Tcase > Tcore WTF? (e4300 - fresh batch)

[r34550r]

hello im new here lookin for some help with understanding that crap :/

Tcase = 47-78
Tjunction = 22-49 (speedfan temps ... TAT shows 15C higher all the time)
Ambient = 18 (late night with window open)
Chipset = 965P
C2D = e4300 L2
CPU Cooler = PENTAGRAM FREEZONE NXC-100 Cu
Frequency = 2400 (9x266)
Load = TAT
Motherboard = P5B
Vcore = 1.285

would apriciate any imput on WTF is going on that Tcase is higher than Cores :/ - its the same in bios too

also i really dunno wich temps i should consider closer to reality...
its not coller couse i remounted it like 5 times already - no change

also when it reports such high values on Tcase cooler is chill like ambient temp .. so i reallyhave no clue what these guys were smokin while makin this cpu batch lol

 

[CompuTronix]

Thank you for organizing your information, as shown in the Guide. Please verify that you've also tested, as shown in the Guide, since the accuracy of my advice depends on it. Also, the guys who write smoke signals for BIOS thermal tables are more responsible for this mess than the guys at Intel! :lol:

Comp 8)

 

[dr_kuli]

My P5B-E (1202 BIOS) is off 15C at least. I've set an offset of -15 in SpeedFan 4.32 and it's more believable now, but still to high at load, I think.

 

[r34550r]

well i tested it with TAT 10mins both cores at 100%

and idle runs some crap in background but nothing that makes cpu spikes higher than 2%

i also updated bios to the newest one (no change in displayed temp)
hehe i know its against laws of physics that Tcase would be higher than cores but it blocks my oc becouse im getting shutdown temp when i stress it with tat @3000+

there is also absolutly no blue screens etc no errors what so ever.

PS.
sorry for my english its self-learnt

 

[dr_kuli]

What makes you to believe Tcase is REAL??? Actually, it's NOT Tcase, as you can see in SpeedFan's advanced config. It's Winbond chip that reads it, NOT INTEL CORE. And only Asus knows, where is the sensor, if it is at all.

 

[r34550r]

you missunderstood me i was pointing out that IT CAN'T be real becose its against laws of physics

im not an expert so i just asumed that winbond needs to take info from some place and most likely in my understanding it would be temp sensor on on IHS
Guess its not the case

so guys any conclusions?

 

[Heyray2]

I'm inclined to suggest that you take the TAT readouts with a grain of salt on the E4300s.

I've spent this entire week watching the temps on my 3.1GHz OC of my E4300 very intently. If the temps were actually what TAT suggests (around 75c max load, 80c max load @ 3.3GHz), I'd be burning my hand on the HSF.

Now I know this is the supposed temperature at the IHS, but that heat has to go somewhere and 80c from the core would be noticeable on the HSF. I'm pretty sure I'm not alone when it comes to feeling this way.

http://forums.hardwarelogic.com/f10/rfi-re-core-2-duo-temperature-4389.html

 

[r34550r]

thx Ray thats what i was thinkin exacly - at first i thought that i mounted my HSF bad but after few remounts it was clear that sumfin is wrong

ill focus more on speed fan if it comes to core readouts coz it shows the same as "core temp 0.94" wich shows my cpu as allendale and 0.95 as conroe

on 0.95 Tjunctionmax is stated as 85* and on 95* as 100*

and temps on 0.95 are simmilar to TAT

speedfan tho shows same as 0.94 and that sick winbond readout about 25* higher than hotest core lol

cheers

 

[Heyray2]

TAT wasn't designed to monitor C2D chips, and while Speedfan and CoreTemp (before 0.95) used data from thermal probes outside the CPU, I think they are a much better gague of temperatures and the performance of your cooling equipment.

Don't get me wrong. TAT can be useful, in that I've never seen another tool that can put more stress on a CPU, so it's great for testing stability and cooling efficiency.

I'm far from being an expert when it comes to overclocking, but I've never had a OC'ed chip die on me yet, and based on how my E4300 is performing, I don't think I'm on the edge of it's thermal limits either.

Good luck on your OC! Continue using common sense, like you have been already, and you'll do just fine.

 

[little_scrapper]

There is absolutely no way to read the actaul temps from the C2D chip via software. ALL programs are simply starting with a baseline temp (a fairly arbitrary number based on which chip) and by using an assumed Tj Max temp of either 85C or 100C doing some calculations and giving you a number. Again this number is based off of a calculation with assumed variable and therefore is and approximation. A refrence point, if you will.

Since TAT uses 100C for the E4300 and Speedfan uses 85C that is why TAT will read +15C on certain settups.

 

[Heyray2]

little_scrapper,

You are exactly right.

Take a simple comparison of my two newest machines, running the exact same MoBo (MSI P6N SLI-FI BIOS v2.0) and differing only in CPU (E4300 and E6600), the E6600 reads identical temps in Speedfan and TAT, while the E4300 reads 15c higher in TAT.

Two chips with very similar design, running in similar voltage ranges. I sincerely doubt the E4300 runs 15c warmer, epecially when the E4300 has a much more efficient HSF (Freezer Pro 7 on E4300 vs CNPS7700 on 6600).

And as quoted from CompuTronix in the C2D Temp Guide:

(E) Any hardware and / or software may misreport Tcase and / or Tjunction temps.

That would certainly be the case with TAT on the E4300.

 

[little_scrapper]

TAT uses Tj max of 100C instead of 85C for the E4300 Hence the +15C temps it gives. It may actually give +15C for all L2 stepping CPU's but Im not sure of this. All E4300's are L2 stepping.

I have actually proven this point by turning off the fan and blocking the sides of the HS until throttling kicked in. kicked in at 83C on Speedfan and like 97-98 on TAT. I also did some calculations by using TAT idle/100%load vs orthos idle/load vs TAT idle/70%load and came to the conclusion that TAT wasnt linear. The temps based on the three different loads indicated taht Speedfan was very linear and that TAT's readings were based on a curve.

 

[dr_kuli]

http://www.alcpu.com/forums/viewforum.php?f=35

In case you're too busy to read all:

According to one of the more interesting threads I've read in Intel's open Dev forums, a very unpleasant picture arrises. Desktop/Server CPUs DON'T have a constant Tjunction temperature, unlike many Mobile CPUs which contains the Tjunction information in bit 30 of MSR 0xEE. This value is either 85/100. Possibly by a mistake, this value seems to exist in the desktop processors as well, but this does not make it valid at all.

TAT was developed for mobile CPU usage. The whole Tjunction minus DTSDelta was good for these CPUs(with small deviations of several degrees above or below), but it is basically useless for desktop CPUs. In desktop CPUs the calculation is PROCHOT# minus DTSDelta, and obviously we cannot read PROCHOT# temp.

And THESE CPUs were without an IHS...

 

[Heyray2]

dr_kuli,

Thank you for the link to that information. That was quite useful and informative.

It seems like the more information we dig up about TAT, the less accurate of a gague is seems to be for C2D OC'ers to use for temp monitoring.

I"ll just stick with my trust SpeedFan and common sense.

 

[dr_kuli]

I"ll just stick with my trust SpeedFan and common sense.

And precise lab thermometer :lol:

 

[r34550r]

heheh so it seems that ill just stres the biach up over 3gh and keep my hand in the pants... i mean on radiator lol

until this crap is stable ill be happy

thx guys again for very good info . and im sure that keeping this post up will help some ppl coz im sure im not the only 1 with temps on tcase higher than cores wich is idiotic

thx again wasted reassor out

 

[little_scrapper]

No.. your the only one.