Page 2 of 2

Re: Raspberry Pi 4 Cooling Testing

Posted: Fri Feb 14, 2020 1:48 am
by kliclick
Hi, just a stupid question : how did you place the 40 mm Noctua fan in/over the armor case since these models have no fan or small dual fans ?
Thanks. (just looking for a quiet and good cooling solution, but don't need an extreme one).

Re: Raspberry Pi 4 Cooling Testing

Posted: Fri Feb 14, 2020 10:00 pm
by LTolledo
More likely you'll need to make your own threaded holes for the mounting.

if you dont have tools for making threaded holes, hot glue may do the trick...

I've measure those dual small fans to be 22db at 30mm from the measuring device. More likely to be lesser as the distance is increased.
Is 22db still "irritatingly loud"?

Re: Raspberry Pi 4 Cooling Testing

Posted: Sat Feb 15, 2020 8:04 am
by Imperf3kt
Many of my spare heatsinks I removed from Intel CPU coolers, simply rely on the fins to thread into - its aluminium after all, a screw will carve its own thread.

But if all you want is a cooling solution and don't need "extreme", then simply forget the fan at all, and prop the Pi on its edge (you'll have to account for cables) and the natural convection of the air around the Pi will cool it more efficiently than laying it flat, in my experience, enough to prevent it from ever throttling (I tested mine with a basic stress test on a 46C degree day and never got my Pi over 82 degrees C, 3 degrees below throttling temperature. In ordinary use, you'll likely never stress the Pi as hard as a stress test will, so it shouldn't even get that hot.

Re: Raspberry Pi 4 Cooling Testing

Posted: Sat Feb 15, 2020 9:32 am
by fruitoftheloom
Imperf3kt wrote:
Sat Feb 15, 2020 8:04 am
Many of my spare heatsinks I removed from Intel CPU coolers, simply rely on the fins to thread into - its aluminium after all, a screw will carve its own thread.

But if all you want is a cooling solution and don't need "extreme", then simply forget the fan at all, and prop the Pi on its edge (you'll have to account for cables) and the natural convection of the air around the Pi will cool it more efficiently than laying it flat, in my experience, enough to prevent it from ever throttling (I tested mine with a basic stress test on a 46C degree day and never got my Pi over 82 degrees C, 3 degrees below throttling temperature. In ordinary use, you'll likely never stress the Pi as hard as a stress test will, so it shouldn't even get that hot.

Now 2020 and after many Raspbian Buster updates overheating, except in a closed plastic case, is largely a non issue:

https://www.seeedstudio.com/blog/2019/1 ... than-ever/

Re: Raspberry Pi 4 Cooling Testing

Posted: Sat Feb 15, 2020 10:56 am
by jahboater
fruitoftheloom wrote:
Sat Feb 15, 2020 9:32 am
Now 2020 and after many Raspbian Buster updates overheating, except in a closed plastic case, is largely a non issue.
I agree.

The RPF engineers have done a great job here.

My first attempt at over-clocking a Pi4 failed at 1650MHz so I have run it at 1600MHz ever since.
Having observed the much lower temperatures with the latest firmware, I tried again yesterday.
Now 1800MHz is completely stable, verified with hours of stress tests (over_voltage=2).
Furthermore, even at that speed, it never reaches throttling temperature under any load!!!

Its outside of a case of course, mounted on edge, no fan.

Perhaps the raspi-config Overclock menu should have some standard settings for the Pi4 like it has for the early Pi's ?