Supbeam K40 Mod thread

Hi all, I decided to create a K40 mod thread to share information on how to make the K40 into a wicked thrower. Since about 20 people bought it here, it feel like this would be relevant.

The K40 is very similar to the TN31 in terms of design, circuitry, and disassembly. Many have stated that Supbeam, despite its funny name, is an OEM company for Thrunite.

So here’s what I did:

  • Dedomed successfully using Vinh’s hotplate + razor blade method
  • Increased the current to 6.5A. I used the following resistors in parallel: 0.050, 0.050, and 0.082 Ohms. Shorting the the resistor pads gave a ridiculously high (>8A) current.
  • Lapped the PCB to a mirror (sorta) finish.
  • Changed the thermal paste to MX-4. (Overkill)

For how to open up the light, please refer to AlexGT’s tutorial: Here is how to mod the ThruNite TN-31 flashlight...

Some differences I would like to point out…

  • The head is not glued. No need to boil the light or use strap wrenches. This was a big relief to me.
  • I can’t get the driver out. After I’ve removed the contact plate, the driver seems to be stuck. Here’s how ImA4Wheelr did it:

Here’s some tips to get you started:

Some say lapping the PCB will improve thermal conductivity because the PCB makes better contact with the base. I started off with 600 grit sand paper and rubbed it 30 times each on a 90 degree rotation. I then moved on to 1000 grit, 1500 grit, and 2000 grit sandpaper and followed the same procedure.

It should look like this:

I then used MX-4 thermal compound to replace the stock white one. Some spread it out on the PCB, but I like to put a pea sized blob on the plate and set the PCB down and press evenly. This avoids microscopic air gaps.


Current tweeking:
The two sense resistors are in parallel: 0.100 and 0.082 Ohms, adding up to 0.045055 Ohms. The stock current was a about 3.455A, so according to Ohm’s law, the voltage is 0.155665025 volts. (Pardon my significant figures.)

You can mix and match resistors to your apetite but keep in mind that they should be closely matched so that the current (and heat) would be spread between them. They do get hot at high currents.

Anyhow, here’s the catch. I found a discrepancy when I was using Ohm’s law to predict the output current. The driver wasn’t perfect, and the output current would vary by the forward voltage. For example, when I shorted the output with my DMM, I read an excess of 8A. When I used a couple 1N5408 diodes, the current would vary by the amount of diodes.

So what I did was used 5 of them to match the forward voltage of an XM-L2 at 5-6A. I them tweeked the resistors accordingly to set the current at 6.5A. I used a 16ga wire across the resistors at first, but that was supposedly too think and the current was to large.

Here’s what I have now:

This may or may not be the hardest part, depening on your experience.
I haven’t tried the gasoline yet, but here’s “the guide” to gasoline dedoming: A Perfect Dedome?

I was too impatient and uses Vinh’s method of warming up the LED and then swiping it off:

Thankfully, it worked really well.

I’ll try to answer any questions, so please don’t hesitate to ask.
If you can sharing your mods as well, that would be great!

Beamshots, here we go: (Oh and the bridge is 975m across.)

I first would like to apologize to the couple. :party:

Swinging the light around:

I used two exposure settings, 1.0 sec/f3.5/ISO200 for the 6th and 7th photos, and 2.0 sec/f5.6/ISO200 for the rest.

More beamshots:

Where and how are you measuring the current? Interesting the stock TN31 uses 2 R082's, so it's different. I measured LED current on the neg. side and seem to be reading low - 5.5A or so but measuring 1,700 lumens, so the amps must be actually higher - dunno what I'm doing wrong.

Also, I'm pretty sure vinh is just shorting the resistors out, just like he does on the TN31's. I measure insane amps at the carrier (3.5A - effective 10.5), but still much lower at the LED.

Waiting on my K40 to come, so planning on doing the same as I've done to the TN31's, but if you see 8+ amps at the LED, I don't want to go there...

That is one fantastic video on the led mods. Thanks all.

For the stock current measurement, I connected my DMM with reasonably thick leads between the driver - and the LED -.

When I was adjusting the current via the diodes, I put the DMM between the driver’s output and the diodes. I must say that there was a large variation between current and the number of diodes, resulting in different forward voltage.

I suppose you can use a LED to set the current, preferably an XM-L because you might burn it up.

Beamshots on the way, it’s almost midnight here and I’ll be right back. :slight_smile:

Oh well, thanks ryansoh4 for the info. Seems like I'm doing it the same way, between the neg. pad of the star and the wire to the driver "LED -" pad. Oh boy - just doesn't make sense - maybe my cheapo DMM's can't do the high amps, though they are rated for 10A and 20A. I got 2 UNI-T's, the 30 and UT50B from FT. Gotta bring a Fluke home from work...

No problem, I wonder if you can measure the differnence… either by the brightness or the tailcap current.

Thin leads or a DMM with a high resistance might be the problem, as the driver’s max voltage is about 4.1-4.4V open circuit, so the leads may be losing some voltage.

great write up Ryan. thanks a lot!

No problem, thanks Nick for the awesome GB. You da man. :smiley:

I’ll get some beamshots as soon as I finish my write-up on a Korean forum. :slight_smile:

Been using 12 gauge short custom built leads for a long time, very reliable, but not been in the 5+ amps range much, though think I've seen 6 amps on a direct drive XML setup.

That’s odd…

You should definately bring the Fluke home. :smiley:

Or bring the light to work. :stuck_out_tongue:

Thanx for posting this ryansoh3. I can’t wait to go back home and fix my K40. :slight_smile:

Very nice write up Ryansoh3. This will be a great resource for K40 and TN31 owners.

I’m thinking these resistor mods would also benefit the S2200, get some amps flowing to that massive MT-G2! Can’t wait to try it out. :slight_smile:

Thanks for the kind words guys, and updated the first post with beamshots!

Hope you like them. :smiley:

Hopefully the K40 driver is different than the TN31. I tried to drive an MT-G2 with the TN31 driver and it would not light. Checked output voltage and it was only 4.xx (Don’t recall exact reading) volts.

Unfortuantely the K40 won’t power the MT-G2 either.

It seems like the same driver, and has a 4.1-4.4V open circuit voltage.

thanks for the write up , and great beam shots by the way. looks like your carrying a light saber.
hey how come your light isnt yellowish tint beam like the rest of the dedomed out there?


As for the tint, it’s definitely warmer than my CW tinted lights.

It may be my camera automatically adjusting the white balance, not sure. :slight_smile:

the original tint, if it’s quite blue to begin with it’s just a lot closer to white when de-domed. If it’s near white to begin with, it get’s warm…into that creamy light brown color. A warm led will look quite yellowish brown when de-domed. Or this has been my experience.

Here’s 5 lights I took a pic of last night. Don’t look at it too hard, it’d been a bad day. (I’m allergic, got stung)

From left, the de-domed HD2010 w/XM-L2 T6, Solarforce S2200 MT-G2, 1D Mag w/triple XM-L2 U2, Solarforce M3 de-domed XM-L U2, Solarforce K3 MT-G2. Notice the narrow beams from the 2 de-domed lights. Also notice the Neutral XM-L2 T6 went more orangy brown while the XM-L U2 is still a bit blue. It’s also plain to see the further throw of the S2200’s deep large reflector as compared to the much shallower reflector of the K3. The little K3 puts out a nice wall of light, and is now actually brighter than the S2200! Modified driver putting out 4.40A at some 2400 lumens! :slight_smile: Yeah, it gets hot!

Excellant comparison shot and nice watermark!

I have an inclination towards CW, so I would have liked the dedomed K40 to be cooler… :smiley: