED for my Copperhead flashlights and they are working great.
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
Sorry guys; I reread my responses to the idea of direct soldering the star and realized that my enthusiasm did not come across (I was typing quickly at work). I definitely like the idea of the jig for centering the star. My initial thought had been to use the screw holes to center the star as well. I need to see if I can even get the sink out of the light though. It took a lot of force to seat it, and I'm not sure I'll be able to secure the body in such a way that it is not damaged while hammering out the sink. I really would like to try powering it from IMR 26650s though, and that will require (I'm almost certain) direct soldering of the star to the sink.

Bucket, I was looking at your thread today for the mag extension you designed for high powered led mods. I have a couple ideas in my head for builds utilising that kind of design. They will be a ways off though as I expect the required parts will be quite expensive.
 

Chodes

Flashlight Enthusiast
Joined
Jun 5, 2008
Messages
1,092
Location
Australia
Just MHO - while I think soldering star would have the best potential outcome, I suspect screws provide the best practicle result. ie, all you need is a small amount of thermal compound, then rely on the screws for pressure.

Option to add 2 more screws for maximum clamp force too.
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
Thanks tobrien!
Chodes, i agree with your point about practicality. I was checking the numbers last night on the thermal conductivity of solder, and it is something like tens of times better than the diamond thermal compound. My question now is whether solder fills in the"gaps" between the metals as well as the thermal compound.
Thanks for all of the positive comments and suggestions guys!

Sent from my DROID RAZR using Tapatalk
 

ledmitter_nli

Flashlight Enthusiast
Joined
Jun 4, 2012
Messages
1,433
Is there a resister somewhere in that circuit? Two direct drive IMR's didn't burn that LED out? (i've done it before)
 
Joined
Jul 2, 2008
Messages
553
Location
Gainesville, Florida
Thanks tobrien!
Chodes, i agree with your point about practicality. I was checking the numbers last night on the thermal conductivity of solder, and it is something like tens of times better than the diamond thermal compound. My question now is whether solder fills in the"gaps" between the metals as well as the thermal compound.
Thanks for all of the positive comments and suggestions guys!

Sent from my DROID RAZR using Tapatalk

When you solder pipes, they become water and airtight. So, you may assume that the solder fills the gap. The key is flux. Where you have flux, you will have solder!
 

darkknightlight

Enlightened
Joined
Oct 28, 2011
Messages
440
Location
USA
When you solder pipes, they become water and airtight. So, you may assume that the solder fills the gap. The key is flux. Where you have flux, you will have solder!

Sweet, then i think i know what i will be doing with future high power builds! Thank you for your input :)

Sent from my DROID RAZR using Tapatalk
 

Bucket

Newly Enlightened
Joined
Jul 17, 2003
Messages
77
I just soldered a noctigon onto one of my hybrid heat sinks. It is cooling down right now. The greatest problem that I had was that the LED die moved while I was setting the star on the sink, so I had to mess with it to put it back where it needed to be and didn't get it exactly right.

Here it is.
knlmKFemmffRIQrfrSPGSqUZiRSBh3ZrbvQIHWhK50Y=w916-h915-no


20131108_125328.jpg
 
Top