Supbeam K50 V2 Mod Thread

I accessed the driver - they use thermal epoxy on the area of the resistors, so you have to pry the cover off. I did it, cleaned up some - it's clearly a labeled K40 driver. I measured 1.7A at 16.4v at the carrier, so theoretical max of 6.4A to the LED but we know it doesn't get that much - lots of loss with this design. Think on a TN32 I'd measure like 10A equivalent to the LED, but measured only 5.5A - 6A at the LED.

Clearly though the resistor area has a stack of 3 resistors, top one is R120. But, there are other SMD components, 3 legs, looks almost like FET's. Also a fine wire/jumper is in this area -- have to be real careful in clearing the epoxy around that wire. These SMD's are not in a K40 or TN31 driver -- different for sure.

I measured lumens and throw - throw is a bit higher than others measured, actually I got over 200 kcd, but man, these things can really throw stock - checked outdoors too. On one K50, the lumens measured was almost dead on with the specs in all modes, but the other light was a bit higher. selfbuilt measured it much lower than the specs -- who knows...

So far, can't get the other threaded joint open - the one at the mag ring. It would make it easier getting that one open to work on the emitter area - I hate working down a big head. Didn't try the strap wrench's yet on it. I suppose next step is to wire up the LED neg. to test stock amps at the LED (no idea what that is), then pull off the stack of resistors and see what they are - doubt I'll get to this yet - maybe someone can beat me to it and post? SmileSmile

Tom, was it easy to pry the driver of the the cover?

Just got my light in today. Did some measurements and most of the teardown. Probably wont do anything more today.

Held the head of the light in one hand (with some grip help) and a strap wrench in the other. On the picture the strap wrench is not properly in place since Im holding the camera and the flashlight.

After cleaning. Mine had a little solder blob on it..

Top side of the driver.

Pill is about 2 mm thick.

Here are emitter amps from lowest to highest mode according to my cheap DMM.

Moon 0,00237 ( second time I measured 0,00245)

0,058

0,53

1,43

2,69

High starts at 5,03 and sinks to about 4,9A in a minute.

Basically 5A to the emitter.

187kcd (In comparison, I measured a stock TK61 to 181kcd with similar "correction increase". That light is driven with 2 amps less at the emitter.With a "stage 1 resistor mod" TK61 measured about 267kcd at 4,7A)

Measured about 1470lumen OTF. (Once again Ive added a bit to readings in order to get a bit closer to others on BLF. In comparison, a "stage 1" resistor modded TK61 showed 1533 at slightly lower amps)

Im still working on my correction numbers. Sometimes I used different values. Finding a balance between ANSI and the higher numbers typically posted on BLF is not easy.

RaceR, how much torque did you apply to get the head off? Thanks for the pics.

Thanks! Cool with the amps measurements. So you are working down the head to get access to the MCPCB? I'd prefer opeing the lower threaded connection, but it's almost always glued on the K40 and TN31's.

For the bottom side where the resistors are, did you have thermal epoxy over all the chips in that area? Mine did - a lot of epoxy, big blob stuck to the inside of the cover, so had to force the cover off by pryin it from the driver. I'm lucky I didn't pull the real fine jumper wire...

Gotta run upstairs to the desktop to post pics...

My measurements:

K50 #1: lumens from hi: 1,656 1,074, throw: 208 kcd (only 2 top modes measured)

K50 #2: lumens from hi: 1,612 1,030 619 248 26 0.6, throw: 207 kcd from 5 meters

I used a combo of Pana PD and PF cells - just good ones had laying around, charged. Who knows - my measurements seem high when always compared against selfbuilt's, but have been in line with rdrfronty's. The throw #'s were impressive though - wow.

Thank you RaceR86. So you are saying that the TK61 produces 267kcd driven at 4.7A, while K50 driven with 5A produces only 187kcd, so much difference?

Thanks for the photos. Are you able to post a picture of the spring side of the driver please?

Here's the pics I took of the spring side:

After cleaning it up being very careful around that wire jumper:

I'd love to know a little more bout what's going on in this area with those 3 leg chips and the jumper.

Update: Ahh, ok. Looked at my driver pics from the K40, and it has only one 3 leg chip and spots for two resistors. This is a definite mod, adding something that looks like a FET where a resistor pad is. Other than that, the drivers look the same, but without going through all the parts carefully.

Thanks for all the reporting and pictures Tom E and RaceR86. RaceR86, when you look at the TK61 and K50 reflectors, does the TK appear to have a more refined finish and or shape? Not asking you to get precise. Just hoping to hear your casual observation.

Does anyone know how these Hall sensor drivers generally work? It appears that each hall sensor has it own set of resistors and such. I'm wondering if all the non-blinky sensors regulate current by providing a specific amount resistance or voltage to the MCU or buck controller or something.

EDIT: It does appear that there are enough traces and pins on what appears to be the MCU, that each hall sensor may be just pulling up or down a specific pin on the MCU. All the different resistors would seem unnecessary if that is that case though. I think when I get mine, I'm going to put a magnet near a sensor and then slightly reduce resistance to one of its resistors to see if emitter current changes.

Ohh -crap, mis-understood. Well, there was some epoxy shown at the edge - cut through and scraped it away with a sharp screwdriver. Then used the same sharp, thin screwdriver tool to slip it in there (jam it in), and twisted it slowly, moving the screwdriver around the edge, again more twisting til eventually she loosened up enough, going slowly. Glad I went slow after seeing that fine wire jumper up in the air...

ImA4Wheeler - think these mag ring chips have been reverse engineered before, here on BLF. I believe there is a simple level reading coming into the MCU, like 0-255 (A-D/PWM) or 0-7 value on 3 pins, something like that, then the MCU handles it to set the output level accordingly. Forgot where or who but recall reading bout it. Might have to check the traces and look up the parts - could be a master controller that feeds the MCU.

From a SupBeam K40:

From a TN31:

Funny it's all the same...

Thanks for the info and interesting pics Tom E. Boy, they do all look the same. Kind of an indicator that the K50 driver won't drive an MT-G2 stock.

Nice and clear, well illuminated, easy to see all those tiny components. Well done!

I am also confused at the difference in throw between the 2 lights you compared. Shouldn’t be such a large disparity. Interesting.

I don’t see the pill (I think this is the main pic). This is pill-less design?

….still waiting for mine….

Yes, pill-less. has the same exact design as the TN31 or K40 - not much heat sinking in the critical pill area because of the mag ring. MCPCB shelf could be thicker like the others, but this K50 has a massive head with chunky fins - better than a TN31.

The big head and fins along with the magnetic control ring and in light charging are all what appealed to me. And the price made it simply irresistible. Do you think the heavy fins do a good job dispersing the heat? Is the reflector also chunky aluminum?

How’s the balance in hand? How about the lens, is the AR coating apparently a high quality? Have you thought to try a lumens test in the box with the lens in place and again without it? Do you think cell choice is important in this light, or will regulation run the light pretty much the same regardless of cells used?

Does there seem to be any operable difference between the 2 lights you received? I like the looks of the text on the one with the Rohs label, for what it’s worth. Might have to glue a brass plate with BLF O-L GB on mine. lol

In your honest opinion, if you had the inclination to keep a light in pure stock form would this be a good choice to do so with? I’ve modified virtually every light I have, think this one might be a good one to leave stock. Crazy thought, huh? lol Maybe hot de-dome it in the light and call it good?

Thanks for bearing with all the questions. Price you pay for getting yours first and getting 2 versions. :stuck_out_tongue:

Yes

I used PF. Light is regulated (at least when stock), so cells does not matter. Measured at 10m. I may be off by 0-7cm from time to time.

Some of you guys (who often share numbers) are always high compared to most others and manufacturers that use ANSI. And on the few cases when I check numbers from other forums, they are lower. I mean, measuring 1200 lumen OTF in a TK61 driven at 2,9-3A that several does here. Im not sure that is possible. Usually I try to be conservative, or what I consider correct. But that means it would seem like the mods I were doing seems bad in comparison to others, so Im working on a balance.

Oh well, the important thing is know where everyone is when it comes to numbers. In general its not that wise to compare one person numbers against another persons.

[quote=bibihang] Thank you RaceR86. So you are saying that the TK61 produces 267kcd driven at 4.7A, while K50 driven with 5A produces _only_ 187kcd, so much difference? [/quote]

Yeah. Actually, it was 4,8A the last time I measured it. Too many numbers in my head. Not really a surprise. K50 barely beats TK61 stock in terms of kcd and its driven much harder. Numbers by selfbuilt verifies that too.

Stock TK61 is driven to about 2,9-3. K50 is driven to 4,9-5A. So when you push the TK61 harder, ofcouse it will beat the K50 in terms of kcd.

TK61 have a considerably thrower beam compared to K50. Which is due to the wider reflector. About 80mm inside vs 73mm on the K50. That makes a difference. Depth is very similar.

I cant really tell. They both look good to me.

Thanks Dale. If you or anyone ever needs larger pictures, let me know. Way easier to see components there compared to the tiny highly compressed pictures I upload.

Nothing to be confused of. See my answer to bibihang. Once you have them both in front of you, you will notice that the extra reflector width makes a difference. Physically and in terms of throw.

On a sidenote. Saw one of your comments in another thread. Fenix makes very good reflectors. Its just that the old-school copper crazed guys from Texas are not able to understand that plastic reflectors can be well made and work great. :D Btw, TK75 also uses plastic reflector. Its not a design fault, or money saving thing. Most people in the industry of hand held devices, cars, or whatever will always take lightweight as long as you do not loose out on performance. Obviously, it works very well in the TK61 and TK75. No chance of shorting the emitter wires either.

Ill leave your other questions to Tom E (seems they were focused towards him as well.) I started the teardown right after opening it up and doing the measurements. :D

Us old school copper lovers in Texas have learned from years of experience that plastic melts. Plastic gets hot and brittle and breaks. We have 100 days of 100º in the summer, stuff falls apart. Plastic sucks. Way of life. Give me aircraft quality aluminum or watch me take my business elsewhere. Really that simple.

I’m not a wuss, I can carry a few ounces. What’s all this “weight” everyone keeps referring to? It’s small, it can only be so heavy for crying out loud! I’ve carried construction materials all my life, the very thought of needing to keep weight down in a light seems very silly to me. I remember when MagLights first came out and took the flashlight market by storm…BECAUSE they were heavy and BECAUSE they were solid and had very little plastic in them! (Their reflectors have always been plastic and always been their weakness, IMHO)

Part of my other questions were just ribbing Tom. :slight_smile: Please feel free to throw in whatever observations you might have as many of us are still on the edge of our seats regarding this light. I’ve never even seen one, so don’t know anything about how it might compare say to a TK61 or any other.

Also interested in the throw comparisons, as my previous best thrower is a MAXToch SN6X-2X that doesn’t have anywhere near the diameter of these big lights but still hits 365Kcd. Still learning, still experimenting, still playing.

And still not happy with my TK61. Easily getting 341Kcd from the Courui. At considerably less monies. Perhaps when I get everything in the TK playing nicely with each other I might change my opinion.

Yes... Tongue Out

+1