Review: BTU PK26

LED Flashlights – General Info 26650 Flashlights
1

Just wanted to get this out quick. I received the BTU PK26 light today for review from RIC. I already provided him some initial impressions and feedback via pm. Though it originally was supposed to arrive a few weeks earlier, there were all sorts of delays: normal production delays, the LED I requested did not become available as soon as expected, then the last issue was stuck in customs for a few days .

So, without further delay, here’s some pictures.

Shipping box:

Product box, with instructions:

Champagne colored Pk26 surrounded by similar sized lights, 40-41 mm head diameter. Notice the 7 trit slots? Ok, I'm gonna fill them. To tell you the truth, I held off ordering them in case the light wound up being a dud, but overall I'm liking it!

Yes, it's a Pocket Rocket. Why? Because it says so:

Some more views:

This gives you an idea of relative length, compare to a Convoy/Roche F6 and Convoy S2+:

Here's a view of the driver. Yes, those two R100's on the left are the suspects to be mounted for more possible amps (speculating...):

Inside the pill, that's epoxy, not grease, and you can barely make out the bottom of the copper MCPCB, plus the rubber switch cover on the right:

Close-up of the driver mounted switch:

The stock LED: NW XHP-35 HI and MCPCB:

I removed the stock 3M isolator and used kapton tape to protect from shorts instead, hoping to get the LED in a little tighter to the reflector:

Here's the two piece battery tube - yes, there's a threaded connection in the middle so the tail end can be removed for a single 26350 cell:

Square threads on both ends:

More Photos

With the NiteCore 40 mm diffuser installed:

Wall shot (couch shot?) w/diffuser. I've always been impressed with this NiteCore diffuser, plus it adds nice rubber protection of the head, securely attaches, easy to get OFF and ON. Notice the nice, smooth white tone you get with a NW. The hot spot is essentially gone:

Same mode (hi) without diffuser:

The mysterious blue button - only viewable in dark places:

The short setup with a 26350 installed. The S2+ shorty is not so short here:

Here's some #'s. Now since this is NW HI version, lumens should be down somewhat, and throw up (what did I just say?). The OP reflector though is gonna knock down throw #'s a little.

Measurements

Ouput Measurements done on a LiitoKala 26650 5000 cell charge to 4.20v. Lumens measured using my PVC based light box with my original LX1330B meter:

  • lo-lo: 9 lumens
  • lo: 170 lumens
  • med: 534 lumens
  • hi: 993 lumens, throw: 21 kcd (290 meters), measured at 5m
  • turbo: 1,600 @start, 1,470 @30 secs (lumens), throw: 32 kcd (358 meters), measured at 5m

  • parasitic drain: measured 0.182 mA on both the LK 26650 and 26350 cells
  • confirmed the light does drop down from turbo to hi at 2 minutes
  • lumens and throw match up very close to the #'s above using a 26350 cell charged to 4.21v (no difference between a high drain 26650 and a 26350 cell)
  • tail amps (clamp meter) using a LK 26650 @4.22v: 6.0A on turbo
  • tail amps (clamp meter) using a 26350 @4.22v: 6.6A on turbo
  • tail amps rises as the cell depletes, while it tries to maintain the hi and turbo level output

Initial Impressions

There are some issues, like the pre-flash, apparently in all production units so far reported, but I'm overall impressed by this little guy. We've seen a ton of lights in the ~40mm head size, all sorts, from the classic 4 AA lights like the NiteCore EA41 to the latest Crelant V4A, then there's the F13 and it's clones, and the X6 and D80. I like the Lumintop SD10 BLF Edition as well with a 32650 cell. But this light is shorter than them all, thinner tube than the head, and still packs a 26650, and with the new LiitoKala class 5000 mAh cells out, this light seems tailor made for them.

Pros:

  • I like the finish and feel. Maybe it's the color, dunno, but feels like quality anodizing
  • form factor is outstanding, fits my hand well
  • The NW HI is a great looking tint - very white, smooth beam pattern. Maybe the OP reflector helps it out.
  • decent performance (hoped for better though) - need to check amps to get a rough idea of how much power the boost driver is taking. The NW HI I have though would theoretically be the lowest lumens LED option
  • the UI is decent, nothing too fancy, easy to learn though
  • solid tailstander
  • love the SS bezel

Cons:

  • pre-flashes when turning on in lo-lo mode - very annoying. From OFF if you double-click, it goes straight to turbo, then 1 click OFF (no pre-flash)
  • not a fan of the right angle driver mounted switch - they tend to bend back with use
  • The digital lockout works, but not as described - RIC cleared that up -- Update: to enter lockout, hold the switch for 3 secs, then hold for 3 secs releases it. Double-click from OFF though is not locket out, though lockout remains in effect.
  • not sure if the copper MCPCB is DTP - suspect it's not because of lack of a name brand
  • no springs for the battery -- this means very limited on cell size, unless you are ok with not screwing the tail completely on. Actually if a spring adds a lot to the length, I might be against the spring and for leaving as-is. The LiitoKala cells seems to fit pretty good - fully threads the tailcap on -- I'm good with that
  • though the threads are square, I did notice some metal shaving on edges, so the machining seems a bit rough, and some particles in the threads, probably metal fragments. I removed what I could and used Nyogel - they feel better now.
  • threads are bare, not anodized - I think anodized threads are much cleaner. In fact I'm not sure if I ever saw square un-anodized threads in a quality light like this
  • Added: The battery level indicator LED's are barely visible - the button cover is not translucent enough. Haven't tested low battery levels yet, but the blue LED is only displayed with the light ON, making it even more difficult to see
  • Added: a 26350 cell has limited ability with this light. When fully charged, it can do the full two minutes on Turbo, but at some point (I tried at 3.76v), the cell can't function in turbo mode -- the amps draw spikes at 12 amps, then turbo drops to lo-lo mode quickly, and the RED indicator is lit, indicating low voltage. The normal 4 modes continue to work fine though after turning OFF/ON the light - blue indicator returns.

Accessories/Notes

2

Tweaks/Light Mods

I broke the light further down by pulling out the driver and LED. Under the MCPCB, found the probable cause of the significant droppage in both the first 1-2 seconds, then over the course of 30 seconds. This is thermal epoxy, but it's way too thick, but made thick because of the center machine marking left on the shelf under the MCPCB:

Here you can see a raised dimple in the center, and in the pic above, you can see the center spot where there's no epoxy. The dimple even appears to have some copper rubbed off on it:

Close up after cleaned up. Also note the extra large wire holes and jagged edges on them:

So, I used some 600 GRIT sandpaper and worked it out flat. Also cleaned up the jagged edges in the holes using an xacto type knife. I use the tail end of a metal chisel and wrap the sandpaper over it. This is how it came out below. Notice the thickness of the shelf? It's pretty thin at ~1.1 mm.

Stock 24 AWG wires replaced with 22 AWG. Not sure how much this helps with a regulated boost driver, and relatively low amps to the LED:

I added small dabs of solder to the front posts of the switch, to give it more support:

Sanded the backside of the copper (confirmed non-DTP by RIC - but he's changing that) to 800 GRIT:

Newly set in place with GC Extreme III thermal grease:

Flat soldered the wires, re-applied new kapton tape, assembled and re-tested.

Lumens: 1,632 @start, 1,561 @30 secs, throw: 35 kcd (374 meters) measured at 5 meters

This is a nice bump of 92 lumens @30 secs, a 6.3% increase, but even more important, the noticeable 1-2 sec drop and significant drop over the first 30 seconds is far less, indicating the heat is getting into the shelf more efficiently.

I noticed also the LED centering piece lip actually protrudes above the reflector base, and suspect that can't be good - photons have to be getting blocked there, effecting lumens but also effecting throw. So I pulled out the centering piece and took a thickness measurement:

Did some mild sanding on the bottom side:

Then took off a fair amount on the top, using 400 GRIT paper:

Cleaned it up, also cleaned the AR lens using TiFFEN Lens Cleaner (no dirt or fingerprints I could tell, but a little smudgy). This is how it looks now - notice the lip of the centering piece no longer protrudes above the reflector base. This pic was taken with the lens in:

Re-assembled, and tested:

Lumens: 1,734 @start, 1,646 @30 secs, throw: 41 kcd (405 meters) measured at 5 meters

This is another 5.4% bump in lumens @30 secs, and a 12% bump total from the original readings. Plus a 28% bump total in kcd.

Now I'm a lot more happy with it's performance. RIC is aware of these QC issues, and will be addressing them. I was able to do these tweaks without any additional cost in parts.

3

Thanks for the numbers. Output is lower than I expected. I’d like to see a stronger driver, DTP MCPCB, and SMO reflector.

4

A stacked sense resistor along the lines of an R140 would give it that additional power, Richard sells XHP-35’s on Cutter DTP copper mcpcbs, and wet/dry sandpaper will make the reflector smooth. What the net result from this ~30 minutes of modding would produce is a mystery…

Thanks Tom, nice looks at a neat light. Still haven’t decided if it’s a must have for me but it feels like it’s heading that way. :wink:

5

Thanks Dale - more minor updates in the OP in the CONS list. Yea, that size resistor should be a nice mild bump - I'll give it a try. Not sure if I'd go for a XHP-35 replacement yet, but the known copper DTP MCPCB would be nice to try. I really like the tint of this NW HI.

Looks like I'll be ordering the trits - don't own any now, and definitely feel this light is a keeper, and worth the expense of adding 7 trits.

I tried out a charged up 26350 this morning and it works very well so far. Amazing compact size, not to mention super stable tailstanding! I'll have to post pics and more #'s on it this evening.

Really like the feature of going direct to turbo from OFF by a double-click - no pre-flash, and 1 click to turn OFF. Think this is the best feature of the UI.

Got to post more impressions of the UI. Clicks for mode changes are little delayed, probably because how they chose to handle double-clicks. I prefer instant responses, prefer to over react than under react in switch based UI's, but that's me. With the light on in a mode, you can click really fast and they are all ignored - a side effect of this method they used, where my Werner style UI's act on every click, and quickly.

Another interesting thing is when you turn the light OFF, the LED goes very low quickly, but then fades to completly off - maybe all in 1-2 secs but probably an effect of the driver.

The threads seem to be getting smoother with use, but the Nyogel I'm sure helped, and should help on the wear&tear of the o-rings.

6

for more power we need good parts (the transistor, Schottky, inductance IHLP-5050FD).
This inductance will be very hot :slight_smile:

Boost Converter Boost Converter

Vin_min = 3.0V Vin_max = 4.32V Vin = 3.7V
Vout = 13.0V Iout = 1.5A f = 0.25MHz
L = 2.2uH ΔILfor Vin_min = 4.26A

7

1470lm and 32kcd seems okay for the size and price.

Interested in seeing runtime/amps/etc.

8

Thanks for the review, I like this light :slight_smile:

I can hardly imagine a XP-sized led putting out 1600 lumen on a non-DTP board not frying itself, so my feeling is that it should be DTP which has become cheap and easily available in China lately.

9

I’m ordered one from banggood :slight_smile:

10

Just pm'ed RIC again to ask if he knows if it's a true DTP. I could check myself, but might require popping off the LED. The fairly high droppage in the 1st 30 secs could be caused by thermal sag, maybe the MCPCB, maybe not a good bond/application for the thermal epoxy. If I were to do some more work on it, I'd use a good thermal grease instead. There's also fairly deep machine marks on the pill top, as partially shown in a detail pic of the LED/MCPCB above. It appears to be a uni-body style pill though.

11

Nice review.

“The right angle driver mounted switch tends to bend back with use”. But this one looks pretty solid, being that it’s a box versus a flat board soldered in couple places. This one looks like there are solder joints on all 4 corners and may hold up quite well.

12

Could be. Was hoping to get some comments on these switch types. They tend to have the two front solder points pull up, sometimes taking the trace up with them. I've had full and partial failures with this type - full means it's pushed/tilted so far back, the switch no longer works. Partial is when the switch is tilted a bit, but still enough contact for it to work. I've also had lights with this switch type that never shown signs of tilt.

13

Did you happen to read the microcontroller part? Couldn’t make out a marking in the pictures and I haven’t opened mine up yet. Hoping I can reprogram mine to remove the flashing, otherwise I’ll be selling it.

14

Nope - intentionally erased or covered. Can check if it's an ATMEL by seeing where grnd and VCC connect. Chances are, like most, it's a PIC. I don't think the flash was written in the firmware intentionally, I could be wrong though.

15

Can You write a program for the boost driver ?

16

Who? Me? No for me, too many unknowns on the hardware parts and design. There's another major part with the #'s hidden, so reverse engineering would be fun, to say the least...

17

Juaquin whont to reprogram

18

since I have received my light I have learned to double click from off to Turbo and then a single click to off. And I have learned how to lock the light out electronically. The light keeps getting better and better. Anything else coming?

question please, dumb one probably. Are these 26350 cells protected cells? If not, does the light have that feature where it will protect the cell? Thank you.

19

No, the 26350's fron CNQ and GearBest are not protected - would make them probably too big to fit well. Protected 26650's are too big, I'm sure, at least to tighten up the end cap completely. Not sure of LVP - probably has it because it has all the support for the colored indicator LED's for battery level - will test this.

Actually on CNQ's PK26 listing page (http://www.cnqualitygoods.com/goods.php?id=2244), it says:

Battery Low-Voltage protection:

When 1/5 power energy left, the flashlight will flash some times ; Then the flashlight will step down to Low Low mode to extend the runtime.

20

Nice review Tom. A little different in a good way from the norm. Looks like you have something else to keep you occupied now. :+1: