Sign in with
Sign up | Sign in
Your question

Air cooling goals

Tags:
  • Heatsinks
  • CPUs
  • Cooling
  • Cases
  • Temperature
  • Overclocking
Last response: in Overclocking
Share
April 22, 2007 10:00:06 PM

Would it be fair to say that if you current dual core AMD cpu is idling at a temperature that is within 1 degree celcius of what your Case temperature is being reported as, then I dont care how good your HSF is, its not going to get much better than that?

That is to say, isnt it pretty impossible for your CPU to idle at a temperature which is lower than your Case temperature on air? Since the CPU is being cooled by case air blowing over the fins on the heat sink, I dont see how it can be expected to cool to any more than the actual temperature of the air blowing over the heat sink.

Im only talking about idle, now, because its obvious that at full load, were not going to be anywere Near the case temperature.. But I still find it useful to look at idle temperatures, because the cooler your idle temperature, the cooler will be your full load temperature. There is a delta which is pretty constant for any given setup. That is to say, if youre measuring a delta of 20 degrees between your idle temp and your full load temp, and your idle/full load temps are 30/50, then if you lower your idle to 25 degrees, your full load temp will likely drop to 45 degrees. The reason your idle temp could be lowered from 30 degrees to 25 is because your Case temp has also cooled.. That is a fairly constant relationship also.

In the end, to all those chasing miracles, whether it be with AS5, or whatever cooler du jour your using, if you are running your cpu idle at the same temp your Case temp, youre doing allright!!

More about : air cooling goals

April 22, 2007 11:13:42 PM

Thanks for that....a lovely piece of random unsolicited advice.
April 22, 2007 11:41:36 PM

Quote:
Thanks for that....a lovely piece of random unsolicited advice.


And, thats exactly what yours is.. Unsolicited advice! thanks, but no thanks.. AT least, mine was in response to alot of threads Ive been reading from seemingly well minded folks trying to cool beyond that which is possible.. If you dont dervive any benefit from my post, then why just dont you leave it alone.. Do you feel compelled to comment on Everything that you find no personal use in? You must be one very busy commenter!!
Related resources
a b à CPUs
a b K Overclocking
April 23, 2007 12:16:44 AM

I'm curious, what is your CPU's temps. I'm just wondering since both of our CPUs are at the same clock speeds, though they are very different CPUs. My idle is 29-30C, my max is 47-50C with the setup in my sig and a very well ventilated case.
April 23, 2007 1:16:58 AM

Quote:
Would it be fair to say that if you current dual core AMD cpu is idling at a temperature that is within 1 degree celcius of what your Case temperature is being reported as, then I dont care how good your HSF is, its not going to get much better than that?

That is to say, isnt it pretty impossible for your CPU to idle at a temperature which is lower than your Case temperature on air? Since the CPU is being cooled by case air blowing over the fins on the heat sink, I dont see how it can be expected to cool to any more than the actual temperature of the air blowing over the heat sink.

Im only talking about idle, now, because its obvious that at full load, were not going to be anywere Near the case temperature.. But I still find it useful to look at idle temperatures, because the cooler your idle temperature, the cooler will be your full load temperature. There is a delta which is pretty constant for any given setup. That is to say, if youre measuring a delta of 20 degrees between your idle temp and your full load temp, and your idle/full load temps are 30/50, then if you lower your idle to 25 degrees, your full load temp will likely drop to 45 degrees. The reason your idle temp could be lowered from 30 degrees to 25 is because your Case temp has also cooled.. That is a fairly constant relationship also.

In the end, to all those chasing miracles, whether it be with AS5, or whatever cooler du jour your using, if you are running your cpu idle at the same temp your Case temp, youre doing allright!!


huh...ok... what is the objective of this thread?
April 23, 2007 1:52:30 AM

Quote:
Thanks for that....a lovely piece of random unsolicited advice.


Umm, this is a comment, not advice.
April 23, 2007 2:12:42 AM

Quote:
Would it be fair to say that if you current dual core AMD cpu is idling at a temperature that is within 1 degree celcius of what your Case temperature is being reported as, then I dont care how good your HSF is, its not going to get much better than that?

That is to say, isnt it pretty impossible for your CPU to idle at a temperature which is lower than your Case temperature on air? Since the CPU is being cooled by case air blowing over the fins on the heat sink, I dont see how it can be expected to cool to any more than the actual temperature of the air blowing over the heat sink.

Im only talking about idle, now, because its obvious that at full load, were not going to be anywere Near the case temperature.. But I still find it useful to look at idle temperatures, because the cooler your idle temperature, the cooler will be your full load temperature. There is a delta which is pretty constant for any given setup. That is to say, if youre measuring a delta of 20 degrees between your idle temp and your full load temp, and your idle/full load temps are 30/50, then if you lower your idle to 25 degrees, your full load temp will likely drop to 45 degrees. The reason your idle temp could be lowered from 30 degrees to 25 is because your Case temp has also cooled.. That is a fairly constant relationship also.

In the end, to all those chasing miracles, whether it be with AS5, or whatever cooler du jour your using, if you are running your cpu idle at the same temp your Case temp, youre doing allright!!


huh...ok... what is the objective of this thread?

There isn't one....
April 23, 2007 3:01:42 AM

Quote:
I'm curious, what is your CPU's temps. I'm just wondering since both of our CPUs are at the same clock speeds, though they are very different CPUs. My idle is 29-30C, my max is 47-50C with the setup in my sig and a very well ventilated case.


At idle, if my case is 28 which it often is, my CPU is 28. At 100% load both cores, its around 50 with full case fully closed. Thats with a masscool HSF which is strictly OE like at best.. Yet, its doing a pretty good job at protecting my overclock, so I cant say im unhappy with it.
April 23, 2007 3:47:24 AM

The delta will be smaller for better heatsinks. It's when the cpu is under full load that the higher end coolers get to shine. They have the ability to dissipate MORE heat. Just look at the reviews and comparisons of coolers, there is a reason some dominate (and it's not the case temp).
April 23, 2007 3:53:44 AM

What are you using to measure CPU and case temperatures? What are the readings you're getting?

I ask specifically because measurement and true value are never the same thing. You have to be aware of sources of error, which generally fall into two categories: precision and bias. It is not difficult to build a temperature measurement sensor that has a very small precision error, but I suspect that the CPU temp readings are subject to a great deal of bias.

It's also important to keep some terminology in mind. Tcase would be the temperature of the CPU heat spreader. This temperature will be lower than the maximum temperature within the CPU. In the same vein, it's important to know where the temperature is being measured. Temperatures reported by Speedfan are generally taken from a motherboard sensor, which is going to be even cooler than Tcase.

The whole point of the above is that the idle CPU temperature could very well be in the 30s while the temperature you're reading says 25. When OCing my chips, I tend to leave a 10C margin of error to the temperature reading. I never let my CPU temp go into the 60s, and I prefer (and do) keep my temps in the mid-40s. For interest, I'm using a PD820@3.36GHz and a X2 3600+@2.09GHz (limited by M2A-VM motherboard).
April 23, 2007 3:55:13 AM

Quote:
The delta will be smaller for better heatsinks. It's when the cpu is under full load that the higher end coolers get to shine. They have the ability to dissipate MORE heat. Just look at the reviews and comparisons of coolers, there is a reason some dominate (and it's not the case temp).


I agree. On the other hand, Ive seen alot of other posts from guys who are very pleased with their high end HSFs being able to keep them at a peak of 50 degrees on their amd X2 100% loaded dual cores, which is what I achieve with the 14 dollar masscool, with an idle near 30, which is simlar to what they report. Got me thinking maybe some of whats at work in making people think AS5 is capbable of things superman would be proud to have attained is at work here too.

>>What are you using to measure CPU and case temperatures? What are the readings you're getting?<<

I use Speedfan, which is confirmed by my mobos own utilities from Biostar, just to confirm their values. The biostar utilities are called Fan Conditioner.. The 'Case Temperature' which is how Speedfan refers to it is reported by teh Biostar utiliity as 'System Temperaure', and must be reading a diode somewhere on the mobo, reported on a chip called IT8712F . At idle, it makes sense to me that the cpu temperature is mirroring this system temp very well.. Sys temp goes up a degree, and so does my cpu's idle temp.. If I open a window and bring cooler air into the room, system temp goes down a degree or two, and sure enough, the CPU temp is quick to follow.
April 23, 2007 4:08:50 AM

Ah... when I use SpeedFan on my P5LD2, the "System Temp" is the measurement of the northbridge. It goes up when I OC (passively cooled). The NB temp easily goes into the 30s.

Even so, on my X2 3600+, some of the reported CPU temps are just not believable. There are 4 temperatures reported by the CPU itself, and at least one from the motherboard. The motherboard tells me I'm at about 40 full load, while the 4 from the CPU stay in the teens. At idle, the 4 CPU sensors indicate below 10C, which is just ridiculous. There is definitely a calibration issue with those.

Remember, again, that the MB sensor will tell you a temperature a few degrees below the peak CPU temperature. I suspect it is less than 10C off, but I doubt it is less than 5. If you have a thermometer, place that in your case while you're running. I don't know what case you have, but I sure hope your air temperature within your case is somewhere in the 20s.
April 23, 2007 4:44:04 AM

Quote:
Ah... when I use SpeedFan on my P5LD2, the "System Temp" is the measurement of the northbridge. It goes up when I OC (passively cooled). The NB temp easily goes into the 30s.

Even so, on my X2 3600+, some of the reported CPU temps are just not believable. There are 4 temperatures reported by the CPU itself, and at least one from the motherboard. The motherboard tells me I'm at about 40 full load, while the 4 from the CPU stay in the teens. At idle, the 4 CPU sensors indicate below 10C, which is just ridiculous. There is definitely a calibration issue with those.

Remember, again, that the MB sensor will tell you a temperature a few degrees below the peak CPU temperature. I suspect it is less than 10C off, but I doubt it is less than 5. If you have a thermometer, place that in your case while you're running. I don't know what case you have, but I sure hope your air temperature within your case is somewhere in the 20s.



Thats largely correct.. If you use speedfan, you MUST confirm its values with a mobo utility you can trust. For instance, the Core value reported by speedfan for me is totally bogus. At this moment, my System temp is being reported as 29, and my CPU temp is being reported as 30.. Both are confirmed by my Biostar utilities so I know they are correct.. However, the Core temp from speedfan is reading 16 degrees C, so I know this is completely bogus, and never quote it.. Why is it bogus? Apparently, my Brisbane core temp is not being interpreted properly as its calibration changed not long ago, and hasnt yet modified properly in speedfan. But, I have no reason to doubt my Biostar hardware utilities which have done a good job at reporting accurate readings from the mobo and cpu. By the way, the biostar Tutility fan control, and Tutility Hardware control's readings are also confirmed by PC Wizard, and other utilities IVE got.. Theres no doubt that the System temps and the CPU temps that are being reported by all these sources that confirm one another are as close to the truth as IM going to get ...

Ive run Orthos on Small FFts, the best cpu stressor, on both cores, for 8 hrs, and the cpu temp settled on 50-51, back and forth, for the whole time. Nary a hiccup was experienced, and IM quite pleased with the OC result, taking me from a stock 1.9ghz to 2.8 ghz completely stable and wihtout problems, all with this 14 dollar hsf. Regardless of how much they spend on hsfs, all the experienced folks Ive spoken with are quite content with a 50 degree cpu peak using the measurements sources iVe reported.
a b à CPUs
a b K Overclocking
April 23, 2007 12:46:49 PM

Yeah, I'm not 100% sure on my temps, but I think that they are pretty accurate. The max is probably a little higher, I haven't done a good, long test yet (the OC I did was so small I didn't worry about stability issues).
One reason I don't completely trust my MB is it doesn't even recognize my processor (this is with a BIOS that came out after the FX-60). Programs like PCWizard 07 call it a Athlon 6000+ or even call it a mobile processor. My motherboard is just about dead (it already has had a major repair), I just hope it will last another year at least. I don't know if AMD forgot to label my processor, or if the MB can't read it.
!