Lapping worth it?

[new2spcr]

Hello,

Summer has hit with almost full force, and the current ambient temps are around 27-28 degrees C in my apartment, and so I'm expecting that my E4400 @ 1 volt, currently idling @ ~ 44 / 46 C will reach close to 50 C idle (!) or thereabouts when ambient temps in late July beginning of August will reach 29-30 C or even more.

I've found that increasing the Nexux 92mm fan speed on the Minja, ~1000 rpm, doesn't help much, but increasing the system fan speed > ~600-620 rpm does lower the idle temps 1-3 degrees or so. But then noise is becoming an issue... Anything above 550 rpm seems too noisy for me.

So lately I've been looking into the possibility of lapping my IHS to gain some better cooling performance.
This is how the IHS + Minja's bottom look, unmounted:





I always apply a tiny TIM blob on the middle of the IHS and let the HSF's pressure spread the blob evenly. Never cared about the credit card trick. Note that the pressure isn't 100% optimal, although I think both cores are being more than sufficiently covered.

What do you guys think? Will lapping improve the temps?

[N7SC]

Those temps you are getting look awfully high, especially for an E4400, especially if it is running at 1 volt. I'd be concerned about why my cpu temp(s) were so high in the first place. Your cooler is letting your cpu temps at idle go up nearly 20 C above ambient. That is unacceptable. 10 degrees C is more like it, even for a very quiet (low fan speed) cooler, I think. Lapping has little effect, it only gives temp drops of 1 to 3 degrees. It is a lot of work. Maybe a more efficient, quiet cpu cooler would be better.

For example, I have an e6600, running at default voltage, which is up to 1.35V, normally around 1.31V, and overclocked to 3GHz (= more heat). Ambient temp in my house reaches up to about 26.5 degrees C (79 to 80 deg. F), and the core temps average between 29 C and 31 C at idle. Now, my cpu cooling is water, but still, after running for several hours, even the large thermal mass of the water gets in thermal equalibrium with the cpu, and temps stay the same. I just think something is wrong with your setup, and your cpu temps are way too high. Maybe a different, more effective cooler would benefit you more than lapping the cpu and heat sink. Have you checked the reviews here at SPCR for a quiet, but highly effective cpu cooler?

For what it is worth, I envy you. Summer has not even begun to hit with full force here, and the outside temps have already already passed 37 deg C (about 99 to 100 F). The heat index (what it feels like, calculated using temp, humidity, etc.) has passed 42 deg. C (108 F). Today begins a much needed "cooling trend," the outside temps are "only" going to get up to around 34 to 35 C, about 94 to 95 F.

[CyberDog]

It depends on the IHS and heatsink. Some are very convex. How's are your case temps?

[N7SC]

A quick scan of some of the comments on the mini-ninja, which I assume is what you have, at Newegg, shows that average idle temps are well below what you are getting. Also, per CyberDog's comments, at least one user did report that the bottom of the minja was quite convex and needed to be lapped. If you have a good, accurate straight edge, you might check it carefully. Maybe it would be worth the time to lap the bottom of the heatsink, if your minja is very convex. Check the CPU too.

[new2spcr]

Those temps you are getting look awfully high, especially for an E4400, especially if it is running at 1 volt. I'd be concerned about why my cpu temp(s) were so high in the first place. Your cooler is letting your cpu temps at idle go up nearly 20 C above ambient. That is unacceptable. 10 degrees C is more like it, even for a very quiet (low fan speed) cooler, I think. Lapping has little effect, it only gives temp drops of 1 to 3 degrees. It is a lot of work. Maybe a more efficient, quiet cpu cooler would be better.

For example, I have an e6600, running at default voltage, which is up to 1.35V, normally around 1.31V, and overclocked to 3GHz (= more heat). Ambient temp in my house reaches up to about 26.5 degrees C (79 to 80 deg. F), and the core temps average between 29 C and 31 C at idle. Now, my cpu cooling is water, but still, after running for several hours, even the large thermal mass of the water gets in thermal equalibrium with the cpu, and temps stay the same. I just think something is wrong with your setup, and your cpu temps are way too high. Maybe a different, more effective cooler would benefit you more than lapping the cpu and heat sink. Have you checked the reviews here at SPCR for a quiet, but highly effective cpu cooler?

For what it is worth, I envy you. Summer has not even begun to hit with full force here, and the outside temps have already already passed 37 deg C (about 99 to 100 F). The heat index (what it feels like, calculated using temp, humidity, etc.) has passed 42 deg. C (108 F). Today begins a much needed "cooling trend," the outside temps are "only" going to get up to around 34 to 35 C, about 94 to 95 F.

Thanks for comments. So I would probably not gain any significant changes by lapping...

Yes, I think something's not right with my system but I can't figure out what it might be.
The Minja's plate is even so I figured it had to be the IHS that's not completely even. But then, judging by the pretty even spreading of TIM, I think the crucial area has a decent contact with the Minja. I've also read at an OC forum that the E4400 is a pretty hot chip. The reported idle temps were around 40 degrees C at similar ambient as here.
As I'm writing this, I've increased the rpm on both fans (system + cpu fan)

CPU fan1: 950 RPM
System fan2: 614 RPM

When fully booted into the OS, it takes just less than 30 minutes to reach 40 C on core 0 and 44 C on core 1.

I've re-seated the Minja in at least three different directions, experimented with the size of the TIM blob... but I'd prefer having it mounted so that the spill-over airflow flows over the mosfets.
The Minja isn't even luke warm to touch! The fins, heatpipes are all cool. I don't know if this is normal.
I didn't observe any great temperature variations by re-seating the Minja, perhaps some 2-3 degrees C drop in the optimum mounting direction.
This is the best I get.

I actually don't know what my temps were with the stock HSF, because at that time I didn't really care. I just assumed everything was ok since the computer was built by a professional. Once I checked the CPU temps in BIOS (I know, it doesn't show the correct numbers) and it was hovering around 32-33 C which i took as good and healthy.

I'm currently monitoring my cores with 'sensors' in Linux, which I think presents accurate data.
Yes, I've read the recommended HSF section and I fell for the Minja.
While I'm happy that I did get some valuable temperature drop by undervolting the chip, I can only imagine, and that irritates me, what would've happened if I left it at stock voltage at these ambient temps.

[new2spcr]

It depends on the IHS and heatsink. Some are very convex. How's are your case temps?

It's been a while since I measured the case temps, but I'd say + - 30 degrees C. I find it little difficult to measure the case temp - where do I put the thermometer?

I usually try to get a rough overview by checking the temps on various HS and chips, such as NB, SB, GPU etc.

Asus Geforce 7300GS is around 53-54 C (passively cooled)
NB HS, Thermalright HR-05, passive - around 55-60 C
Southbridge HS passive - around 50 C
Harddrive, passive - around 30 C

@ N7SC - Yep, I've read the Newegg reviews as well. And plenty of others. It seems lots of folks are getting somewhere around +- 30 C idle, stock voltage and overclocked.

I might look into lapping the Minja, but judging by the photos of the TIM's footprint, wouldn't you say the contact is pretty good?

Current temps: 42 + 45 C!

[new2spcr]

Regarding a new HSF....Hm. I feel the TRUE and its counterparts are little to pricey. For that price wouldn't it be better to buy a more energy efficient 45 nm processor?

Edit
Temp. update:

Increased the fan speed even more -

Temps:

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +42.0°C

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +44.0°C

it8718-isa-0290

CPU FAN Speed: 1184 RPM
CHASSIS FAN Speed: 733 RPM
CPU Temperature: +34.0°C (this is probably the same as the one that BIOS is reporting)
MB Temperature: +41.0°C

Funny, the fanmate allows the 1200 rpm Nexus fan to spin at +1500 rpm but I've settled for ~ 1200 rpm for about 10 minutes.

GPU - 50 C. Increasing the system fan helped couple of degrees.

[N7SC]

When you remove and reseat the heatsink, are you cleaning the old TIM off properly, off of both the heatsink and the IHS? Are you making sure the pins that lock the minja to the motherboard are seated properly? The comments on Newegg seemed to reflect that the plastic pins were difficult for some to get installed properly. Make sure that they are turned in the proper direction when you push them down and that they lock securely and hold the heatsink down with enough force against the processor.

What TIM are you using?

[new2spcr]

When you remove and reseat the heatsink, are you cleaning the old TIM off properly, off of both the heatsink and the IHS? Are you making sure the pins that lock the minja to the motherboard are seated properly? The comments on Newegg seemed to reflect that the plastic pins were difficult for some to get installed properly. Make sure that they are turned in the proper direction when you push them down and that they lock securely and hold the heatsink down with enough force against the processor.

What TIM are you using?

Yes, I always make sure the surfaces are wiped clean with isopropylalcohol, no lint, dust or fingerprints are left.
The pins are installed correctly, I usually take out the mobo to double check.
The TIM in question is Arctic Cooling MX.

[CyberDog]

does the heatsink ever get any warm? My TRUE is definately warm and my temp readings are pretty much same as yours. Except I have different cpu and overclocked. But the point is that if your CPU is warm so should be your sink also.

[new2spcr]

does the heatsink ever get any warm? My TRUE is definately warm and my temp readings are pretty much same as yours. Except I have different cpu and overclocked. But the point is that if your CPU is warm so should be your sink also.

Hi CyberDog, my HS (the heat pipes) is slightly warm to touch when I run Orthos but not when the chip is idle.
I have re-mounted the HSF three times now. Tried every possible direction, also tried the credit card & TIM trick - same temps.

I guess the IHS and Minja don't play well together...

30 minutes of Orthos give me 56 & 57 C.
Exhaust temp (near system fan) is 36 C

Maximum fan speed affects the core temps but only 1-2 degrees. Something's definitely wrong here.

[new2spcr]

I'm beginning to believe the thermal diodes aren't either calibrated or the analog -> digital conversion is giving off digits. And /or the lm_sensors in Linux aren't giving the correct values. Or a mix of all three?

Quote:

"(C) Tcase and Tjunction are both acquired from Thermal Diodes. Tcase and Tjunction analog to digital (A to D) conversions are executed by separate devices in different locations. BIOS Calibrations from motherboard manufacturers, Factory Calibrations from Intel, and popular temperature utilities are frequently inaccurate."

"(E) Intel provides only partial documentation for Tjunction Max on desktop processors. For Throttling and thermal Shutdown protection, Intel uses the Digital Thermal Sensors (DTS) to monitor Delta to Tjunction Max, which is a relative value that varies from Core to Core, and is not an absolute temperature."

and...


"(F) Tjunction Max must be known to calculate absolute Core temperature, which is Tjunction. Popular temperature monitoring utilities may incorrectly estimate undisclosed Tjunction Max values, which results in excessive Core temperatures and inconsistent Tcase to Tjunction Gradients among Ci7 and C2 variants."

Unquote

http://www.tomshardware.co.uk/forum/pag ... _11_0.html


With lm_sensors in Linux, I noticed the critical limit for cpu temperatures are way over the E4400 max temp of 61 degrees C stated by Intel. I don't recall those max numbers right now but they're closer to 90 C or even more.
Perhaps this has something to do with the core temperatures being inaccurately calculated.

Yesterday, with ambient room temperatures hitting ~28-29 C, idling core temps hit 45 & 48 degrees C after couple of hours of inactivity. That can't be right.
Minja was correctly installed. CPU fan speed ~850 rpm.
The Minja's heatpipes close to its base are not warm to touch at all - Perhaps normal body temperature was more like it.
Wouldn't I have noticed if the core temps were hitting close to 50 C? CPU temps in BIOS report 38-39 C.

Temperature inside the case (measured with a analog thermometer lying on the bottom of the case) was 30-31 degrees C.

EDIT:

I just tried the Intel stock heatsink (note: it belongs to E4700 and it's thinner than the stock HS for E4400) to see what it'd do to the temps -
The idle core temps went up to 50/51 within 1-2 minutes and pretty much stayed there.
Fan was controlled by Asus Q-fan, it didn't ramp up until I ran Orthos. Within 1 minute of Orthos, core temps hit 60 C.
At that point, I quit running the software.

I doubt the stock HS (okay, it belongs to the E4700, which I assume is slightly cooler than mine given its thinner appearance compared to the stock HS for E4400.. but my chip is undervolted) wouldn't manage to keep my E4400 @ 1 volt within safe limits. It doesn't make any sense that the cores would reach maximum temps within 1-2 minutes even with the stock HS - normal computer usage with stock HS should allow both cores to stress for longer periods of time than that. Especially idle temps wouldn't hit 50 degrees C even with the stock HS!
I pulled out the CPU and studied the IHS's surface - it's pretty flat and even.

So, if stock HSF gives terrible idle /load temps and an aftermarket HSF gives similar results, I'm inclined to conclude that there's nothing wrong with neither the IHS or HSF hardware. It gotta be a calibration problem.

[N7SC]

. . . I'm inclined to conclude that there's nothing wrong with neither the IHS or HSF hardware. It gotta be a calibration problem.

So far, the data seems to point in that direction. But, I would be very careful and try to find a way to verify with direct readings of the temp of the IHS or something like that.

Can you get a thermal sensor of some kind or an accurate thermocouple in there, up against the side of the IHS to see whether or not it agrees with the values that your software is reporting?

I do seem to remember that some of the original Core 2 Duos had thermal sensors that were difficult to read correctly. And, if I remember correctly, the E4XXX chips were the worst.

But, don't jump to conclusions and risk frying your cpu. Gather some more data.

[new2spcr]

. . . I'm inclined to conclude that there's nothing wrong with neither the IHS or HSF hardware. It gotta be a calibration problem.

So far, the data seems to point in that direction. But, I would be very careful and try to find a way to verify with direct readings of the temp of the IHS or something like that.

Can you get a thermal sensor of some kind or an accurate thermocouple in there, up against the side of the IHS to see whether or not it agrees with the values that your software is reporting?

I do seem to remember that some of the original Core 2 Duos had thermal sensors that were difficult to read correctly. And, if I remember correctly, the E4XXX chips were the worst.

But, don't jump to conclusions and risk frying your cpu. Gather some more data.

Hi N7SC,
Unfortunately I don't have access to those thermal diodes you mentioned. I have a pen-style infrared thermal reader but I doubt I can measure the IHS temperatures with it.
I've found it is rather difficult to get an accurate reading even on a big size HS mainly because the IR-sensor needs a surface to read, not "pocketed" air or mix of the two... and pressing the start button moves the pen slightly too!

I googled up more information about lm_sensors for Linux and it seems this particular software sometimes gives inaccurate readings depending on version:

http://bbs.archlinux.org/viewtopic.php?id=49442

Not sure though if the latest kernel & lm_sensors remedied the problem. As far as I can remember, lm_sensors has always given me same results.

I'm little confused though, because the first line (CPU, I assume it's t-case) that comes from lm_sensors corresponds well with the BIOS's temperature reading:

CPU Temperature: +36.0°C (high = +90.0°C, crit = +125.0°C)

Which indicates lm_sensors is accurate, well, as accurate as the hardware sensor on my mobo that reports t-case.
(note the "high" value of 90C. it's way higher than 61.4 C t-case stated by Intel)

Then, we have the lm_sensors output of both cores:

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +43.0°C (high = +74.0°C, crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +46.0°C (high = +74.0°C, crit = +100.0°C)

So yes, it is little dangerous to assume that lm_sensors would give accurate t-case readings but inaccurate t-junction readings.
So maybe lm_sensor is doing its job but the thermal diodes for the cores aren't calibrated well, which you indicated and especially for E4XXX chips.

This could be confirmed if I could install Windows on my machine and try Realtemp or hwmonitor. I might look into that option.

[Atmosper]

The temperatures you are getting seem high at idle but I wouldn't lose sleep over it. You won't fry anything at those temps.
CPUs can take 50°C all day long, all year long. I certainly wouldn't spend money on a new heatsink, much less a new CPU.

I have more than a decade of experience with overclocking and serious cooling. The only time I've ever broken something with high temperatures is when I accidentally ran an Athlon 1400 without any heatsink what so ever.

You have a long way to go before you reach critical temperatures and the CPU starts throtteling.
If you feel those temperatures are high, just up the fan speeds for the (at least in Sweden, short) summer.

[N7SC]

new2spcr,

Here is a link to an article titled Core i7 and Core 2 Temperature Guide: http://www.tomshardware.com/forum/22174 ... ture-guide If you haven't seen it yet, it contains some good info. Also check with intel.com for data on the thermal sensors in the E4400, I think, IIRC, that all of the Allendale chips have either less sensors than the Conroes or inaccurate sensors, don't recall which.

I think the above guide is written for Windows systems, but the info it contains may be of some help. I may comment some more later, but just got up and can not think clearly yet, sorry.

And again, I envy you guys in Sweden:

. . . just up the fan speeds for the (at least in Sweden, short) summer.

We're going to be above 90 F or about 32.2 C here in Florida until at least mid-September, maybe much longer, say until mid-October. And will likely spend many days well above 90/32.2, most of July, August, and half of Sept. above 95F/35C. Worst of all, the grass in the lawn grows fastest now, meaning that someone (guess who, in my case) has to mow it every week, in that heat and sun.

[new2spcr]

The temperatures you are getting seem high at idle but I wouldn't lose sleep over it. You won't fry anything at those temps.
CPUs can take 50°C all day long, all year long. I certainly wouldn't spend money on a new heatsink, much less a new CPU.

I have more than a decade of experience with overclocking and serious cooling. The only time I've ever broken something with high temperatures is when I accidentally ran an Athlon 1400 without any heatsink what so ever.

You have a long way to go before you reach critical temperatures and the CPU starts throtteling.
If you feel those temperatures are high, just up the fan speeds for the (at least in Sweden, short) summer.

Hi,

I am pretty sure my cpu won't fry, it's just that I feel uncomfortable with those high idling temps. I admit that I envy those who have very low idling temps, like lots of folks are reporting that their chips run just couple of degrees over ambient.
I doubt though that all of those reports are correct. I suspect many of them are simply reading the T-case temp.

[new2spcr]

new2spcr,

Here is a link to an article titled Core i7 and Core 2 Temperature Guide: http://www.tomshardware.com/forum/22174 ... ture-guide If you haven't seen it yet, it contains some good info. Also check with intel.com for data on the thermal sensors in the E4400, I think, IIRC, that all of the Allendale chips have either less sensors than the Conroes or inaccurate sensors, don't recall which.

Hi again. I've seen that link, it's also the same link I posted here. It's a good one to bookmark. I don't fully get what the article says but I get the general hang of it.

Intel only mentions the maximum t-case for E4400:

http://www.intel.com/cd/channel/reselle ... 299986.htm

I think the above guide is written for Windows systems, but the info it contains may be of some help. I may comment some more later, but just got up and can not think clearly yet, sorry.

Thanks for your efforts and thoughts, I appreciate it - you don't need to put more energy into this.
I googled up some more information and it seems many people are having inaccurate temp. reports:

http://techreport.com/forums/viewtopic.php?f=4&t=57735

And again, I envy you guys in Sweden:

. . . just up the fan speeds for the (at least in Sweden, short) summer.

We're going to be above 90 F or about 32.2 C here in Florida until at least mid-September, maybe much longer, say until mid-October. And will likely spend many days well above 90/32.2, most of July, August, and half of Sept. above 95F/35C. Worst of all, the grass in the lawn grows fastest now, meaning that someone (guess who, in my case) has to mow it every week, in that heat and sun.

32-35 C... I feel your pain.
Swedish summers are short, thank God, but they are much warmer now than they were 10 years ago... at least that's my feeling. It's 21:00 now and sun is still up and strong. Room temperature is 28 C. When I go out, it feels like I'm in the Mediterranean. We expect sunset just before 23:00. 4-5 hours later, we have sunrise... so summer is short but very intensive. How are the nights in Florida? Humidity? Thankfully, we don't have any humidity to speak of. I don't envy those who live in FE Asia where temps can reach 35-40 C with atmosphere saturated with H2O...

[LodeHacker]

You guys never been to Dubai? I spent the whole summer there (2005), 45C from morning to evening... I remember it too well. There you don't care about noise, I actually had a big desk fan blow at my PC to keep it and myself cool... Got a sunburn on top of it

[new2spcr]

You guys never been to Dubai? I spent the whole summer there (2005), 45C from morning to evening... I remember it too well. There you don't care about noise, I actually had a big desk fan blow at my PC to keep it and myself cool... Got a sunburn on top of it

What, they don't use airconditioners?? I would be much afraid of the HD temps...

[LodeHacker]

You guys never been to Dubai? I spent the whole summer there (2005), 45C from morning to evening... I remember it too well. There you don't care about noise, I actually had a big desk fan blow at my PC to keep it and myself cool... Got a sunburn on top of it

What, they don't use airconditioners?? I would be much afraid of the HD temps...

Unlike our neighbours, we had an airconditioner. It didn't work though

[N7SC]

I'm still not comfortable with just assuming that your software is misreading your temps unless you can get some direct reading from the side of the IHS or the side of the bottom of the Minja to provide better data. God forbid something is actually wrong and your software is reporting accurate temps - you would have almost no safety margin for your system.

For example, do you, or any friends, have a digital multimeter that can read temps when used with a type JK thermocouple? That is what I used to verify the software readings when I initially built my system. You may be able to get the probe in and around everything to touch the side of the IHS and take a reading there. If your temps are really as high as your software is reporting, you'll find out quickly.

Another option, since there is a lot of software for Windows that reads the thermal diodes in the processors: do you have another hard drive or can you borrow one, to make a windows XP boot drive? Use that to start up the system, then download and use any of the windows temp utilities, like Coretemp, Speedfan, etc.

It does get very humid here, in Florida, at times during the summer. It feels awful just to go outside. The electric bill for the airconditioning can get outrageous in the summer. Then there are the strong thunder storms and the potential for severe damage from lightning strikes, and my house has been hit before, as has my neighbor's, the golf course right behind me, the empty lot across the street and one house down, etc. Not to mention possible, but rare, hail damage. And rattlesnakes (not very rare here), coral snakes, poisonous toads (if my dogs get hold of one . . .), mosquitos, gnats, mold on the patio, mud on the dogs paws and my shoes, etc.

I have long entertained the idea of moving to Europe, and Sweden is one of the countries that I consider. But the fact of the very long days in summer is something I have not considered properly. That would drive me insane very quickly as I have a hard time sleeping here when there is a full moon. Full sun almost 20 hours a day would put me in an asylum in short order.

Isn't Dubai where they are building a hotel with an airconditioned beach? Wouldn't it just be cheaper to move or vacation somewhere more comfortable in the first place?

[walle]

Swedish summers are short, thank God...

...and I wish for you a speedy recovery.

[new2spcr]

I'm still not comfortable with just assuming that your software is misreading your temps unless you can get some direct reading from the side of the IHS or the side of the bottom of the Minja to provide better data. God forbid something is actually wrong and your software is reporting accurate temps - you would have almost no safety margin for your system.

For example, do you, or any friends, have a digital multimeter that can read temps when used with a type JK thermocouple? That is what I used to verify the software readings when I initially built my system. You may be able to get the probe in and around everything to touch the side of the IHS and take a reading there. If your temps are really as high as your software is reporting, you'll find out quickly.

Another option, since there is a lot of software for Windows that reads the thermal diodes in the processors: do you have another hard drive or can you borrow one, to make a windows XP boot drive? Use that to start up the system, then download and use any of the windows temp utilities, like Coretemp, Speedfan, etc.

It does get very humid here, in Florida, at times during the summer. It feels awful just to go outside. The electric bill for the airconditioning can get outrageous in the summer. Then there are the strong thunder storms and the potential for severe damage from lightning strikes, and my house has been hit before, as has my neighbor's, the golf course right behind me, the empty lot across the street and one house down, etc. Not to mention possible, but rare, hail damage. And rattlesnakes (not very rare here), coral snakes, poisonous toads (if my dogs get hold of one . . .), mosquitos, gnats, mold on the patio, mud on the dogs paws and my shoes, etc.

I have long entertained the idea of moving to Europe, and Sweden is one of the countries that I consider. But the fact of the very long days in summer is something I have not considered properly. That would drive me insane very quickly as I have a hard time sleeping here when there is a full moon. Full sun almost 20 hours a day would put me in an asylum in short order.

But the difference between core temps and IHS is pretty big, no? If I assume BIOS reports accurately, the IHS's temp is around 40 C, and that relatively small heat output would be taken care of by the Minja so fast that I wouldn't be able to feel it by touching its base. I can't access the base area by the way, only the heatpipes at the bottom and those are cool.
I might look into the possibility to access the IHS the HSF area is really cramped and I also don't have that tool you mentioned.

Coretemps/lm_sensors in Linux report both coretemps and what I think is the IHS, or the t-case temp. My core temps, idling, went up to 47 + 50 degrees C yesterday (!) but IHS was around 40 C. Mobo temp (whatever that sensor is measuring) reported 46 C.
Since Intel never released maximum temps for C2D, the t-junction max, I can only guess what's safe or not. But if the IHS sensor reports correctly, I still have around 20 C or so of margin since my chip can take 61.4 C (t-case). Also, I think the Minja performs well, since I only get around 10 C rise from idling temps if I run Orthos. If the contact between Minja and IHS were terrible, I think I'd get bigger differences than that.

The other option you mentioned is feasible. I have spare HD's and a W2K cd I could experiment with. I might look into that.

Hail, snakes, humid, storms, mold, gnats, bugs, toads...
I thought Florida was a haven, a place to where old folks move when they get tired of earthquake prone areas like California? I've never been to Florida but from what I've seen on the internet it surely looked nice.

Yes, it is little difficult to sleep during the summers...blinds and drapery help some but the rays seem to find its way into the bedroom no matter what I do. I bought some, what I thought would be, nice sun ray-blocking-drapery from IKEA but either the product description contained lies or it was tested during the winter.
Sweden is overrated. You also must consider the very long and cold winters, it's dark when you wake up and dark when you leave the office. I like the winters but I'm little odd.

[new2spcr]

*Major update*

So I took a spare HD and installed W2K on it.
Downloaded Orthos (newest version), Real temp (version 2.5 and 3.00), HWmonitor (newest version) and Core Temp.

Idling these are the temps I'm getting (ambient room temp: 28 degrees C):

HWmonitor (no indication of Tj-max, but I believe it's 100 C. *T-junction max is set as 100C. This has been verifired by increasing the T-junction max in RealTemp 3.00, and then I get same numbers as HWmonitor):
Core 0 - 50 degrees C
Core 1 - 52 "

Real Temp 2.5 (t-junction max setting is 85 C):
Core 0 - 30 C
Core 1 - 32 C
Distance to TJ-max: 55 C, 53 C.

Core Temp: (TJ-max: 95 C)
Core 0 - 42
Core 1 - 44


Stress test:
A session of Orthos 15 minutes, Large FFT etc:

HWmonitor:
Core 0 - 57 C
Core 1 -58 C

Real Temp 3.00 (T-junction max setting is 90 C in this version):
Core 0 - 47
Core 1 - 48

Real Temp 2.5 (TJ-max - 85C)
Core 0 - 42
Core 1 -43
Distance to Tj-max: 44, 43

Core Temp (Tj-max 95 C)
Core 1 - 52
Core 2 - 53


These numbers are weird. Don't know which one to trust. They are all guessing the TJ-max values thus spitting out different results... the TJ-max value varies from 85-95-100 degrees C!
I also tried running orthos with Vcore Auto [~1.32 V] set in BIOS, 5 minutes = ~ 10 degrees C over previous load temps @ 1 volt!
This demonstrates how valuable undervolting really is.

Lm_sensors in Linux most likely has around 100 degrees C set as TJ-max (in spite of it saying 74 degrees C "high" and "critical" 100 C) *this has also been verified by increasing TJ-max in RealTemp 3.00 to 100 C and I get the same output as Lm_Sensors), so it's unlikely that my temperature readings are accurate...I've read reports concerning the E4400 chip's TJ-max being less than 90 C. So, either Real Temp 2.5 or 3.00 is giving the correct numbers AFIWB (As Far as I Want to Believe). I could probably slice off at least 10 degrees or even 15 degrees from the output I'm seeing in lm_sensors.

So...I think my idle temps are 35 & 37 degrees C if TJ-max is assumed being 90 degrees C for my chip. Moreover, it doesn't make any sense that the cores are idling at 2-4 degrees over ambient temps (Real Temp 2.5) with aircooling.


Any ideas?

Edit 1: minor misspellings
Edit 2: added my hasted conclusions
Edit 3: did a new test to fill in some blanks

[N7SC]

Hmmm . . . I have to think about this data for a while, and check up on some stuff that I have long forgotten...will get back in a day or so.

Are your CPU Thermal Monitor 2 and EIST turned on in the BIOS? If they are, then your CPU should either throttle down or fully shut down if it gets too close to overheat, and that will protect you until you get the matter of correct temp readings sorted out.

Meanwhile, I tend to trust the readings from CoreTemp the most. Are you using the latest version, which I think is 0.99.4?

[new2spcr]

Hmmm . . . I have to think about this data for a while, and check up on some stuff that I have long forgotten...will get back in a day or so.

Are your CPU Thermal Monitor 2 and EIST turned on in the BIOS? If they are, then your CPU should either throttle down or fully shut down if it gets too close to overheat, and that will protect you until you get the matter of correct temp readings sorted out.

Meanwhile, I tend to trust the readings from CoreTemp the most. Are you using the latest version, which I think is 0.99.4?

Yeah, I have cpu throttling enabled in BIOS, but I don't remember if they called it CPU thermal monitor or EIST.

The Core Temp version is 0.99x.
It sounds pretty reasonable that I should consider trusting Core Temp (TJ-max 95 C) as well, since ambient temps are pretty high and thermal rise, idle, from ambient is around 15 C which isn't that bad.

The TDP for E4400 is 65 watt & max T-case 61.4 which I think can translate into a lower T-junction max... at least lower than 100 C. It'd be interesting to know the thermal relationship between, if there is any, the TDP and T-junction max. But I doubt there is an easy formula to calculate the numbers. If there were, the great Intel Temperature Riddle would've been solved long ago.
But I'm guessing wildly now.

Needless to say, it became pretty obvious to me the actual/real CPU temperature is so very difficult to measure. Just fiddling around little with RealTemp 3 shows the TJ-max set in various temperature reading software varies a lot, even between same software but different versions and also I think some of them automatically adjusts TJ-max depending on the processor. Right now, I'm running E4600 2.6 ghz at work and RealTemp 3.0 on this machine has a set TJ-max of 100 C.

What I think should be taken with a grain of salt is all the temperature reports that can be gathered from the net be it from overclockers or underclockers reporting good temps or bad temps etc.
I think the focus should rather lay on how much (approximate of course) thermal margin one has to TJ-max.

[N7SC]

It is difficult to understand why there are no reliable programs to read TJ Max, it is a value set in a register by Intel at the time of the chip's manufacture. It is a simple matter to write a program in assembler that will read the register. I tend to trust CoreTemp.

While my memory is pretty bad, I vaguely recall that readings are obtained as follows:
TCase is read via the hardware on the motherboard, and is subject to errors or variation due to differing hardware, different BIOS code, etc.

TJ Max is read directly from a hard coded value in a processor register. At least in those programs that actually read it, which I think includes CoreTemp. I have my concerns about RealTemp because it seems to default to 90C, and lets the user adjust the value. Not all Intel chips have a TJ Max of 90C.

TJunction is read, I think, if I recall correctly, as a displacement downward from TJ Max, but not directly. That is, if you put CoreTemp in the mode where it is displaying Delta to TJ Max, you are seeing what it is actually reading. A "real" temperature is simply calculated by the software based on the TJ Max value and the Delta.

Your TJ Max, based on the T-case Max for the E4400 should be at least 66.4C, so you can comfortably use 65C as a TJ Max. If that assumption is correct, and I think it is both correct and conservative (thus safe), you should be quite safe as none of the stress temps you posted come close enough to this for concern. Further, your BIOS, acting on processor temp readings, never shut down the computer. Just don't overclock it until you have gotten direct temp reading from the IHS to calibrate your software with.

[new2spcr]

It is difficult to understand why there are no reliable programs to read TJ Max, it is a value set in a register by Intel at the time of the chip's manufacture. It is a simple matter to write a program in assembler that will read the register. I tend to trust CoreTemp.

It seems T-junction is neither fixed or readable by any software, some processors even seem to have different TJ-max between the cores.

All we can do is guess the t-junction max and starting from there, use a formula to calculate the distance to TJ-max. Which, obviously, will give good or bad values depending on what TJ-max level has been set in the software.

While my memory is pretty bad, I vaguely recall that readings are obtained as follows:
TCase is read via the hardware on the motherboard, and is subject to errors or variation due to differing hardware, different BIOS code, etc.

TJ Max is read directly from a hard coded value in a processor register. At least in those programs that actually read it, which I think includes CoreTemp. I have my concerns about RealTemp because it seems to default to 90C, and lets the user adjust the value. Not all Intel chips have a TJ Max of 90C.

TJunction is read, I think, if I recall correctly, as a displacement downward from TJ Max, but not directly. That is, if you put CoreTemp in the mode where it is displaying Delta to TJ Max, you are seeing what it is actually reading. A "real" temperature is simply calculated by the software based on the TJ Max value and the Delta.

Yes, CPU values in BIOS (probably t-case) is known to be incorrect due to many reasons involved.
I've asked ASUS if the CPU temp reported in BIOS indeed is T-Case, but they failed to respond.
Asus home page doesn't mention anything about calibrated temperature sensor in the BIOS update section, so... to trust them or to not trust them...

But from what I've read, and if I've understood it correctly, TJ-max isn't being read - it's supposed to be an unknown factor for desktop chips...only Intel would know why.
I don't fully grasp the concept, but I think you're right because it implies one has to carefully investigate what the probable /closest TJ-max is for the chip (for example by observing when the processor finally throttles under load). But Intel never disclosed any information about TJ-max for most of their chips except for only few:

http://www.anandtech.com/weblog/showpost.aspx?i=478

and:

http://www.evga.com/forums/tm.asp?m=645260

Whatever "target TJ" is... nobody seems to know if Target TJ is same as TJ-max. For my chip, it's 70 C, which is ~ 10 C over T-case max. I don't know if that makes any sense.

I discovered today that Core Temp allows you to manipulate the TJ-max as well.

Your TJ Max, based on the T-case Max for the E4400 should be at least 66.4C, so you can comfortably use 65C as a TJ Max. If that assumption is correct, and I think it is both correct and conservative (thus safe), you should be quite safe as none of the stress temps you posted come close enough to this for concern. Further, your BIOS, acting on processor temp readings, never shut down the computer. Just don't overclock it until you have gotten direct temp reading from the IHS to calibrate your software with.

I have only experimented with 85-100 C TJ-max values... but yeah, I think 65 or less than 70 as max temp reported by the software under load will give me some security margin.
The Delta values to TJ-max, no matter if I state 100 or 90 degrees C are all same though, and I'd expect the same if TJ-max was set as low as 65 C.

To be honest, since T-Case max for my (and many other) chip is readily available, I'm looking for a software that reports this value accurately. Then this confusing nightmare would end.

[N7SC]

Hail, snakes, humid, storms, mold, gnats, bugs, toads...
I thought Florida was a haven, a place to where old folks move when they get tired of earthquake prone areas like California? I've never been to Florida but from what I've seen on the internet it surely looked nice.

As I brace myself to face the heat, sun, and tall difficult-to-mow grass (tall, because it has been raining for days and that has both prevented me from mowing when it needed it, and made it grow faster), I am reminded of the $1000 US, or more that I have spent in the past 8 years for veterinary care of my three Great Danes, just for swollen paws, and infections resulting from bites from fire ants, ground (flightless) wasps, and other assorted Florida insects.

Don't believe the stuff you see on the internet about Florida. The only way it becomes even nearly habitable is to plow it flat, clear it off, and cover it with concrete and paving. I do love native Florida, and don't like how it is cleared and destroyed to make it habitable, so please don't think I like what they do to my home state. But, I don't want you to get a false impression about what it is like here either.

Interesting fact: When the Spanish and Portugese explorers first came here, long ago, they were not able to penetrate too far from the beaches due to the enormous amount of mosquitos that lived here. They actually had to sleep on the beaches, covered in, or buried in the sand to avoid being nearly eaten alive by the mosquitos and gnats.