c/w POLL: A Brainstorm.

Following the success of the CPU Temperatures POLL I have decided a c/w POLL would be much more accurate.
But before I decide to post one I thought I'd let everyone contribute there ideas in a sort of Brainstorming session.

For instance do you want Diode/Thermistor options in the POLL as well?

Bearing in mind the limitations are 10 items per POLL.
I'll look forward to hearing your ideas!

M_N

Good idea

Make sure you put a good explination of exacty what c/w is

NO ideas?

We can't do a poll on C/W. Nobody can give the correct wattage values for their CPUs, guesstimates are no good & online conversion utilities are as accurate as your weekly guess on the LOTTO, so by doing this it will just confuse the new comers & create more fiction than fact.

The more inaccurate figures you input into the formula the further it is from the truth so lets just stay with the CPU temp readings being the half truths they already are.

Sonny said:

We can't do a poll on C/W. Nobody can give the correct wattage values for their CPUs, guesstimates are no good & online conversion utilities are as accurate as your weekly guess on the LOTTO, so by doing this it will just confuse the new comers & create more fiction than fact.

The more inaccurate figures you input into the formula the further it is from the truth so lets just stay with the CPU temp readings being the half truths they already are.

Click to expand...

Good point, and one reason why I posted this Thread.
However inaccurate they may be it would be interesting to have a look at (like the standard temperatures) and if we find a medium of measuring then at least the variations will be reasonably accurate.

If anyone wants the Poll I could put it together, whether its any use or not is another question.

M_N

How do you measure C/W, anyway?

I'm interested in this pole. I think that although we wouldn't be able to achieve accurate results, we could achieve precise results. For instance, we could all decide on a single conversion tool/utility to calculate the power output (wattage) of our processors. We could also emphasize that our cpu temp is taken after an hour of priming or something like that.

There is the problem, however, of achieving precise ambient temperature readings, plus the problem of inaccurate motherboard temperature readings. People would have to take their best guess at the actual temp of their processor, based on what others have found, using the same board.

If we could find a way to minimize these problems, I would be all for it.

-------------------------------------------------------------------
Nerdlogic, to measure C/W, you need to know the ambient temperature (the temperature of the air (asuming you're using air cooling) that enters you heatsink's fan), your CPU's temperature, and your CPU's wattage at that given temperature. After you look up on some data sheets and find a somewhat reliable value for your CPU's wattage under load, and after you record your cpu's temp under load, you're ready to go. You take the difference between the ambient air temperautre and your cpu load temperature ("delta t") and divide it by the wattage of your processor under load. The value you get is your HSF's °C/W rating, or at least it will give you a ballpark idea of what it is.

Wouldn't there be a number of factors that affect the CPU's wattage? Is there any other way to get this reading than looking it up on a chart?

Maximus Nickus said:

However inaccurate they may be it would be interesting to have a look at (like the standard temperatures) and if we find a medium of measuring then at least the variations will be reasonably accurate.

Click to expand...

With the differing ways of measuring temps, the temp diode still can't be calibrated reasonably well & thermistors are just plain funny, add to that the inconsistent actual voltage being run to the CPU with the "I got lower temps after reseating my HSF" routine is there really some order to this confusion?

I know I'm being really negative about this but this will confuse a lot of begginers & a good number of more experienced users. We already had a spell with people linking to a site that supposedly calculated your C/W & getting twice the performance of what JoeC. was getting with more accurate testing.

Cheers Nick C.

Rezin777 said:

Wouldn't there be a number of factors that affect the CPU's wattage? Is there any other way to get this reading than looking it up on a chart?

Click to expand...

Yes, there are. There are some sites out there with wattage "calculators" that let you enter in specific parameters like your cpu model, the voltage you're running it at, and the the speed it's running at. However, these are considered by most to be highly inaccurate, and with that said, so are most of the charts out there.

The more I think about this pole, the more I realize that the results just wouldn't mean anything. There would just be too many inconsistencies. Arg, I liked this idea though...

nerdlogic,
C/W is typically not a direct measurement but a calculation based on the wattage and the temperature rise of the device in question.
If your CPU dissipates 50 watts and the temperature rise on the CPU is 10 degrees then your C/W is 0.2 degrees/W.
I hope this helps.

LA

This thread discusses an accurate measure of wattage calculation. Even if wattage can't be calculated with 100% accuracy, even approximate C/W ratings can be very useful. As a rule of thumb, your processor's wattage is going to be no lower than 50, and no higher than 150. A rather large range, however many calculation programs report them to exceed this range by far. Just my $.02.