heatsink thickness
- Thread starter Axkiker
- Start date
Help Support Candle Power Flashlight Forum
bshanahan14rulz
Flashlight Enthusiast
yes insomuch as it increases the volume of the heatsink.
when considering a heatsink, you want something that can eat up the heat applied to it. Even better, you want it to eat up that heat and then release it into the air.
If the sink is going into a flashlight, I'd pick a metal that can grab a lot of heat. Copper is a good candidate, but the trade-off is that it is very heavy. Aluminum is much lighter but can't hold as much heat or move it as quickly. Both of these should be transfering their heat to the casing of the flashlight.
If it is going to be in open air, pick something with a lot of surface area. If the air will be moving past it in a predictable direction, stacked fin would probably be the way to go. If it is going to be sitting in stagnant air, pick a pinned heatsink. I've heard good things about anodized and/or radial pin heat sinks where the pins fan out.
So, I guess in summary and in my opinion, you should consider in this order the volume, surface area, and material when choosing a heatsink.
when considering a heatsink, you want something that can eat up the heat applied to it. Even better, you want it to eat up that heat and then release it into the air.
If the sink is going into a flashlight, I'd pick a metal that can grab a lot of heat. Copper is a good candidate, but the trade-off is that it is very heavy. Aluminum is much lighter but can't hold as much heat or move it as quickly. Both of these should be transfering their heat to the casing of the flashlight.
If it is going to be in open air, pick something with a lot of surface area. If the air will be moving past it in a predictable direction, stacked fin would probably be the way to go. If it is going to be sitting in stagnant air, pick a pinned heatsink. I've heard good things about anodized and/or radial pin heat sinks where the pins fan out.
So, I guess in summary and in my opinion, you should consider in this order the volume, surface area, and material when choosing a heatsink.
yes it does. i found that an 1/8" aluminum baseplate is a good thickness for an aluminum natural convection heatsink. this was the result of experimentation with a free online heatsink calculator, which can be found here. When I built the actual heatsink the results of the online tool turned out to be accurate.
results from the online tool also support the idea that a copper baseplate is a significant improvement over aluminum, but that copper fins have little additional value over aluminum.
it's definitely a useful tool.
results from the online tool also support the idea that a copper baseplate is a significant improvement over aluminum, but that copper fins have little additional value over aluminum.
it's definitely a useful tool.
HarryN
Flashlight Enthusiast
Perhaps a slight change of terminology will help visualize what is actually happening and make the answer more clear.
What is commonly referred to as a "heat sink" in flashlights is really a "heat spreader" This simple device of course cannot cause heat energy to "disappear" or "sink" into another dimension.
The goal is to move the heat from the center of the LED thermal pad to the walls of the flashlight as quickly as possible.
Imagine that that you can take your beautiful flashlight, and cut it right down the middle - right through the LED. (pretty disturbing visual of course). Now, imagine that light is emitting in all directions (lambertian pattern) from the bottom of the LED through a "clear" heat spreader, and you would like to get as much of that light as possible to the walls. That is more or less what is happening to the heat in your flashlight.
You can see that if the distance to the wall is very short, then "most" of the light reaches the wall very quickly.
If your heat spreader had only 2 dimensions, then the ideal heat spreader would be approximately 1/2 as thick as the distace to to the wall. Since a "real" heat spreader has 3 dimensions, then this thickness can be reduced to approx. 1/3 of this distance.
In an actual flashlight, the interface from the heat spreader to the wall is not perfect, and the walls are often thinner than you would make if thermal optimization was the deciding factor, so a heat spreder thickness of even 1/6th of this distance is still adequate using either Al or Cu. up to about 10 watts.
Once you significantly exceed 10 watts of heat, a normal size, passively cooled flashlight being hand held by the user cannot dissipate enough heat to hold it continuously. That is, unless they are also holding and drinking something ice cold. Yes, I am serious.
What is commonly referred to as a "heat sink" in flashlights is really a "heat spreader" This simple device of course cannot cause heat energy to "disappear" or "sink" into another dimension.
The goal is to move the heat from the center of the LED thermal pad to the walls of the flashlight as quickly as possible.
Imagine that that you can take your beautiful flashlight, and cut it right down the middle - right through the LED. (pretty disturbing visual of course). Now, imagine that light is emitting in all directions (lambertian pattern) from the bottom of the LED through a "clear" heat spreader, and you would like to get as much of that light as possible to the walls. That is more or less what is happening to the heat in your flashlight.
You can see that if the distance to the wall is very short, then "most" of the light reaches the wall very quickly.
If your heat spreader had only 2 dimensions, then the ideal heat spreader would be approximately 1/2 as thick as the distace to to the wall. Since a "real" heat spreader has 3 dimensions, then this thickness can be reduced to approx. 1/3 of this distance.
In an actual flashlight, the interface from the heat spreader to the wall is not perfect, and the walls are often thinner than you would make if thermal optimization was the deciding factor, so a heat spreder thickness of even 1/6th of this distance is still adequate using either Al or Cu. up to about 10 watts.
Once you significantly exceed 10 watts of heat, a normal size, passively cooled flashlight being hand held by the user cannot dissipate enough heat to hold it continuously. That is, unless they are also holding and drinking something ice cold. Yes, I am serious.
First some clarifications on some of the items previously posted:
1) Copper has much higher thermal conductivity versus aluminum.
2) Aluminum has higher specific heat than Copper (i.e. Aluminum can store more energy by weight in the form of heat versus Aluminum).
The required "thickness" of the heat-sink per-se is really a factor of how the heat is dissipated into the environment. Imagine a 2" * 2" square heat sink (copper or aluminum), of some given thickness, with a power LED mounted onto one side right in the middle. Now imagine the other side of the heat sink covered in 1" high rods, 1/16" in diameter (a typical high performance heat sink).
If the primary part of the heat sink was aluminum and it was say 1/16" thick, then most of the heat would be transferred to the environment by mainly the rods directly on the other side of the heat sink and rapidly decreasing as you moved away from the center.
If you made the heat sink thicker, i.e. 1/4" or even better, 1/4" and and copper, then the temperature across the plate would start to even out and the rods on the heatsink over much of its area would contribute to cooling. The ones in the middle would still be the hottest, but the ones on the outside would be contributing. The thicker the heatsink, the better, up to a point. You will quickly get to diminishing returns and could even get into a situation (would need to do some math) where you could start getting lower dissipation.
Some of my flashlights have a big heatsink on the front and no transfer into the body with a plastic body. I use old Thermaltake round CPU coolers without the fans. I can dissipate quite a bit of heat as these are pretty beefy.
If you are just moving heat to the body of the flashlight, I think that was covered well on the last post. You quickly get into diminishing returns. You may be able to move heat to the body of the flashlight with less resistance with a thicker piece of metal for transfer, but since the body is thin, the heat will not transfer will not be great along the length of the body and hence the part right near the transfer will get hot, but the temp comes down quickly along the body. In this case, body thickness is more critical to good heat transfer to the environment.
Semiman
1) Copper has much higher thermal conductivity versus aluminum.
2) Aluminum has higher specific heat than Copper (i.e. Aluminum can store more energy by weight in the form of heat versus Aluminum).
The required "thickness" of the heat-sink per-se is really a factor of how the heat is dissipated into the environment. Imagine a 2" * 2" square heat sink (copper or aluminum), of some given thickness, with a power LED mounted onto one side right in the middle. Now imagine the other side of the heat sink covered in 1" high rods, 1/16" in diameter (a typical high performance heat sink).
If the primary part of the heat sink was aluminum and it was say 1/16" thick, then most of the heat would be transferred to the environment by mainly the rods directly on the other side of the heat sink and rapidly decreasing as you moved away from the center.
If you made the heat sink thicker, i.e. 1/4" or even better, 1/4" and and copper, then the temperature across the plate would start to even out and the rods on the heatsink over much of its area would contribute to cooling. The ones in the middle would still be the hottest, but the ones on the outside would be contributing. The thicker the heatsink, the better, up to a point. You will quickly get to diminishing returns and could even get into a situation (would need to do some math) where you could start getting lower dissipation.
Some of my flashlights have a big heatsink on the front and no transfer into the body with a plastic body. I use old Thermaltake round CPU coolers without the fans. I can dissipate quite a bit of heat as these are pretty beefy.
If you are just moving heat to the body of the flashlight, I think that was covered well on the last post. You quickly get into diminishing returns. You may be able to move heat to the body of the flashlight with less resistance with a thicker piece of metal for transfer, but since the body is thin, the heat will not transfer will not be great along the length of the body and hence the part right near the transfer will get hot, but the temp comes down quickly along the body. In this case, body thickness is more critical to good heat transfer to the environment.
Semiman
blasterman
Flashlight Enthusiast
- Joined
- Jul 17, 2008
- Messages
- 1,802
230 vs 130 I believe.
I also don't see why it's relevant. If I attach a heat source to a solid chunk of metal, and that chunk of metal were theoretically insulated on all sides so that it weren't allowed to radiate heat, then it doesn't matter what's it made of or how big it is. It will eventually over-heat. The heatsink ultimatley needs radiating area to remove the heat.
When dealing with power LEDs, and volumes of metal in ranges of only an ounce or two, then the 'heat battery' effect of a heat-sink doesn't apply but for only a short period of time. Surface area to radiate heat trumps everything else. Obviously if you make the heatsink too thin then the finite thermal conductivity acts as a bottleneck, but if it's too thick and doesn't have sufficient radiating area then the same problem occurs.
In any respect, what I'm getting at is surface area to radiate heat is really the most important factor when dealing with these size heat-sinks. Making them a bit thicker won't fix the problem.
The point is, that having a lot of surface area to radiate heat is of little consequence if you do not have an efficient path to conduct heat to that whole surface area.....which was the point I was making in my post. That is why I described a heatsink with heat dissipating rods on one side connected to a plate base. Perhaps I was not clear enough.
Thickness of the walls of the flashlight and/or heatsink area of the flashlight can be critical in ensuring that the areas of flashlight intended to dissipate heat to the outside world really are doing their job. If you look at the base of any CPU cooler, they are relatively thick, and more recently, copper has become the preferred material, especially on the ones that maximize surface area by having tons of small fins. They use copper to ensure that as much of that surface area is truly transmitting heat to the outside world as possible.
Semiman
One important issue: if you care about heatsink size then you should care about how much of the heat produced by the led goes in the heatsink. Just using some screws to fix the led to the heatsink is not a good ideea. You need some "cooling paste" like arctic silver or any other CPU cooling paste.
jtr1962
Flashaholic
Very useful indeed. I spent the better part of yesterday playing around with both the online calculator and their downloadable software. The forced air heat sinks interested me more as I use those for cooling thermoelectric modules. The calculators returned results which pretty much agree with reality. Their downloadable software lets you do cool things like input static pressure versus airflow for fans. The software uses this along with the pressure drop in the heat sink to calculate the flow rate, and in turn the temperature rise. Very useful for doing "what if?" scenarios like trying different fans with different heat sinks, or even designing your own heat sinks. It's a big time saver to simulate things first, find out what works well, and what doesn't.
Correct that copper fins don't provide much improvement over aluminum. For one forced air heat sink I tried the improvement was only about 8%. The value of a copper baseplate depends of course upon the size of your heat source. For a very small source a copper baseplate is a tremendous improvement. When the sizes of the heat source and base are close the improvement is more modest. For example, I did a 68mmx83mm heat sink which had a source size of 40mm x 40mm. While the copper baseplate cut the temperature rise there by close to 40%, it only contributed a small amount to the overall thermal resistance. The net improvement was only something like 6%.
Last edited:
Copper for fins start to become important as the thickness of the fins is reduced. The thinner they are, the bigger the benefit in using copper. Modern CPU coolers have lots of very thin copper fins to maximize surface area under forced air cooling.
Semiman
Well there IS a question over how thick the spreader under the pad must be. If it's very thin the thermal resistance can be high even if there is plenty of surface area overall.
For example, using PCB exposed copper area for cooling. Copper is a great conductor for sure. However, on a common 1oz pcb the thermal resistance of the copper foil is so high (due to its lack of thickness) that area outside of a few mm provides negligible additional cooling. Of course a thermal spreading slug typically has a pretty decent thickness for structural reasons and because there's no motivation to thin it out. But it also has to transport the heat much further, aluminum requires greater thicknesses, and people tend to shoot for crazy low thermal resistances so it's worthwhile to calculate.
For example, using PCB exposed copper area for cooling. Copper is a great conductor for sure. However, on a common 1oz pcb the thermal resistance of the copper foil is so high (due to its lack of thickness) that area outside of a few mm provides negligible additional cooling. Of course a thermal spreading slug typically has a pretty decent thickness for structural reasons and because there's no motivation to thin it out. But it also has to transport the heat much further, aluminum requires greater thicknesses, and people tend to shoot for crazy low thermal resistances so it's worthwhile to calculate.
Walterk
Enlightened
Would be great to use thermal-imaging techniques on flashlights...
Armed_Forces
Enlightened
- Joined
- Oct 19, 2008
- Messages
- 488
Please SEARCH before posting the first thought that pops into your head!
CPF is a treasure trove of information and most topics have been discussed in detail. There's very little left uncovered.
