Pic is coming, my phone battery died.
Hyprmtr, the original resistors stay in place.
Werner, the tailcap current will be lower on 4x26650, probably around 4A. Total input power is right around 60W regardless of three or four cells.
I have not tested for the existence of a low battery warn/cutoff yet. I probably should have did that before I modded it 
Edit: tested the tailcap current on 4x26650: 3.97A. Pretty close to my guesstimate above. OTF virtually the same at 5140 lumens.
Edit2: total weight loaded with 4x26650: 1342g J)
Added a pic of the resistor mod to the first comment. This one, showing the tenting of two R22 resistors.
Beautiful pic! I made my order with digikey yesterday. I ordered both resistors.
As soon as I made my online order, I called them to see if I could get the parts overnighted so I couls have the weekend to work on it. Ive got to say, the people over there at digikey are the most polite customer service people I have ever delt with! It was a pleasure.
Thanks agian, Relic and vestureofblood, for being the same!
Ken
Glad you could join us Ken.
Thanks again for another very useful thread Relic.
Relic, do you suppose similar mod can be made on the Ak-47, the flashlight :bigsmile: , by lowering similar resistor ?
I don’t remember a post on this mod for ak47 …
By changing the whole driver board is a $20 venture… I have some five of them to mod for people. So I thought a bag of chip resistors may be a better bet, assuming longevity is preserved.
Hi BarrelMount, I’m not too familiar with the AK-47. There is an extensive thread on here about that light, specifically to do with modding it after breaking it.
To resistor mod the existing driver should be fairly easy, but I’m not sure how much farther that driver can be pushed. For lights you plan to give to others, resistor mods can potentially cause you some grief down the road in the form of “hey that big light died, can you fix it again?”. Or worse.
If you want to try a resistor mod, take some pics of the driver and we can see about trying something you should test thoroughly before handing out.
Yes, but the thread was about changing the driver, not modifying it.
Good point on the mod for others. That’s wisdom! Thx.
I have two J18 and mod’d both. Just now, I pull the old OEM drivers out and look at them. Wow, the third resistor is not populated on one board while the other board has that resistor pulled off to fuse to a transistor (Q1). Chinese mfr, huh?
That points out the need to look at several driver boards. I’ll do that tomorrow and perhaps start a new thread. Thx a bunch!
The J18 board is identical to the ak47 board, except for the resistors. Based on the two J18s, they don’t use the third resistor spot leaving it blank. The ak47 board uses that spot but, as I remember, missing the first resistor R500. More later. Thx.
What happen if use only one R22 resistor?
Most likely more current than the driver can handle.
Also, that R22 may need to be bigger than 1/2 Watt, as it will get most of the current.
It would be nice to have a box of multiple value sub-ohm chip resistors, say 200 of them.
Relic, I think it’s Q4 transistor that will really heat up. That one needs heat sink.
This is what I’m curious about. How can I tell which transistor is which? Also, is there a way I can test and/or identify exactly what kind of transistor each one is?
Ken
The markings are etched off. There are two MOSFETS and two diodes that get hot. Those are the ones I added the heatsinking to.
Q4 does not get very hot (that’s the mode controlling MOSFET, PWM generator). Maybe 60C or so. The others will hit 100C in a minute or less with no heatsink.
Hard to tell as we don’t know the stock resistance of the stacked up resistors…
If it are two r620 stacked you would lower the total resistance as follows
1÷(1÷0.62+1÷0.62+1÷0.62+1÷0.62+1÷1.5)=0.14
1÷(1÷0.62+1÷0.62+1÷0.62+1÷0.62+1÷1.5+1÷0.44)=0.106
1÷(1÷0.62+1÷0.62+1÷0.62+1÷0.62+1÷1.5+1÷0.22)=0.086
And this would increase current…
I for sure would measure resistance and then calculate how much current flows through every resistor before I would choose which resistor to use…
That should kick it up to near 6.9amp, which is quite toasty!
R22 —-> 6.9amp Requiv=0.12 ohm
R44 —-> 5.2amp Requiv=0.16 ohm
R50 —-> 4.9amp Requiv=0.17 ohm
Open —-> 3.2amp Requiv=0.26 ohm
Assuming linearity… close enough.
3-stage cascading current source amplifier ? Where Q4, the lowest current, is the initial stage?
All these boards must have come from one basic circuitry, and they just mod the values to fit.
Werner, there are not stacked resistors, did I forget to go back and change that? OK, fixed now.
You can calculate the % change, but based on my use of 0.044 Ohms giving a 50% increase in tailcap current, I would expect around 7A+. I would not expect the driver to handle it for very long.
Oops…
Then I totally confirm if stock 3.2 then with one R22 6.9A and a small resistor will be absolutly cooked as more than the half current has to flow through it…
I had great success with the “tented” resistor mod! It “was” really bright! Yes I did say “was”. Haha! One of my heat sinks came loose and shorted out on the driver. Needless to say, I will never trust the double sided tape that comes on the heat sinks agian. I did see that one of the transistors had blown out. Another one looked burned too. The numbers are all gone on top of them so there really was no way of telling what kind they were. I did swap the burned ones out with some from and old PC I had. It does work in all modes, but very dim.
So, ultimate mod coming up. I will install (12) XM-L2 U2 leds. Driven by HyperBoost. Hopefully, will be completed within a week. Just waiting on parts!
Ken
That’s too bad Ken. I used Fujik on mine, which I trust a bit more. Actually it holds quite well once cured. The MOSFETs are probably easily identified from other drivers that do not have the marks removed. I bet they use the same parts on most of the drivers.
Here’s one idea that I’ve had, and almost tried. Maybe crazy, but if you rewire the emitters for 3S4P, you can direct drive them from 3x26650.
Even crazier: If you rewire them for 4S3P you may be able to direct drive them from 4x26650.
I have no idea what the emitter current would be, but I suspect a lot (especially the 4x26650 option). In fact it may result in more frying J)
The only reason I didn’t do the 3x26650 DD option was the resistor modded driver got results over 5000 lumens.
The 4S3P interests me. I might play around with that tonight.
I plan on wiring the new U2 leds in a 12S1P configuration with the Hyperboost. I will run 4 cells at 1600ma to start. I might run 1000ma and check tail cap current first though. Any other suggestions with my planned mod?
Ken