Question about TR-J20 modding

I’ve modded a few of these. Best results so far was going with a BLF driver, put the emitters on noctigons, de-dome half of them for some extra throw and a more neutral tint. Wire them 3s4p. I also put a side switch in one of them.

I don’t know if that driver would power XHP-70’s but I wouldn’t count on it and I wouldn’t see the advantage in doing so anyway. More floody I guess but you would get a nicer beam pattern with just going 12x XM-L2’s and the heat would be spread out a little better as well that way.

Was thinking along the lines of aspherics actually, but how to mount them securely inside is another issue.

The largest driver i’m aware of yet is the 46mm SRK driver, not sure if that’s the one you’re referring to?

Unfortunately side switch modding is not possible for me, too much expertise required out of a completely amateur haha. Does the stock driver put out a fixed 18V by the way, or is there a way around that?

I was assuming the complete opposite regards xhp70 vs xml2’s. Aren’t xhp70’s 4 xml’s clustered together? That would mean more luminance in a smaller area than 4 separate xml’s spaced apart, unless there’s something i’ve missed out? And it seems to me that XHP’s need 30W to put out 4022lm compared to XM-L’s which do 1040lm at 10W. Was in favor of XHP because of this greater efficiency.

Aspheric lenses would be pretty close to impossible to get to work on a light like this. You would need one for each emitter, they would each have to be perfectly positioned so they all come to the same point. They would not work with XHP’s at all. That would be if you wanted throw. If you wanted flood Aspherics could do that but you would lose a ton of lumens that way.

The size of the driver isn’t what’s important here, other than ease of fitting it in. What you need is a driver that will provide the correct current maxed out. The cheapest way is to mod the existing driver. I don’t know how many amps per emitter that will create, but it’s pretty good from what I remember, not close to max though.

If you want to max out the emitters and give them the most power they can handle then you can use this driver. You take everything off of the driver that comes with the light and use that wiped clean stock driver as a contact board and attach this smaller driver to it.

This will allow you to get up into the 5amps per emitter range with good batteries. Of course with that much current the emitters must be mounted onto noctigons or sinkpads or they will die a very quick death, not even seconds, more like a fast flash of light and gone.

XHP’s are four XM-L2’s but three of them in this light would be all flood. Why? Because although they are bunched together they still have the exact same surface area as four XM-L2’s and you would only have three small reflectors. If you put an XHP-70 behind a small reflector you have four times the surface area, none of which is even perfectly centered. (throw is calculated by surface intensity of the emitter by diameter of the reflector). Maximum throw is created by using the smallest emitter with the largest reflector. An XPH-70 is a large surface area and when you put it behind a small reflector it’s very floody with a large center hot spot.

Keep this in mind when using FET (direct drive) drivers. Each XM-L2 or XP-G2(3) or XP-L2 can handle all the current from 1 battery without frying, so long as it’s on a noctigon or sinkpad. So if you have three batteries in series (like on this light) then you can drive Three of any of these emitters in series and they can take it. Not Two, that would kill them and not Four it wouldn’t be very bright. (three batteries in series drives 3 emitters in series to their max (2 drives 2 and 4 drives 4, etc) So you have to match the number of emitters in series to the number of batteries in series. How do you get 12 emitters to work with three batteries? That’s were Parallel comes in. You do Four sets of Three emitters in Series. (3s4p) This essentially means that you are still driving each emitter with a single battery. That FET driver I posted is “direct drive” with low resistance. This means that it’s throwing pretty much all the power that the battery has at the emitter. Using this driver you are able to come close to maxing out each emitter. It’s a little bit short of max, but only because it’s next to impossible to use heavy enough wire to make that happen. So you end up using wires that are not as thick as you could and the thinner wires create some resistance of their own.

Now do you want to really go this far with this light? That’s up to you. It’s not exactly a “beginner” mod and it’s not cheap, you can be almost guaranteed that it will take more than a few tries and lots of troubleshooting to get it right. but it’s far from impossible and would be a heck of learning experience.

If you decide to go this route let me know and I’d be glad to talk you through it. Feel free to PM me anytime with questions.

Can anybody tell me the diameter and thickness of the glass lens in this light?

yeah, throw is the way i’m going, the awkward tri-beam at short ranges notwithstanding. I know the 70 isn’t the best platform to get throw out of at all, but at 4000+ lumens and with an aspheric focused to infinity, i don’t know, maybe that’ll do something? Would this make the case for XHP70’s any stronger?

Let me get this straight, because i’m not sure that i understand how it works. Using 2 batteries would fry only the 1st emitter in the string because more amps would be drawn out of the batteries to equate to the V=IR equation, and when using 4, fewer amps are drawn out? Also, do tailcap readings only reflect the amps drawn out of the battery the negative lead from the multimeter is contacting?

I remember reading a post about using the stock driver as a contact board( was it yours?), but don’t know how that’s done. Is it connecting both drivers by means of soldering wires to them?

Thanks, i greatly appreciate that, it’d be very helpful, but i wouldn’t try until i master the basics of electrical theory, which i completely lack now! (Got some R470’s on the way by the way)

I remember seeing a post on the thickness and width of the lens, but i can’t seem to find it at the moment. i’ll get mine out and measure it in a moment.

Diameter is 84mm and thickness is 2.75mm. Not AR-coated. Wish i knew where to get one on the cheap

You would have more than an awkward tri-beam. You would have three sets of four beams. The XHP-70 looks like it’s a single emitter but because it’s actually 4 XM-L2’s jammed together they are still separated by enough space that under an aspheric they will come out as four beams that never do intersect. So even if you could get the three emitters to intersect perfectly you would still have four beams. I imagine they would get pretty close eventually IF you could somehow manage to line them all up perfectly but it still wouldn’t be the most throw from the light because the XM-L2 is not the best throwing emitter under even the best conditions. I have built one multi-aspheric light and that was with 4xXP-G2’s, which throw further than XM-L2’s and it was still a waste of time because the lenses needed to be relatively small. Theoretically if you wanted to turn this light into an aspheric thrower your best bet would be to use a single de-domed XP-G2 S4, right smack in the middle and one huge aspheric lens. When you are going for max throw it’s the same with reflectors or lenses, the bigger the better and multiple small ones don’t cut it.

I’m not saying it would be a complete waste of time, it would make for a neat light but it would be an unbelievable amount of work creating the housings to get those three beams focuses and it would still not get you the most throw or a very nice looking beam.

On the batteries frying the emitters. If you put Two batteries and wire them to Two emitters in Series you are fine. Wire them to Two emitters parallel and it will fry both of them. I suspect that if you were to somehow manage to wire 12emitters all parallel to two batteries it might not fry them if you got the wires exactly perfect so that none of them got any more power than the others. This could create enough battery sag to save the emitters, but I wouldn’t count on being able to get the wires perfect and once one emitter fried the next would go and the next and the next.

And yes when you take a tail-cap reading you are only reading the draw from that end. A more accurate way is to read the current at each emitter. I am no expert on this part at all. Although I have built hundreds of lights I haven’t taken a current reading in years and I don’t think I ever did one correctly anyway. I know roughly what you get from each configuration and try not to post too much on exact amp draws or even exact lumens or KCD. I’ll do so in a general way but I stay away from making too specific of a claim. I give ballpark figures and that’s about it. People have been trying for years to come up with a perfect measurement system for these things and I leave that to others.

Yes, using the driver as a contact board means running wires from your old board to your new one and then on to the emitters. After you take all the stuff off the old board you use your DMM to isolate a positive and negative and solder the wires there.

I’m not even close to an electronics expert, others here are way past me on that part. However, even though I’m no expert on the electronics part, I’ve become fairly expert on the actually building of lights and even building and modding drivers. I guess what I’m saying is that you don’t need to know every detail of how the electronics works to build a great light. For some people here the electronics is the most fun, for others it’s the software design. For me it’s just the actual construction of the lights and seeing the final product.

Modding the existing driver is great way to start and good luck.

Oh yeah, that’d be painfully strange( strangely i’d had such a dream once about this), but i’d really been planning on a dual aspheric, a pre-collimator positioned just far enough to get all the light onto the main lens, which will be a 100mm FT lens, which is focused to infinity. That’d make it 3 circular beams. Unless i get the new XHP70.2’s?

I would go down the XPG way, but what i understand is that it’d be not particularly bright and have a very tiny spot, and at the distance it will throw to, one would need visual aids to see where it hits. My rationale behond XHP70’s (and the dual aspheric hence) is partly due to their massive output, so even though due to their large area they wouldn’t ever throw as far as an XPG, they’d still go quite a distance and light up a broad area, but could illuminate a varied area size by adjusting the main lens. If this idea i have in mind can achieve the sort of results i’m anticipating that’d be awesome.

Knowing how aluminum is a poorer conductor of heat but better at dissipating heat than copper, would it be necessary to maintain a minimal distance between the sinkpads, if i stick with XHP70’s? And are the Arctic Alumina grease and paste the best to secure the sinkpads to the heatsink shelf?

So it is because of the voltage drop across each LED in series as opposed to the full voltage from the batteries across each LED in parallel that makes the difference, is that correct?

By getting the wires perfect, do you mean using the correct gage and length?

Lastly, how do i use the DMM to isolate a positive and negative? Aren’t the + and -ve poles on the drivers marked already?

I’m not so sure i’d want to go as far as being an electronics expert myself, but i’d be more than content to reach your level of expertise. Inspired!

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.

Thanks!

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?

Yep that pretty much sums it up.

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.

“0-10+ Amp capable LED Flashlight driver board 17mm, Cree XHP70, XHP50, XML2, XPL”

The size of the driver doesn’t usually have anything to do with how much power it can handle does it?

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.