Supbeam K50 V2 Mod Thread

ImA4Wheelr · July 31, 2014, 4:38am

I didn't get a chance to charge up or find 4 matched low voltage cells to check. I will try tomorrow.

It looks like the carrier is only supposed to go in one way. It appears that the thicker side would go towards the tail as the tail has 3 contact surfaces and the thicker side of the cell holder does too. It sure would be easy to accidentally put this carrier in backwards. Not sure what that would do, but be careful folks. Hopefully, they designed the light so that such an incident will cause no damage.

Pavithra_uk · July 31, 2014, 4:52am

I think this driver can mod to use third party driver too such as NANJG with magnetic ring still functioning

I found two control pins. one goes to Buck FET controller and other to tiny FET that connected with sense resistors and LED -V

Can someone help me to find purpose of these two pins if I tell the test method ? because I don’t have a K50

I suspect second FET is used to Dimming. If its true, dimming modes should off once that FET bypassed.

ryansoh3 · July 31, 2014, 6:40am

Hmm I just got my K50 from the GB and I can’t seem to get the bezel off?

Edit: Never mind, it just took a lot of force.

Edit 2: Make sure you don’t injure yourselves while opening the bezel, I did use gloves but some of my skin peeled off and stressed out my thumb.

DB_Custom · July 31, 2014, 1:34pm

Never heard of anyone injuring themselves while removing a flashlight bezel. Must be seriously tight!

ToyKeeper found that a carpeted floor can be a big help. Push the light bezel first into the carpet and twist.

I have a section of 1/2” thick floor cushion, the stuff they put at a work station to stand on. Kind of a foamy rubber. I use that to press the light into and grip the bezel, works quite well.

ryansoh3 · July 31, 2014, 1:45pm

I used the back of a mouse pad for traction, but all I could see it doing was ruin the mousepad. I don’t have any carpets around here, but maybe I can find an ESD rubber mat.
I seriously thought the bezel was glued, but looking at the reports here it was just tight, very tight.

_the · July 31, 2014, 1:49pm

ryansoh3 · July 31, 2014, 2:00pm

Had to use Google translate to figure out what that was, but I’ll have access to Ikea in two weeks.

ryansoh3 · July 31, 2014, 4:30pm

I used a dab of hot glue to fill the void between the rubber boot and the clicky switch. I didn’t like the mushy feeling, so now the switch has a very solid firm feel.

RaceR86’s photo:

Has anyone been able to access the mode selector switch? I would like to disassemble it and add some lube and maybe reduce the clickiness of it. I prefer the tougher ring with subtler indents like in the K40.

Tom_E · July 31, 2014, 4:37pm

Ryan - you mean the mag ring? I tried with strap wrench's - No Go. She is sure glued tight shut. So much for no glue -- I've gotten that threaded joint open on TN31's and it takes quite an effort, usually. They use a pretty good LockTite.

I would like to open for two reasons:

lube it up with Nyogel as I've done before - definite improvement
easier to work on the MCPCB, etc.

ryansoh3 · July 31, 2014, 4:48pm

Yup, the mode selector ring. I remember noticing two bumps which looked like the ends of two screws looking inside the driver compartment.

I’m not sure if the screws are from the MCPCB, as the screws that hold it down are fairly short. I wonder if this indicates the threads are fastened by the two screws instead of loctite.

Tom_E · July 31, 2014, 4:59pm

Yes - those screw ends are definitely from the MCPCB. It's all really the same design going on between the K40, TN31's, etc.

ryansoh3 · July 31, 2014, 5:03pm

Ah I see. So it’s just loctite holding the threads together with the MCPCB is removed?

Old-Lumens · July 31, 2014, 5:08pm

I wish mine would come, so I could tear it completely down and take photos of everything... Sooner or later, maybe...

Ouchyfoot · July 31, 2014, 6:40pm

I’m going to wait until all the smoke clears before I touch mine.

Tom_E · July 31, 2014, 10:07pm

The threaded connection is to remove the head from the body - the MCPCB is totally separate. Like this - should be the same as this TN31 I worked on:

penumbra · July 31, 2014, 10:17pm

maybe 2 weeks waiting OL

mine too on china this time

is this light interesting with one XPG2? for modding.

ryansoh3 · July 31, 2014, 11:59pm

Thanks for the pics Tom, I also had a TN31 that opened right there because there was no glue. It was indeed quite easy to mod.

ryansoh3 · August 1, 2014, 2:51pm

I’ve done some basic current and voltage measurements of the stock driver.

Level 1:

Level 2:

Level 3:

Level 4:

Level 5:

Level 6: It peaked at 5.2A but drops to 5.0A in 5 seconds and keeps going down.

Open circuit voltage, same at all modes except stand-by and strobe (Not enough to drive an MT-G2):

ImA4Wheelr · August 1, 2014, 3:00pm

^ Excellent info and pics. Thank you ryansoh3 :)

EDIT: We gotta figure out if there are any way to adjust the voltage and also why the current drops like that. Was you emitter in the test well heat sinked?

ryansoh3 · August 1, 2014, 3:05pm

Glad to help.

Figuring out how to control the max output voltage would be great for MT-G2 LED’s. I’ve contacted Supbeam many times about selling TN30 (3x XM-L2) boards, but they wouldn’t budge. Maybe a 50 unit GB would work?

The current drop is probably due to the driver itself overheating and individual components not being able to handle the heat. My guess is that as the sense resistors heat up, their resistance increases and the output current decreases, which is probably why Supbeam put a blob of thermal glue between the driver cap and the resistors.

I was using 1N4007 diodes I use for general purpose testing. I had three of them in series, which emulates the forward voltage of the XM-L2 LED.