Direct Drive MT-G2 with Two Samsung 20R's

I first tried the MT-G2 with a fasttech TR-0124B driver, a TR-3T6 copy. I used a fasttech 10 ohm pot wired parallel to the original sense resistor as seen here on the driver. DST using 3 18650's (Finished)
The heat sink had a copper center section that I used to reflow the entire noctigon 20mm board with the MT-G2 on to.
.

With 2 NCR18650 2900mah laptop pulls, the driver done 4.5 amps to the MT-G2. Well, I wasn’t all that impressed with the driver. So I added another battery for a total of 3 and checked it again. The driver gave 8 amps to the led. Now I’am impressed. I didn’t test it more than about 30 seconds, the aluminum copper Maglite heat sink I made was blistering hot, no where near enough heat sink out of the light. I have no idea if the driver will survive at that current for very long. I didn’t notice anything getting hot besides the 10 ohm pot and it never got to hot to touch but that was only at 30 seconds of use.
The mag I intended to but this in is a 2D so as I was trying to figure out what I was going to do about getting 3 li-ions in a Maglite. I thought to myself, I don’t want to do that. I just got 2 of those Samsung 20R’s from RMM, Hey I’ll try those. I decide to take pics this time, so you couldn’t say it didn’t happen. :bigsmile:
.

.
That was the reading I got directly after plug it up direct drive.
.

.
This is what I got after 30 seconds. I don’t know if the cells where just starting to heat up and let the current flow or if it was the led heating up and dropping the vf. 9 amps direct drive with cells that have a very good voltage curve under that current. That should do just fine with no driver at all though modes would be nice.
.
At that current in 30 seconds the heat sink was getting so hot I couldn’t even think about touch it. It took a few minutes before I could even begin to touch the thing again. I have never seen a led produce so much heat in that amount of time. I set up another meter with a thermal probe and taped it to the heat sink, just to see how hot it was getting.
.

.
That was right around 30 seconds of being on and by the way that’s fahrenheit. You could boil water on that thing. I’am not real sure the Maglite body is even going to cope with this amount heat for any length of time. I guess I’ll find out. :open_mouth:

Do the words "massive heat sink" sound familliar?Wink An MT-G2 at 8-10 amps would need one, to last very long. The big heat sink, is only to, well, sink heat... It can't get out of the body fast enough to keep it stable, so a really big heat sink will just extend the inevitable, but that extends run time. Try a 4D maglite, with a tail cap switch and make the area from the top of the tube, all the way back thrub the first two cell places, into all heat sink. Now you got a little run time before the led melts off the star.Smile

One of these days, I need to al least do a lux test between an SST-90 at 10 amps and an MT-G2 at 9 amps.

I for one would be interested in that test as well OL.
.

That’s about the only way I see it too. There’s going to have to be more heat sink in this thing some where or back off the current, and who wants to do that. :_(
I original thought I would be able to use it for a few minutes at a time with out getting to hot, but now I think a minute may be about tops. Funny thing you mentioned the led sliding off, I kept checking the MT-G2 ever time I turned the power off to make sure it was still there. :bigsmile:

In my Solarforce M8, with the MT-G2 reflowed to 3 ounces of copper on a Noctigon (all reflowed together) and using 2 Samsung 20R cells, direct drive I was seeing 12A and 4140 OTF lumens at start-up, with the numbers falling like a bag of cement out of a helicopter.

I’m running it at 5.98A now on 2 Panasonic NCR18650PF cells through a Qlite, for 2322 OTF at 30 seconds, 2636 OTF at start-up. The MT-G2 is now sliced and diced. :slight_smile:

I like it. I haven’t succeeded in cooling the MTG-2 for continued running at that current either. Even 5 amps creates a tremendous amount of heat.

Gosh, maybe we should listen to Cree and keep em at 3A or below!

Naaaaaaaaaaa! :stuck_out_tongue:

Just use em when it’s real cold out, not a problem! manxbuggy1 and rdrfronty brought some lights over and we went out and played, maybe 80 lights between us. At 40º and windy my K3 at 4.74A and 2311 OTF wouldn’t even warm my hands!

12 amps from 2 cells. What charged voltage on the cells did you start out with? Did you warm the cells up first? I wouldn’t even have thought that was possible from djozz graph. crash-testing a MT-G2 on a copper Noctigon, graph done, mod done and repaired :-)
Not saying it didn’t happen, I just wouldn’t have thought it was possible with 2 cells and voltage sag. That’s impressive DBCstm!
Mine was actually increasing in current, I didn’t have enough heat sink to see where it stopped at.

Thanks for the info. I was going to PM you ask how well your light coped with the heat.
If all the copper you used want get it done, then what will?

Active Cooling, not passive. If one would forgo having a waterproof light, then a larger body light, (3x18560 or 4x18650), with a large head could be used and add a cooling fan.

Original led shelf

Copper CPU heat sink underneath

Cooling fan under the fins if the heat sink

Metal separation plate, so there would be inlet air and outlet air

Holes drilled in the head at both inlet and outlet.

With a 2S/2P battery set-up, you could just use a 6v fan. Plenty of them in round format, to fit into a head

Best off with a head/body made by someone with a lathe

I was going to do similar with an all copper light and 7xSST-90 leds, but I could never afford to make it.

fan

It could be simplified by drilling holes in the original led plate and letting the fan blow thru it and up around the reflector and exhaust out holes near the bezel area. For a custom one off light, who cares about the holes really?

Or, everything could be made so the fan area is sealed off so moisture only gets there and use a sealed fan.

I put together a direct drive last night with 2 king kong ICR’s to a MT-G2 It’s in a body from a 4xt6 no pill so the heat transfers directly to the head and out of the light. Works great outside but it’s well below freezing outside right now. Orange peel reflector 3,000 lumens, excellent flood and the tint is super. Perfect for walking around the backyard with (five acres) but I wouldn’t take it on long hikes. It was still at 3,000 lumens when I came inside. The head was hot but not the batteries so that’s good.

I put two of the Samsung 20r’s in a ZY-T08 MT-G2 that has a lustefire driver, it was popping the protection on Panasonic protected batteries after 30 seconds or so on high. With the Sansungs it was way more lumens at startup but dropped down so fast that it wasn’t worth running them. So I’m running Sanyo unprotected and it’s great.

I like to test em on rested cells. My SoShine S1 Max V3 charges these to 4.21V, after a rest period they stabilize at 4.18V. This was just done recently, as I haven’t had my lightbox long, and the weather has been chilly so the cells were probably under 70ºF when I started the test.

I also took the body off the M8 and tested it with 2 Powerizer 26650 cells. The amperage was almost identical to the 20R’s with slightly more lumens at start-up, but falling the same. The 20R’s seem to be a near equal to the larger 26650 and it’s one of the best out there.

I tested 4 different cells with the MT-G2 in the M8 in it’s regulated form, with interesting results…
Sanyo lap pulls 5.92A 2205 OTF at 30 sec.
Efest IMR V2 5.96A 2260 OTF at 30 sec.
Sam 20R 5.98A 2087 OTF at 30 sec.
Pan PF 5.98A 2322 OTF at 30 sec.

The Panasonic does what the best Hi Discharge cells do but have considerably better sag control with their 2900mAh capacity. So that’s what stay’s in em. The Sanyo lap pulls at 1.08 ea. hold their own quite well too, best bargain cell I ever bought! (HP battery pack, 12 of em for 12.96 shipped)


Sliced and diced

OL, that ought to do it. :bigsmile:
That’s probably a little more time than I would want to invest at the moment, but who knows if I cant find some happy medium. Where have all you MT-G2 users been hiding?
I might end up making a copper aluminum head extension, similar to what OL was suggest with the 4D. Just use the 2D and a massive extension.
.
DBCstm, did the tint shift noticeable after your dome fell off? :bigsmile:

I too would be very interested in that test. Perhaps even a SBT-90. With almost half the Vf of the MT-G2, the watts and heat would be almost half at 9-10A. If lumens are close to each other, the SST-90/SBT-90 might be a better bet. Is there sinkpads/noctigons available for SST-90's?

The MT-G2 should definitely win in the lumens department, but potentially higher lux and cleaner/tighter hotspot from the single die sst90 could be beneficial in some lights. A comparison between a 5000k sst/sbt90 vs a 5000k mt-g2 (domed and trimmed) would be very interesting!

The dome is still on 007, it’s just been sliced off directly above the die surface, and the phosphor extensions around the dies diced and removed. This gives a bit tighter hot spot, a little warmer tint, and better throw from the hybrid reflector in the Solarforce M8.

All in all, a nice experiment. Tint is about the same, maybe a bit more towards green than pink, as my triple Nichia 219. It looks merely warm by itself, but side by side next to the Nichia, it’s not as rosy a warm.

The M8 has a separate pill from the rest of the light…the pill is the mid-section that the battery tube and head screw onto, with the Noctigon thermally pasted directly to a shelf on the top end. I’ve added 1/2” of copper under this shelf, which of course is still not enough.

My plan is to put any money I receive at Christmas towards a fully copper mid section. VOB will be doing this for me. A bit wider in diameter, with thinner fins than the stock finned mid-section and about 10mm longer as well. The Noctigon will then reflow directly to this finned copper mid-section and it should be good for me to go direct drive. I will be utilizing a 20A switch…the smart switch, from mattaus in Australia. Should be good for the full 12A and we’ll see how the Panasonics handle that.

Might be pertinent to have VOB machine the copper section to fit his new HD2010 extension tubes and go with 2 26650’s instead of the 18650’s for better sag protection and longer run times. Hmmmmm……

The sliced and diced emitter in the baked M8 pill…

“Pill” in the M8, notice that the HA TypeIII anodize baked into an olive green military look, which compliments their choice of font quite well!

The M8 baked, in shorty configuration using one tube

Hi Dale. I thought the sliced and trimmed MT-G2 had a cooler tint than the original unmodified MT-G2? How does the tint of the sliced and trimmed MT-G2 compare against a dedomed XML2?

That is pure flashlight porn there! (running off to bathroom) 8)

here is a copper sink that i was going to use in a monster build but never got a chance yet to start lol but that’s what u need :slight_smile:

!!

holes drilled for the wires and drilled and tapped to hold your star in place. with a set screw to hold the slug in place in side a maglite body
!!

Ive been doing direct drive MT-G2 for some time.

With fresh 20R, im seeing around 9A on startup and uptowards/around 9,5A after a bit (15-20 sec?) runtime (on the tail).

Thread and post here.

If you want sustained high output, the Sony 2600mAh 26650 is better than the 20Rs.

Ill start a beamshot thread some day this week comparing that light and several other high output lights, already have the pictures. :)

Like a big fuel tank for heat, huh Nitro? :wink:


Race, the Powerizer 3600mAh cell does even better for total amperage, just saw today that while the 26700 Moli doesn’t make as high amperage at start-up, it holds it longer than the Powerizer. A trade-off. Kind of tight squeezing those Sony’s into places made for 20R’s……