Ad
News

Researchers easily cracked RFID car alarm system

Published on January 30, 2005

Researchers from Johns Hopkins University could crack car alarm system used in millions of cars, based on radio-transmitter chips that make possible high-security car keys and swipe-by gasoline passes. Read more

Zone Alarm claims it's cracked spyware problem

Published on July 18, 2005

Zone Labs claimed the latest version of Zone Alarm, version 6.0, will prevent malware and spyware from gaining a foothold on a PC. Read more

Forward car collision warning system tested

Published on April 01, 2001

An automotive industry group working with US Federal safety officials has produced and is testing a forward collision warning system. Read more

Jeeves Alarm Clock Wakes Sir With A Gentle Message

Published on March 09, 2007

Would sir care to stop hitting his alarm clock? Read more

Last Reviews & Articles

Round Up: Five Powerful, Light Ultraportables

Published on November 05, 2008

Executives, road warriors, and gadget geeks all lust after ultraportable notebooks. Five of these amazing machines battle it out in this roundup. Read more

Core i7: 4-Way CrossFire, 3-way SLI, Paradise?

Published on November 04, 2008

For the first time ever, gamers are being treated to a reasonably priced platform enabling the best that AMD and Nvidia have to offer. Is Core i7 a gamer’s nirvana or does the processor serve up more of the same? Read more

Editor's Corner: Overclocking Core i7

Published on November 04, 2008

After discovering Intel's Overspeed Protection in yesterday's first-look at Core i7's performance, we're back to set the record straight on the overclocking headroom of the flagship 965 Extreme and entry-level Core i7 920. Read more

Intel's Core i7: Blazing Fast, But Crippled O/C

Published on November 03, 2008

Core i7--previously referred to as Nehalem--requires new motherboards, coolers and memory. Its performance is compelling and means AMD is falling behind even further, but Intel is putting in some speed bumps that will impact overclocking enthusiasts. Read more

  Tom's Hardware UK and Ireland Forums » CPU & Components » Other Components » Tj Max 100C - Cause for Alarm?
 

Tj Max 100C - Cause for Alarm?

Advanced Search

There are 369 identified and unidentified users. To see the list of identified users, Click here



Word :   Username :  
 
Bottom
Author
 Thread : Tj Max 100C - Cause for Alarm?
 
Profile: stranger
More Information

Hi,

I hope I've posted this in the right section! :)

I've been having this recurring problem - but should I be alarmed?

At least 5 or 6 times a week, my desktop suddenly becomes non-responsive. That is, none of my programmes/applications work or espond at all - though slightly curiously, I can still surf the web. I then have to turn my computer off - as it just freezes if I try to reboot. After that it works fine again - until the next time it becomes non-responsive. And so on.

I am 99-100% certain it is nothing to do with viruses, spyware, adware, malware, etc - as I've ran near endless scans & checks (including online ones) and nothing of any importance is ever found. And when I go to Administrative Tools/Event Viewer, no warnings or alerts are flagged up. In which case, it strikes me as fairly logical to think that it might be something to do with Temperatures and/or Voltages.

Yes, all my fans seem to be whizzing around how they should.

So, I installed Core Temp 0.99.1. On average - and as I write, with the PC having been on for a couple of hours - Core #0 is 30C and Core #1 is 31C - which seems OK to me. However, my "Tjunction" temperature is around 100C - which, to my (albeit non-expert) ears, does sound rather on the high side. Maybe alarmingly so.

Next, I installed SpeedFan 4.34 - and, yes, there is indeed a warning about the Core temperature being 55C. I then enabled Automatic Fan Speed - but nothing seemed to change. And I am uncertain as to how to properly use or configure SpeedFan. I have sent an email to the author of SpeedFan outlining my problem (or potential problem).

Finally, I installed RealTemp 2.70 and it told me this:

Core Temperature - 26C, 27C

Distance to TJmax - 69C, 68C

Minimum temperature - 26C, 26C

Maximum Temperature - 28C, 34C

My Tech Specs are: XP Home+Vista Ultimate (32-bit) \ Intel E6750 @ 3.47 \ Abit IP-35 Pro \ Antec 900 Ultimate Gamer \ Arctic Cooling Freezer 7 Pro \ 2gb Corsair Twin2X2048 XMS Xtreme Performance (800mhz) \ XFX 8800GTS 320mb \ Samsung SpinPoint T 500gb \ Plextor PX-740A \ Corsair CMPSUHX 620w PSU \ Speedtouch 510 v6.

By the way, all this is happening on XP Home (my 'default' O/S) and I am currently NOT running it Overclocked with the Abit uGuru software - that is, everything is at 'defaults' (including my Graphics Card).

Have you got any tips or advise on what I should do next? If you have, that would be much appreciated!

Anyway, I look forward to hearing from you,

DANNY.

Related Product

Register or log in to remove.

Profile: old hand
More Information

TJMax is the temperature value at which your CPU cores will begin throttling in order to prevent overheating. You can also think of it as the maximum temperature your CPU will run before it tries to save itself. The cores can actually run hotter before dying, but if you have thermal throttling enabled, then TJMax is as high as it can go before lowering the speed. 100°C is actually not accurate, it should be 85°C for that processor (or whatever Real Temp says, maybe 95°C). I would trust Real Temp, but not the other programs you listed. Real Temp is built based off of real-world testing where Core Temp and Speed Fan are based on guesswork.

 

As for your problem... have you tried a repair install of the OS? Also consider that the hard drive could be going bad or that you may have a faulty power supply. Finally, test your memory with Memtest86+


Message edited by qwertycopter on 08-04-2008 at 02:14:45 PM
Profile: member
More Information

yuck. where to begin...well Tjuntion is always 100c. It is on every computer i've seen... i dont remember why...maybe cause it cant get a proper reading so it goes to 100? either way its nothing to worry about and not your problem.

I'd like to think it has to do with how you're dual booting/ partitioning the drive. Did you try using any diagnostic tools to rule out hardware?

Memtest86+ for your memory
prime95 for cpu (even though its not oc'ed just run it to see if you get any error)
do a disk cleanup/defrag/chkdsk for the HDD and then run HDtune and do an error scan.

I like SIW, system information for windows, better than SpeedFan (more accurate) and its free. Oh and how hot is your GPU? might wanna check that temp. but yeah not sure bout the whole dual boot thing cause ive never gone through this process before. when did it start happeneing? might want to try a system restore if it wasnt too long ago.

Profile: Forum Veteran
More Information

First off... there is no tjmax that the software can detect. Intel does not publish what the tjmax is. Any program (coretemp/realtemp/speedfan/hd monitor) that reads core temps rely on a tjmax. If you ever used those programs, you should notice different temps it reports. That is because the tjmax is a guess, an assumption of what it is.

Also, you really shouldn't rely on core temps for idle temps. One reason why the tjmax does exist, is that it is part of a process for thermal throttling and thermal shutdown.

It's perhaps best to go by the Tcase sensor, that your MB app will read, what your bios health monitor reads, and speed fan should also see that sensor. Intel does publish the thermal spec, which would be the Tcase max that you shouldn't exceed.

Profile: old hand
More Information

Grimmy,
Don't lump Real Temp in there. UncleWebb used actual real world measurements to determine TJMax. It varies, either 85 or 95 on the Core products. I believe it has to do with the process size (45nm vs 65nm). The way he determined this was:

 

Core Temp = TJMax - DTS

 

thus

 

TJMax = Core Temp + DTS

 

He found the Core Temperature at a given time by measuring with an infrared thermometer, accurate to within 1°C. So the formula became:

 

TJMax = Measured Temp + DTS

 

DTS (Digital Thermal Sensor) is reported by the core, so it's possible to calculate to within 1°C of the actual TJMax value.

 

UncleWebb has tested numerous Core chips and a few other people have verified his results with the same testing method.

 

Programs like Speed Fan and Core Temp have only guessed at the value. This guess is based on the TJMax value for Intel's mobile Core products which was publish as 105°C. Intel has not published the values for its desktop line processors, but testing has shown it is not as high as the mobile chips.


Message edited by qwertycopter on 08-04-2008 at 05:38:36 PM
Profile: Forum Veteran
More Information

What I'm trying to get across... It seems people assume that when they install and use a temp program, to read core temps, that it is correct. That the program actually gets a tjmax from the CPU. Which it does NOT.

 

I never said anything about UncleWebb being wrong, but since you brought him up, in order to use ANY program that gives core readings, you need to calibrate it. (Even RealTemp)

 

Edit:

 

And if you ever used the "Test Sensor" on RealTemp, read what it says at the very bottom of the message when its done testing.


Message edited by Grimmy on 08-04-2008 at 06:40:30 PM
Profile: old hand
More Information

Grimmy,
I misunderstood your original point. You're right, TJMAX is a fixed value, an upper limit of temperature that the CPU wants to avoid crossing. This value is not read from a sensor, it's a static number set to facilitate throttling.

 

You're right, the DTS reading is not linear. In RealTemp you need to calibrate it for idle temperatures because the reading is wrong otherwise. And of course there is a degree of inaccuracy with RealTemp, but it's the closest approximation of the programs in question when used properly.

 

RealTemp is the most accurate "guess," as you put it, at TJMax. It can be assumed beyond reasonable doubt to be an accurate guess, within 1-2C... much better than Core Temp which assumes a TJMAX that's perhaps 15C too high. With proper calibration you can get a fairly accurate idle temperature as well.

 

All that said, load temperature is really the only one to be concerned about. My point is that RealTemp isn't just guessing at TJMax, it's an educated guess, one you can trust.


Message edited by qwertycopter on 08-05-2008 at 12:52:26 AM
Profile: Forum Veteran
More Information

Misunderstandings happen, so I just wanted to clarify.

I didn't mean to make realtemp look bad, and needed it to clarify the different tjmax's of different programs that everyone would see a difference.

But I'm glad we came to an agreement. :D


  Tom's Hardware UK and Ireland Forums » CPU & Components » Other Components » Tj Max 100C - Cause for Alarm?

Go to:
 

Google ads