Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
Post deleted.
Last edited:
SureFire KL4 MC-E mod
- Thread starter Justin Case
- Start date
Help Support Candle Power Flashlight Forum
Robocop
Mammoth Killer
Very slick mod....I have an older KL4 (TW4) that is already opened as I swapped the 5 watt out some time back. I have seen others who have done this however I believe they had to do some mill work inside the bezel. If you were able to do this without any work to the original heat sink I may also try this mod.
Have you powered it long enough to see if it transfers heat very well? I am not sure if others have modded their circuit when doing this mod however maybe you can eventually find a way to get a little more to the emitter.
Have you powered it long enough to see if it transfers heat very well? I am not sure if others have modded their circuit when doing this mod however maybe you can eventually find a way to get a little more to the emitter.
Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
I've run it for 30 min straight, driven by 1xAW17670 in the original SF L4 body (old E2 Executive twisty non-lockout tailcap). Initial tail current draw was about 1.2A. This is the same tail current draw I measured in an E1L body using 1xIMR16340.
The KL4 head gets warm after about a minute and is noticeably hot within about 5 min. By 15 min or so, the head is burning hot to the touch but there is no tint shift. In some previous thermocouple tests on a KT1 and KT4, a burning hot flashlight surface measured to around 130F (54C), and the heat sink temp wasn't excessive (162F max for a room temp ambient test -- could be different if you start out at 100F ambient).
At the end of the 30 min run, the hot spot lux measurement was ~96% of the initial value. So thermal management seems sufficient. In contrast, with the stock Luxeon V, the hot spot lux value continuously drifted downward. It dropped about 8% within just a few minutes IIRC. If you look at the photo of the Lux V emitter slug in my original post, I think that the thermal interface to the heat sink was poor at best. The LED was poorly secured to the heat sink. In fact, if you pressed down on the Lux V case, the LED would actually sink perceptibly downward. But no thermal paste got pushed out, so I doubt that there was much paste present to couple the slug to the sink. And even if there had been enough paste present, it would seem that the gap between the slug and heat sink was overly wide and thus the amount of thermal paste would have been overly thick anyway.
To get more current sent to the LED, you could use a different driver. TaskLED's Fatman looks like one candidate. The stock driver is about 0.87" diameter. The Fatman is 0.8" diam. If you want to mod the stock driver, then you'd have to change the sense resistor value to get more current out.
At some undetermined point, more current to the LED might be counterproductive due to excessive heat generation. If you use a 1000mA boost driver, for example, the MC-E is going to draw around 6.6W of power, vs about 4W for 640mA. IMO, that's a lot of extra wattage that is mostly going to turn into waste heat. The small KL4 head is already running hot as-is.
Update: I made a full one hour run using 1xAW17670. The cell started running out of gas very soon after the 60 min mark (the lux values started dropping clearly faster than for the first hour, although not a drastic fall-off), so I terminated the test so as not to overdischarge the cell. Even so, the OCV immediately upon removal from the light was 3.37V.
After a short delay, I put a thermocouple to the side of the light and measured 130F (felt burning hot). The light was very difficult to hold in my hand at the end of the test. Presumably, the heat sink was even hotter and the LED junction temp hotter still. However, I noted no tint shift and the hot spot lux at the 60 min mark was still 95% of the initial value.
The KL4 head gets warm after about a minute and is noticeably hot within about 5 min. By 15 min or so, the head is burning hot to the touch but there is no tint shift. In some previous thermocouple tests on a KT1 and KT4, a burning hot flashlight surface measured to around 130F (54C), and the heat sink temp wasn't excessive (162F max for a room temp ambient test -- could be different if you start out at 100F ambient).
At the end of the 30 min run, the hot spot lux measurement was ~96% of the initial value. So thermal management seems sufficient. In contrast, with the stock Luxeon V, the hot spot lux value continuously drifted downward. It dropped about 8% within just a few minutes IIRC. If you look at the photo of the Lux V emitter slug in my original post, I think that the thermal interface to the heat sink was poor at best. The LED was poorly secured to the heat sink. In fact, if you pressed down on the Lux V case, the LED would actually sink perceptibly downward. But no thermal paste got pushed out, so I doubt that there was much paste present to couple the slug to the sink. And even if there had been enough paste present, it would seem that the gap between the slug and heat sink was overly wide and thus the amount of thermal paste would have been overly thick anyway.
To get more current sent to the LED, you could use a different driver. TaskLED's Fatman looks like one candidate. The stock driver is about 0.87" diameter. The Fatman is 0.8" diam. If you want to mod the stock driver, then you'd have to change the sense resistor value to get more current out.
At some undetermined point, more current to the LED might be counterproductive due to excessive heat generation. If you use a 1000mA boost driver, for example, the MC-E is going to draw around 6.6W of power, vs about 4W for 640mA. IMO, that's a lot of extra wattage that is mostly going to turn into waste heat. The small KL4 head is already running hot as-is.
Update: I made a full one hour run using 1xAW17670. The cell started running out of gas very soon after the 60 min mark (the lux values started dropping clearly faster than for the first hour, although not a drastic fall-off), so I terminated the test so as not to overdischarge the cell. Even so, the OCV immediately upon removal from the light was 3.37V.
After a short delay, I put a thermocouple to the side of the light and measured 130F (felt burning hot). The light was very difficult to hold in my hand at the end of the test. Presumably, the heat sink was even hotter and the LED junction temp hotter still. However, I noted no tint shift and the hot spot lux at the 60 min mark was still 95% of the initial value.
Last edited:
Robocop
Mammoth Killer
I am looking for an MC-E emitter online right now and I believe I may attempt this exact mod. I will want to stay with the one cell body most likely as I really like the VG-1 setup I have now....thanks for the extra information.
Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
Post deleted.
Last edited:
Robocop
Mammoth Killer
I am doing a little research now trying to figure out just what tint I would like as well as the best voltage for a single cell application. I have ordered a few things from KD in the past however will most likely not order from them again....I simply do not have the patience to wait so long for my order.
Wish me luck and thanks again for the pics.
Wish me luck and thanks again for the pics.
Nailbender has some M bin WG for sale. He's also got warm whites if you like those. I too would never order emitters from KD or DX again but not because of shipping times. I've ordered emitters from them a few times in the past, IMO the bin info they post doesn't mean squat, more like they post the bin info that seems to popular. The times I've got emitters from them did not seem to match the bin info posted at all. Luckily these were never used for flashaholics/anyone here. After a few trys I won't try again.
Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
Post deleted.
Last edited:
Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
FWIW, I measured Vf for the 2S2P K-bin MC-E at 6.0V@320mA per core. For another KL4 mod, I measured Vf for a 2S2P M-bin MC-E at 5.9V@332mA per core. For both KL4s, the stock Luxeon V Vf measured 6.3V. I estimate that the KL4 driver efficiency is 85%-90%.
Last edited:
Latest posts
-
-
-
What was your prep for today?
- Latest: Hooked on Fenix
-
