From IOS you can get 22 gauge and 18 gauge at a good price. I've been buying 20 gauge silicone from a rather pricey eBay source in the US - like $10 w/shipping for 30 inches of black, 30 inches of red.
Mine gets hot in the same place - think AlexGT mentioned the same thing. It's a good sign the heat is getting out because this resistor mod is really cranking up the amps. I notice it in 3 mins, hoping it would disperse better over time.
1C is cool white, technically, and the stock star is copper with a direct thermal path - bigger than SinkPAD's, thicker -- actually nicest copper star I've seen. I emailed ThruNite a couple of weeks ago for the tint, and they replied XM-L2 U2 1C - I posted their reply somewhere...
I have already dedomed my SupBeam K40. It also has a thick copper star similar to the Thrunite TN31. Like Toppel, I just removed the copper star and ‘marinated’ it in gasoline to get the dome off and I am very happy with the results. Here’s the picture I took comparing it with my 2 Convoy C8s. The leftmost is a Convoy C8 T6 3C with 10*7135, the middle is a Convoy C8 T6 3C with 10*7135 and dedomed, the rightmost is my SupBeam K40 dedomed:
I’m doing the resistor mod next using the resistors I won from RaceR86. J)
I've been using that exact R082 part from DigiKey - they are like $1 each. I've done 3 TN31's so far with resistor mod and de-domed, and I measured all 3 at about 345-350 kcd, though my #'s are a little low, so think slewflash measured one of those at 360-370 kcd I believe.
Actually even the expensive light meters ($150-$400) are rated at +/- 5% in accuracy, so there's potentially a lot of slop in these #'s.
I went ahead and used my smallest valued resistor, the R120, in modding my K40. The results were just amazing. The previous tail cap reading was 1350 mA at 12.4 volts. The reading now is 1990 mA at 12.4 volts. 47.4% increase. :bigsmile:
It seems almost everyone is only adding one R082 resister to get to about 5amps. Why not add 2 and get to about 7amps? It seems the light has adequate heatsinking.