Exactly how much heat can a p60 dropin dissapate?
- Thread starter Anto
- Start date
Help Support Candle Power Flashlight Forum
Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
Some simplifying assumptions:
- The P60 drop-in is 100% aluminum. Since I don't know the actual mass of aluminum in a typical drop-in, I will assume 100 g. The specific heat capacity of aluminum is 900 J/kg-K.
- The starting temp is 25C and the max temp is 150C (the max junction temp spec for the XP-E)
- The G2 is a perfect insulator so all of the heat generated accumulates in the drop-in.
Thus, if we have a 100g aluminum drop-in/heat sink and can accept a max of 125C temp rise, that gives 3.125 Wh. If we assume that the LED is 20% efficient in converting electricity to heat and the driver is 90% efficient, then for one hour of run time,
waste heat from driver + waste heat from LED = 3.13W
or
0.1x + 0.8*(0.9*x) = 3.13
where x is the total power generated by the driver.
Thus, x = 3.82W. That means the LED gets 3.44W (90% of 3.82W) for an hour, and the drop-in dissipates a total of 3.13W as heat (0.38W from the driver and 2.75W from the LED). To run an XP-E at 3.44W, you'd exceed the datasheet max of 3.4V at 700mA drive current (2.38W). Extrapolating, you might be able to run the XP-E at at about 900mA-1000mA drive current (which my eyeball says is around 3.6V-3.7V, which give an LED power draw close to 3.44W) for an hour before hitting 150C.
However, running any LED right to the max junction temp is a recipe for a short LED life. Suppose instead that you shoot for a max junction temp of 110C, or a temp rise of 85C.
Now you get 2.12 Wh of power to dissipate. Again, assuming a driver efficiency of 90% and an LED efficiency of 80%, we get x = 2.59W of power from the driver and 2.33W sent to the XP-E. Looking at the XP-E's Vf-If curve, 2.33W corresponds to about 3.4V/700mA (2.38W) for one hour. Still looks surprisingly good.
The fly in the ointment here is that there is a junction to solder point resistance of 9C/W for the XP-E. So if the aluminum heat sink is at 110C after one our at an LED power of 2.33W, the LED junction is at 110C+9*2.33 = 131C.
I personally think that Tj should be kept under 100C. So let's iterate to find out what LED power will do that for us for one hour of operation.
My guess is that a max aluminum sink temp of 80C (55C temp rise) might do it. I get 1.375 Wh, or about 1.68W power from the driver and 1.51W sent to the LED. Thus, Tj = 80+9*1.51 ~ 94C. Close enough. At 1.51W for one hour, the XP-E Vf-If curve gives about 3.3V at 500mA drive current.
If you want to use a different aluminum mass instead of 100g or a different run time than one hour, those are just scaling issues. I used one hour for ease of calculation.
Also, this is all based on the paper datasheet specs. Real life, I've measured lower Vfs for my XP-Es (e.g., 3.2V at 700mA drive current).
It would seem that SureFire is even more conservative with their P60L. Based on the advertised 80 lumen output and the use of a Seoul P4 LED (I'll assume a U flux bin with output of 91 lumens at 350 mA), I'd estimate that they are driving the LED at about 300mA. And even then, after about 15 min of steadily decreasing output in an all-Nitrolon G2, the P60L finally settles in at around 60 lumens, which corresponds to around 250 mA.
- The P60 drop-in is 100% aluminum. Since I don't know the actual mass of aluminum in a typical drop-in, I will assume 100 g. The specific heat capacity of aluminum is 900 J/kg-K.
- The starting temp is 25C and the max temp is 150C (the max junction temp spec for the XP-E)
- The G2 is a perfect insulator so all of the heat generated accumulates in the drop-in.
Thus, if we have a 100g aluminum drop-in/heat sink and can accept a max of 125C temp rise, that gives 3.125 Wh. If we assume that the LED is 20% efficient in converting electricity to heat and the driver is 90% efficient, then for one hour of run time,
waste heat from driver + waste heat from LED = 3.13W
or
0.1x + 0.8*(0.9*x) = 3.13
where x is the total power generated by the driver.
Thus, x = 3.82W. That means the LED gets 3.44W (90% of 3.82W) for an hour, and the drop-in dissipates a total of 3.13W as heat (0.38W from the driver and 2.75W from the LED). To run an XP-E at 3.44W, you'd exceed the datasheet max of 3.4V at 700mA drive current (2.38W). Extrapolating, you might be able to run the XP-E at at about 900mA-1000mA drive current (which my eyeball says is around 3.6V-3.7V, which give an LED power draw close to 3.44W) for an hour before hitting 150C.
However, running any LED right to the max junction temp is a recipe for a short LED life. Suppose instead that you shoot for a max junction temp of 110C, or a temp rise of 85C.
Now you get 2.12 Wh of power to dissipate. Again, assuming a driver efficiency of 90% and an LED efficiency of 80%, we get x = 2.59W of power from the driver and 2.33W sent to the XP-E. Looking at the XP-E's Vf-If curve, 2.33W corresponds to about 3.4V/700mA (2.38W) for one hour. Still looks surprisingly good.
The fly in the ointment here is that there is a junction to solder point resistance of 9C/W for the XP-E. So if the aluminum heat sink is at 110C after one our at an LED power of 2.33W, the LED junction is at 110C+9*2.33 = 131C.
I personally think that Tj should be kept under 100C. So let's iterate to find out what LED power will do that for us for one hour of operation.
My guess is that a max aluminum sink temp of 80C (55C temp rise) might do it. I get 1.375 Wh, or about 1.68W power from the driver and 1.51W sent to the LED. Thus, Tj = 80+9*1.51 ~ 94C. Close enough. At 1.51W for one hour, the XP-E Vf-If curve gives about 3.3V at 500mA drive current.
If you want to use a different aluminum mass instead of 100g or a different run time than one hour, those are just scaling issues. I used one hour for ease of calculation.
Also, this is all based on the paper datasheet specs. Real life, I've measured lower Vfs for my XP-Es (e.g., 3.2V at 700mA drive current).
It would seem that SureFire is even more conservative with their P60L. Based on the advertised 80 lumen output and the use of a Seoul P4 LED (I'll assume a U flux bin with output of 91 lumens at 350 mA), I'd estimate that they are driving the LED at about 300mA. And even then, after about 15 min of steadily decreasing output in an all-Nitrolon G2, the P60L finally settles in at around 60 lumens, which corresponds to around 250 mA.
Last edited:
carbine15
Flashlight Enthusiast
Nerd.
just kidding
just kidding
Packhorse
Flashlight Enthusiast
<---That's what my brain did when I read Justin Case post #2.Bullzeyebill
Flashaholic
JC, nice rundown. Thanks. (Nerd?- anybody who is involved with flashlights and other lighting like we are on CPF is a Nerd, and thank God, cause us'n Nerds are hav'n fun.) 
rizky_p
Flashlight Enthusiast
that is no need for such word, we are here because we all are 
i learn something today, thanks Justin.
i learn something today, thanks Justin.
leukos
Flashlight Enthusiast
Thanks, Justin Case. I was curious what some of the current thresholds were for drop-ins in nitrolon lights. I realize the 700ma calculation is not an end all be all, but it sounds like a good benchmark to keep below. Thanks! :thumbsup:
Justin Case
Flashlight Enthusiast
- Joined
- Mar 19, 2008
- Messages
- 3,797
I just weighed a DX11836 drop-in (outer spring removed) on a postal scale and got 0.7 oz, or about 20g. That's 1/5 lighter than the 100g that I had guessed above.
So to keep the junction temp below 100C, I would drive the XP-E drop-in at about 450mA-500mA for no longer than 12 min, not 60 min.
Interestingly, this seems to agree quite well now with the SureFire P60L design in terms of the run time after which the thermal sensor kicks in and cuts the P60L's output.
So to keep the junction temp below 100C, I would drive the XP-E drop-in at about 450mA-500mA for no longer than 12 min, not 60 min.
Interestingly, this seems to agree quite well now with the SureFire P60L design in terms of the run time after which the thermal sensor kicks in and cuts the P60L's output.
