Bad CPU thermal sensor?

Status
Not open for further replies.
I apologize for the length of this, but I want to give as much info as possible as this problem has been ongoing for quite some time. Thanks in advance for any advice you may offer. First, my system specs:

Epox 4PDA2+ Motherboard
Intel Pentium 4 2.6 GHz proc
Thermaltake Silent<something> 480W PSU
1 GB PC3200 RAM
MSI Nvidia 6600GT Video (previously All-In-Wonder Radeon 9800 Pro)
Thermaltake CPU Heatsink & Fan (previously stock heatsink & fan)
1 SATA hard drive, 1 IDE hard drive, 1 DVD-R drive, 1 CD drive
Sound Blaster Audigy Gamer card

Ok, for about the past 6 months to a year, this system has had problems cold booting. It produces an alternating high-low beeping for approximately 30 seconds, but eventually boots. Once the machine was up and running it ran -perfectly-, which is why I've probably let the problem drag on for so long. Well, the problem has been worsening recently and finally came to a head when I decided to shut down the system (my big mistake) to install a new power strip. Of course the same problem occurred so I decided to look into it further. However, most recently after booting and while playing a game, the system had some kind of failure and after some electrical 'whining' noises, shut down. So finally I decided to look at this seriously.

I began with the RAM, seeing as the post code that occurs the longest during the beeping phase is 'C1' which I believe indicates some kind of RAM problem. I tried each of the 512 MB RAM sticks in each of the slots individually and still had the problem. I also tried a completely separate stick of RAM and had the same problem. When the computer boots, I also tried a bootable DOS RAM checking utility which found no problems. So, I think I can eliminate RAM as the problem.

Next I decided to look into video. As I mentioned I previously had installed the ATI All-In-Wonder 9800 Pro. I tried booting the system with just the video plugged in, same problem. Also, somewhere along my troubleshooting way, my video began booting with garbage characters and skewed text. I plugged in an older Nvidia card (4x00 something) and by chance, the system booted with no beeps whatsoever. So, I figured video must be the problem. I later found a scorch mark on the power plug going into the ATI card and have since RMA-ed the card for repair. So, I ordered the NVidia 6600GT card and am having the same beeping problem.

Ok, next look at the CPU and motherboard. My CPU has always run 'hot', at around 75C or greater (its gone as high as 125C). Only recently have I found that this is ridiculously hot and never thought anything of it because I didn't notice any performance problems. The temperature gauge also alternates wildly varying hot temperatures with dropping completely to zero. As a precaution, I thought it would be good to replace the heatsink and fan. I (under the supervision of a IT professional) replaced the stock heatsink with a copper heatsink and fan with Artic Silver thermal paste. The system still reports about the same temperature and still drops to zero occasionally. My 'system' temp is typically around 35C, which I believe to be accurate. Additionally, the CPU heatsink is cool to the touch.

Currently, the system randomly may boot or not from a cold boot. It has taken to mixing the high-low beeps with another steady recurring beep and frequently shows the post code 'AF'. If the boot consists of both beeps, the machine will not proceed any further. However, after resetting a few times, and getting just the high-low beep, the system will boot. Now, unfortunately the system runs extremely slow as if the mobo or CPU is throttling itself because it thinks its too hot. I've tried utilities like Motherboard Monitor and they report near the same temperature as the BIOS. I can't imagine that the CPU is actually running at this temperature. Wouldn't it have burned up long ago??

So this is where I sit as of right now. I'm not sure what else to try. I'm running out of options. Here are some other random things I've tried:

- Tried taking the system into another room and plugging solely the power cable into the outlet. This would seem to eliminate the power strip and power outlet being the problem. However, I have the same beeping problem. Though, curiously I -have- noticed different behaviors when I've taken the system to a different house. Either it beeps more frequently at the other place or at my apartment building. I live in a fairly new apartment building (~10 years old) so I can't imagine incoming power would be an issue, but I can't rule anything out.
- Tried swapping out the PSU for an older, 350W power supply. Same problem with just having the mobo and video card plugged in.
- Tried replacing the power cable from the PSU to the wall, same problem.

So, again I'm running out of ideas here. I'm so tempted to just get a new motherboard and CPU and be done with it (not my first choice obviously) but I keep getting this 'dangling carrot' feeling that if I just replace one more thing that it will magically start working. Any advice you guys can offer is most appreciated.
 
I would guess that the mobo is bad, but I don't know. If the mobo was bad I think it could read temps wrong or fry a graphics card, and have boot problems. I think your cpu would be fried at those temps, so I doubt it actually is that hot.
I think the mobo is the most likely candidate.
 
Was it a whine or a whistle? Sounds like maybe a leaky capacitor. Sometimes the caps will go out with a whine/whistle. Check out the ones on the mobo to see if you can see any leaking around the top. Sometimes it's the ones in the PSU but if you tried a different one then that's probably been eliminated.
 
It sounds like a bad temp sensor. I believe you will need a new mobo. The processor would have cooked at around 85-90c so you know those readings are bad. Also, the processor will throttle back at high temps so that's what is causing the slowdown. If you can find the same mobo somewhere like eBay, you could just swap it out with no changes to Windows. If you were thinking about upgrading, this would be the time. :cool:
 
DeLaRocha I have pretty much the EXACT same problem with the temperature sensor. I upgraded the cpu to a 3.0 and still have the issue. While swapping the cpu I noted the sensor which is a small black chip in the middle of the cpu socket, which looked undamaged. Mine reports 0 degrees at boot and later fluctuates to 85C. New cooler installed and system temp at 33C. So there is no way this temp sensor is accurate, and iss causing us the temp alarm boot nightmare. Anyone think this sensor can be repaired or disabled? You can make sure your bios temp throttle is set to 15% at least, so the PC is usable.
 
If it is an INTEL processor, the TCC circuit will shut the VCore down when the core reaches 135C which I suspect would happen before the temp TC reached 85C
 
Thanks for your replies! This is the second intel processor I've tried, and has even better cooling...the sensor is obviously lying. I never had a thermal shutdown while running, only the boot time fiasco. It developed over time as well, after running fine for a long long time.

I don't think the sensor can be bent, it looks like a transistor soldered right to the board. It's not like the kind that's on a wire.

I'm using a RAID array on SATA on this board. I tried to flash to the 12/03 bios long ago but it caused the disks to be unreadable. I could go this route, but I would have to somehow back up and restore to avoid an entire reinstall. This machine does a ton,..music editing, gaming, and software development with all the tools, so I don't want to start from scratch. Any suggestions on how to backup and restore a huge disk would be appreciated!

I'm wondering if the sensor is reading any actual temp at all, and if some kind of thin insulation layer would help.
 
Status
Not open for further replies.
Back