Calculate heat dissipation from processor

Click For Summary

Discussion Overview

The discussion revolves around calculating heat dissipation from computer components, specifically processors and hard disk drives (HDDs). Participants explore various methods for estimating power dissipation, including theoretical calculations and practical measurements, while considering different states of operation such as load and idle.

Discussion Character

  • Exploratory
  • Technical explanation
  • Debate/contested
  • Mathematical reasoning
  • Experimental/applied

Main Points Raised

  • Some participants suggest using the formula P=I^2R for calculating power dissipation, questioning what values to use for current (I) and resistance (R).
  • Others propose looking at the power supply's current and voltage to determine power dissipation more accurately.
  • One participant mentions the importance of consulting the processor's datasheet for load and idle states to find relevant power values.
  • Another participant shares a link to a processor specification, indicating a maximum power dissipation of 65W, while expressing surprise at the high power requirement.
  • Some participants discuss the challenges of accurately measuring power dissipation, especially for overclocked processors or those not running at full load, suggesting methods like using a clamp meter on the motherboard's EPS rail.
  • One participant notes that thermal design power (TDP) can provide a rough estimate of power dissipation, but emphasizes that actual power may vary.
  • A later reply discusses the complexity of measuring power dissipation for multiple components on a motherboard, suggesting the use of TDP values and practical measurement techniques for individual components.

Areas of Agreement / Disagreement

Participants express a range of views on how to calculate heat dissipation, with no consensus on a single method. Some agree on the utility of TDP values, while others emphasize the need for practical measurement techniques. The discussion remains unresolved regarding the best approach to accurately determine power dissipation.

Contextual Notes

Limitations include the dependence on TDP values, potential inaccuracies in measurements, and the complexity of accounting for inefficiencies in power supply systems. There are also unresolved questions about the specific conditions under which measurements should be taken.

no183
Messages
15
Reaction score
0
Hey guys. i am on a little project here, may i know hot to calculate heat dissipation from processor and HDD ? is it using P=I^2R will do ? if so what is my R and I?
 
Engineering news on Phys.org


no183 said:
Hey guys. i am on a little project here, may i know hot to calculate heat dissipation from processor and HDD ? is it using P=I^2R will do ? if so what is my R and I?

The better way is to look at the power supply current and voltage for the parts you are interested in. Get their power dissipated from their input I and V.
 


hmm, let say a processor, while it has the state of load and idle, how do i calculate it?
 


no183 said:
hmm, let say a processor, while it has the state of load and idle, how do i calculate it?

Look it up in the datasheet for the processor.
 


hmm, where do i need to get that btw?
 


From the manufacturers website - it can't be calculated from scratch.
 


Yeah, Berke - that's the reason PC processors have basically stopped getting faster. People aren't going to buy a pc with a water cooled chiller attached to it.
 
  • #10
Based on my time working as an engineer working for http://www.coolitsystems.com/" , here's what I've learned:

This is/ can be actually really difficult to determine depending on what your reasons for wanting to know what the power of a proc. is.

If you don't need a terribly accurate guess, say within 15% going by the thermal design power (TDP) of the processor will probably be fine. Note, this number will only be valid when the processor is loaded at 100%, even then the actual power will be a little bit lower than the TDP.

If, however, you need a more accurate power or you have overclocked the processor, or it isn't running at 100% then you need to get crafty.

You can get a more accurate measure of the power of the processor by probing the current going into the EPS rail of the motherboard. There will be a number of yellow and black wires going from the power supply to the motherboard near the processor. You need to cut the mesh sleeve of the wires and use a clamp meter on the yellow wire, make sure you get all of them and none of the black wires (this is the reason for cutting the sleeve). This supplies 12V to the switching power supply that steps to voltage down to about 1.3V for use by the proc.

Of course, this number won't be the actual power going into the processor, you need to account for inefficiencies in the aforementioned switching power supply, say 85%-90%. This will get you to within about 5% of the actual power dissipation.

If you need to get any more accurate than this you would need to probe the current of every single ground pin on the processor. This is not a practical solution.

Depending on the processor expect power level from about 50W for a low power desktop processor all the way up to 220W for a high end desktop processor.
 
Last edited by a moderator:
  • #11
looks like i found the correct person, is like this, i want to do a finite element analysis on heat transfer at the motherboard. the heat dissipated involve processor, N/S Bridge heat sink, RAM and VGA card, other negligible.. any idea?
 
  • #12
That's a stinky problem.

The only thing you can do is take the TDP values for each of the parts independently. Unfortunately, there isn't really a good way to measure the actual power of each. You can however measure the power of every element of the motherboard together. To do this, cut the sleeve on the large motherboard power connector and use a clamp meter on all the red, yellow and orange wires in the bundle. The red ones are 5V, yellow is 12V and orange is 3.3V. You have to measure the current of each rail individually to calculate the total power.

The VGA can be done in a similar fashion as the processor if it is a medium to high power card, there will be a cable going to the end of the card furthest from the I/O (it might also be on the side near the end). Cut the sleeve on that and probe the yellow wires. Again, they will be 12V. For a low end card (that doesn't have this cable) all you can know is that it will be less than 75W (the power limit for a PCI-E bus)

Sorry I can't offer anything more than that though.
 

Similar threads

  • · Replies 53 ·
2
Replies
53
Views
5K
Replies
11
Views
3K
  • · Replies 23 ·
Replies
23
Views
8K
  • · Replies 15 ·
Replies
15
Views
3K
  • · Replies 30 ·
2
Replies
30
Views
2K
  • · Replies 21 ·
Replies
21
Views
8K
Replies
8
Views
3K
  • · Replies 14 ·
Replies
14
Views
5K
  • · Replies 10 ·
Replies
10
Views
2K
  • · Replies 5 ·
Replies
5
Views
6K