Have you hooked it up to a XM-L2? MT-G2? Any idea what it puts out on a single cell? Double cell? Need information! 
Thanks Comfy, appreciate you looking into this one, will probably be using it in the near future…
I think RMM has already done some testing and found it to be pretty constant as a buck driver over a range of cells. As with all dd drivers once parasitic resistance is accounted for, output depends mostly on the cells and arrangement.
Edit - 
What am I thinking? A buck driver is anything but DD.
Next-mode memory?? Man! Well, it does reset back to low but it takes way too long, around 6-8 seconds. I'll have to poke around, which of the two pairs of capacitors are controlling the reset time isn't obvious yet; anybody else see something I'm overlooking?
As far as I can tell the stars don't do anything at all.
I did just a quick temporary test to see if it worked, didn't measure anything. But with 8.4v and a single XML2 the LED wasn't destroyed, so it's definitely not direct drive like the old version was.
It is not direct drive, it is definitely a buck driver. In stock form my pulls about 1.8 amps from two cells going into an XM-L2, also works with three cells (reduced current draw by approximately 1/3 as expected). I haven't tried four. It also works with an MT-G2 with both two and three cells but I haven't tried cranking it up yet with that one (too expensive to waste!).
I tried to jumper over the sense resistors and in about half a second it burned out my XM-L (was half expecting it to). Hooked up a trimpot and cranked it up to 2.65A at the input with two cells going into an XM-L2... held it there for three minutes and nothing burned up. The current stayed fairly stable the whole way through, rising up to around 2.69A by the end of the test. I was still able to hold onto the driver board the whole way through, although it was getting warm. At right around three minutes the driver stepped back down to high from turbo.
This looks very promising! I don't have more time tonight to play with it but I am going to try for 3A at the input side with two cells... around 6 amps going into an XM-L2 with some regulation is a thing of beauty.
Indeed RMM! A thing of beauty for sure, and thanks for putting this through the ringer. Watching diligently…
zeremefico has tried to do resistor mod on the new DRY driver and get good direct drive result. Think I am going to post the link over here to concentrate all the related threads together.
Single XML2
1 cell = NO WORKY! :(
2S (8.27v)
L = 0.19A
M = 1.07A
H = 2.47A
T = 3.49A
3S (12.41v)
L = 0.19A
M = 1.06A
H = 2.46A
T = 3.48A
4S (16.60v)
L = 0.20A
M = 1.07A
H = 2.47A
T = 3.50A
It really needs resistor mod. Since when 3.50A is “low” for us? :bigsmile:
Sure, 3.5A is low for a single LED, but if it does the same 3.5A into 2 LEDs in series it's not bad. And if it still does 3.5A into 3 LEDs in series...
In this thread, an MT-G2 was run with 3S by one member and half fried. Another did the same, but for only like 20 seconds and his emitter survived. So is the verdict that this driver does not buck well with the sensor bank shorted?
Buck drivers that freak out and connect the full input voltage to the outputs when the sense resistors are shorted isn't uncommon. https://budgetlightforum.com/t/-/20635
It's likely that the MTG2 can survive 12v in the lower modes, when the switching frequency is high and duty cycle (on time vs. off time) is low (~30-40%). LEDs with little bond wires that just act like fast-blow fuses are less tolerant. MTG2s are tough SOBs.
What you say Comfy makes sense to me. Your 4S to xml2 measurements above say alot too.
With 3 series XMLs and 3 series cells, shorting the resistors should make it act like the older driver. If one LED won't fry when direct drive off one cell, 3 won't fry when run direct by 3 cells. Knowwhatimean?
But for other LED setups (1x MTG2 with 3 cells, 3x XML with 4 cells), it'll need to remain in buck mode, which means no shorting across the sense resistors. I don't know if it has a falls-off-the-cliff point as the resistance is reduced like the S1100 driver, it'll need some careful experimentation to see how far it can be pushed.
I hear you. Like you said before the last post, (paraphrasing) the jury is still out on this driver.
The old version would not buck in 'turbo', this new one is much more flexible. I still have no idea what the stars are for. |(
I checked the circuit of the photo with the QX5241 data sheet and found that the right side of the board was the same as the standard circuit on the data sheet. The uC should be Microchip uC or similar which Pin 4 is Vcc. I wonder the wiring of the FET at the left side which look like not be wired parallel with the right one.
FET on the left has the drain going to ground, one on the right goes to LED-. But they're in series - source for the one on the right goes through the vias to the battery side, to the other set of vias at the left FET's source pin.
LED- > drain > source > source > drain > ground
specs from taobao.com
PCB diameter 32MM, double panel! Thickness 1.6MM!
On the circuit to do a lot of cooling holes! Dissipate heat better! Positive MSO tube heat dissipated by the inductor!
With reverse polarity protection! Output open! Short circuit protection! LED does not take damage!
Positive and negative output LED does not come together to burn! Barbara's multi-protection circuits!
LED: the single T6, U2, U3, MTG, ST-50, ST-90. 3 12V can concatenate strings T6, U2.
Dimming circuit with four groups! Especially for DIY! (Two five-pointed star shorted together to achieve the appropriate exchange group)
Default: 100% 30% 5% Baoshan SOS
A group of 5% 30% 100%
Group 2 100% 30% Strobe
Group 3 100% 50%
No matter which gear closed in around 5 seconds! Automatically reverts to the first gear!
Because a large board area, power walking fast heat bulk! Work to stabilize! 2.8-3A current is definitely not a small circuit board can be stability ratio.
Current default :2.8-3A, ensure stable! After a long test!
I just installed the new Dry driver in a Tangsfire 3XU2
The emitters are XM-L2 U2 on Noctigons, and everything is wired with 18AWG wire.
With three fully charged KingKong ICR26650 I’m getting 3.4A at the tail in turbo. I was hoping for a little better. I’d really like to see 4A. I’m considering trying the old 20mm Dry driver to see what it’ll give me. Is that driver known for Squealing.
I’ll charge up my E-Moli cells to see if there’s any improvement. Unfortunately the cells are so long I can’t even engage the first thread of the tailcap.