AMD Processors
Decrease font size
Increase font size
Topic Title: Opteron 165 - stock fan speed
Topic Summary: Stock fan slower than that of single core Athlon64 2.0GHz Venice?!
Created On: 09/16/2007 07:37 AM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 09/16/2007 07:37 AM
User is offline View Users Profile Print this message

Author Icon
_Wolf
Junior Member

Posts: 9
Joined: 09/09/2007

Just upgraded the 939-system, Athlon64 2.0GHz Venice replaced with dual-core Opteron 165. Motherboard Gigabyte K8-NSC now supports all 4GB in dual channel mode at the default speed of 200MHz (DDR400), BIOS previously updated from F5 to F8 to support Opteron.

When installing Opteron 165, I used the original AMD copper-aluminium heatsink and fan.
I haven't fully loaded the CPU yet, and the idle temperature reported in BIOS is about 30-34'C (depending on the ambient temp, increase over 27'C with Athlon64) but what is curious is that the fan is exactly the same as the one previously used on Athlon64 but its speed is slower.
Where I had 3400 RPM with Athlon64, now it is about 3000 RPM with Opteron.

What are the normal idle temperatures for Opteron 165 (not overclocked, working at FSB 1600, clock speed 9 x 200 MHz, 1.3 V)?

If the speed of 3000 RPM for the CPU fan is below the average, would it be a good idea to replace this fan with something faster while the heatsink is still mounted over the CPU?

(Maybe the speed does not tell the whole story, could it be the case that due to improved design, the fan that came with Opteron 165 provides bigger air flow in comparison to the Athlon64 fan while working at somewhat lower speed?)

I would be grateful if someone would report their experiences (and measurements :-)
 09/16/2007 08:02 AM
User is offline View Users Profile Print this message

Author Icon
Stern63
Member

Posts: 156
Joined: 08/08/2007

Hi Wolf! I think your idle temp and fan speed is normal, this is what I got with the supplied fan which is regulated by a temp sensor. This doesn't mean it is good, or that this crude control scheme will work when the cpu is put under stress.

For me it was too noisy when idle, and insufficient under load.

Check out the problems I had and the solutions I found in this thread (assuming you would like to keep the stock cooler): http://forums.amd.com/forum/me...id=87156&enterthread=y
 09/16/2007 01:54 PM
User is offline View Users Profile Print this message

Author Icon
_Wolf
Junior Member

Posts: 9
Joined: 09/09/2007

Hi Stern,
Thanks for the link, that was an interesting reading for sure.
If I understand well, dual core Opterons and Athlons come with temperature sensors that control the speed of their fans. Meaning that 3000 RPM can go up when the CPU temperature rises? Are these sensors actually controlled through the BIOS/motherboard (the fourth wire?).
In that case, any 60mm fan with this sensos would be suitable for mounting on the heatsink (provided it is capable of speeds > 3000 RPM)?

Edited: 09/18/2007 at 04:09 PM by _Wolf
 09/16/2007 08:54 PM
User is offline View Users Profile Print this message

Author Icon
Stern63
Member

Posts: 156
Joined: 08/08/2007

The temp sensor control is a simple thermistor arrangement as I understand it, implying that fan speed vary with the temp of the air flow that goes through the fan when connected to a fan header that supplies constant 12V.

Naturally, I can only guess about the design of the original fan due to lack of documentation but this seems consistent with observations during my attempts to get the supplied fan to work.

I can see no use whatsoever for a crude thermistor-based fan speed control in a cpu cooler for a high powered system where intelligent fan control is almost mandatory.

A possibly complicating factor:
on my MB (asus a8n sli) there are only three-pin fan headers: ground, voltage and rpm sensing. Maybe that can partly explain my problems getting the supplied fan to do what I expected it to. As I said, no fan speed regulating utilities could control the fan speed of the supplied fan.

The pins on four-pin fans are supposed to be used in a way that doesn't seem entirely backwards compatble with three-pin fan headers: http://www.andigilog.com/newsp...%20vs%20DC%20fan3.pdf

Especially not if a temp sensor is included in the circuit.

It would be really nice if one of the AMD technicians could butt into one on the numerous threads on the subject of overheating Opterons and explain how the stock cooler was designed to prevent it.
 09/18/2007 04:16 PM
User is offline View Users Profile Print this message

Author Icon
_Wolf
Junior Member

Posts: 9
Joined: 09/09/2007

Another excellent reading Stern.
This document even contains the snapshot from the computational fluid dynamics model of the tower pc case showing the variation of temperatures as a function of the air-flow.
Back to your comment, I agree too that it would be more than beneficial to have AMD engineers from time to time explaining hot topics on this forum, but they must be too busy at work to have time to educate us amateurs :-)
Statistics
112018 users are registered to the AMD Processors forum.
There are currently 0 users logged in.

FuseTalk Hosting Executive Plan v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.



Contact AMD Terms and Conditions ©2007 Advanced Micro Devices, Inc. Privacy Trademark information