Always use the thinnest possible layer of thermal paste/grease, even the very best is not great at conducting heat. What it’s best at doing is removing air pockets. Those tiny air pockets are insulators and the point of putting down paste/grease is to get rid of that air. People debate endlessly on what is “best”, but arctic silver is very good. There are products that are “better” but it’s debatable if it’s worth spending any extra money on getting the very best. I use Wakefield Deltabond 153 but it’s $40 and probably not any significant amount better than Arctic silver. More important is making sure you use a very thin layer and that you have mated the two surfaces in a way that doesn’t allow any large gaps at all. If the area you are mounting the emitter to is rough it’s important to sand it down as smooth as possible. I use various grades of sandpaper until it’s basically polished. You want as much metal to metal contact as possible between the bottom of the sinkpad/noctigon and the heatsink. If both were absolutely perfect you wouldn’t need thermal paste at all, you could hold them in place with screws.
When you put the emitters in series it takes twice the current to get the same amount of amps to two emitters. When you put the emitters parallel you get roughly the same amps to the emitter no matter how many emitters you wire that way. I say roughly because it’s a bigger drain on the battery if you have a bunch of them wired in parallel, but usually not enough to consider wiring them safely hoping on thermal sag to save them. How should I put this? Hmm. OK, Wiring them in parallel is the same as if you were to run a separate set of wires to each emitter, think of it that way, because it’s essentially the same thing. It would take a lot of emitters wired that way to drop the amps quick enough to save them from burning out. It’s not really worth trying to do it and trust me, I’ve tried it and lost emitters a bunch of times. I’ve also tried running thinner wires hoping the loss of current to the resistance from the small wires would stop them from frying. All I did with that one was to cook my wires. So it’s best just to remember that each emitter can handle a full load from one battery. In the case of XHP’s and MT-G2’s it’s TWO batteries though, because they are 6 volt emitters. It’s just with XPG’s XML’s and XPL’s that it’s a one to one ratio. It’s convenient that it works out this way though. It makes it much easier to figure out the correct way to wire them.
It makes it more difficult when it comes to wiring XHP’s with a FET driver though. For instance your light won’t work right in the 3xbattery mode using XHP’s and a DD FET. Three batteries run into a single XHP or a group of XHP’s parallel is too much and will kill them. So if you do go with XHP’s you have to hope that modding that driver will actually work with XHP’s and I wouldn’t count on it. The difference is that the XHP’s are 6volts and I don’t know how that driver will react. From my experience using a wide range of stock drivers to attempt to build 3xMT-G2 lights I wouldn’t hold out much hope. I think ComfyChair might be the only guy I know who managed to get a Stock driver to work for any length of time with 3xMT-G2 (also a 6volt emitter). I tried at least a dozen different drivers and they all died after a short while. It wasn’t until we started building these FET DD drivers that people started having any real luck building triples with these 6volt emitters. I know a bunch of guys who tried it who failed.
I built a Triple XHP70 diving light the other night and used the FET driver but that was because the diving light uses 2x26650 and I could run all three emitters parallel. If it was a three battery light it would not have worked and I don’t know of any stock driver that will reliably drive 3xXHP70 although there must be because they sell a few triple XHP70 lights. I haven’t had a chance to tear one apart yet and see what they did. For all I know they might just be using a FET driver too.
So thinnest possible layer of Arctic Alumina, and both surfaces must be smooth as possible. Got it! I guess the J20’s shouldn’t be needing any sanding down then?
So basically i’d be better off just doing a resistor mod if sticking with the stock driver, and if i want to do any 6V or 12V emitters i’d be better off using another driver, is that what you’re saying?
In the J20, since it by default uses 3 batteries in series, unless i were using single 12V LEDs, 2 6V or 4 3V emitters in each parallel string, it wouldn’t work, correct?
Thanks! By the way, would you happen to know anything about this driver on eBay? I’m not sure if direct links are allowed so i’m just pasting the full listing title.
You can post direct links here no problem. Affiliate links where you make money from sales are different but I’m not even sure where that stands now. Even those were fine for the longest time and I they were debating changing the rules but I don’t know if they ever did.
And that is the driver that I was talking about. That’s one of the ones that I build and it was developed and tested by the guys right here on this forum. I call it the BLF FET driver because it really is a BLF driver. It’s the result of a ton of work by the modding crew here. Richard at mountain electronics sells them but Ventures of Blood who is selling the one on e-bay is also a great guy and a member in long standing here as well. In fact I think he’s been around here longer than Richard (BLF name RMM). That driver will easily handle the current you need and then some. It will give you the most power possible to each emitter without breaking a sweat.
That’s amazing, how does one get started on building a driver?
That’s good to know. Are there any disadvantages to a driver with more modes? I’m thinking of choosing one with memory, 5 modes (Moonlight to Turbo). As to the other options i see on the mtn website, i’m completely overwhelmed! No idea what to choose.
If i haven’t got it misunderstood, this driver can work with up to 3 batteries. So it can work with all 3,6 and 12V LEDs i assume?
By the way, how does a contact board + repalcement driver setup look like? Would you mind posting a picture of such a setup, because i can’t imagine what one looks like, or how i can fit that into the J20 at the moment? Thanks.
I simply stripped the existing driver and piggybacked in an FET with Zener mod, put 4 of the 9V variant MT-G2 on individual 20mm Noctigons then a Ledil reflector in it’s holder on top of each of those. I’m getting right at 15,000 lumens on 3 of the big Trustfire 32650’s.
To handle the heat, I put a 1/4” thick aluminum spacer on the emitter shelf and made a massive head filling heat sink to spread the heat more efficiently to the finned area.
Original switch is handling it fine.
So yeah, 3S cells to 4P 9V MT-G2 emitters. Definitely not a throwy light, but a very nice tint and just a huge amount of output.
Edit: With the cells rested and not fully charged I’m seeing 20.85A on the 3S cells at the tail end.
Five modes is nice. No particular advantage to any of them, really just personal preference. I like to keep them simple and don’t really go for all the hidden mode stuff, but lots of people seem to like that too.
These FET drivers will work with pretty much any amount of batteries and any emitters so long as you match the correct current to the right emitters.
I don’t have any pics but it’s fairly simple. The existing driver just ends up with two wires hanging from it with the new driver hanging just below it. You can “pot” the new driver. Which just means you encase it in something like a thermal epoxy and stick it to the old driver. This helps to keep the new driver running cooler. But you don’t have to do that part really.
The finned head section is hollow, a retaining ring holds the driver in from inside the head instead of below… there are 4 screws originally through the emitter shelf into the reflector, so I machined the inside heat sink to have a hole to fit around the holes and allow passage of the wires through to the emitters, the base doesn’t screw all the way into the head by about a 1/4” so the heat sink sits on top of the base such that when the parts are screwed back together the heat sink is pressed up against the bottom of the emitter shelf. I flared the hole in the sink out as it drops down to the driver so it clears everything easily and directs heat outward.
The first picture shows the additional 1/4” thick aluminum plate under the emitters. I used a sheet of copper to join all the positives on the mcpcbs with an 18ga lead soldered into the sheet. 18ga leads leapfrog around the perimeter to join all the negatives.
The light is rather large, so this heat sink is bigger than the pictures make it appear.
That sounds insanely awesome! Must look one hell of a lightsaber haha. Just one thing though, is it then impossible to do 6V XHP’s with the piggybacked driver in the J20? If so, is that because it is direct drive without a voltage step up/down function? Sorry if this is a stupid question by the way.
Would copper do just as fine? I mean it does conduct heat much quicker, although it also dissipates much slower than aluminum, from what i’ve read. If i were to run the light at turbo for a few minutes would the material make a difference?
Damn, that makes me want to learn something about CNC machining! The things i’d do with it… maybe go overkill and do a 10” LED spotlight! But i guess by then there wouldn’t be much sense having the relatively minute XHP35HI, but rather an HID setup…
…. you know if you hadn’t told me that i’d have done it the exact opposite way!
Is there a good thermal epoxy at a reasonable price? I reckon i’ll need that if i’m going to go all out on this. And the fact that i haven’t got any CNC lathes to mill some custom heatsinks or whatnot
You can see the white epoxy potting my little 17mm FET driver to the big contact board, on the left in the second and third pictures. I figured with this big heavy light it’s inevitable that it falls or gets dropped so I potted it for durability.
Since I am using 3 cells I wanted to match the emitter voltage. Pushing the 9V MT-G2 pretty hard it has about a 11.3V forward voltage so it takes the 3 cell configuration nicely. I’m doing the same in a BTU Shocker as well, with 3 of the same emitter and using 18650’s in the beefed up carrier.
You could run 4 6V emitters, 2S2P and they should do fine. There’s all sorts of ways to configure a light this size of course. Richard replaced the XM-L’s with XHP-50’s and I made him an extension tube for a 4th cell along with the big heat sink. He got 39,000 lumens and it would run with the sink to kill cells. Initially, without the sink, it was too hot in about 30 seconds. So yeah, there’s all sorts of things you can do with this huge beefy light.
Copper works better but is so much heavier. In a light this size, already quite heavy, copper would make it a brute! And I do love me some copper, so I’m not saying that out of hand. I build a triple XHP-50.2 from scratch and the copper pill was 3.17 pounds when I started out.
Was looking for 15,000 lumens out of this one, it “only” does 11,833.
Definitely Some works of art being made out of the TR-J20! I’ve made it into my new project light. Started upgrading the mosfets but I botched my first board by lifting a pad. Got a replacement driver from Trustfire and it looks like the new ones have 6 mosfets instead of 5. With one of them being stacked. Replaced all 6 of the Mosfets with IRLR8726TRLPBF from MTN electronics. Did the resistor mod and it already seems brighter than before.
Trying to get as much power as I can get out of the stock driver and running 12 XHP50.2 I’m looking for some good output while still keeping it efficient.
