CPU internal thermal diode calibration

I found the following reference about calibrating the temperature measurement from the thermal diode of the CPU.
http://www.arcticsilver.com/diode_calibration.htm
I have applied the described method to calibrate the CPU diode (Temperature Sensor 2: PII diode in Speedfan) on my K7N2-Delta2-LSR. The offset was 7C, i.e. it was measuring 7C below the actual temperature. Just wanted to let others know, although the offset value may not be the same on every K7N2-Delta2-LSR, you can repeat the procedure yourself to do your own calibration. I used the system temperature (temperature sensor 1) as inlet temparature to my fan (that may not be accurate too but, I think, it is still a good estimate for inlet temperature). For the the two different CPU speeds, I used 12 and 8 multipliers at 200 FSB, 1.65 Vcore, and fan speed at ~4000rpm, i.e. ran the CPU at 2.4 and 1.6 GHz for the two measurements. Prime95 was used to load the CPU and temparature was monitored with Speedfan.
After calibration (i.e. 7C offset), my idle CPU temp is 40C and loaded 57C at 2.4 GHz. These temperatures are, I think, completely acceptable with a cheap cooler.
The relevant specs of this system is the following.
MSI K7N2-Delta2-LSR (Bios B.5)
Athlon XP-M 2500 -> OC: 2.4 GHz (12x200, 1.65v)
Kingwin KCU-7015 cooler with Arctic Silver 5
Single channel 512 MB PC3200 DDR 2.5-3-3-8 (running as DDR400)
300W Generic PSU
I will also apply the same method to my other system with the following specs. Let me know if anyone wants to know the results for that system too.
MSI GNB-Max-FISR
Pentium 3.06HT -> OC:3.45 (23x150, 1.55v)
Zalman CNPS-7000A-AlCu cooler with Arctic Silver 5
Dual channel 1 GB PC2700 DDR 2-3-3-6 (running as DDR300)
350W PSU Fortron
Cheers

Brichy,
Here are the calibration data for GNB-Max. Please consider that your calibration could be slightly different.
The offset was -2C, i.e. Speedfan was reading 2C above the actual temperature. Again, I used the system temperature as inlet temperature to the fan, this may not be exactly accurate but should still be a good estimate for inlet temperature. Given that GNB-Max is not the most overclocking friendly motherboard (but the most stable I have ever owned), I used FSB=133 and FSB=150, with 1.55 Vcore, and fan speed at ~2500rpm, i.e. ran the CPU at 3.067 and 3.45 GHz for the two measurements. Prime95 was used to load the CPU and temparature was monitored with Speedfan as before.
After calibration (i.e. -2C offset), my idle CPU temp is 39C and loaded 60C at 3.45 GHz. Interestingly, BIOS is giving 4C more than what Speedfan was reading before calibration, i.e. it is off (higher) by 6C.
The system specs are as follows. Please let me know if more info needed.
MSI GNB-Max-FISR
Pentium 3.06HT -> OC:3.45 (23x150, 1.55v)
Zalman CNPS-7000A-AlCu cooler with Arctic Silver 5
Dual channel 1 GB PC2700 DDR 2-3-3-6 (running as DDR300)
350W PSU Fortron

Similar Messages

KT3 Ultra2 & AthlonXP thermal diode

Hello,
I'm not quite sure whether the board does really read the internal cpu temperature from the thermal diode or not. The board reviews I've read didn't come to a clear solution on this topic.
So I'd like to know
a) if the temperatures read by the winbond chip are 'correct' and
b) if the 'auto shutdown on high cpu temperature' bios option works as it's supposed to.
Thanks,
KAi

Well,
this is what I found on mbreview.com :
Quote
Starting with the socket, you'll notice that it lacks a temperature probe we've all gotten used to. Could this mean that for Athlon XPs, we can finally make use of the internal thermal diode? Not exactly. Temperatures are not read off the diode, but rather, the socket itself monitors the CPU, much like the old days. Although the readings aren't 100% accurate, they are closer to the true temperature of the CPU.
If this is correct it would mean that the 'thermal protection circuit' uses the XP thermal diode and the hardware monitor function uses some other (external) sensor.
Can we expect this to change with a new bios version? I don't see why it shouldn't be possible to display the internal temperature when it's possible to set the max. temp for auto shutdown.
Any official statement on this? (I'm mainly interested in the KT3 Ultra2 if this makes a difference)

Ondie Thermal Diode on K7N2

Hi,
I've got an MSI K7N2 Nforce2 Board here, with an AMD XP2200+ installed.
I want to check that the temp reading I am getting in MBM5 is the one from the XP's core-internal thermal diode.
Also, I have three measurements in MBM5, is one of them CPU, one case, and one Northbridge?
And finally.... Does the board have a motherboard based thermal monitoring diode (under the chip type thing)...as I couldn't find one and was wondering if it was just too small to see, or absent?
Thanks for your time.

Quote
Originally posted by JoeF
Hi,
I've got an MSI K7N2 Nforce2 Board here, with an AMD XP2200+ installed.
I want to check that the temp reading I am getting in MBM5 is the one from the XP's core-internal thermal diode.
Also, I have three measurements in MBM5, is one of them CPU, one case, and one Northbridge?
And finally.... Does the board have a motherboard based thermal monitoring diode (under the chip type thing)...as I couldn't find one and was wondering if it was just too small to see, or absent?
Thanks for your time.
You should be able to check in your bios for correct temperature readings--probably under "PC Health" or some other designation in the bios.
MBM does not yet support nf2 boards from what I gather.
My own motherboard comes with its own hardware and software monitoring, so it's no problem form me. One thing I have noticed in reading various forums, some nf2 boards seem not to utilize the on-chip diode but rather use socket thermastores instead (I believe the Asus boards do this, or at least that they use both a socket therm & on-chip diode--sure about the socket therm but don't know whether the Asus boards make good use of the on-chip diode, frankly.)
At any rate, if your chip is monitored via the on-chip diode you *will see* temps which average 8-12 degrees C *above* what you saw with previous socket therms, so don't be alarmed. The fact is that the on-chip diode is providing the correct core temp info, whereas the older socket thermastores were not.

Thermal Diode vs Thermal sensor ???

Hi there,
i have just fitted a coolermaster cooldrive to my pc. I am using the included thermal sensor wires to monitor the surface temo of my WD 1200JB hard drive. I am consistantly receiving at least 5 degree higher temp readouts from SpeedFan over the integrated thermal wires of the coolermaster unit on the hard drive. Ok so here it is.. which is the correct temp ? And where is the thermal diode sourcing its temps from the hard drive.. ?? Inetrnally or what ???
Thanking all in advance...

Quote
Originally posted by BigPoppa
Speedfan does def pick up HDD temps. I have mine showing the HDD temp.
I am trying to establish which temp is more accurate, Speedfans Thermal Diode or the thermal sensor wire that i have attached to the upper surface of the HDD and that is being monitored by the coolermaster unit.
+
Yes speedfan does give temps for the HD's but I'm not sure where it senses it from, I would tend to think that your thermal diode would be more acurate.

Are the NEO2 boards using internal CPU thermal diodes?

Temps change very rapidly and are high from start up but dropp after a while. Is it safe to assume that the temperatures is beeing read from the acctual cpu, and not the die?

The winbond temp sensor IC is updated from thermal registers in the A64 cpu.
And yes it's kind of internal diode reading on a64 systems .
Any work the cpu does will cause a tempeature change immediatly .
It's up to the temp sensor windows application to how often the winbond IC
is query'd to give a temp readout in windows.
Looking at the temp in bios it's updated every second .
It's definatly not like the NF2 motherboards that had a sensor diode in the cpu socket that needed time to get heated up from the
underside of the cpu.

CPU fan/thermal pad

I am trying to purchase a new thermal pad, but there are so many. I am trying to pull up a picture of the fan from the HP parts store website but am un able to do so. Does anyone have a picture of the CPU fan for an HP Pavillion dv7 model # 7135us. The part # is 682100-001. Thanks Janelle

With a Duron 800 and a Thermalright SK-5 and a 60 MM fan blowing about 12 CFM i was sitting around 120 F. I installed a Valcano 9 and the temp whent up to 131 F. I thougth i installed it wrong so i took it off, cleaned it and re aplied some Arctic Alumnia and the temp was stilla round 130 F. I took it off, put the SK-5 back on and it dropped to around 124 F. I took the 80 MM fan off the Valcano 9 and placed it on the SK-5. The temp whent down to around 118 F. Needless to say the Valcano 9 heatsink is a peice of . Or my first generation old ass SK-5 is badder asser. But i dought that seeing the I was getting the same temps with a $8 CoolerMaster.
Now with my Athlon 1700+ @ 195x10 I hover around 127 - 131 F with the Thermaltake fan from the Valcano 9 at its lowest setting, compleatly silent, spinning about 2500 RPM's.
I attached the sensor from the Valcano 9 to the bottem of the Proc whith some thermal tape. With a little re-wiring i ran it into my Vantec NXP 201. What ever it reports is usally around 3 - 5 C lower then the Bios. Usually around 48 C. I thought the Bios temps where way off. But now lookign at it i'm not so sure they are. If the temp in the Bios is being taken fromthe core of the Proc and the temp on the Vantec Nesus is being taken from the bottom of the proc, and theres only a free degrees difrence, i'm not sure the Bios is wrong.
This is on an ILSR BTW..

CPU B Thermal Sensor Weirdness

I'm posting this again as there was some coding in the text of the applefan file that wouldn't allow me to post it. Moderators, please delete the other post.
I have an overheating problem that I'm working on. I have a PowerPC G5 Dual 2Ghz/4Gb Ram/Raptor boot disk/Photoshop scratch disk/Radeon 9800 Pro. About a month ago, the power supply failed (for the second time in three years!). After I replaced it, the computer would repeatedly freeze and had to be restarted via the power button. Here is what I've done so far:
Wiped boot drive and upgraded to Leopard;
Ran various utilities, ie, Disk Utility, Tech Tool (no problems reported);
Ran Diskwarrior (no problems reported);
Here is where it gets interesting:
Ran Apple Hardware Quick Test (no problems reported);
Ran Apple Hardware Extended Test (received error message "25TH/1/3: CPU B AD7417AD1");
Installed iStat Pro and Temperature Monitor;
Reset the SMU;
Looked into System/Library/Extentions/AppleFan.kext and inside the package's contents I found a folder called 'MacOS' and inside is a file called 'AppleFan'. I've deleted a huge amount of gibberish, but this is what is inside the file:
AppleFanfan-polling-periodfan-speed-tablefan-speedup-delayfan-slowdown-delayfan- hysteresis-tempplatform-getTempcompatibleadm1030AppleFan::doI2CWrite write failed, retrying...
AppleFan::doI2CWrite cannot write to I2C!!
AppleFan::setProperties SOMETHING IS VERY WRONGdefault-paramsAppleFan::getCPUTemp bad params!!
AppleFan::getCPUTemp failed to retreive CPU temp!!
AppleFan::doI2CRead read failed, retrying...
AppleFan::doI2CRead cannot read from I2C!!
AppleFan failed to open bus for setting fan speed
AppleFan failed to write to Tmin/Trange register!
AppleFan failed to write to fan speed register
AppleFan::getRemoteTemp bad arguments
AppleFan::getRemoteTemp cannot open I2C!!
AppleFan::getRemoteTemp failed to read temperature reg!
AppleFan::setRestartMode failed to open bus
AppleFan::sPMNotify invalid target
AppleFan::saveADM1030State failed to open bus!!
AppleFan::initHW unable to save ADM1030State!!
AppleFan::initHW failed to open I2C bus
AppleFan::restoreADM1030State failed to open bus!!
AppleFan::setFanSpeed FATAL ERROR FETCHING REMOTE CHANNEL TEMP
AppleFan::doUpdate ERROR FETCHING CPU TEMP
regAppleFan::start failed to fetch provider's reg property!
PPCI2CInterface.i2c-uni-nAppleFan::start failed to find UniN I2C interface!
AppleCPUThermoAppleFan::start failed to find CPU thermistor driver!
AAPL,phandleAppleFan::start AppleCPUThermo is attached to wrong thermo!!
AppleFan::start failed to initialize operating parameters!
IOPlatformPlugin::start failed to allocate thread callout
AppleFan::start failed to initialize ADM1030!
709:J:”;<;î;„<)<j<¶<◊==H=x=ß=‘>Yˇˇˇˇcom.apple.driver.AppleFan1.0.10f1ˇˇˇˇ"∞#"§Eˇ ˇˇA"†D‘ˇˇˇA"îßUˇˇˇA"êßTˇˇˇA"Ñ©UˇˇˇA"Ä©TˇˇˇA"\†UˇˇˇA"P†VLˇˇˇA"D©]ˇˇˇA"<ß]ˇˇˇA",M‘
Ouch!
At any given time, CPU B's temperature is shown almost 50 degrees Fahrenheit cooler than CPU A. The ambient temperature around both CPUs is within a few degrees. And sure enough, I took a flashlight and looked in the grill and CPU A's intake and exhaust fans are spinning away. CPU B's fans are turning very slowly. The computer 'thinks' CPU B is cool, and doesn't spin the fans up. So, CPU B is overheating and hangs the whole machine. When I open the case and feel the radiators around the CPUs, B is much hotter than A.
Interestingly enough, last night I turned on the A/C in the room and put a Vornado house fan blowing into the front grill, and the computer ran like a watch. As a matter of fact, I ran the Apple Hardware Extended Test during this time and no errors were reported. As I write this, I can reach down, turn off the Vornado, and in about ten minutes the computer will freeze up.
I've going to guess that when the power supply fried, it did something to the temperature sensor on CPU B. So, I'm trying to figure out what to do next. I could get a new CPU, but I can buy a used example of this model computer for about what the processor will cost. And who knows what the condition of the power supply will be in another used machine!
Anyway, does anyone know how to set the intake/exhaust fans on an individual CPU to full speed? Short of putting an additional fan in front of CPU B, locking its intake/exhaust fans on high seems like it might work. Or at least keep the machine working until I can save my pennies and buy a new machine.
Any other suggestions?

Well, for some reason, the website won't let me post the code from inside the file. Here is the last bit of my question:
At any given time, CPU B's temperature is shown almost 50 degrees Fahrenheit cooler than CPU A. The ambient temperature around both CPUs is within a few degrees. And sure enough, I took a flashlight and looked in the grill and CPU A's intake and exhaust fans are spinning away. CPU B's fans are turning very slowly. The computer 'thinks' CPU B is cool, and doesn't spin the fans up. So, CPU B is overheating and hangs the whole machine. When I open the case and feel the radiators around the CPUs, B is much hotter than A.
Interestingly enough, last night I turned on the A/C in the room and put a Vornado house fan blowing into the front grill, and the computer ran like a watch. As a matter of fact, I ran the Apple Hardware Extended Test during this time and no errors were reported. As I write this, I can reach down, turn off the Vornado, and in about ten minutes the computer will freeze up.
I've going to guess that when the power supply fried, it did something to the temperature sensor on CPU B. So, I'm trying to figure out what to do next. I could get a new CPU, but I can buy a used example of this model computer for about what the processor will cost. And who knows what the condition of the power supply will be in another used machine!
Anyway, does anyone know how to set the intake/exhaust fans on an individual CPU to full speed? Short of putting an additional fan in front of CPU B, locking its intake/exhaust fans on high seems like it might work. Or at least keep the machine working until I can save my pennies and buy a new machine.
Any other suggestions?

K7N2 Delta2-LSR S2K Bus disconnect support ?

Does anybody know if there is built in support for S2K Bus disconnect in K7N2 Delta2-LSR? I've read numerous posts on the Internet, that enabling this feature helps dropping down Athlon XP temps by 6-10 degrees. However, I could not find such option in the BIOS, so I had to use a software utility to do that. It works, but I'm not sure if it has side effects, and it's definitely not the most...elegant solution...
And my other question is: Does this board have support for the internal CPU thermal diode ? As I was assembling my system, I found an external thermal sensor in the CPU Socket, while there is no such in other nForce 2 boards... I've heard thet these external sensors are not accurate at all, and that the actual core temperature is higher by 10-15 degrees...Is that true?

kamelot:
I use Speedfan to monitor the temperatures of K7N2-Delta2-LSR. After Speedfan detects the sensors in K7N2-Delta2 and if you go to Configure->Advanced->Chip:Winbond83627THF, you will see that Thermal Sensor Diode 2 corresponds to "PII diode", which means that it is most probably connected to the internal thermal diode of the processor. The thermal fluctuations with the processor load is instantaneous and steep on this sensor, which tells me that it is indeed the internal diode. Otherwise, there should have been some detectable lag between the thermal response and processor load. I could not find any direct reference to this information but this is what I found by testing and experimenting with Speedfan.
Regarding the accuracy issue, the absolute temperature value can be off but the delta temperatures should be correct. In the end, absolute temperature reading is just a voltage plus/minus a reference value, reference can change and so the temperature reading. My idle CPU temperature is lower than System temperature, which I think is not correct. But I do not mind this so much because the loaded CPU temperature is about 15C higher then idle and I believe this is the sign of efficient CPU cooling.
K7N2-Delta2-LSR
Athlon XP-M 2500 -> OC: 2.4 GHz (12x200, 1.65v)
Kingwin KCU-7015 cooler (cheap but works well, and when fan speed is controlled by BIOS it is very silent)
Single channel 512 MB PC3200 DDR 2.5-3-3-8

Ominous but not 100% confirmation that there is no temp sensor on the NX6600GT-VTD128

User Support Tue, Jan 25, 2005 at 3:52PM
To: -------------------
Dear Customer,
There is no sensor on the video card for temperature.
Sincerely,
Technical Support Division
MSI Computer Corp.
MSI Platinum Series Motherboards -- In Pursuit of Quality Perfection.
http://www.msicomputer.com
1.626.913.0828
1.626.581.7721 Fax
Due to high volume of cases daily, we may not be able to answer question
promptly. Please kindly provide your name, phone number, model number, state
in which you are calling from, a brief description of your problems, and we
will try to reply your issue as promptly as possible. Thanks
The information transmitted is intended only for the person or entity to
which it is addressed and contained confidential and privileged material.
Any review, retransmission, dissemination or other use of, or taking of any
action in reliance upon, this information by persons or entities other than
the intended recipient are prohibited. If you receive this in error, please
contact the sender and delete the material from any computer.
-----Original Message-----
From: -----------------
Sent: Sunday, January 23, 2005 11:36 PM
To: User Support
Subject: Re: 3 -- Customer Problem Description Form r/c2
I appreciate the response but my questions were not completely answered.
I know the sensor is disabled, but does the sensor PHYSICALLY exist on the
PCB itself?
On Wed, 19 Jan 2005 17:14:54 -0800, User Support
wrote:
>
>
>
> Dear Customer,
>
> The answer to your questions is no, this video card does not support
> temperature monitoring.
>
> Sincerely,
> Technical Support Division
> MSI Computer Corp.
> MSI Platinum Series Motherboards -- In Pursuit of Quality Perfection.
> http://www.msicomputer.com 1.626.913.0828
> 1.626.581.7721 Fax
> -----------------------------
> Due to high volume of cases daily, we may not be able to answer question
> promptly. Please kindly provide your name, phone number, model number,
state
> in which you are calling from, a brief description of your problems, and
we
> will try to reply your issue as promptly as possible. Thanks
>
> -------------------------------------
> The information transmitted is intended only for the person or entity
> to which it is addressed and contained confidential and privileged
> material. Any review, retransmission, dissemination or other use of,
> or taking of any action in reliance upon, this information by persons
> or entities other than the intended recipient are prohibited. If you
> receive this in error, please contact the sender and delete the
> material from any computer.

Strange indeed, since I always thought that the nvidia chip itself contains an internal thermal diode. But maybe MSI has just "forgotten" to wire the corresponding pin? What about other card manufacturers?
To be honest, offering "D.O.T" on the one hand but no temp monitoring on the other one is ridiulous in my eyes, I'll go favor another brand for the next videocard I purchase.
cu,
Jackie

Macbook pro 13" CPU A Diode temperature

Hi all,
I have recentlypurchased MBP13 inches (Intel Core i5 2.4 ghz, 500 G HDD, Internationalkb). I observed a difference in CPU heating behaviour compared to mypreviously used MBP 15" (2008 version).
So I did the following test:
1. CPU Temperature after 10 minutes idle (no activity except hardware monitor; user load 1-3% as measured by istat pro) :
observations:
CPU A temp diode: 42 degree C.
CPU core 1: 40 deg C
CPU core 2: 41 deg C
CPU core 3: 40 deg C
CPU core 4: 41 deg C
2. CPU temperature after 10 minutes of constant intense CPU activity ( 100% CPU load) given on both cores of the CPU
(by typing command in two terminal windows: yes > /dev/null)
CPU A temp diode : 93 degree C.
CPU core 1: 99 deg C
CPU core 2: 99 deg C
CPU core 3: 95 deg C
CPU core 4: 98 deg C
I have got two questions:
1. Is this a normal behaviour of i5 CPU heating ?
2. Why does the hardware monitor show 4 cores of CPU when in fact Intel i5 is a dual core processor ?
Thanks

Well, I talked to Apple support and tried to find out whether my observed temperature ranges vs CPU load were normal. Surprisingly, they told me that they haven't got any publicly available information on this phenomenon.
I have also inserted a plot of temperature rise vs CPU load here.
So it sounded like Apple does not want to come transparent about this issue with their customers. They only advised me to contact the nearest service center and see what is their opinion. Now in the absence of any precise and standard information about temperature ranges, I don't know what would be the basis of opinion from service center.
@pixelized - I suppose i7 overheating is increasingly discussed. But mine is i5. I guess the temperature rise vs CPU load should be different between the two processors. I have never observed temperatures like 120 C, even after 10 mins of 100% CPU load on both cores or by watching videos. Ambient room temperature was 18C in my case.
If anyone has obtained observations of temperature rise on i5, I would appreciate if you share them here.

Can a bubble in thermal paste make CPU temp range 39ºC-60ºC?

Hello.
I've read the postits above and they say 45 ºC idle is very likely the heatsink spreader is not properly seated or the temperature is wrong.
But, what if idle temp is about 39 on the heatsink (meassured with a termometer) and the bios says CPU temp is 39? Does that mean I did well with thermal paste and heatsink? Does that mean the motherboard is reading the righ temperature?
It could look like I am asking a very easy question, but I feel it is not. Just go on reading.
I need to know if BIOS does actually provide the rigth temp because:
When the system heats up due to CPU under heavy load, CPU temp goes up to 50-55 ºC sometimes even close to 60 ºC. Could that mean there is an air bubble within the thermal paste but it is too small to interfere when cold enough (40ºC) but big enough when warm (50ºC) lowering heatsink performance and rising CPU temp up to 60 ºC?
Also sys temp is very high. It never gets below 50 ºC but I do not know where that sys temp is taken, and as I have the NB-fan connector conected to the power supply fan, I do not know what that sys temp means. :(
Room temp is around 25 ºC.
It does boot OS and I can actually play games (heavy load for graphics card and CPU). All readings I take were got from CoreCenter application at stock settings (Vcore 1.42 - FSB 201 - Mem volt 2.5 - AGP volt 1.5).
I am afraid of having done a nice job with thermal paste. I know (just read it here) I could easily fry the CPU if not, and I feel 55-60 ºC is too much, isn't it? Of course if BIOS readings are right.
BTW. Here follow details on the system:
CPU: AMD 64 3000+ NewCastle
CPU Family/Model/Step: 15.12.0
Thermal compound: Artic Silver 5
CPU Heatsink: Thermaltake Silent Boost K
Motherboard: MSI K8N Platinum
Motherboard BIOS: 1.4 (08/26/2004)
Graphics Card: Sapphire ATI Radeon 9800 SE AIW (not modded in any way)
Power Supply: Enermax 465 W [EG-465AX-VE(W)FMA]
Case: Lian Li PC-V1000
Mem, hdd, dvd-rw, floppy and D-bracket are present but did not feel necesary to provide details on them.
Thanks

Well, if your heatsink is cold and your CPU temp on BIOS is hot there are two possible reassons:
[list=1]
BIOS 1.4 does not guarantee accurate CPU temp readings
Heatsink is not properly seated on the CPU or thermal paste has bubbles within[/list=1]
But that is not my situation. I do know CPU readings are accurate when it shows 39 ºC because I actually used a termometer in contact with the heatsink to meassure it.
What I am not sure is if that accuracy is so when the CPU is hot (>50ºC) as it is when cold (

CPU Temp Diode at 82 degrees C - Fire fire!!

So then,
I'm using Safari, Mail is open and i'm listening to iTunes. The normal set up for me and my Macbook. The most intensive program I use it Word and it's closed when not needed.
But my MB seems to run hot hot hot all the time - well most of the time.
The CPU A Temp Diode says it's running at 82 degrees Celcius which seems mighty high to me. My fans are going mental.
Apart from taking it to get the once over at the apple shop is there anything I can do/not do in the meantime??
Thanks to anyone who can help!

Ok, thanks Joey.
It's quiet as a mouse now and acting like a 'normal' MB.
When it plays up again i'll check the Activity Monitor and see what's what.
Idea....
Could it have anything to do with a plug in I have for Mail? It seems to quieten down when I close Mail. I think.
I have a plug in to make the 'Mail In' Sound and I also have one to fetch Hotmail into my Mail account.
Could these be the problem?
Thanks again!

Forced Sleep, Thermal runaways, and how hot is your cpu?

I've been dealing with a problem for a while on my PowerMac G5 dual 2.7. Under certain conditions is goes to sleep on it's own, usually when under heavy load, but also sometimes after waking up. The problem has gotten progressively worse.
After reading thorough this excellent forum I discover to look for Thermal runaways. Sure enough, the system log confirms that is my problem.
Based on advice I have gleened from here, I have so far tried:
1) Updating the system from the 10.4.3 combo download from the apple site.
2) Resetting the SMU.
3) Replace both of the .plist files for Energy Saver.
4) Run Apple Hardware test. (it showed nothing)
So far the only thing that has worked has been to change the pocessor setting in Energy Saver to "reduced".
While going through this process I downloaded a nifty little program called Temperature Monitor, and I have been running it constantly to keep a decent log.
Yesterday I took my Mac to the Genius bar at my local Apple store. We were unable to replicate the problem in the 15 minutes we had it running, althought it showed about the same temp range as when I've had it running at home for hours. If running (even sleeping) for a couple of hours I can reproduce the problem easily.
Here lies my delimma. My processors seem to be running hot, and CPU B runs about 15 degress hotter than CPU A.
Here's some data (averages are estimated based on weeks of data):
CPU A min temp 90.4 max temp 187.0 average 168
CPU B min temp 95.5 max temp 209.7 average 175
All other temeratures seem to be in line with downloaded data I got from another freeware called ThermographX. The ThermographX data seems to indicate that my processor temps should be in the 90F - 137F range. Mine are clearly in the 155F-190F range, with CPU B always 10-25 degrees hotter than CPU A.
So does anyone know what temp the CPUs should be running at?
One last thing. The data from my system log (see below) appears to give some temps (T_cur=115 >= (T_max:88 + sleepOffset:20). The question is, what do these numbers mean?
Dec 21 18:17:09 tolladays-power-mac-g5 kernel[0]: PowerMac7,3 Thermal Manager: Thermal Runaway Detected: System Will Sleep
Dec 21 18:17:09 tolladays-power-mac-g5 kernel[0]: PM73 T_cur=115 >= (T_max:88 + sleepOffset:20)
G5 Dual 2.7 Mac OS X (10.4.3)

My case was very similar to yours. I came to a solution so simple that it's even funny.
I experienced high temperatures on my Dual Powermac G5 2.5 (late 2004), around 90ºC even when doing simple tasks such as browsing the web. When doing cpu-intensive tasks (such as encoding a video with iDVD) the temperature went up to 105ºC or more, causing the machine to go into sleep mode and writing that "thermal runaway" entry in the logs.
I ran the Apple hardware test (which I recommend you to run if you haven't) and it gave me the following error code:
2STH/1/2: CPU A AD7417 AD1
According to http://docs.info.apple.com/article.html?artnum=86492 there seemed to be a problem with my CPU A thermal sensor.
Since the mere thought of taking my powermac for repair is completely unbearable to me, I bought a new SATA hard disk drive with the purpose of installing OS X from scratch, therefore discarding any software related problems.
When I opened the G5 case I noticed large amounts of dust in front of the CPUs air inflow grid. And that was the cause of my problems. I cleaned the inside of my powermac with a vacuum cleaner, paying special attention to ventilation grids and fans.
I ran the hardware test again and it gave me no errors. Additionally, now my powermac runs at lower temperatures (around 55-60ºC when not doing cpu-intensive tasks) and I haven't had any more thermal runaways.
My solution was very simple and perhaps obvious, but it worked. I hope it helps.
Dual Powermac G5 2.5 & Tibook 1.0 DVI Mac OS X (10.4.4)

Thermal CPU reading

Hello,
Can you please confirm that the KT3 Ultra (not the 2) can read the processor's on-die thermal diode ot ig it is an external thermistor that is used.
By advance, thank you.

Quote
Originally posted by Repoman
I too have a question concerning this. Approximately how much higher of a reading would we get with the on-die diode as opposed to an off-die diode? Appreciate any response!
I don't think it fair to say or assume that one temp sensor yields a higher or lower reading than the other. Truth is, the on-board sensor varies up to +/- 20%. The XP's on-die sensor has been within 3C of actual on every one I've checked.

Expected internal/CPU temps for MEGA 865?

What are the typical CPU/internal temperatores for the MEGA865?
Mine is running about 55C for the CPU and 41C case temp.
Is that high or ok?
I haven't tested at full load yet.

Try Here
http://www.komplett.co.uk/k/kl.asp

CPU internal thermal diode calibration

Similar Messages

Maybe you are looking for