PINE64
another heatsink question - changing the pad - Printable Version

+- PINE64 (https://forum.pine64.org)
+-- Forum: ROCKPRO64 (https://forum.pine64.org/forumdisplay.php?fid=98)
+--- Forum: RockPro64 Hardware and Accessories (https://forum.pine64.org/forumdisplay.php?fid=102)
+--- Thread: another heatsink question - changing the pad (/showthread.php?tid=9889)



another heatsink question - changing the pad - Mentaluproar - 05-18-2020

I'm working on an enclosure for the rockpro64.  The new version has provisions for active cooling but I'm having trouble with that and in testing found that the stock tall heatsink is almost enough.  It can run full throttle and just barely need throttling.  I'm thinking if I tried replacing the thermal pad with either some more exotic material or even just use thermal paste it would be enough.  Unfortunately the DRAM under the heatsink isn't as tall as the SOC so using a thermal paste wouldn't work for the DRAM.  

So I have a few questions.  Does the DRAM actually need cooled or is it just located so close to the SOC it was a matter of "why not?"  What have you tried to improve passive cooling?


RE: another heatsink question - changing the pad - Gienek - 06-07-2020

I had the same dilemma some time ago. In order to decide about connecting DRAMs to the heatsink of CPU, I have removed the heatsink and measured temperature on CPU and memory chips simultaneously.
I have tried to generate load on memory, but the temperature of the DRAMs has been constantly more than 10 degrees of Celsius lower than that of CPU. I think that connecting the DRAMs to the same heatsink with CPU is actually heating DRAMs. Not cooling them.
Thus, I have connected the tall heatsink to CPU only. I have used ARCTIC MX-4 thermal paste. The new one, with 8.5 W/(mK). I am surprised, how well it works. My ‘NAS’ has no enclosure, yet. CPU temperature under load never goes beyond 65 degrees of Celsius. Previously, without any pad or paste, it used to reach 74 degrees. I do not think, I will ever mount a heatsink without heat-conducting paste or pad. Not even for short test.
Still, I would not dare to put my RP64 to an enclosure without a fan. My enclosure will have fan with variable velocity. When I finally make it Wink .


RE: another heatsink question - changing the pad - kuleszdl - 06-11-2020

Well, the official large heatsink is not bad but not sufficient for heavy-duty usage. I ran mine with heavy load (compiling kernels and stuff for hours) without case and it got quite hot if you ask me. I think an even larger heatsink or (better) some kind of heatpipe solution would be better.

Did someone try putting heavy loads on the Rockpro64 running in the "Premium Aluminium" case?


RE: another heatsink question - changing the pad - MotoFufu9k - 12-19-2020

I have been compiling QT on 4 threads on the the RockPro64 with the Aluminium case for >10 hours, haven't reached higher than 65°C yet and if often goes down to 60°C in between. In theory pointing a fan at the case should help quite a lot as well.

Just keep in mind: My case put too much mounting pressure on the CPU, so I needed washers (ca. 0.2mm) to get the thing to boot.