Heat sinking for an x-drive?

[starlight]

Thanks guys I think I understand what is needed.

 

[upaa27]

You can either epoxy a chunk (or finned, obviously) of copper or aluminum- copper being preferred. Or you could epoxy it right to the host like you said- a third option would be to epoxy a heatsink to the driver, then said heatsink to the host. That might make it a bit easier to remove.

He said it was a lab laser, not a hand held.

Alan

This is what I was referring to.

 

[grainde]

You can either epoxy a chunk (or finned, obviously)

Why finned "obviously"? If it is inside the host it doesnt really matter if it is finned or not, unless you have a fan. Yes we can discuss surface area and convection/radiation, but we also have to think of overall mass, which bears a direct relation to the specific heat capacity. This will be greater if the heatsink is not finned.

Anyway I have done some tests with the X-drive on my power supply at 8.4 V and 2.4 A. Both sides start getting hot after 40 seconds and are very hot after 70 seconds. I had a piece of copper non finned on the IC side and then added a second larger piece of Al to the other side. The driver ran cooler over a 70 sec run. :beer:

 

[djQUAN]

Here's how I heatsinked the drivers in my 9mm build.

I used these from FT:
https://www.fasttech.com/p/1285500

wedged them in between the drivers and the heatsink and soldered wires between the two drivers holding them together.

Here's a pic of my setup.

 

[rhd]

I see most drivers attached to diode directly.

"most" ?

That surprises me. I don't see this a lot with builds on this forum.

Why finned "obviously"? If it is inside the host it doesnt really matter if it is finned or not, unless you have a fan. Yes we can discuss surface area and convection/radiation, but we also have to think of overall mass, which bears a direct relation to the specific heat capacity. This will be greater if the heatsink is not finned.

An excellent point, and one that many builds on this forum seem to forget. Sure, there is air inside a host, but there is no air FLOW, so a fined heatsink is irrelevant, and in fact likely to be LESS useful.

---

To answer the OP's question, the correct approach is to thermal epoxy the IC(s) to the host's heatsink itself. Then you are properly dissipating heat from the driver into the same system that is already designed for dissipating heat to the outside world. Other approaches simply store some of that heat and extend the initial interval in your duty cycle, but don't actually improve the duty cycle itself.

On some drivers, it may actually be more advantageous to thermal epoxy the inductor to the heatsink, rather than the IC. I've had some anecdotal evidence to suggest that this can in some situations cool the driver better than heatsinking the IC itself. For some perspective, a 3.5A buck driver that is also small, is probably stuck using an inductor with a DC resistance of at least 100 mOhm. That's a drop of 350mV across the inductor, which is more than you're likely to see dropped by the IC and the sense resistor combined. In other words, the inductors in some of these high current drivers become significant (often the most significant) sources of heat within the driver.