Tested: SureFire Bezels and Heat Dissipation for LEDs

[RobertM]

After reading numerous threads on CPF about SF bezels regarding heat dissipation for LEDs, I decided to put some to the test. I was especially curious to see how the M2 bezel would do since it tends to be frowned upon for LED use.


TEST 1: BEZEL COMPARISON

The Setup:

3 bezels for testing: SF 6PL, SF M2 (latest, hexagonal), and SF G3 (Nitrolon, hexagonal)
1 SF G3 body
1 SF P60L module
2 Fluke 87 DMM's
2x AW 17500 cells

The Experiment:

The P60L was ran for 10 minutes continuously in each of the three bezels. The P60L was allowed to cool for an extended period of time between tests and the AW 17500 cells were fully charged prior to each test.

One Fluke 87V with a temperature probe taped to the bezel monitored exterior bezel temperate thought the test(s). The other Fluke 87 measured current at the tailcap (tailcap removed).

Since the SF P60L will reduce power with heat, I feel that the current measurement is a decent indicator of what is going on inside temperature wise.

Test Results:

Data was collected in 15 second intervals for 10 minutes. Here is the collected data, graphed:

In addition, as soon as the 10 minute test concluded, I immediately removed the P60L from the body/bezel and took a temperature reading of the actual P60L module. Here are the results:

Using 6PL bezel: 98.5 °F
Using M2 bezel: 115.5 °F
Using G3 bezel: 117.5 °F

Difference between final bezel temperature and LED module temperature:
Using 6PL bezel: 03.7%
Using M2 bezel: 28.9%
Using G3 bezel: 08.3%


TEST 2: BODY COMPARISON

I've now performed a second test as per the request of some CPF members.

The Setup:

1 SF 6PL bezel
1 SF G3 body
1 SF 6PL body
1 SF P60L module
2 Fluke 87 DMM's
2x AW 17500 cells

The Experiment:

This time around, the bezel remained constant and the body is what was tested. The 6PL bezel was tested on both the 6PL aluminum body and the G3 Nitrolon body. The test was ran identical to the above tests with one difference being that the 2x AW 17500 cells were external to the body and connected via magnets, test cables, and DMM leads. Due to this difference, I would look at this test and the above test independently of each other.

Test Results:

Once again, as soon as the 10 minute test concluded, I immediately removed the P60L from the body/bezel and took a temperature reading of the actual P60L module. Here are the results:

Using 6PL body: 94.3 °F
Using G3 body: 105.0 °F


TEST 3: SF P60L VS. MALKOFF M60

I have now performed a third test as per the request of CPF members and ElectronGuru. ElectronGuru was very kind and sent me a Malkoff M60 for testing purposes.

The Setup:

1 SF 6PL bezel
1 SF M2 bezel
1 SF 6PL body
1 Malkoff M60 module
1 Fluke 87 DMM
2x AW 17500 cells

The Experiment:

Due to the heat sinking differences of a SF P60L and a Malkoff M60, this test was to determine whether or not the M2 bezel's performance would be closer to that of the 6PL's with an LED drop-in that utilizes the flashlight's body/bezel for heat sinking (unlike the SF P60L). The 6PL body was tested with both the 6PL aluminum bezel and the M2 aluminum shock-isolating bezel. The test was ran identical to the above tests with the 2x AW 17500 cells being external to the body and connected via magnets and test cables.

Test Results:

Similar to the other two tests, as soon as the 10 minute test concluded, I immediately removed the M60 from the body/bezel and took a temperature reading of the actual M60 module. Here are the results:

Using 6PL bezel: 99.2 °F
Using M2 bezel: 103.4 °F

Difference between final bezel temperature and LED module temperature:
Using 6PL bezel: 00.5%
Using M2 bezel: 09.3%


FINAL CONCLUSIONS:

First, the 6PL bezel obviously kept the LED the coolest while the M2 was almost as bad as the G3's Nitrolon bezel.

Second, exterior bezel temperature is NOT a good indicator of what is going on inside. The data would lead one to believe that the M2 bezel's mass kept the exterior cooler while the LED was still heating up inside.

In the body test, it would appear that having an aluminum body to transfer heat away from the bezel definitely makes a difference. The aluminum bezel remained cooler when attached to the 6PL body compared to the G3 body and the P60L was cooler upon removing it from the bezel/body at the conclusion of the test.

In the Malkoff M60 test, there were two interesting findings:
First, the M60 clearly transfers heat away for the module much better than the P60L based on the closer bezel-to-module temperatures in the data above.
Second, the M60 still ran hotter with the M2 bezel, but the M2's heat sinking performance was greatly improved when using an LED module that heat-sinks directly with the body/bezel rather than through the reflector (SF P60L).


Please feel free to discuss and ask questions about the experiment. Hopefully I covered everything.

-Robert

 

[Marduke]

How about letting each setup run for 10 minutes, then quickly take the head off and measure the module temp directly in an identical manner.

 

[RobertM]

How about letting each setup run for 10 minutes, then quickly take the head off and measure the module temp directly in an identical manner.

Actually, I did:

In addition, as soon as the 10 minute test concluded, I immediately removed the P60L from the body/bezel and took a temperature reading of the actual P60L module. Here are the results:

6PL: 98.5 °F
M2: 115.5 °F
G3: 117.5 °F

Thanks,
Robert

 

[GreyShark]

Interesting. I would expect the much thinner Z44 bezel to dump heat pretty quick after the heat source is removed whereas I would expect the relatively greater thermal mass of the M2 to hold heat longer after a heat source was removed. Based on my experience with casting metals I would expect the beefier M2 to keep the temperature lower longer than the Z44 as it would have to absorb a whole lot more energy to raise its temperature to the same degree.

If possible I would also be interested in the results of a test using a Malkoff drop in since there's a Malkoff that outperforms the P60L in any way imaginable and their heat path is through the body of the light.

Casual observation has shown me that the biggest effect seems to come from whether or not you're actually holding the light. The temperature is drastically reduced if the light is held in the hand.

 

[Marduke]

Actually, I did:



Thanks,
Robert

Whoops, missed when those readings came from.

Moving along, nothing to see here...

 

[milox]

Thank You for great info.
I'm running P60L in M2 bezel, this is very interesting data.

 

[Size15's]

It's great to see your experimental results.
I'm interested to know whether the Classic Z32 Shock Isolated Bezel performs obviously different to the "M2" bezel.

 

[Justin Case]

Repeat the test but use a metal 9P body instead of the G3 body. Or use a 6P body and 2x16340 instead of the 2x17500.

Then use some aluminum foil between the P60L and the neck of the metal body, and repeat the test once more.

 

[Size15's]

Repeat the test but use a metal 9P body instead of the G3 body. Or use a 6P body and 2x16340 instead of the 2x17500.

The Nitrolon body likely represents more onerous conditions giving an indication of how each bezel performance in relative isolation.

I agree that comparing the Aluminium and Nitrolon bodies is an interesting experiment, especially since it would be interesting to see whether there is an obvious benefit from having the body act as a larger heat sink with hopefullly a better thermal path from the bezel (or whether the the metal body does not actually play this role in a significant way)

 

[Owen]

Can you try this with a drop-in like a Malkoff, Dereelight, Solarforce, etc. that makes contact with the body?
Since the P60L uses its reflector as a heat path, with a P60L and a Nitrolon or shock-isolated bezel, the heat really has nowhere to go, while it directly contacts the rim behind the lens in the 6PL bezel, which allows the head to act as a heatsink to some extent.
The results with the M2 bezel may be much better when the heat is transferred through direct contact with the body at the base of the module, due to that bezel's greater mass and surface area.

 

[kramer5150]

Nice little test. thanks for re-confirming what many M2 owners suspected. Makes me feel reassured using my M2 primarily as an incan host.

Curious how different the results would be by foil wrapping the module tight in each host. Without foil the M2 bezel completely isolates and suspends the module inside the host with no body contact. Thats how it achieves shock resistance.

Curioius #2...
Wouldn't a more effective measurement method be to compare the temperature of the bezel with that of the module? The closer the two are, the more effective the host is at conducting heat...??

 

[Sgt. LED]

It's very good to have hard numbers to back up what I and others here have been saying for some time now.


Thank you for doing these tests.

Will you be doing more combo's for us?

 

[Justin Case]

The Nitrolon body likely represents more onerous conditions giving an indication of how each bezel performance in relative isolation.

I agree that comparing the Aluminium and Nitrolon bodies is an interesting experiment, especially since it would be interesting to see whether there is an obvious benefit from having the body act as a larger heat sink with hopefullly a better thermal path from the bezel (or whether the the metal body does not actually play this role in a significant way)

Sure, but the point is to find out if the metal body basically makes the bezel performance differences moot.

 

[MrGman]

this is good data. It would be ideal if you could put a small thermocouple some where on the P60L module that was not directly in contact with the bezel or host casing, and track the 2 temperatures of the bezel and the module directly at the same time, to show the differential as they warm up. the thermal lag from the module to the bezel shows the actual thermal resistance of the path out to the bezel. Using an LED module that consumes more power and would get hotter would also help to show the differential better.

But in the end we already know the answer. A good snug fit of an aluminum housing to the module and a good aluminum bezel on top of that, all to transfer heat away from the module and spread it to the outside world via the casing/host of the flashlight is the way best way to go. Malkoff Modules in a snug fitting 6P still seem to be the best at that. G

 

[RobertM]

The Nitrolon body likely represents more onerous conditions giving an indication of how each bezel performance in relative isolation.

I agree that comparing the Aluminium and Nitrolon bodies is an interesting experiment, especially since it would be interesting to see whether there is an obvious benefit from having the body act as a larger heat sink with hopefullly a better thermal path from the bezel (or whether the the metal body does not actually play this role in a significant way)

I agree that it would be interesting to see if there is any noticeable difference. I may try to run the tests again using the 6PL body and 2 IMR16340 cells that I have. I must admit though, the tests are painstaking with having to hold the pair of test leads to the battery and body perfectly still for the whole ten minutes :sick2:

Can you try this with a drop-in like a Malkoff, Dereelight, Solarforce, etc. that makes contact with the body?
...

Unfortunately, I don't have any Malkoffs to test or any other LEDs modules for that matter. I've sold off all of my other P60 LED drop-ins other than the P60L than came in my 6PL. But I agree, a Malkoff would be a very interesting test.

It's very good to have hard numbers to back up what I and others here have been saying for some time now.


Thank you for doing these tests.

Will you be doing more combo's for us?

No problem on the tests. I was curious myself and though that CPF would be interested in the results as well. As stated above, I may try the suggestion of running the tests again in my 6PL body. Any other ideas?

this is good data. It would be ideal if you could put a small thermocouple some where on the P60L module that was not directly in contact with the bezel or host casing, and track the 2 temperatures of the bezel and the module directly at the same time, to show the differential as they warm up. the thermal lag from the module to the bezel shows the actual thermal resistance of the path out to the bezel. Using an LED module that consumes more power and would get hotter would also help to show the differential better.

But in the end we already know the answer. A good snug fit of an aluminum housing to the module and a good aluminum bezel on top of that, all to transfer heat away from the module and spread it to the outside world via the casing/host of the flashlight is the way best way to go. Malkoff Modules in a snug fitting 6P still seem to be the best at that. G

I've been racking my brain trying to figure out how I could get the temperature probe inside against the P60L module while it's running, but haven't come up with a solution :thinking:. Even if the batteries are removed from the body and connected to the module via wires externally, you still can't get the probe up against the aluminum portion of the module.

-Robert

 

[MrGman]

"I've been racking my brain trying to figure out how I could get the temperature probe inside against the P60L module while it's running, but haven't come up with a solution :thinking:. Even if the batteries are removed from the body and connected to the module via wires externally, you still can't get the probe up against the aluminum portion of the module.

-Robert "


you couldn't use the "probe" you have, you would have to use a little tiny thermocouple head on its own set of wires and bond it to the P60 module in an appropriate place. Not something that comes with a low budget DMM kit. Or buy it from a scientific equipment supply store.

 

[Size15's]

I must admit though, the tests are painstaking with having to hold the pair of test leads to the battery and body perfectly still for the whole ten minutes :sick2:
-Robert

Have you considered a rig of some kind to remove yourself from the process?
I used to work for a testing lab and we were very inventive when it came to designing ourselves out of the tests - especially if they were repetitive or hard work or repetitive.

 

[RobertM]

As per some requests, the original post has been edited and now includes a new test which compares the aluminum and Nitrolon bodies.

Size15's, I took your advice and was able to figure out a way to remove myself from the experiment. This time I didn't have to physically hold the DMM leads to the flashlight for the durration of the test--much better. :thumbsup:

Let me know what you guys think of the new test results.

-Robert

 

[Norm]

It would be interesting to see how my method stacks up against the foil wrapped drop ins.
http://www.candlepowerforums.com/vb/showthread.php?t=230854
Norm

 

[Size15's]

That's good to see - the 6P does appear to draw away more heat than the G3 body.

Apart from the curiousity of how much difference having the batteries inside the body makes, my curiosity now is whether the hand acts as a further heat sink resulting in a lower set of temperatures.
A little ironic given that I've previously expressed the desire to see human involvement in the testing removed!