Help needed! Serious computer problem!

[grapefruit]

I've recently started getting serious problems with my computer and everytime I find and fix a problem, more problems come up and it's only getting worse! I've been having this computer for 6 years by now and never had even a single freeze or bluescreen, ever.

A week ago, when I turned it off normally, I noticed several processes failing during Windows' shutdown. I didn't think anything of it at first: the thing is old and often runs low on memory, causing similar bugs. Anyway, the next day when, the thing wouldn't turn on at all. The power LED would light up, the CPU and power supply fans would start, but that's about it. Nothing on screen, nothing loading, and it would turn off on its own after a few minutes.

So I unplugged it, opened the casing, and cleaned it throughoutly, thinking that dust might be clogging the CPU fan (happened once a few years ago and that was the problem). Still wouldn't turn on though.

I started testing the hardware. Sometimes it would boot, but would eventually power off on its own after a short while. Turns out it was a faulty RAM stick. Works fine with only 1 stick, but slow as bloody *censored*, obviously.

I went and bought a new pair of RAM sticks to replace the old ones. The first time I turn it on after that, it boots fine, works perfectly, even lots faster than it originally was. It shows no sign of being defective whatsoever, so I decide to test the new RAM to check stability. At first, no issues, everything goes fine... then suddently, a whole bunch of new symptoms appear. First, the computer becomes incredibly laggy and unresponsive for no apparent reason then when I click around, it completely freezes and beeps continuously. I decide to put the ram sticks in other slots, just in case...
I push the power button again... the CPU's fan is doing a weird crackling sound but it is running... before Windows finishes loading, after logging on, another huge lag appears. The task bar won't show up when I hover the mouse on the bottom of the screen, no program will run, and I get error messages when trying to shut down.
Third attempt at powering on the system: This time nothing happens when I press the power button. But it ends up booting many seconds after pressing it a few repeated times. As soon as the Windows boot screen appears, everything completely freezes.

Any help would be incredibly appreciated!!!

Model: Medion M3 Composer 4100
CPU: Intel Pentium4 @ 2.8Ghz
RAM: 2x512Mb DDR @ 400mhz (originally 2x256Mb DDR @ 400mhz)
HD: Seagate 160Gb

Note: Before anyone says "Your pc is too old, buy a new one": I don't have the money for that.

[Computer_Commando]

I have almost the same model, a 5100; found it in the trash with everything missing except the case, motherboard & CPU. After rebuilding it, it's now my primary computer.

Seeing as it's the original hard drive, I would test it first, it's more than 5 years old.  Get the Seagate SeaTools for DOS, burn a CD and boot from it.  If you think you can keep the computer running for a while, try SeaTools for Windows.
http://www.tacktech.com/display.cfm?ttid=287

[grapefruit]

The hard drive does work properly, I'm using it right now with no issue at all.
Unfortunately I dont think I can keep the comp running long enough for that. I'll check if there is a way to check a secondary drive with these tools though, and connect the drive to my laptop to perform said tests.

[grapefruit]

Update: I was able to perform a long generic test on the hard drive using seatools. The test was successful, no problems were detected.

[Computer_Commando]

That's good!  Lets see if something is overheating or voltages are out of spec or the fan is spinning too slowly.
Try this:  http://www.cpuid.com/softwares/hwmonitor.html
Capture a screenshot & post it here (Under Additional Options, below Message, for upload)

[grapefruit]

It would actually help if I was at least able to start the computer up and manage to get it to run long enough to run that... As it is now, I'm lucky if I even get to the user screen...
I do know CPU temperatures get high though, I've been monitoring them on a regular basis since the last few years, but they stabilize long before it becomes hazardous.

Gonna rebuild the comp now to initialize all hardware connections, and try to boot up/run the tool. I'll update later with the results, if it has worked.

[grapefruit]

Update: Okay! I have both good and bad news.
The good news: I was able to start up the computer and run under safe mode for a good while without any problem. I was able to perform a full scan using MalwareBytes, and run both HWmonitor and Speedfan.

The bad news: Right after finishing the scan, the continuous beep came. However this time it did NOT freeze. It just became veeery slow, with several unresponsive programs, and that beep. Unfortunately, that was before I could save any screencap of HWmonitor and Speedfan. When trying to open MSPaint I'd get this error message:
"The application or DLL c:/WINDOWS/AppPatch/acGeneral.DLL is not a valid Windows image. Please check this against your installation diskette."
And when trying to bring the alt/ctrl/del Window, I'd get this error message:
"taskmgr.exe failed to initialize properly."

The reason why I brought up Speedfan is because HWmonitor didn't show much data at all. Only one temperature to be exact (the hard drive one), nothing else, while speedfan showed many, as well as the speed of the fans. From what I have seen, the CPU temperature is a bit high, as expected, around 79 celcius. Other temps were either below 50 or just plain bogus (I had a -56....). The CPU fan was said to be over 4000 rpm. However, there were 2 other fan data at 0rpm. The only other fan in the computer is the power supply one. So either it is not working (either defective, or temps too low to need to work), either it's bogus data once again.

I have searched for BIOS beep codes to find the meaning of that one long neverending beep, to no avail. I use a Phoenix BIOS. They have a huge range of detailed beep codes, but nothing is said about a single continuous beep. For most other BIOS though, these mean a power supply problem.

[JJ 3000]

Find a power supply of equal or greater wattage. Swap it with yours and see if your problems persist.

[Salmon Trout]

CPU: Intel Pentium4 @ 2.8Ghz

the CPU temperature is a bit high, as expected, around 79 celcius.

That is not "a bit high". That is very high indeed. If it is a true reading, you have a hardware failure situation. This may well explain the other symptoms.

[grapefruit]

That is not "a bit high". That is very high indeed. If it is a true reading, you have a hardware failure situation. This may well explain the other symptoms.

I've been having this reading (and sometimes higher, during intensive use) for years and I've never had any problem. I still need to put some new thermal paste on though, to make sure the cooling is optimal. If it keeps crashing after that, I'll monitor temperatures until said crash to see if it could be in any way related.


I do suspect the power supply too though, since these problems began after replacing the RAM sticks with newer and greater ones. I'll try running the pc with the soundcard removed, in a attempt to compensate for a possible lack of wattage to feed all cards on the mother board. If it does work, well I guess there are good chances of having found the culprit.

I will update with results of these tests.

[BC_Programmer]

I've been having this reading (and sometimes higher, during intensive use) for years and I've never had any problem.

I don't understand. That hardly means the temperature is fine. The Maximum operating temperature for that processor is 65 degrees. The fact that "you've been running that way for years without problems" is not a logical way of eliminating it as a possibility, if anything, you've pretty much confirmed this is the problem. Thermal Damage, to my understanding, is cumulative.

[Salmon Trout]

Thermal Damage, to my understanding, is cumulative.

Cooked cpu and nearby motherboard components is my guess.

[Computer_Commando]

...I have searched for BIOS beep codes to find the meaning of that one long neverending beep, to no avail. I use a Phoenix BIOS. They have a huge range of detailed beep codes, but nothing is said about a single continuous beep. For most other BIOS though, these mean a power supply problem.

It's not a BIOS beep code, but likely an Overtemp alarm.
My 5100's H/W Monitor is disabled in the BIOS; I assume the 4100 is the same.  My BIOS is Phoenix-Award; yours is likely the same.
Since you said:  "...the CPU's fan is doing a weird crackling sound but it is running...".  I suspect the fan bearings are shot, get a new fan.
Remove CPU fan, clean off old thermal paste, reapply new thermal paste, install new CPU fan.

[grapefruit]

So I applied new thermal paste a few minutes ago. Works fine so far, the temperature dropped greatly and the fan stopped running at full speed non-stop. The crackling in the fan stopped, I assume it was just set loose.
I am writing this from said faulty computer while monitoring the system, stable so far. Also, somehow this time HWmonitor shows more than one value, so I am posting a screencap of it now.

Screencap'd right after booting, before doing anything else. VTIN and SYSTIN values seem bogus, everything else seems accurate.

[recovering disk space - old attachment deleted by admin]

[Computer_Commando]

So I applied new thermal paste a few minutes ago. Works fine so far, the temperature dropped greatly and the fan stopped running at full speed non-stop. The crackling in the fan stopped, I assume it was just set loose.
I am writing this from said faulty computer while monitoring the system, stable so far. Also, somehow this time HWmonitor shows more than one value, so I am posting a screencap of it now.
Screencap'd right after booting, before doing anything else. VTIN and SYSTIN values seem bogus, everything else seems accurate.

Loose fan will allow temps to get high.  Run it for a while & see what happens, but I think you've solved it.  65C is a good number.
Every time he releases a new version, he messes up the code for the older sensors.  I send him bug reports from time to time.  Look at the +12V reading, we know it's wrong!
Winbond W83627THF is supported:  http://www.cpuid.com/softwares/hwmonitor/supported-hardwares.html