easy explanation asked for aluminium/Copper/silver, pls

yellow · Apr 4, 2013

lets look at the typical (led) light setup:
Aluminium housing + some material as mounting plate for the led (in direct contact to housing to pass on heat)

now, looking at the heat capacity data on wiki,
silver is 0.2 J/Kg
copper 0.4 J/Kg
aluminium 0.9 J/Kg
"but" the text says that this is the amount of energy needed to raise the temp of the material for 1 degree.

I somehow think to remember that the common view is: copper is better than aluminium,
but dont these numbers mean, that aluminium needs double the Energy to heat up the same?

What is my error and what material is best at the thermal part position?
(weight/cost not yet of interest)
:thinking:

AnAppleSnail · Apr 4, 2013

1. How big is a kg of aluminum, vs a kg of copper? In other words, a copper flashlight and its aluminum-sized twin will weigh about 3x as much. Unless your flashlight is infinite in size (Or only used at peak power for short duration), it's more important to remove heat than to store it.
2. Thermal conductivity is important for continuous operation. But the limiting factor in most flashlight designs is not thermal conductivity, but the rate at which the flashlight can export heat to the environment at a reasonable temperature. This parameter is a combination of surface area, emissivity, and especially environmental temperature and air circulation, and other conductive factors.

I find that my well-designed flashlights below 10W never get especially hot in my hand. But any of my lights can get quite warm at 3W sitting on a table. There is no magic material that will make good thermal design. You have to plan well.

Reduce thermal resistance at each bottleneck (Thermal paste, very flat polished surfaces, etc). Increase heat rejection to the environment (Bigger, fins, held in hand). Those two things are far more important than aluminum vs. copper. Thermal interfaces range from something like 1 deg C/W for good junctions, to 10x or more that for the removal of heat from the flashlight.

StarHalo · Apr 4, 2013

The number you're looking for is thermal conductivity:

gadget_lover · Apr 4, 2013

AnAppleSnail nailed it. Try to imagine heat to be like water, and the LED as a mountain spring filling a pond. You can keep the pond from flooding if you soak up the water with a sponge, and that's like having a massive heat sink. Eventually the sponge is as wet as it can be, and can hold no more. The heat sink does the same thing and eventually reaches the same temperature as the LED. If the sponge has a place for the water to drain to, the pond will not flood. If something cools down the heat sink, it will continue to cool the LED.

The heat sink assembly is a conduit to move the heat away from the LED to a cooler area. Without a heat sink, a 3 watt LED running at full power will be damaged in seconds. A heat sink in an insulated light (plastic, for instance) will pull the heat away from the LED for several seconds or even minutes before it reaches a temperature where the LED will be damaged.

Daniel

Colonel Sanders · Apr 4, 2013

Here's an anecdotal example of an actual operating difference between aluminum and copper. I have two Mac's Tri-EDC hosts. One is aluminum and the other copper. Same size, same shape, and otherwise identical. On the highest mode (2.8a draw) the copper host takes MUCH longer to feel hot to the hand as compared to the aluminum host. Now it will get hot, mind you, but it takes far longer. I'd say 3 times as long would be about right. I attribute this to the copper host having a far greater capacity to store heat before exporting it to my hand.

So, if I want to run the light on high for 5-7 minutes straight, I can comfortably do this with the copper host but not with the aluminum host. This is a huge plus in my book.

The drawback to the copper, in my view, is that it certainly sits heavier in the pocket (but feels much sturdier and solid in the hand.)

gadget_lover · Apr 4, 2013

There is often confusion about the benefit of a body that is slow to heat up. Some think that the slower transfer of heat means that the heat sink is more efficient. Others contend that the faster the heat gets to the surface of the flashlight, the more effective it is.

The truth is that the name of the game is to keep the LED below the temperature where it breaks down. If the surface of the flashlight is slow to heat then it's because the transfer of heat is slow. It is reasonable to also expect that it's slow to transfer the heat away from the LED.

The slower heat is removed from the LED, the more the temperature rises in the LED.

It is my opinion that the slower the flashlight body heats, the poorer the heat management.

Daniel

Colonel Sanders · Apr 5, 2013

I agree with what Daniel just said 100%. However, we also must understand that when a host has more mass, it can remove more heat from the LED before transferring it to the atmosphere at a great rate. There is more heat storage capacity. The copper is pulling the heat from the LED just fine (so does aluminum obviously) but it takes longer for the copper host to "fill up" with heat and start to spill to the outside world.

SemiMan · Apr 5, 2013

Colonel Sanders said:
Here's an anecdotal example of an actual operating difference between aluminum and copper. I have two Mac's Tri-EDC hosts. One is aluminum and the other copper. Same size, same shape, and otherwise identical. On the highest mode (2.8a draw) the copper host takes MUCH longer to feel hot to the hand as compared to the aluminum host. Now it will get hot, mind you, but it takes far longer. I'd say 3 times as long would be about right. I attribute this to the copper host having a far greater capacity to store heat before exporting it to my hand.

So, if I want to run the light on high for 5-7 minutes straight, I can comfortably do this with the copper host but not with the aluminum host. This is a huge plus in my book.

The drawback to the copper, in my view, is that it certainly sits heavier in the pocket (but feels much sturdier and solid in the hand.)

The copper one has 3.3 times the density, but also has only 44% of the heat capacity by mass. The end result is that the copper version has 50% greater thermal mass for exactly the same physical dimensions hence based purely on thermal mass, it will take 50% longer to get to the same temperature.

That said, the better conductivity of the copper will tend to improve its heat sink capabilities and hence over time more energy goes into the air and surrounding and if you are holding it in your hand, the copper version will be much better at transferring heat into your hand without getting warm.

Semiman

SemiMan · Apr 5, 2013

gadget_lover said:
There is often confusion about the benefit of a body that is slow to heat up. Some think that the slower transfer of heat means that the heat sink is more efficient. Others contend that the faster the heat gets to the surface of the flashlight, the more effective it is.

The truth is that the name of the game is to keep the LED below the temperature where it breaks down. If the surface of the flashlight is slow to heat then it's because the transfer of heat is slow. It is reasonable to also expect that it's slow to transfer the heat away from the LED.

The slower heat is removed from the LED, the more the temperature rises in the LED.

It is my opinion that the slower the flashlight body heats, the poorer the heat management.

Daniel

In many cases, especially a well designed flashlight, your opinion would be completely wrong.

I will put it into electrical terms as most people understand it better:

- Thermal resistance = electrical resistance
- Thermal capacity (or mass) = electrical capacitance
- Temperature differential = voltage differential
- Temperature = absolute voltage

Any element in a thermal system has both thermal resistance and thermal capacitance.

Let's say you try to dump 2 watts of energy into a small capacitor through a defined resistance. The capacitor will quickly rise in voltage. As the resistance is define, the voltage whatever it may be, will also go up quickly. Now if you have a big capacitor, the voltage on the cap will go up slowly as will the voltage on the other side of the resistor (I fixed power).

Now let's say I attached a 2W dissipating LED directly to a big copper block (or aluminum) ... just a big block. The high thermal mass of the block means that the block will be very slow to rise in temperature. At the same time, the high heat transfer capabilities of the material will work to keep the temperature of the block somewhat homogenous though obviously there will be a differential. After some period of time, the block of metal will hit equilibrium with the heat going into it and the heat being transferred out of it. With the exception of a small differential through the block (it has a high conductivity), the LED temp never gets much higher than the surface temperature.

So let's take a copper (or aluminum) block that is 1/8 the volume (and 1/4 the surface area). The block will heat up much faster until it reaches equilibrium. However, because it has 1/4 the surface area, that equilibrium temperature at the surface w.r.t. room temperature is now 4 times higher. The thermal resistance between the surface of this smaller block and the LED may be 1/2, but since the conductivity of the materials is so high, it make little difference in terms of a thermal differential. Hence in this scenario, the increase in LED temperature will be 4x.

The only thing you can surmise from a hot bodied flashlight is that what is inside is even hotter.

A flashlight body with no fins will heat up much faster (and get much hotter overall), then one with a ton of fins with the end result being a hotter LED during all the operation.

If you "could" make a graphite bodied flashlight (graphite conducts fantastic in the alignment of the strands), then the outside would be hot quicker and the LED would potentially be cooler in equilibrium. However, with very little thermal mass, if you had brief "turbo" modes, during the turbo mode, the LED would be much hotter than an aluminum or copper body that could absorb (like a capacitor) the thermal energy spike.

Thermally enhanced plastics would get hot quicker than aluminum or copper as well, however, due to their low thermal mass and higher thermal resistance, the LED will absolutely be hotter both in transient and equilibrium conditions.

Where a slowly heating body would be a concern would be with materials with a high ratio of thermal mass to thermal resistance. The obvious one is non thermally enhanced plastic bodies. All that happens there is the LED gets really hot, then the interior gets really hot, which make the differential across the plastic body enough to transfer the heat being generated.

The other case is less obvious, titanium flashlights. Titanium on a volume basis has a similar thermal mass to aluminum. On the other hand it is a poor thermal conductor, about 1/7 that of aluminum. Because of that there will be a greater thermal gradient across the material for equivalent energy transfer. On the other hand the thermal mass is high so it will take time before the exterior heats up. It is not an indication of poor thermal management per se, just poor material choice.

Semiman

AnAppleSnail · Apr 5, 2013

And still in most flashlight applications, the bottleneck is in keeping the outer surface of the flashlight cool enough that it can accept heat from the interior of the light and maintain an at-LED steady-state temperature that is acceptable. In other words, no matter how big the thermal path is, the traffic jam comes at the edge of the flashlight.

SemiMan · Apr 5, 2013

I think that is true in a lot of cases, but depending on the design tradeoffs, it may not be worse than the other thermal resistances combined. You don't have thermal bottlenecks per se., all resistances add in series. If you have some air movement through usage that can make a big difference too.

Semiman

gadget_lover · Apr 5, 2013

SemiMan said:
In many cases, especially a well designed flashlight, your opinion would be completely wrong.

You are quite entitled to your opinion, but I find that it does not match my experiences or measurements. No matter what your material, the heat is being generated continuously inside the light and has to leave it somehow. A large mass without an effective way to shed heat only delays the point in time where the LED overheats.

One of my assumptions for a well designed flash light is that it can be used continuously until the battery dies. That may be 1, 2 or 3 hours. If I continuously generate heat from a 3 watt LED for an hour and the flashlight never warms up, then the light either has a lot of surface area or it's baking hot inside.

The best way to measure the effectiveness of the cooling is to measure the output of the LED as soon as it is lit and then after running continuously for a little while. If it dims appreciably the cooling is suspect.

BTW, I like this bit...

It is not an indication of poor thermal management per se, just poor material choice.

Material choice IS part of thermal management. Poor choices mean poor performance.

Daniel

yellow · Apr 6, 2013

... I am still not getting the explanations ...

make it more detailed and easy, for just a theoretical "discussion"

1* 18650 led flashlight ... led mounted on thermal plate ... thermal plate pressed into aluminium "barrel" body ...
so the size of the plate is roughly one coin
(so weight and/or cost is not a factor)

be it from:
titanium --> extremely bad for the led
aluminium --> gets heat away better
copper/brass --> gets heat away best (and at the same time heats up less than the alum?)

or does simply not make sense to decide between alum <-> copper in that setup?
:thinking:

AnAppleSnail · Apr 6, 2013

Titanium isn't so bad. Most of the modders find that below about 5W in aP60 host, most metals work well if your junctions are good. Bad junctions can lead to much more thermal resistance (degrees hot per watt) than metal choice does. Degrees hot per watt is tough to calculate, but it's more important than thermal conductivity of the bulk material. The details aren't so simple, and you'll have to think.

In useable design, you'll have a " self sufficient" power level where the light stays cool unaided, and a higher mode where it needs help (fans, water, a hand). These levels represent different assumptions about design, which is part of why you're getting conflicting answers. My experience with P60 hosts is that very good thermal contact is more important than exotic materials. It's also cheaper than trying to learn to braze-coat copper on to allow goofy thermal connections. The next most important thing is thermal mass in the host, in my experience.

So how does one craft good thermal junctions of low thermal resistance? A coin of amy metal probably has 1 or less C/W resistance. No matter the metal (within reason). Machine-flat, smooth mating surfaces. Interference fit and thermal grease. Large, smooth areas of excellent thermal contact. This is why shimming or wrapping P60 hosts is key.

I suggest a thermal-reservoir (heat sink, literally) just behind the LED. This warms ups with the LED, smoothing out temperature spikes, and passes heat to the light. Copper and aluminum are cost effective. Has anyone tried silver? Gold? Those gains might be marginal. This slug is in excellent contact with the body, and serves as a low-resistance path. Surfaces are key. Possibly more so than material, at reasonable power levels. At high power levels, you have to get it ALL right.

SemiMan · Apr 6, 2013

gadget_lover said:
You are quite entitled to your opinion, but I find that it does not match my experiences or measurements. No matter what your material, the heat is being generated continuously inside the light and has to leave it somehow. A large mass without an effective way to shed heat only delays the point in time where the LED overheats.

One of my assumptions for a well designed flash light is that it can be used continuously until the battery dies. That may be 1, 2 or 3 hours. If I continuously generate heat from a 3 watt LED for an hour and the flashlight never warms up, then the light either has a lot of surface area or it's baking hot inside.

It is not an opinion, it is physics.

A well designed flashlight (thermally) with a high thermal mass will heat up slowly. If it heated up faster, that would just mean it had less thermal mass. If it had less thermal mass, then it would probably also have more thermal resistance. A flashlight exterior that gets quickly hot could just as easily be a bad design as a good design.

Semiman

Russel · Apr 13, 2013

gadget_lover said:
There is often confusion about the benefit of a body that is slow to heat up. Some think that the slower transfer of heat means that the heat sink is more efficient. Others contend that the faster the heat gets to the surface of the flashlight, the more effective it is.

The truth is that the name of the game is to keep the LED below the temperature where it breaks down. If the surface of the flashlight is slow to heat then it's because the transfer of heat is slow. It is reasonable to also expect that it's slow to transfer the heat away from the LED.

The slower heat is removed from the LED, the more the temperature rises in the LED.

It is my opinion that the slower the flashlight body heats, the poorer the heat management.[...]

The body of a flashlight can be slow to heat up and but also transfer heat faster than another flashlight up to a point depending upon the construction.

Consider two flashlights, same LED, same body, same heat sink. One heat sink made of copper, one of aluminum. The LED would be cooler with the copper heat sink due to the higher thermal conductivity. But, because the copper takes about [50 percent more] heat to raise a given volume of it one degree compared to aluminum, the body of the flashlight would remain cooler until the heat sink reaches thermal saturation.

So, the copper heat sink would keep the LED cooler and it would take longer for the body to heat up, that is until the flashlight reached thermal saturation. At that point the heat dissipation of the flashlight body becomes the bottleneck.

If the flashlight is used intermittently for short periods of time, the body of a flashlight with a copper heat sink would be cooler than the same flashlight with an aluminum heat sink of the same dimensions. (Assuming that the copper heat sink hasn't reached thermal saturation.) Of course, the flashlight with the copper heat sink would take longer to reach ambient temperature, when turned off, than the flashlight with the aluminum heat sink because of the lower thermal difference between the body and ambient.

Russel · Apr 13, 2013

yellow said:
lets look at the typical (led) light setup:
Aluminium housing + some material as mounting plate for the led (in direct contact to housing to pass on heat)

now, looking at the heat capacity data on wiki,
silver is 0.2 J/Kg
copper 0.4 J/Kg
aluminium 0.9 J/Kg
"but" the text says that this is the amount of energy needed to raise the temp of the material for 1 degree.

I somehow think to remember that the common view is: copper is better than aluminium,
but dont these numbers mean, that aluminium needs double the Energy to heat up the same?

What is my error and what material is best at the thermal part position?
(weight/cost not yet of interest)
:thinking:

How much does a cubic centimeter of each metal weight?

Ag
1.00kg = 95.3 cm^3
1000/95.3 = 10.49g/cm^3

Cu
1.00kg = 112 cm^3
1000/112 = 8.93 g/cm^3

Aluminum
1.00kg = 370 cm^3
1000/370 = 2.7 g/cm^3

Taking the heat capacity of each metal from the above quote, what is the heat capacity of a cubic centimeter of each metal?

Sliver .2 * .01049 Kg = .0021
Copper .4 * .00893 Kg = .0036
Aluminum .9 * .0027 Kg = .0024

Although copper isn't as high as I thought, it does have a higher heat capacity per given volume than that aluminum or silver. Of course, silver has the best thermal conductivity, with copper next and aluminum last.

Thermal conductivity:

Silver 429
Copper 401
Aluminum 205

AnAppleSnail · Apr 13, 2013

The next step is to compare the thermal resistance of 1mm of these metals to the necessary junctions in the light. The one internal to the LED, the thermal bridge from LED to heatsink, heatsink to host, and host to air. Characterizing these for a given host allows a good comparison of where your priorities are for improvements.

Russel · Apr 13, 2013

I think the ideal configuration, provided that good mechanical thermal joints can be obtained, is the LED on a coin sized piece of silver, maybe the size of a nickel, to maximize thermal transfer from a small surface area, the LED, to a larger sized area, the sliver "coin". The silver "coin" then mounted on a copper heat sink that is in good thermal contact with the flashlight body.

That would leave four main thermal connections (excluding the internal LED junction to the LED case). The thermal connection between the LED case and the silver coin, between the coin and the copper heat sink, the joint between the heat sink and the flashlight body, and finally, between the body and ambient air. When working with the older AMD processors that have a relatively small surface areas (about the size of you pinky finger nail) I found that adding a slab of silver between the CPU and heat sink would help conduct the relatively high heat of the CPU from a small area to a larger area of a copper heat sink better than having the CPU directly in contact with the copper heat sink, despite adding another thermal joint. I would expect, given all the mechanical surfaces are well finished and flat with good mechanical connection, that this would be true with a high powered LED.

Of course, an elephant still remains in the room. Once the flashlight parts reach thermal saturation, the entire heat dissipation of the flashlight is limited by the ability of the flashlight body to dissipate heat into the ambient air. If you are talking about a 1 to 3 watt flashlight, this shouldn't be a problem. If you are talking about a 20 watt flashlight, this is a problem unless you are only using the light intermittently, or it is a dive light that is used under water.

Russel · Apr 14, 2013

yellow said:
... I am still not getting the explanations ...

make it more detailed and easy, for just a theoretical "discussion"

1* 18650 led flashlight ... led mounted on thermal plate ... thermal plate pressed into aluminium "barrel" body ...
so the size of the plate is roughly one coin
(so weight and/or cost is not a factor)

be it from:
titanium --> extremely bad for the led
aluminium --> gets heat away better
copper/brass --> gets heat away best (and at the same time heats up less than the alum?)

or does simply not make sense to decide between alum <-> copper in that setup?
:thinking:

I would recommend a copper thermal plate. I will take more heat (by about 50%) than aluminum to raise it's temperature by the same amount (it has greater thermal inertia), and it conducts heat much better. A copper thermal plate will weigh more than an aluminum one of the same size and shape, but with a smaller flashlight that shouldn't be something to be concerned about. I would recommend aluminum next and brass last. Although brass contains copper, it doesn't conduct heat as well as aluminum.

It isn't that copper heats up less than aluminum. It take more heat to raise the temperature of a given volume of copper. One way of thinking about it is that one cubic inch of copper weighs a lot more than one cubic inch of aluminum. It weighs more because it has more mass. It basically has more material occupying the same space. Because there is a lot more mass, it takes more heat to raise the temperature of a cubic inch of copper than it does aluminum. Now, if instead you were talking about a given weight, such as a Kg of copper and a Kg of aluminum (meaning the same mass), the aluminum would take more heat to raise it's temperature because it has a higher heat capacity, but a Kg of aluminum would occupy much more volume. So, if volume (size) isn't a big issue, aluminum might be a better choice. It has good thermal conductivity and good heat capacity. But, if you are limited by volume (size) such as in a small flashlight, then the higher density of copper gives you more thermal inertia and better thermal conductivity.

easy explanation asked for aluminium/Copper/silver, pls

Flashlight Enthusiast

Flashlight Enthusiast

Flashaholic

Flashaholic

Flashlight Enthusiast

Flashaholic

Flashlight Enthusiast

Banned

Banned

Flashlight Enthusiast

Banned

Flashaholic

Flashlight Enthusiast

Flashlight Enthusiast

Banned

Enlightened

Enlightened

Flashlight Enthusiast

Enlightened

Enlightened

Similar threads