why is my cpu getting hotter? x2 3800

[d3 npb]

My lapped x2-3800 with a freezer pro 64 or whatever used to run at 24C idle, 37C load, oc'd to 2.6 on air cooling. Ambient temp was actually 25!. Granted, that was before the heat wave we've been having, but now it's running around 42C idle, 49C load. This has been a very, very slow process of the temps increasing over the last 4-5 months or so, however long it has been. I clean my case air filters regularly, and have inspected the hsf visually, and do not see any dust buildup whatsoever. I had considered the possibility of a loose connector/bad interface, but it hasnt been moved, bumped, or anything. Could the clip have loosened up? Might I just need to open it up, alcohol the crap out of it to take all the as5 off and start fresh? I am going to try that, but I just dont understand what would cause this gradual increase in temps, back to stock hsf levels, and dont want to 'fix' it only to have it gradually creep on back up to these high levels.. gotta keep my 2.6 overclock working nice and cool! Is it the interface material? It's not actually as5 it's antec's imitation as5.. (doh!) Thanks

 

[BassBomb]

it has been getting hotter lately, reseating the heatsink is definatley a good idea.. when its off uc an also dust it out for sure too

 

[d3 npb]

yeah i'm gonna dust-off the whole thing but using the double height (for dual fan) filters on the front of my case to keep dust out, I dont have much of a dust issue. I guess I'll just have to hope for the best with a re-seat, and also cry a little since my $300 cpu is so cheap to buy now.

 

[Ayah]

Measure the changes in deltaT(Difference between ambient and idle/load temps) instead of just what the load temp is at the time.
It could be due to a rise in ambient temperature, dust buildup; reducing the effectiveness of airflow and heat dissipation from the heatsink.

Your heatsink might've gotton loose, thermal paste dried, etc. etc.

And even the thermal probe might've worn out or something of that sort.

 

[potato28]

Originally posted by: Ayah
Measure the changes in deltaT(Difference between ambient and idle/load temps) instead of just what the load temp is at the time.
It could be due to a rise in ambient temperature, dust buildup; reducing the effectiveness of airflow and heat dissipation from the heatsink.

Your heatsink might've gotton loose, thermal paste dried, etc. etc.

And even the thermal probe might've worn out or something of that sort.

This is true. The delta for my 3200+ Whinchester is 7C all year-round, even tho during the summer-time the idle is in the mid 40's. A good setup for testing this is Prime95 and Speedfan for 10 minutes.

 

[lopri]

Umm.. I might be making a statement that goes against modern science, but sometimes 25C ambient in January and the same ambient in July could be different.

 

[lopri]

Originally posted by: d3 npb
My lapped x2-3800 with a freezer pro 64 or whatever used to run at 24C idle, 37C load, oc'd to 2.6 on air cooling. Ambient temp was actually 25!

Oh wait, I'm not the only one who goes against science!

 

[d3 npb]

that's not against science.. ever pick up a piece of metal and find that it feels cold, whereas other objects do not? Some objects literally are colder in the same ambient temp. It's entirely possible given the hsf material and airflow to have the idle temp 1 deg less than ambient.. (wind chill factor? i'm sure it applies on some level as well.. pls dont flame me for saying that, just grasp the idea). And since the ambient probe is in a different location, its more likely that the air has less flow there, and therefore more built up heat. (Ambient temp taken from mobo)

Potato and Ayah, thx for the suggestions and info. I'm doing the clean/reset/etc tonight, and will use the difference between ambient and idle/load temps instead of just what the load temp is, and do a comparo with my ac duct pumped into the case =)

 

[plonk420]

i say dust it off. my AXP 2400 was rising higher and higher (was averaging 60-63C this spring)... one day it jumped to 67-69C, so i decided to reseat the HSF. i was taking it apart when i noticed all the dust. brought out the air-in-a-can, gave it a fair number of blasts outside, took it back in, and temps were back around when i'd first reseated it a few years ago (i have to take the entire mobo out; no mobo tray) ~45C. i can procrastinate yet again ^_^


edit: reread message.. WTF?! idle 25, load 37?!? and complaining about load 47-49C? =P only start worrying if it hits 55C+ and save your worries for more important things.... >_>

 

[Operandi]

Originally posted by: d3 npb
that's not against science.. ever pick up a piece of metal and find that it feels cold, whereas other objects do not? Some objects literally are colder in the same ambient temp. It's entirely possible given the hsf material and airflow to have the idle temp 1 deg less than ambient.. (wind chill factor? i'm sure it applies on some level as well.. pls dont flame me for saying that, just grasp the idea). And since the ambient probe is in a different location, its more likely that the air has less flow there, and therefore more built up heat. (Ambient temp taken from mobo)

Potato and Ayah, thx for the suggestions and info. I'm doing the clean/reset/etc tonight, and will use the difference between ambient and idle/load temps instead of just what the load temp is, and do a comparo with my ac duct pumped into the case =)

Wow... I guess you must have skipped/failed basic science or just haven't got there yet.

Your CPU will never be less then idle with air cooling, that's a physical impossibility. If those are the readings you are getting they are obviously flawed and can't be relied on in the first place.

 

[d3 npb]

I assume you mean the cpu will never be at 'less than ambient' and not 'less then idle'. It's not an impossibility, just proven by theory in most normal cases to be as such. In this case, however, as I implied, the ambient temp reported by the mobo probe may have been an elevated number due to lower airflow in that area of the case, and hence more heat buildup, and not 'true' ambient temp reportings.

Also, when you skipped or failed basic english/grammar, it is less 'than' and not less 'then', as you wrote.

Why do people on here and every other BB insist on being such asses? Ask for help and get picked apart. Thanks to those who truly intend to help.

 

[Ayah]

Let us know if the temps drop (or even go up).

 

[MadScientist]

The AC Freezer 64 Pro HS is easy to install, but the reviewer at Frosty Tech found that it's easy to misalign the the 3 plastic black tabs on the RM with the rention clip on the HS.
With my MB and case I found it was impossible to see if the retention clip had fully engaged the plastic tabs, since it was so close to the bottom of the PSU. My next case will have a slide out MB tray.
I would remount your HS and check that all 3 plastic tabs are engaged.

 

[d3 npb]

per Ayah: Measure the changes in deltaT(Difference between ambient and idle/load temps) instead of just what the load temp is at the time.

At the time of my initial measurements the delta was 15, and now is 22 1/3. Whether The increase in ambient creates a non-linear increase in that measurement, I've no clue.
Ambient of original: 22.2 C
Ambient of last test: 26 C

Anyhow, this is a dead issue for now.. if anyone cares to know the results, I'll post them tomorrow.

 

[Operandi]

Originally posted by: d3 npb
I assume you mean the cpu will never be at 'less than ambient' and not 'less then idle'. It's not an impossibility, just proven by theory in most normal cases to be as such. In this case, however, as I implied, the ambient temp reported by the mobo probe may have been an elevated number due to lower airflow in that area of the case, and hence more heat buildup, and not 'true' ambient temp reportings.

Also, when you skipped or failed basic english/grammar, it is less 'than' and not less 'then', as you wrote.

Why do people on here and every other BB insist on being such asses? Ask for help and get picked apart. Thanks to those who truly intend to help.

It's not a theory; it's a law of physics. The CPU (assuming air cooling) will never be cooler then ambient that surrounds it, all the CFM will never change that.

 

[flashbacck]

Originally posted by: Operandi

Originally posted by: d3 npb
I assume you mean the cpu will never be at 'less than ambient' and not 'less then idle'. It's not an impossibility, just proven by theory in most normal cases to be as such. In this case, however, as I implied, the ambient temp reported by the mobo probe may have been an elevated number due to lower airflow in that area of the case, and hence more heat buildup, and not 'true' ambient temp reportings.

Also, when you skipped or failed basic english/grammar, it is less 'than' and not less 'then', as you wrote.

Why do people on here and every other BB insist on being such asses? Ask for help and get picked apart. Thanks to those who truly intend to help.

It's not a theory; it's a law of physics. The CPU (assuming air cooling) will never be cooler then ambient that surrounds it, all the CFM will never change that.

QFT. Even with water cooling, the CPU will not get colder than ambient.

that's not against science.. ever pick up a piece of metal and find that it feels cold, whereas other objects do not? Some objects literally are colder in the same ambient temp.

The reason metal feels cooler is because it draws heat from your hand better than other objects.

 

[lumbus]

I have a similar problem with my X2 4200+. I built it almost exactly a year ago, so room temperature should've been about the same. It was very quiet to begin with, but over the last few months, the stock cooler has started to spin a lot faster creating much more noise... 5,500 - 6,000 rpm! This is at a CPU temperature of 40 C idle, which is more than before, but still not dangerously high or anything... Is it normal for it to be spinning that fast at a temperature that low?

I am a little annoyed with my NeoPower PSU for being very conservative with its fan RPM. The PSU seems very hot to the touch and probably contributes a lot to raising the case temperature (and it's right next to the CPU of course). The case is an Antec Super Lanboy.

I've been too lazy to reseat the HSF so far as it looks like it's clean and attached well... But I guess I will have to do that too.

 

[Ayah]

Originally posted by: lumbus
I have a similar problem with my X2 4200+. I built it almost exactly a year ago, so room temperature should've been about the same. It was very quiet to begin with, but over the last few months, the stock cooler has started to spin a lot faster creating much more noise... 5,500 - 6,000 rpm! This is at a CPU temperature of 40 C idle, which is more than before, but still not dangerously high or anything... Is it normal for it to be spinning that fast at a temperature that low?

I am a little annoyed with my NeoPower PSU for being very conservative with its fan RPM. The PSU seems very hot to the touch and probably contributes a lot to raising the case temperature (and it's right next to the CPU of course). The case is an Antec Super Lanboy.

I've been too lazy to reseat the HSF so far as it looks like it's clean and attached well... But I guess I will have to do that too.

Have you enabled CnQ' through your BIOS?
You should make sure that there isn't dust clogging up the fan, heatsink, vents, etc.
I don't remember exactly about the connectors on the SLB's two fans. If they're 4 pin molex, you should connect them to a standard molex instead of the ghetto "fan only" connectors that the NeoPower provides. The NeoPower is a good PSU but it's fan control is seriously lacking.

 

[Howard]

Originally posted by: flashbacck

Originally posted by: Operandi

Originally posted by: d3 npb
I assume you mean the cpu will never be at 'less than ambient' and not 'less then idle'. It's not an impossibility, just proven by theory in most normal cases to be as such. In this case, however, as I implied, the ambient temp reported by the mobo probe may have been an elevated number due to lower airflow in that area of the case, and hence more heat buildup, and not 'true' ambient temp reportings.

Also, when you skipped or failed basic english/grammar, it is less 'than' and not less 'then', as you wrote.

Why do people on here and every other BB insist on being such asses? Ask for help and get picked apart. Thanks to those who truly intend to help.

It's not a theory; it's a law of physics. The CPU (assuming air cooling) will never be cooler then ambient that surrounds it, all the CFM will never change that.

QFT. Even with water cooling, the CPU will not get colder than ambient.

that's not against science.. ever pick up a piece of metal and find that it feels cold, whereas other objects do not? Some objects literally are colder in the same ambient temp.

The reason metal feels cooler is because it draws heat from your hand better than other objects.

Correct. It's a function of thermal conductivity and heat capacity, I believe.

 

[lumbus]

Originally posted by: Ayah

Originally posted by: lumbus
I have a similar problem with my X2 4200+. I built it almost exactly a year ago, so room temperature should've been about the same. It was very quiet to begin with, but over the last few months, the stock cooler has started to spin a lot faster creating much more noise... 5,500 - 6,000 rpm! This is at a CPU temperature of 40 C idle, which is more than before, but still not dangerously high or anything... Is it normal for it to be spinning that fast at a temperature that low?

I am a little annoyed with my NeoPower PSU for being very conservative with its fan RPM. The PSU seems very hot to the touch and probably contributes a lot to raising the case temperature (and it's right next to the CPU of course). The case is an Antec Super Lanboy.

I've been too lazy to reseat the HSF so far as it looks like it's clean and attached well... But I guess I will have to do that too.

Have you enabled CnQ' through your BIOS?
You should make sure that there isn't dust clogging up the fan, heatsink, vents, etc.
I don't remember exactly about the connectors on the SLB's two fans. If they're 4 pin molex, you should connect them to a standard molex instead of the ghetto "fan only" connectors that the NeoPower provides. The NeoPower is a good PSU but it's fan control is seriously lacking.

Thanks for the suggestions! I forgot to mention that I enabled C'n'Q in BIOS and Windows the other day. It works as a charm as far as controlling frequency and voltage, but it didn't do crap for temperature or RPM (surprisingly).

Then, just now, I noticed that Q-Fan was disabled in BIOS! Doh! I don't know how that happened, it must have reset itself at some point. The RPM instantly dropped to significantly and now it's idling at 2,100 RPM, still at 40C. My CPU temperature seems to be a constant that doesn't depend on the fan RPM. No, I'm sure it would be a different story under load (just getting ready to test that).

BTW, it's amazing how much halfing the RPM does for the noise level, I've noticed that before. The noise must be an exponential function of the RPM. Is there a rough formula or does it depend on the fan?

 

[Ayah]

Depends ALOT on the fan. The motor, the positioning, the attachment of the fan, any objects in the way of the airflow, the blades (angle/size/shape, etc.)

 

[lumbus]

Originally posted by: lumbusThen, just now, I noticed that Q-Fan was disabled in BIOS! Doh! I don't know how that happened, it must have reset itself at some point. The RPM instantly dropped to significantly and now it's idling at 2,100 RPM, still at 40C. My CPU temperature seems to be a constant that doesn't depend on the fan RPM. No, I'm sure it would be a different story under load (just getting ready to test that).

FULL LOAD:
Room temp: 84 F / 29 C
CPU temp: 52 C
Fan RPM: 4,200

 

[d3 npb]

yah yah i know its a law of physics and I know why metal feels colder.. but what you keep ignoring is the fact that I'm not talking about the ambient temp surrounding the cpu, but the ambient temp at the thermistor labelled as the 'case' or 'ambient' temp. Also if you see my prev posts you'll see that the 'true' (room) ambient was around 22C, which certainly allows for the case's ambient to be at 25C and the CPU at 24C, given differences in air circulation. Stop trying to correct me and just help or move along.

Thanks~

 

[Operandi]

Originally posted by: d3 npb
yah yah i know its a law of physics and I know why metal feels colder.. but what you keep ignoring is the fact that I'm not talking about the ambient temp surrounding the cpu, but the ambient temp at the thermistor labelled as the 'case' or 'ambient' temp. Also if you see my prev posts you'll see that the 'true' (room) ambient was around 22C, which certainly allows for the case's ambient to be at 25C and the CPU at 24C, given differences in air circulation. Stop trying to correct me and just help or move along.

Thanks~

Unless your CPU is outside the case it's not possible.

You can believe whatever you want, I won't post in here anymore.

 

[d3 npb]

So you say it's not possible to have 22C air pumping into a case, and have a thermistor in a low airflow area soaking up dissipated heat to 25C, and have my cpu cooled by the 22C air to a nice 24C? Given that there are 2 80mm fans blowing air into the case towards the cpu, the large hsf fan (perpendicular to mobo, assisting airflow in direction towards back of case) and the 120mm fan directly behind the cpu/hsf sucking out hot air, I think this is entirely possible. And my cpu is actually INSIDE the case. What are you smoking? Its possible my cpu could be 22 and the ambient 25, due to the airflow. Whatever.