MiSTer FPGA Forum

Has anyone done research on whether a fan works more efficiently blowing onto the chip/heatsink or pulling hot air away? My theory is that a fan pulling hot air from heatsink would work most efficiently in a setup that has a case while a fan blowing onto the heatsink would work best for no case. Just wanted to see if anyone had collected data to support that.

I know MiSTerAddons.com did a 40mm fan head-to-head but that’s comparing different fans all blowing onto heatsink.

https://www.retrorgb.com/mister-ultimat ... otout.html

At the moment and while a friend of mine prints me a 40mm fan support plate, I have the Mister completely open and with a 200mm vertical fan (800rpm) just aside. The results are pretty good clearly dropping the temperatures of the heatsink on touch.

This is not the best scenario, yet it may serve to anyone else who wants to keep things cool at least temporarily.

I used this page for my information:
https://electronics.stackexchange.com/q ... ck-or-blow


TLDR version: sucking (blowing up) seems to work better

It depends on the airflow across the heatsink.
If you are using it without a case, blowing down on the heatsink is more efficient; blowing up allows too much free air to bypass the heatsink.

However, if you have a fully-enclosed case, the air path is different.
Blowing down on the heat sink may just circulate hot air inside, if the exhaust/inlet vents are too distant. On the other hand, blowing up would cause those vents to be inlets, and as long as the incoming air crosses the heatsink, this may be superior.

But you'd have to run tests for each case style.

In my own personal testing you want the air flow blowing into the heatsink.

However it should be pointed out the Cyclone SoC on the DE10-Nano is not like other single board computers such as the Raspberry Pi, you certainly do not need a strong airflow to cool, just airflow over the heatsink.

...Also, most fan blades are directional, so you need to make sure that the fan is spinning in the correct direction; just leaving the fan in place but spinning in the opposite direction isn't the same as turning it over.

dshadoff wrote: ↑Sun Jun 07, 2020 5:50 pm Blowing down on the heat sink may just circulate hot air inside, if the exhaust/inlet vents are too distant. On the other hand, blowing up would cause those vents to be inlets, and as long as the incoming air crosses the heatsink, this may be superior.

This is exactly what I was thinking. I don't like the idea of blowing air onto the heatsink since it might get dusty in there but that seems to be the most efficient given that I don't have a case.

It may be more efficient, but you are ignoring that there is no containment, and thus stagnation of hot air without a case. So effective cooling matters less than in an encased context.

Speculation never ends!

Fan placement, efficiency, etc... These are all pretty deep subjects if you want to get technical.

I don't think the DE10-Nano's processor gets hot enough to even worry about this that much. You will see better cooling results depending on the thermal tape and the kind of heatsink you use. I chose a heatsink with broader gaps and thicker fins because I knew the fan I was getting was going to not be mounted flush to the fins. Heatsink I used for reference - https://www.amazon.com/gp/product/B01JB ... UTF8&psc=1 - This was more because I wanted decent ambient air space in between the fins since less pressure and flow will be applied to the relative surface area. Additionally, the higher thermal mass of each fin makes it so the change in temperature will ideally be slower going up and down.

What is good placement depends upon all kinds of factors. Is it in a case or is it open air? What is your heatsink like? Is your fan optimized for airflow or static pressure with it's fin design? Are you using 5v or 3.3v for your fan speed control?

Currently, blowing directly on the heatsink is actually very optimal in my opinion, since that's where the primary amount of heat is being generated, and you are using the fins of the heatsink more efficiently than if you were just blowing through. Right now it's practically an "active" cooling solution. If you were just blowing through the case and there happened to be a heatsink there, that would basically be a "passive" cooling solution.

FWIW, as a newcomer, I’m currently facing an issue where my unit is rebooting/shutting down when enclosed in an Ultimate MiSTer PCB case. If I remove the panels, it works fine. The Cyclone has a heat sink on it, and my fan is facing label down.

Based on this thread, I may flip the fan label side up and see what happens, since someone asked this very question about the fan orientation in the thread I opened in the Newcomer forum. I’ll be sure to report my findings.