I want copper copper copper! (well... maybe not)

Dr_Evil said:

You have to heat up the outer portion so you can slide the other part in. That's a messy job.

Click to expand...

Wolfman29 said:

Or better yet, thermal fit I have been thinking about using thermal fits lately. Basically, it is as it sounds - stick the outside part in the oven, the inner part in the freezer. Then the inner part will shrink and the outer part will expand. You then rapidly remove them from wherever they were (after a sufficient time) and then put them together. Once they both reach ambient temperature, their fit will be amazing.

Anyway, it is looking like copper is going to be the best option for most heatsinks. It is up there with steel and iron (very close) in specific heat due to its high density, and it has the second best thermal conductivity. Because we both expect the heatsink to dissipate some of its heat to the host but also to sink it, I would think we would want the best of both worlds, i.e. copper.

Click to expand...

rhd said:

I have a running theory that there is a major misunderstanding of what we're actually doing when we create "heatsinks" for our builds. There's no doubt that the common wisdom is that copper is the ideal choice (within the realm of affordable metals) for heatsinks that need to dissipate heat away from sensitive electronics - things like CPUs, GPUs, RAM chips, etc. I agree with this common wisdom, but I propose that it does not translate into copper being a better choice for the majority the host-heatsinks that we use in laser builds. Even ignoring price entirely, I'm going to suggest that steel, iron, or nickel would be better options. Think I'm crazy?

There are two primary functions for the heatsinks we use inside our builds.

Function (A) is to transfer heat from the diode to the host walls, where the host can then dissipate the heat off to the air.
Function (B) is to absorb a whole lot of heat from the diode during your duty cycle.​

For the vast majority of hosts we build in, Function (B) is more of a priority than Function (A). While transfering diode heat to the outside air is important, the amount of heat you can actually dissipate this way is limited:

• Many of our hosts are themselves made of metals that are HORRIBLE conductors like steel. So once the heat reaches the edge of your heatsink, it hits a major bottle neck.
• Even when a host is aluminum (and thus able to conduct fairly well), there's generally not that much surface area for it to use in transferring the heat into air.
• The hosts themselves are generally fairly thin, and don't add substantially to the heat absorbing mass of your build. It's not rare for a heatsink to itself weigh more than your host!

In other words, given a choice between a heatsink that can conduct the heat to the host walls more quickly, and a heatsink that can absorb a whole lot more heat, in most builds you would prioritize the heatsink that could absorb more heat. Exceptions might be hosts made of highly finned copper (or maybe really really highly finned aluminum), where you realistically COULD actually dissipate lots of heat from host walls to outside air rapidly. Here's why it is ridiculous that people will create TINY builds, and then prioritize using "copper! copper! copper!" for the heatsink:

Thermal conductivity is a completely different property from a metal's specific heat capacity!

Sometimes metals will do really well or really poorly at both, but the two properties are not necessarily linked. While copper is known to a fantastic choice for finned heatsinks (for CPUs, etc), that knowledge does not automatically translate into our application, where what we really need is the ability to RETAIN or ABSORB a whole lot of heat!

• Copper has a relatively crappy specific heat capacity, at 0.39 kJ/kg K
• Aluminum actually does much better, at 0.91 kJ/kg K

Now, what makes up for copper's poor specific heat capacity, is that it is a really dense metal. So adjusted for volume, copper still performs about half again as well as aluminum, but it's nothing like the 2x thermal conductivity advantage that many people point to.

Here's the real kicker though. Neither copper nor aluminum are the ideal choices (even within what is reasonably affordable) for heatsinks where specific heat capacity is the real issue. I selected a handful of reasonably attainable metals (and a few that I was just curious about) and I created a chart. I researched their specific heat capacity per kg, and then adjusted based on their typical density to get their specific heat capacity per unit of volume instead of mass.

I've highlighted, in blue, the clear winners. Not copper, not aluminum. Steel, iron, and nickel. Long story short, we're doing this backwards. We should be using copper for all of our Aixiz modules, and we should be using steel, iron, or nickel for our heatsinks. The exception, being in builds where you have a host made of material that conducts really well (basically aluminum or copper) and has some sort of substantial fin structure to increase the surface area with the air. Short of that, we should be waging war on "heat absorption" rather than "heat transfer".

Spending money on Aixiz's copper 3.8mm, 5.6mm and 9mm diode modules (made by forum vet PontiacG5) to carry the heat TO your heatsink makes a whole lot more sense than trying to make the heatsink itself out of an expensive metal that can hold less heat than steel

Click to expand...