Am I the only one who wants to run an XHP with a single cell!?!?!?

Looks great! How did it turn out?

Thanks for all the comments. I guess I’m having BLF post issues… I checked on this thread for a couple of days after I posted it and it showed I didn’t have any comments. Now today, (3 weeks later) I get a notice I have a new comment, then I check it out and see that there were several comments, some made the same day I posted… go figure.

No, I’m afraid I can’t boast of those skills, but I was hoping to somhow make that driver fit by maybe somehow modifying the pill… I still have relatively little experience doing flashlight mods but I’m trying new things often.

I modified a Uniquefire 1405 with a fet DD and noctigon XHP70 from Mountain Electronics and it SHINES!! It’s brighter then my car’s headlights! I figure a Uniquefire 1504 with high drain 26650 and an XHP50 properly driven would outshine the MT-G2 in the Shadow JM35, plus it would give me the zoom that the Shadow is missing.

Right now I have the 1504 running an XP-L HI with an Astrolux S1 A17DD-L and it’s a laser beam! In fact, I’ve started swapping all of my XM-L2’s with XP-L HI’s at 4000k. Nothing better than a factory De-dome.

I’ve been researching and realizing the “size as the limiting factor” aspect of a 17mm boost driver.

EDIT: B.T.W., will the DQG 26650 ever be sold again? I would like one of those.

Way back when, those “3-18V buck/boost” drivers would not boost with only one cell. They would however run a 3V LED from one cell. The boost part only started working if used with two or more cells. Have they changed them recently?

As hobbyist we can reverse engineer and shouldn’t worry about getting sued unless it’s SF. We have really nice list forming up now:
Acebeam H10 MT-G2
Fenix PD40 MT-G2
Shadow JM35 MT-G2 and XHP-50
Zebralight SC600 Mk III XHP-35
Zebralight SC63 XHP-35
Zebralight SC600Fd III XHP-50
Armytek Viking XHP-35
Olight L5 Pro led ?

Not listed is the DQG triple because it just didn’t quite get there.

What we have here is eight lights from 6 companies. Probably looking at least 6 board designs we can cross reference the %#$ out of them until BLF gets the prize. I know we want that Holy Grail the 17mm boards, but realistically only a few 18650’s can handle this. We are probably firmly in 26650-26700 battery territory so we now have about 24mm’s of area to work with. Of the lights listed I’ve got the Fenix PD40 and she is sealed up tight with glue. Vinh has already taken apart that and the Zebralight SC600 Mk III http://www.candlepowerforums.com/vb/showthread.php?415443-SC600vn-Best-18650-Light-Ever My personal short list is the Olight L5Pro, Armytek Viking, and the Zebralight SC600MkIII.

You’ll need to desolder one of the emitter wires to measure led current so that you can find out what kind of efficiency the driver has and what output. With a boost driver tail cap current doesn’t tell you much other than how fast the cell is draining.

The stock output is only 1.5A on turbo. With the stock XPG2’s combined series vF that’s about 14.5-15 watts. Doubling the output current and asking it to do 30W output results in a small fire.

It uses limiting resistors to set the modes. Four levels, four output channels, turned on one at a time to get L/M/H/T. Low uses some weird setup different from the others which we never fully figured out. Mid is controlled by the 2R0 resistor, high by the R500, turbo by the R120. Decreasing any of those resistors gives higher output on that resistor’s mode.

Yes Comfy, but the measured 1.48A I just pulled with my clamp meter at the emitter was also at 9.71Vf.
This would be very decent running a 9V MT-G2 at that 1.48A. :wink:

Edit: A spec sheet equivalent of ~1656 lumens at 5000K on the Q0 9V. Note an Vf of 9.25 at 2A.

Minor semantic quibble, but IIRC these are actually current sense resistors & not limiting resistors. Limiting resistors are used like the one shown here. - the resistors used in that DQG 26650 driver are used to produce a small but measurable voltage drop from which the driver determines current by ohms law. Then the driver adjusts it’s output to reach the “set current” (eg it’s really just trying to match a specific voltage drop across the resistors).

EDIT: oops, meant to respond to this too.

As I mentioned in the other thread we need very high resolution pictures. If someone who’s friendly with Vinh could ask for very high resolution shots showing component markings that would be good! Providing clear, high resolution pictures of any of those flashlights’ drivers would potentially be useful. (With the goal being legible markings as mentioned.) Failing that - decent photos with accurate written lists of component markings can also get the job done.

Yes, lower output voltage makes the boost driver’s life much easier. Anything you can do to get Vout closer to Vin helps (or, lower Vout will let the driver tolerate higher output current - it’s the output power that determines how hard the input side of the driver has to work).

What LED are you testing with that has that 9.71 vF? Have you measured input volts (under load) & current, to compare with the output side, to get an idea of efficiency?

No, it really doesn’t. There’s no sensing or feedback anywhere in there (except possibly in the ‘low’ mode channel, it’s not like the others). The MCU-thing sends battery voltage to a gate pin, FET turns on, current flows from LED- to resistor to FET to GND.

The Tiny DQG 26650 has 3 XP-G2 emitters wired in series. Hence, it pulls 9.71V at the emitter wires on the mcpcb. Likewise, with a loop soldered in between the negative lead and pad on the board, it pulled 1.48A from an Efest Purple 3500mAh cell rested at 4.20V.

I didn’t measure voltage at the tail under load, but already listed current draw previously in this thread I believe. Do I really want to know this? :wink:

What input current to get that ~15W output Dale? Still the same 6A input? Generally boost circuits are most efficient when the percentage voltage differential is minimized which is easier to do with more LEDs and more cells, at least initially. As the battery drains the voltage differential gets even worse with multiple cells but the percentage differential is never worse than with a single cell and a single led.

I was asking because I wanted to know, and the input power is something I don’t think I ever got around to measuring.

You have a really special driver there! Around 6 watts input power to get 14.5 watts output! Scale it up and the planet’s energy problems disappear in a stroke.

I get 4.31A at the tail on the Purple Efest, but can’t get a voltage reading under load. This battery tube is a one piece design, I don’t have enough hands to loop a lead from retaining ring to battery neg AND hold DMM probes for a voltage read. Tried.

I don’t understand what I’m seeing. I got my 8 yr old to touch the probes of the DMM to the cell while I held a 12ga solid lead to positive and a 12ga Turnigy lead to negative.

Before I turned it on, the DMM showed 4.08V. Turning it on in Turbo showed me 1.71V? I don’t get it? Can it really be knocking this fat 26650 down that hard?

Edit: When I took measurements the other day and posted current draw and such, it was on my JM26 with XHP-70 and the DQG Tiny boost driver. The measurements I’m taking today are actually IN a DQG Tiny triple XP-G2.

Thanks comfy, my mistake. It’s been a while since we looked at that thing. If I had taken the time to read the resistor values you listed I would have realized that you were correct!

Actually I’m still waiting on a delivery of switches, which makes me wonder if I should be opening a dispute soon. I’ll have to check the order date.

I rigged up new leads on my DMM tonight and took some measurements. Sony 50a 26650 = 4.04v
Tail = 3.95a
Led = 1.61a

One thing I did notice about this driver is on low mode it flashes (assuming) to indicate low voltage but does not flash on the other modes???

If it flashes a single time in low that’s probably the indicator for switching mode groups.

The driver you linked to is recommending 2 or more batts. Did you do something to it to make it run on one?