>Heinz
>
>Glad to hear that my serious side worked for you. I just called on of
>the serious computer jockeys here who is always double clocking
>CPUs ect. He says try to keep heat rise to about 10 degrees above
>ambient temp. He said there is no good way to tell what max is.
>It all depends on the quality of the components.
You can attach the temp probe directly to the CPU case (CPU package not the
computer's case) and that will read within a degree or two of the CPU's
junction temperature - which is the critical temperature for silicon based
semiconductors.. You can also tap into some CPU fans' thermoprobes to get a
good temp reading - but that might require signal conditioing.
Standard commercial grade CPUs are spec'd at 75 degree C. If you run hotter
than that the part may slow down to the point that it stops working
correctly. The 10 degrees above ambient is a good general measure - it's
also the manufacturers' guideline for normal airflow. Although it will
running out of spec, most Intel and other CPUs should run just fine at even
higher temps - but marginally designed boards will likely become flakey
during elevated temp operation because bus drivers and the like also slow
down.
The CPUs are spec'd at 75 degrees C for reliability and performance. Each
degree C rise above the maximum spec'd case temperature decreases the
expected life of the part. I wouldn't worry about case temps up to 85
degrees C because you'll replace the PC before the expected end of life due
to higher temp operation.
FWIW, 20 years ago we used to design semiconductors differently for
industrial and mil temp ranges (when we could). Mostly the same parts are
used but simply derated in both performance and life specs.
Henry
Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
>
>Glad to hear that my serious side worked for you. I just called on of
>the serious computer jockeys here who is always double clocking
>CPUs ect. He says try to keep heat rise to about 10 degrees above
>ambient temp. He said there is no good way to tell what max is.
>It all depends on the quality of the components.
You can attach the temp probe directly to the CPU case (CPU package not the
computer's case) and that will read within a degree or two of the CPU's
junction temperature - which is the critical temperature for silicon based
semiconductors.. You can also tap into some CPU fans' thermoprobes to get a
good temp reading - but that might require signal conditioing.
Standard commercial grade CPUs are spec'd at 75 degree C. If you run hotter
than that the part may slow down to the point that it stops working
correctly. The 10 degrees above ambient is a good general measure - it's
also the manufacturers' guideline for normal airflow. Although it will
running out of spec, most Intel and other CPUs should run just fine at even
higher temps - but marginally designed boards will likely become flakey
during elevated temp operation because bus drivers and the like also slow
down.
The CPUs are spec'd at 75 degrees C for reliability and performance. Each
degree C rise above the maximum spec'd case temperature decreases the
expected life of the part. I wouldn't worry about case temps up to 85
degrees C because you'll replace the PC before the expected end of life due
to higher temp operation.
FWIW, 20 years ago we used to design semiconductors differently for
industrial and mil temp ranges (when we could). Mostly the same parts are
used but simply derated in both performance and life specs.
Henry
Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
