Review: DRY 3*XM-L 3*18650 4-mode neutral white

I got myself the 4-mode DRY 3*XM-L in neutral white; 97$ incl shipping from CN Quality Goods. And I'm a bit disappointed.

PRO: Bright. Compact. Good machining, lubed. Comes with Spare parts.

CON: Unhappy with the driver: Not as bright as I hoped and with noticeable PWM, even on high. Waste of diameter. Slight donut.

(More images in this thread.)

Well, yes, it is brighter (flux) than any other light I have - it's my first triple XM-L. Still, I expected more. Just did some measurements: It's (in terms of flux/total brightness)

  • 30% brighter than my 5*XP-G build (~800mA each)
  • 55% brighter than my Fenix TK35
  • about twice as bright as good typical budget XM-L lights.

Spot intensity is 12kcd, that's a NEMA-throw of 220m (note that NEMA specifies a rather generous definition of throw, realistic values are lower.)

The currents on freshly charged batteries (protected) are 2.15A / 0.86A / 0.22A, going down to 1.8A after a few minutes of use.

I haven't yet opened it up completely and taken a look at the driver, but it seems it is a serial direct drive with PWM to get the average current down - even on high mode! The modes are

  • Low, 5% PWM, 260 Hz (260 and 206 are no typo)
  • Medium, 20%, PWM 206 Hz
  • High, 50% PWM, 206 Hz
  • Strobe, 50% PWM, 10.3 Hz

So on high the ~4A for direct drive are reduced to ~2A average by (noticeable) PWM, not really a good way of current regulation. A real constant 2A would be significantly more efficient, i.e. yield more light.
Relative to high mode, the levels are 10%/40%/100%/100% Strobe actually looks brighter than high. There's no mode memory, if you switch it off for 4-5 seconds, it'll start at low again. That's a useable setup, but may be a matter of taste.

I clearly do not like the driver. There's a 3-mode driver version, specced 4A (just direct drive without PWM on high I guess), and it seems there will be a 5-mode version soon with L-M-H-Strobe-Turbo - I guess that just adds a 'no-pwm-direct-drive' @ ~4A. I hope I can get one of those drivers, even though that isn't my idea of a good driver either.
I am thinking about modding it to a parallel setup with a 24*AMC7135 driver instead.

The body is nice, well machined, and quite compact for a 3*18650 (15cm long). I wouldn't want to carry a light with 3*18650 in a row... Tube diameter is 47mm, maybe not that comfortable for small hands. Threads were lubed. Since the tailcap protrudes a bit, it doesn't tailstand completely stable, but stable enough.
For that price an AR coated front glass would have been nice. I just ordered one for 3$.

Included are a spare clicky and spare o-ring, a GITD tailcap to exchange with the default black one, and a nice lanyard.

The head diameter is 58mm, but the actual parabolic reflector areas do not really use that diameter well; much of the front area and thus some throw capability is wasted.

It does get hot on high, but not as much as the 4A version I suppose.

The tint is neutral as expected; I'd have preferred a slightly cooler neutral tone, but that's a matter of taste. However rendering of natural colors is much better with neutral white; in cool while most things look much more pale (though blue colors look more intense with the cool white lights).

The beam is relatively wide and floody as expected, with a slight donut effect.

UPDATE

I took a look at the driver; as I expected: It's just a big FET, PWM-controlled by a mode controller IC. Unfortunately not an ATtiny.

The front glass is 2.5mm thick. My AR-coated 2.0mm thick replacement rattles :(

Update 2

Got the new Turbo driver. All modes 207 Hz now,

  • Low, 4.3%
  • Medium, 20%
  • High, 50%
  • Turbo, 90% (so still PWM... but barely noticeable at 90%)

And with mode memory now... There's an EEPROM chip on the upper PCB (IC1). I didn't get a battery contact board with it (even though I ordered one) and found soldering that one somewhat unpleasant.

Update 3

Did some measurements again with the new turbo driver:
Spot intensity 20 kcd, (NEMA-)throw 280m.
Modes are 6%/28%/68%/100%(turbo) with some estimated 2300lm on turbo.

There's the possibility that this host has been used for a Triple Q5. http://item.taobao.com/item.htm?id=10849356995&ad_id=&am_id=&cm_id=&pm_id=

I'm not completely sure but I think 2100 measured 2.2A on his 4 Mode? He asked RIC about it and it was told to him that the current draw will improve with unprotected cells.

My question, shouldn't a good driver (supposedly regulated @ 2.8A-3A) draw the same current, no matter the cells, protected or unprotected?

The taobao seller describes something interesting here, :

http://item.taobao.com/item.htm?id=12200961733&ad_id=&am_id=&cm_id=&pm_id=

Not a current draw of 2.8A-3A, but of 2.5A-2.8A. And that the runtime on max is a half-hour using Sanyo 2600mAh. Isn't that looking like a inefficient driver?

I clearly wouldn't want to get 2.5A or 3A that way. That meant 5A or 6A with 50% PWM...

Yes, a better driver would really limit the current to, say, 2.8A - as long as the battery can provide it. This is what linear drivers (like those 8*AMC7135) do. And that's why I'm thinking about a 3*8*AMC7135 mod.

Have you thought about this solution?

You may need to add a resistor in line as the XML has a much lower Vf than the p7. This was for a triple emitter mag but applies to the DRY triple too.
(this comes from CPF user “download”)

Excellent review DrJones. Thanks. I suspect most if not all 4-mode buyers are going to want that 5-mode driver. I certainly want it. About the 4-5 seconds, is that the minimum amount of time you have to leave the light off for it to forget its last mode? That's a little long for my liking. I prefer 1.5 - 2 s.

Thank you for the review!

That driver is certainly a joke. For that price it starts to look like nothing special. The manufacturing costs for that body can't be high, really.

Where have you found one?

I just made some measurements and added them into the 1st post.

@Ford Prefect: Yes, I know this setup, but didn't think of a resistor for an additional voltage drop. Good idea. That drop will become unwanted when the battery voltage goes down though. Hm, I'll think a bit more about that.

@peteybaby: Yes.

@ Hikelite: KD

Aaaah, too bad this has that bad PWM :|

Thanks for the review!!

Greetings DrJones,

Thank you for the review. Sorry to hear the driver doesn't live up to the potential it could have been.

Does anyone know if there is a prefered driver by KD, DX, MF, etc that could be used and physically fit without modifications? Looking to buy this but would prefer a more efficient driver or one that can sustain higher lumen output for a longer time as DrJones suggests. I might go the route of just buying the host and utilizing a driver that someone can recommend.

Best Regards.

Thanks for the review Dr Jones! I've been wanting to read a detailed review on this one, I appreciate you doing it.

Frontpage'd and Sticky'd.

I see you have updated the current readings. :(

Thanks for the review.

Honestly, there's nothing worse than a product failing to meet expectations. Especially a product that costs twice as much as the rest of the 3x XM-L lights. Which btw, aren't exactly budget in the first place. ($50+)

You would think that at this price point, something as vital as the driver would be top notch.

Your bad news is my good news, so thanks for taking one for the team!

If you look at the spreadsheet, you'll see that this Dry is $250 cheaper than any other 3x 18650 side-by-side triple XM-L, at the moment at least. Pocketability isn't an issue for some people but it is for me, so the Dry is currently my only choice.

I do not think that the ability of having 3x 18650 side by side justifies the price of those flashlights neither the DRY's. Technically it's nothing special, and with nothing more complex than a 3 x 18650 end-to-end.

What improvements could be brought to the DRY right now :

1. Serious Driver

2. Serious battery holder

What else could it be? Another front glass lens, AR coated, and some sort of knurling, my hands sweat on this body and it slips off my hand.

I have already reported pretty long ago what the 4-mode can/cannot do with measurements etc.... No big surprises bros... LOL!

Something like NW 4-mode is 168 lux vs SR3800 160.... and since it's not current regulated it slips. I also did mention countless times that the DD version also slips, so after it drops till about 3.9V when the triple 18650s starts to rock and roll on the plateau, it is in the 3 plus amps region maybe still close to 4A, so ready to rock and roll in cool winter countries. Tested with ice somemore etc....

Anyway those who are interested in a cool similar product can try the Jetbeam RRT- triple XM-L, its usd275 from HKE shipped, good for international customers. Not as discounted as the TK70, nevertheless just pop 3 x 18650 inside and off you go. Guarantee, no PWM and many many modes for you to try on the nice ring. :D

BTW, the TK70 can use the Kaidomain's 32600s. No issue, tested like 4 hours in Turbo (but split into like 10-15 min runs). Also stress tested it On and Off many times. The driver can take it, and like many others suspected, the drivers are the same as TK70S...... just that they give you an additional TK50 tube. If you are still skeptical, just try the $20/pr Ultrafire LiFePO4 32600s. So that's another option.

The TK70 does about 220 lux ceiling reflected, and DRY about 245-250 lux. Cool countries would be even higher, i tested with ice (freezing point) and it was 290. I'll try to get the UCL lens figures for you guys soon..... (busy with testing the TK70 with Li-ion)

BTW, throw for CW 3-mode is about 400 for 8.17m = 26k. It outthrows my UF-980L at 19k measured on the same meter.

Just took 3 weddings in 3 consecutive days....recovering. This morning's wedding started at 4am, and just ended at 4.30pm. LOL! Zzzzz....

Seriously if you ask me to try the Sky Ray lottery or paying more for the DRYs, i'd choose the latter.

Seriously there are absolutely none NW/WW triple XM-Ls. If you can DIY and afford the time to R&D + CNC a 3 x 18650 side by side, then you are good and more power to ya...... I'll say it in a bad way even as an end-consumer, no seller in a right frame of mind would sell the DRY at $45 because it is "nothing technically special". Remember, you can only get the DRY NW now at a higher price at SB plus international shipping via USPS First Class . Maybe we shall wait for the next emitter....

Remember, you have support, plus the driver is just a meagre 5 bucks to buy. I couldn't even get MF to send me a SR3800 driver (LOL!), i just hooked it up to a 13.6V open circuit full charged SLA and let her rip with DRY ICE. (Later then "R&D" and DD it with 3 x 18650, because the SLA did blow the LEDs twice well it just glowed a wee bit like 5 lumens, and twice i just knocked it hard and it came back on. Welcome to Sky Ray lottery, some spent $110 and never got a good flashlight, third time might be lucky and you finally get a working 1600L OTF light for $160+ gone, might as well just spend it on a TK70 2200L ANSI OTF in the first place)

Since this is truly budgetlights forum, I shall think of it this way for you budgetholics.....

SR3800 is $52 and is playing roulette plus more or less it would flame out. Driver support is nil (forget about the TF 3T6 driver, read my comments on the SR3800 review thread). 160 lux vs DRY's 250. 160/250 = 0.64. $52/0.64 = $81 if you wanna reach DRY's CW 3-mode output. Seriously I don't see any problemo.... But please, do not gun the light for very long in "turbo DD", there is nothing "long runtime" about the DRY in DD. Ric very open about this and is specifically noted in the website. This is probably a bit better in heat levels between this and a Uniquefire UF-2100. (that one heats up really fast due to little mass for heatsinking)....

I thought this too after my Sky Ray 818 disaster.. But well.. My Dry came DOA.. 2 loose solder points and a "nonworking isolator" on top... after fixing, its working fine.. Seems all chinese lamps are lottery with no exception :/ But better crap assembly what I can fix, than a crap driver that dies early...

True, but it's useful to see DrJones' test results showing how the driver works, and showing that on High, the 4-mode is run at 50% duty cycle, and that on all modes, the LEDs are driven at 4A, which is inefficient. Those are all minuses, but there's not yet any competition with a similar form factor and price. BTW, DrJones, did you use a scope to do your measurements? I just bought a $50 DMM a few days ago, but I was considering a handheld scopemeter until I saw the prices.

Poor you, 2100. I photographed exactly two weddings before I gave up because of the stress.

HikeLite, you're probably right that the price is too high for the technology used in the light, but there's not much competition yet with that form factor. What I told myself is that I used to spend $60-70 on little Fenix lights with < 200 lm. If I can spend $85 (+shipping) on 1500 lm or more, that's a good deal to me anyway. And it's still pocketable.

Totally agree with you on your recommended improvements.

@peteybaby: Yes, used a handheld scope :)