Intel 200 ユーザーズマニュアル

30  

 

Thermal and Mechanical Design Guidelines 

To determine the required heatsink performance, a heatsink solution provider would 

need to determine Ψ

JS

 performance for the selected TIM and mechanical load 

configuration.  If the heatsink solution were designed to work with a TIM material 

performing at Ψ

JS

 ≤ 0.50 °C/W, solving for Equation 3 from above, the performance of 

the heatsink would be:  

Ψ

SA

 = 

Ψ

JA

 

− Ψ

JS

 = 1.75 

− 0.50 = 1.25 °C/W 

The heatsink temperature requirement can be obtained from Equation 4.  

T

S-TOP-MAX 

= (

Ψ

jA 

− Ψ

jS 

− Ψ

HS_BASE

) 

× P

D 

+ T

A

 = (1.25 – 0.30) × 20 + 55 = 74 °C 

The local ambient temperature T

A

 is the temperature of the ambient air surrounding 

the processor.  For a passive heatsink, T

A

 is defined as the heatsink approach air 

temperature; for an actively cooled heatsink, it is the temperature of inlet air to the 

active cooling fan.  

It is worthwhile to determine the local ambient temperature in the chassis around the 

processor to understand the effect it may have on the die temperature.   

T

A

 is best measured by averaging temperature measurements at multiple locations in 

the heatsink inlet airflow.  This method helps reduce error and eliminate minor spatial 

variations in temperature.  The following guidelines are meant to enable accurate 

determination of the localized air temperature around the processor during system 

thermal testing.   

For active heatsinks, it is important to avoid taking measurement in the dead flow 

zone that usually develops above the fan hub and hub spokes.  Measurements should 

be taken at four different locations uniformly placed at the center of the annulus 

formed by the fan hub and the fan housing to evaluate the uniformity of the air 

temperature at the fan inlet.  The thermocouples should be placed approximately 

3 mm to 8 mm [0.1 to 0.3 in] above the fan hub vertically and halfway between the 

fan hub and the fan housing horizontally as shown in Figure 11 (avoiding the hub 

spokes).  Using an open bench to characterize an active heatsink can be useful, and 

usually ensures more uniform temperatures at the fan inlet.  However, additional tests 

that include a solid barrier above the test motherboard surface can help evaluate the 

potential impact of the chassis.  This barrier is typically clear Plexiglas*, extending at 

least 100 mm [4 in] in all directions beyond the edge of the thermal solution.  Typical 

distance from the motherboard to the barrier is 81 mm [3.2 in].  For even more 

realistic airflow, the motherboard should be populated with significant elements like 

memory cards, graphic card, and chipset heatsink.  If a barrier is used, the 

thermocouple can be taped directly to the barrier with a clear tape at the horizontal 

location as previously described, half way between the fan hub and the fan housing.  

If a variable speed fan is used, it may be useful to add a thermocouple taped to the 

barrier above the location of the temperature sensor used by the fan to check its 

speed setting against air temperature.  When measuring T

A

 in a chassis with a live 

motherboard, add-in cards, and other system components, it is likely that the T

A