The Thermal Death of The CPU: New Mobo guidelines ?  | |
December 29th, 2001, 08:04 AM
|
#1 (permalink)
| | Prof. of DooGlian Studies
Join Date: Oct 2001 Location: Nr. GroundZero NYC
Posts: 5,511
| The Thermal Death of The CPU: New Mobo guidelines ?
How do modern Mobo's deal with thermal emergencies ?
They don't according to a film compiled at AnandTech.
Worse, the thermal sensors under CPU can significantly under-report the true CPU Temp.
AMD has developed some guidelines--if I read right at this hour--to provide a protective circuit within its next generation Mobo's so that the Power to the System is immediately cut-off shutting down the Computer .
Of course, any safety mechanism is not without a price and part of it is the loss of any (I suppose unsaved) data. Of course, I would imagine most people would prefer to save the CPU in a "thermal emergency".
Question: Is this a "Real" problem or just some media hype-? The test film pulled off the fan mid-stream to show the effects of fan-failure which present-day heat monitoring devices didn't catch in time to prevent the CPU from going up in smoke.
Put more bluntly if the blades of a $4 fan stop spinning, the $130 CPU is toast. I suppose there can also be collateral damage in its surroundings but not an area I am knowledgable in.
Read about the new AMD guidelines for a protective circuit in Mobo's. Kinda sleepy now and your eyes may be better than mine.....
D  G
Last edited by MegalosSkylaki : December 29th, 2001 at 08:06 AM.
|
| |
December 29th, 2001, 08:42 AM
|
#2 (permalink)
| | Ultimate Member
Join Date: Oct 2001 Location: Queen Creek, AZ
Posts: 1,480
|
Personally I like Intels approach to overheats by slowing down the processor but at least AMD is going to do something about the overheats...
I think the video was a bit extreme in causing an over heat but the same thing would happen only slower if the fan died... |
| |
December 29th, 2001, 09:14 AM
|
#3 (permalink)
| | nuisance since 1968
Join Date: Oct 2001 Location: ɐqɟs
Posts: 10,457
| Quote:
How do modern Mobo's deal with thermal emergencies ?
They don't according to a film compiled at AnandTech.
| I think you mean Tom's Hardware Doog. |
| |
December 29th, 2001, 09:25 AM
|
#4 (permalink)
| | Prof. of DooGlian Studies
Join Date: Oct 2001 Location: Nr. GroundZero NYC
Posts: 5,511
| |
| |
December 29th, 2001, 09:33 AM
|
#5 (permalink)
| | Senior Member
Join Date: Oct 2001 Location: Abilene, TX USA
Posts: 832
|
That's nothing new... I have an EP-8KHA+ that shuts down and save the processor in all overheat situations except removing the hsf completeyl. If you just unplug the fan it catces it in time...
__________________
My wife is a hottie.
Last edited by Derek79602+ : December 30th, 2001 at 09:55 AM.
|
| |
December 30th, 2001, 01:46 AM
|
#6 (permalink)
| | Prof. of DooGlian Studies
Join Date: Oct 2001 Location: Nr. GroundZero NYC
Posts: 5,511
|
How does it do that, DEREK , do you know ?
Heck, you could be drawing a fat consultant's fee right now !
Seriously, see if you can figure out (maybe from the documentation) what's going on.
I think the Tom's article with it rather extreme way of of demonstratinf fan failure, was trying to make the point that existing safeguards--if you click through the whole article--were indequate in that they were too slow.
My guess their counterarguement would be that the increase in temp on the EPOX was too slow to mimic total fan failure.
However, in the real world , how often is total and immediate fan failure? Kind of hard to provide for every posible contingency.Are there meters or idiot lights out there that can give imediate warning? Problem is the unattended computer.
Anyway, DEREK, you get that fat consultant's contract--and I get 10% as agent  !
PROTECTIVECIRCUITD  G |
| |
December 30th, 2001, 09:25 AM
|
#7 (permalink)
| | Ultimate Member
Join Date: Oct 2001 Location: Hamilton, On, Ca
Posts: 2,620
|
Meglo, Most modern mobo's will catch the thermal problem at a selectable amount (not on by default, you have to enable it on mine anyway (8KHA+)) Mine is set to warn at 55c, shut down the pc fast at 60c,
if the blades stop spinning most mobo's or even MBM will start scream about a fan failure, oh and fan failure is nothing, temp rises about 1c every 2 seconds, from my current temp that would give me 10 seconds from when the high low siren starts going, to close what I could before it shut down.
This is all done through the bios, and most ATX AMD boards have this, right back to the K6-2 line.
The "Protective Cicuit" that AMD is talking aboot is in the CPU, and they posted the electrical schematic on how to hook it up, this will protect against heatsink removal while running.
In the end, if the heatsink falls off then it obviously wasn't installed correctly. (or there was a defect on the clip on the mobo) which should be warrenty. |
| |
December 30th, 2001, 09:58 AM
|
#8 (permalink)
| | Senior Member
Join Date: Oct 2001 Location: Abilene, TX USA
Posts: 832
|
What he said.....
It's in the bios, and someone tested it... wish I could find the article. They unplugged the fan, and that method always produced sssatisfactory results. When they removed the hsf altogether, they fried 1 out of two cpu's they tried it with. |
| |
December 30th, 2001, 02:36 PM
|
#9 (permalink)
| | Prof. of DooGlian Studies
Join Date: Oct 2001 Location: Nr. GroundZero NYC
Posts: 5,511
|
Heck, Derek & Wiz, somebody over at Toms/AMD don't know about it so get those hi-priced "consultants" Jobs  --and I get 10% as the "agent"  !
DOOOOG |
| |
January 6th, 2002, 01:05 AM
|
#10 (permalink)
| | Canuck
Join Date: Oct 2001 Location: Langley, BC, Canada
Posts: 3,603
|  must remenice (sp?) about the DOOOG.... 
__________________
- Freaky
|
| | | Thread Tools | Search this Thread | | | | |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | | | | Most Active Discussions | | | | | Recent Discussions  | | | | | |