Jump to content

Aluminum Heatsink Case for Odroid-C2

mhoney

By mhoney
in Off-topic

 Share

Recommended Posts

tkaiser

tkaiser

Posted
Posted

What do you guys think?

 

Looks nice, is pretty expensive and I don't see the necessity especially given the huge amount of efforts to replace the heatsink (others reported that earlier)

 

I don't know what I'm doing wrong but I repeated the test done by your tester and get rather different numbers (room temp ~23°C, stock heatsink):

 

 

root@odroidc2:~# armbianmonitor -m
Stop monitoring using [ctrl]-[c]
Time        CPU    load %cpu %sys %usr %nice %io %irq   CPU
20:08:13:  500MHz  0.01   1%   0%   0%   0%   0%   0% 38.0°C
20:08:18:  500MHz  0.01   1%   0%   0%   0%   0%   0% 38.0°C
20:08:23:  500MHz  0.01   1%   0%   0%   0%   0%   0% 38.0°C
20:08:28:  500MHz  0.00   1%   0%   0%   0%   0%   0% 39.0°C
20:08:33:  500MHz  0.00   1%   0%   0%   0%   0%   0% 38.0°C
20:08:38:  500MHz  0.00   1%   0%   0%   0%   0%   0% 38.0°C
20:08:43:  500MHz  0.00   1%   0%   0%   0%   0%   0% 38.0°C
20:08:48:  500MHz  0.00   1%   0%   0%   0%   0%   0% 38.0°C
20:08:53:  500MHz  0.00   1%   0%   0%   0%   0%   0% 38.0°C
20:08:58:  500MHz  0.00   1%   0%   0%   0%   0%   0% 38.0°C
20:09:04: 1536MHz  0.32   1%   0%   1%   0%   0%   0% 44.0°C
20:09:09: 1536MHz  0.62   3%   0%   2%   0%   0%   0% 45.0°C
20:09:14: 1536MHz  0.89   4%   0%   3%   0%   0%   0% 46.0°C
20:09:19: 1536MHz  1.14   5%   0%   4%   0%   0%   0% 46.0°C
20:09:24: 1536MHz  1.37   6%   0%   6%   0%   0%   0% 46.0°C
20:09:30: 1536MHz  1.58   7%   0%   7%   0%   0%   0% 46.0°C
20:09:35: 1536MHz  1.77   8%   0%   8%   0%   0%   0% 46.0°C
20:09:40: 1536MHz  1.95  10%   0%   9%   0%   0%   0% 47.0°C
20:09:45: 1536MHz  2.11  11%   0%  10%   0%   0%   0% 47.0°C
20:09:51: 1536MHz  2.26  12%   0%  11%   0%   0%   0% 47.0°C
20:09:56: 1536MHz  2.40  13%   0%  12%   0%   0%   0% 47.0°C
20:10:01: 1536MHz  2.53  14%   0%  13%   0%   0%   0% 48.0°C
20:10:06: 1536MHz  2.65  15%   0%  14%   0%   0%   0% 47.0°C
20:10:12: 1536MHz  2.76  16%   0%  15%   0%   0%   0% 47.0°C
20:10:17: 1536MHz  2.86  17%   0%  16%   0%   0%   0% 48.0°C
20:10:22: 1536MHz  3.03  17%   0%  17%   0%   0%   0% 48.0°C
20:10:27: 1536MHz  3.18  18%   0%  18%   0%   0%   0% 48.0°C
20:10:32: 1536MHz  3.24  19%   0%  19%   0%   0%   0% 48.0°C
20:10:38: 1536MHz  3.30  20%   0%  20%   0%   0%   0% 48.0°C
20:10:43: 1536MHz  3.36  21%   0%  20%   0%   0%   0% 48.0°C
20:10:48: 1536MHz  3.41  22%   0%  21%   0%   0%   0% 49.0°C
20:10:53: 1536MHz  3.46  23%   0%  22%   0%   0%   0% 49.0°C
20:10:59: 1536MHz  3.50  23%   0%  23%   0%   0%   0% 49.0°C
20:11:04: 1536MHz  3.54  24%   0%  24%   0%   0%   0% 49.0°C
20:11:09: 1536MHz  3.58  25%   0%  24%   0%   0%   0% 49.0°C
20:11:14: 1536MHz  3.61  26%   0%  25%   0%   0%   0% 49.0°C
20:11:20: 1536MHz  3.64  26%   0%  26%   0%   0%   0% 49.0°C
20:11:25: 1536MHz  3.67  27%   0%  27%   0%   0%   0% 50.0°C
20:11:30: 1536MHz  3.70  28%   0%  27%   0%   0%   0% 49.0°C
20:11:35: 1536MHz  3.72  28%   0%  28%   0%   0%   0% 50.0°C
20:11:40: 1536MHz  3.74  29%   0%  29%   0%   0%   0% 50.0°C
20:11:46: 1536MHz  3.77  30%   0%  29%   0%   0%   0% 50.0°C
20:11:51: 1536MHz  3.78  30%   0%  30%   0%   0%   0% 50.0°C
20:11:56: 1536MHz  3.80  31%   0%  31%   0%   0%   0% 50.0°C
20:12:01: 1536MHz  3.82  32%   0%  31%   0%   0%   0% 51.0°C
20:12:07: 1536MHz  3.83  32%   0%  32%   0%   0%   0% 51.0°C
20:12:12: 1536MHz  3.85  33%   0%  32%   0%   0%   0% 51.0°C
20:12:17: 1536MHz  3.86  33%   0%  33%   0%   0%   0% 51.0°C
20:12:22: 1536MHz  3.95  34%   0%  34%   0%   0%   0% 51.0°C
20:12:27: 1536MHz  3.96  35%   0%  34%   0%   0%   0% 51.0°C
20:12:33: 1536MHz  3.96  35%   0%  35%   0%   0%   0% 52.0°C
20:12:38: 1536MHz  3.96  36%   0%  35%   0%   0%   0% 51.0°C
20:12:43: 1536MHz  3.97  36%   0%  36%   0%   0%   0% 52.0°C
20:12:48: 1536MHz  3.97  37%   0%  36%   0%   0%   0% 52.0°C
20:12:54: 1536MHz  3.97  37%   0%  37%   0%   0%   0% 52.0°C
20:12:59: 1536MHz  3.97  38%   0%  37%   0%   0%   0% 52.0°C
20:13:04: 1536MHz  3.98  38%   0%  38%   0%   0%   0% 51.0°C
20:13:09: 1536MHz  3.98  39%   0%  38%   0%   0%   0% 52.0°C
20:13:15: 1536MHz  3.98  39%   0%  39%   0%   0%   0% 52.0°C
20:13:20: 1536MHz  3.98  40%   0%  39%   0%   0%   0% 53.0°C
20:13:25: 1536MHz  3.98  40%   0%  40%   0%   0%   0% 52.0°C
20:13:30: 1536MHz  3.98  41%   0%  40%   0%   0%   0% 52.0°C
20:13:35: 1536MHz  3.99  41%   0%  41%   0%   0%   0% 53.0°C
20:13:41: 1536MHz  3.99  42%   0%  41%   0%   0%   0% 53.0°C
20:13:46: 1536MHz  4.07  42%   0%  42%   0%   0%   0% 53.0°C
20:13:51: 1536MHz  4.06  43%   0%  42%   0%   0%   0% 53.0°C
20:13:56: 1536MHz  4.06  43%   0%  43%   0%   0%   0% 53.0°C
20:14:02: 1536MHz  4.05  43%   0%  43%   0%   0%   0% 53.0°C
20:14:07: 1536MHz  4.05  44%   0%  43%   0%   0%   0% 53.0°C
20:14:12: 1536MHz  4.04  44%   0%  44%   0%   0%   0% 53.0°C
20:14:17: 1536MHz  4.04  45%   0%  44%   0%   0%   0% 53.0°C
20:14:23: 1536MHz  4.11  45%   0%  45%   0%   0%   0% 54.0°C
20:14:28: 1536MHz  4.10  45%   0%  45%   0%   0%   0% 54.0°C
20:14:33: 1536MHz  4.09  46%   0%  45%   0%   0%   0% 54.0°C
20:14:38: 1536MHz  4.08  46%   0%  46%   0%   0%   0% 54.0°C
20:14:44: 1536MHz  4.08  47%   0%  46%   0%   0%   0% 54.0°C
20:14:49: 1536MHz  4.07  47%   0%  47%   0%   0%   0% 54.0°C
20:14:54: 1536MHz  4.07  47%   0%  47%   0%   0%   0% 54.0°C
20:14:59: 1536MHz  4.06  48%   0%  47%   0%   0%   0% 54.0°C
20:15:04: 1536MHz  4.06  48%   0%  48%   0%   0%   0% 54.0°C
20:15:10: 1536MHz  4.05  48%   0%  48%   0%   0%   0% 54.0°C
20:15:15: 1536MHz  4.05  49%   0%  48%   0%   0%   0% 54.0°C
20:15:20: 1536MHz  4.04  49%   0%  49%   0%   0%   0% 54.0°C
20:15:25: 1536MHz  4.04  49%   0%  49%   0%   0%   0% 54.0°C
20:15:31: 1536MHz  4.04  50%   0%  49%   0%   0%   0% 54.0°C
20:15:36: 1536MHz  4.03  50%   0%  50%   0%   0%   0% 55.0°C
20:15:41: 1536MHz  4.03  50%   0%  50%   0%   0%   0% 55.0°C
20:15:46: 1536MHz  4.03  51%   0%  50%   0%   0%   0% 55.0°C
20:15:52: 1536MHz  4.03  51%   0%  51%   0%   0%   0% 55.0°C
20:15:57: 1536MHz  4.02  51%   0%  51%   0%   0%   0% 55.0°C
20:16:02: 1536MHz  4.02  52%   0%  51%   0%   0%   0% 54.0°C
20:16:07: 1536MHz  4.02  52%   0%  52%   0%   0%   0% 54.0°C
20:16:13: 1536MHz  4.02  52%   0%  52%   0%   0%   0% 54.0°C
20:16:18: 1536MHz  4.09  53%   0%  52%   0%   0%   0% 55.0°C
20:16:23: 1536MHz  4.08  53%   0%  53%   0%   0%   0% 55.0°C
20:16:28: 1536MHz  4.08  53%   0%  53%   0%   0%   0% 54.0°C
20:16:33: 1536MHz  4.07  53%   0%  53%   0%   0%   0% 55.0°C
20:16:39: 1536MHz  4.06  54%   0%  53%   0%   0%   0% 55.0°C
20:16:44: 1536MHz  4.06  54%   0%  54%   0%   0%   0% 55.0°C
20:16:49: 1536MHz  4.05  54%   0%  54%   0%   0%   0% 55.0°C
20:16:54: 1536MHz  4.05  55%   0%  54%   0%   0%   0% 55.0°C
20:16:59: 1536MHz  4.05  55%   0%  55%   0%   0%   0% 55.0°C
20:17:05: 1536MHz  4.04  55%   0%  55%   0%   0%   0% 56.0°C
20:17:10: 1536MHz  4.04  55%   0%  55%   0%   0%   0% 56.0°C
20:17:15: 1536MHz  4.04  56%   0%  55%   0%   0%   0% 55.0°C
20:17:20: 1536MHz  4.03  56%   0%  56%   0%   0%   0% 56.0°C
20:17:26: 1536MHz  4.03  56%   0%  56%   0%   0%   0% 55.0°C
20:17:31: 1536MHz  4.03  56%   0%  56%   0%   0%   0% 55.0°C
20:17:36: 1536MHz  4.03  57%   0%  56%   0%   0%   0% 55.0°C
20:17:41: 1536MHz  4.02  57%   0%  57%   0%   0%   0% 56.0°C
20:17:47: 1536MHz  4.02  57%   0%  57%   0%   0%   0% 56.0°C
20:17:52: 1536MHz  4.02  57%   0%  57%   0%   0%   0% 56.0°C
20:17:57: 1536MHz  4.02  58%   0%  57%   0%   0%   0% 56.0°C
20:18:02: 1536MHz  4.02  58%   0%  58%   0%   0%   0% 56.0°C
20:18:07: 1536MHz  4.02  58%   0%  58%   0%   0%   0% 56.0°C
20:18:13: 1536MHz  4.01  58%   0%  58%   0%   0%   0% 56.0°C
20:18:18: 1536MHz  4.01  58%   0%  58%   0%   0%   0% 56.0°C
20:18:23: 1536MHz  4.08  59%   0%  58%   0%   0%   0% 56.0°C
20:18:28: 1536MHz  4.08  59%   0%  59%   0%   0%   0% 56.0°C
20:18:34: 1536MHz  4.07  59%   0%  59%   0%   0%   0% 56.0°C
20:18:39: 1536MHz  4.07  59%   0%  59%   0%   0%   0% 56.0°C
20:18:44: 1536MHz  4.06  60%   0%  59%   0%   0%   0% 56.0°C
20:18:49: 1536MHz  4.06  60%   0%  60%   0%   0%   0% 56.0°C
20:18:54: 1536MHz  4.05  60%   0%  60%   0%   0%   0% 56.0°C
20:19:00: 1536MHz  4.05  60%   0%  60%   0%   0%   0% 56.0°C
20:19:05:  500MHz  3.72  60%   0%  60%   0%   0%   0% 49.0°C
20:19:10:  500MHz  3.42  60%   0%  59%   0%   0%   0% 48.0°C
20:19:15:  500MHz  3.15  59%   0%  59%   0%   0%   0% 48.0°C
20:19:20:  500MHz  2.90  59%   0%  59%   0%   0%   0% 48.0°C
20:19:25:  500MHz  2.67  59%   0%  59%   0%   0%   0% 48.0°C 

 

 

 

38°C idle and 56°C with the light 'stress' workload. What am I'm doing wrong?! Might settings matter?

 

  ___      _           _     _    ____ ____  

 / _ \  __| |_ __ ___ (_) __| |  / ___|___ \ 

| | | |/ _` | '__/ _ \| |/ _` | | |     __) |

| |_| | (_| | | | (_) | | (_| | | |___ / __/ 

 \___/ \__,_|_|  \___/|_|\__,_|  \____|_____|

                                             

 

Welcome to ARMBIAN Ubuntu 16.04.1 LTS 3.14.79-odroidc2 

System load:   0.00            Up time:       6 min

IP:            192.168.83.81 

CPU temp:      39°C           

Usage of /:    32% of 7.0G   

 

 

[ 5 updates to install: apt-get upgrade ]

 

Last login: Wed Oct 12 20:03:41 2016 from 192.168.83.91

tk@odroidc2:~$ stress -c $(nproc --all) -t 600

stress: info: [1525] dispatching hogs: 4 cpu, 0 io, 0 vm, 0 hdd

stress: info: [1525] successful run completed in 600s

tk@odroidc2:~$ uname -a

Linux odroidc2 3.14.79-odroidc2 #8 SMP PREEMPT Tue Sep 13 23:00:48 CEST 2016 aarch64 aarch64 aarch64 GNU/Linux

tk@odroidc2:~$ cat /etc/armbian-release 

# PLEASE DO NOT EDIT THIS FILE

BOARD=odroidc2

BOARD_NAME="Odroid C2"

VERSION=5.16

LINUXFAMILY=odroidc2

BRANCH=default

ARCH=arm64

 

BTW: Even running cpuburn-a53 doesn't lead to throttling on ODROID-C2 with default heatsink (temperature after 10 minutes just 75°C and SoC still happily clocking at the highest clockspeed).

 

No thanks, I'm pretty satisfied with the stock heatsink (Hardkernel did a good job here). But here at Armbian we always do the necessary fine-tuning for the boards we support and that might make a difference?

1
Link to comment
Share on other sites
arox

arox

Posted
Posted

No thanks, I'm pretty satisfied with the stock heatsink (Hardkernel did a good job here).

Yes, it would be nice to have the heatsink in the enclosure as in certain intel fan-less device (or simply boosters), but for practical reasons it is not a good idea for odroid-c2. It would be more useful for boards with the CPU on the opposite side of the PCB than connectors and GPIO (as nanopi neo).

Link to comment
Share on other sites
tkaiser

tkaiser

Posted
Posted

for practical reasons it is not a good idea for odroid-c2. It would be more useful for boards with the CPU on the opposite side of the PCB than connectors and GPIO (as nanopi neo).

 

The original Banana Pi and Pro had the 'hot stuff' on the opposite side and I know of at least one person making a copper enclosure for BPi (taking care of SoC, DRAM and PMIC, the latter being the hottest chip). Unfortunately he never got back to me with actual temperature numbers (maybe because they were not that much better compared to el cheapo heatsinks?)

 

The whole thing could work with ODROIDs too but for this price I would expect that DRAM is also considered (like it's done on Spreedbox -- just scroll down a little). When I did some consumption/performance tests the last weeks I was pretty surprised how heat affected consumption (don't find @ssvb's link where that is explained in detail). IIRC I tested with ODROID-C2 and got a variation of ~300mW depending on temperature (same cpufreq, same performance results, just cramping the board in a tiny enclosure vs. blowing air over the heatsink using a 12V fan). And the same 'problem' applies not only to CPU cores or SoCs but to DRAM too especially when clocked pretty high. So if I would use such an enclosure/heatsink combo it has to consider cooling DRAM too :)

 

Anyway: ODROID C1/C1+ and C2 are two examples for the few SBC out there where the vendor already took care about heat dissipation and where it requires a lot of efforts to remove the stock heatsink. Providing such an enclosure for one of these boards seems a bit strange :)

 

Edit: Found ssvb's link: http://asic-soc.blogspot.de/2008/03/process-variations-and-static-timing.html

Link to comment
Share on other sites
mhoney

mhoney

Posted
  • Author
Posted

Thanks for the input guys.  Besides the Odroid XU4, which SBC do you think would benefit the most from an aluminum heat sink enclosure?  We just finished up the RPi 3 enclosure and have a few more in the works, but we'd like to get the communities input on where an aluminum enclosure would be most beneficial.

Link to comment
Share on other sites
tkaiser

tkaiser

Posted
Posted

Yeah, a silent cooling solution for XU4 would be great. In case you haven't already please take a look at the various NanoPi: http://wiki.friendlyarm.com/wiki/index.php/Main_Page#NanoPC.2FPi_Series

 

Yesterday I examined M1 and M3 a bit and came to the conclusion that if done correctly it should be possible to design a single enclosure/heatsink combo for M1, M2, M3 (and maybe also their upcoming A64 variant) if only the SoC should stay cool. If DRAM and PMIC (M2, M3 and A64) is also considered different enclosure tops are needed. BTW: pcDuino 4 Nano is an M1 OEM so you've another board already covered when supporting M1.

 

Important: M2 and M3 should be able to connect to LCD, access to Micro USB is mandatory and if done right a ribbon cable could expose all 40 GPIO pins too! Now that some fantastic forum members improved camera drivers access to FPC connector on these boards is also important.

 

Such an enclosure would also be interesting for NanoPi NEO and Air but I doubt people will spend more than 5 bucks here.

 

Banana Pi M2+, M3, M64 and M2 Ultra would also be interesting for you (the last 3 maybe able to share one solution). M2+ overheats badly since vendor forgot voltage regulation and the SoCs of the three others all need improved heat dissipation.

 

We have another 2 overheating candidates based on Actions Semi S500 but they come already with heatsinks and are not that popular.

 

From the remaining devices you find at http://www.armbian.com/download/ only Pine64 is prone to overheating but this board might be too large for an all aluminum enclosure approach. Please keep us updated and feel free to ask if further questions occur :)

Link to comment
Share on other sites
Guest
This topic is now closed to further replies.
 Share
Go to topic listing
×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines

  I accept