BSODs/Freezes/AutoRestarts Trouble-shooting Starting Point.
BSODs, freezing and random reboots in WinXP can be caused by faulty hardware as often as software or driver conflicts these days. This post will try to highlight some of the common hardware problems, how to troubleshoot and possible solutions.
Quick Word about Software Conflicts:
With WinXP maturing as an OS and the assistance of WHQL, software developers are running into less and less conflicts with the XP OS in general. With that being said even mature 3rd party software goes thru revisions and in some cases a rush to market for competitive reasons and full compatibility testing was not carefully completed.
System Restore: By default WinXP has system restore turned on, in most cases this is how you would leave it. XP will automatically create a system restore point for you when it detects a significant system event like a new program install. But this does not always happen when you update a currently residing application. It is strongly recommended you create your Restore Point before you install or update any application. To do this: Start/All Programs/Accessories/System Tools/System Restore/Create Restore Point.
If you are suddenly experiencing BSOD’s/reboots/freezing the first question to ask yourself is , “What changed?” Did you install a new program or a revision to a program? At this point your best bet is to go back and uninstall what you just did.
For a list of common WinXP BSOD Codes go here:
http://www.geekstogo.com/forum/Windows-XP-Blue-Screen-of-Death-STOP-Codes-t43519.html
These codes are important and can help to direct your attention to the problem. A common problem with reboots/BSODs is a harddrive error or the beginnings of a harddrive failure. Sometimes a virus infection can cause a weak or aging HDD to fail or become inconsistent. A virus after all is just a piece of code which tells the computer to perform some function, if this function is to write blocks and blocks of code to the HDD, a weak HDD may not be able to keep up.
With that being said many HDD diagnostic programs, like Chkdsk or Fixdsk can put enough strain on a weak HDD to cause it's complete failure.
Other Common Hardware issues causing BSODs:
1) Failing or over taxed PSU: This is especially common after adding a new and fairly powerful graphics card or any new hardware actually.
A newer CPU like an Athlon 64/FX and the P4/P-D series uses in excess of 100Watts from the 12Volt rail alone from the PSU. A good GFX card can use 50Watts or more from the 5Volt rail.
Symptoms of an over taxed PSU are random BSODs, but especially under load like gaming or graphics editing or even watching a DVD.
Go here to check your PSU requirements:
PSUCalculator
ALL PSU are rated by maximum watts and this is NOT what they were intended to operate at continuously. You must take 20-30% off this for a good Brand Name PSU and 30-40% for a generic crappy. Also what the above link calculated you should add another 20-30% for safe operating and expansion headroom.
An Example calculation: Say the psucalculator gave you a wattage need of 250Watts. You buy a good Brand name Thermaltake. So you would buy a wattage of: 250X1.3= 325W minimum. Now add 30% for headroom and expansion (new stuff for the PC). 325X1.3=422.5, so you need 425W unit.
See how easy it is to under rate your PSU? Look for a PSU with lots of amps on the 12Volt output. This is what runs the CPU, motors (HDD, DVD/CD/FF) and fans. An AMD64 series or a 3.0+GHZ P4 will use 10 amps off the 12Volt rail alone. Here is a nice little list of PSU specs:
PSULIST
2) Overheating CPU/GPU or RAM: Check your temps in BIOS under PC Health or in Windows software like Everest/Motherboard Monitor etc. Ensure all fans are running and clean. Especially the HSF on the CPU and GPU. Ensure you have good thermal contact and have used thermal paste on the CPU before installing a new HSF. All else fails, take the side of the case off and put a big table fan blowing on the whole PC.
Temperature limits: AMD XP, Duron and Sempron (older core types) CPUs are around 85-90C. Athlon 64 it's a little lower at about 65C. FX series may go 5C higher. Newer Sempron cores (Paris/Palermo) 65-70C.
Pentium 2/3 < 1.1GHz range from 70-85C, but most fall in around 75C.
PIII Xeon 55-60C
P4 (1.3-2.5GHz) 65-75C (core dependent)
P-D 62-68C
For more of a complete list go here:CPUTempLimits
3) Bad or failing RAM: Remove, clean the copper fingers with an eraser and reseat the RAM. Run with one stick then the other. Run Memtest or Memtest86. Replace RAM.
Indicators for points 1) and 2): Running 3D games cause BSOD/Freezes. If the crash occurs almost immediately then 1) is most likely. If the game runs several minutes then it can be either 2) or 1) and 2). If you are running a 3D game and your FPS really starts to drop this is a strong indicator of point 2). This happens as the CPU or GPU automatically throttles back to prevent complete overheating and self destruction.
Indicators for point 3): Opening a new program or closing an active program. BSOD Stop code 02E.
If you are experiencing auto restarts/reboots: check the Start-up and Recovery options and "turn off auto restart on error". This will allow you to see the BSOD and the Stop code and can help to trouble-shoot the problem.
To do this: Right click on My Computer/Properties/Advanced tab/Startup and Recovery Settings in System Failure, uncheck Automatically Restart.
If this is already unchecked and the PC still reboots automatically this is a strong indicator of a overworking PSU. If the "Power Good" signal falls below 5V the CPU will automatically reset. This is the same effect as if you pressed the reset button on the front panel.
Minidumps: Often the only way to diagnose a BSOD problem is to analyze your minidumps. These are files Windows creates and saves to the hardware drive in the event of a system crash. Minidumps must be enabled on your PC first before Windows will generate these. To enable these:Right click on My Computer/Properties/Advanced tab/Startup and Recovery Settings. At the bottom in "Write debugging Information", scroll the window to Small memory dump (64K).
Advanced Techniques:
BSODs and Freezes are typical symptoms of an unstable OCed system. In order to correct for this it is practically mandatory for Ocers to increase the voltage on CPU, called Vcore, the RAM, called Vdimm and even the northbridge chipset.
You might say, “ But I haven’t Oced my system so this doesn’t pertain to me.”
My response or reasoning is this: The extra speed demanded of the CPU/RAM/Northbridge circuits in an Ocing attempt produces more heat in each of these components. This extra heat causes the internal semiconductor junction bandgap to shift to in energy and thus voltage. Without going into semiconductor transistor device theory, which is beyond the intended scope of this thread, suffice it to say the component (CPU, RAM etc) starts making “State” errors. These are 1’s and 0’s logic level errors. The binary logic that is the very “thing” that a PC is. By adding voltage to the Vcore or Vdimm we reinforce the junction bandgap voltage and 1 and 0 logic state errors are eliminated.
How does this affect me when I don’t OC my computer? Time and heat (i.e. aging) has a slow and accumulated effect of degrading the junction bandgap energy level. This occurs because the actual “dopant” atoms used to create the semiconductor junction in the first place begin to diffuse away from the junction location and penetrate deeper into the semiconductor lattice. The diffusion rate is driven by heat energy. In other words age is wearing out the proper operation of the CPU or RAM or northbridge itself!
To compensate for this we can increase the voltage on some of the key components in the PC, just like Ocers and create a stable system again.
WARNING: By adding voltage to these components we are also adding heat. Add heat we increase the aging rate. We need to add more voltage. You can see where this is going!
I recommend you add only the minimum voltage step available within your BIOS.
Start with Vdimm first. Reboot and see how it goes. If this does not help then add a little to Vcore and finally if required add a little to the northbridge chipset.
Hope this helps.
Cheers. :wave:
BSODs, freezing and random reboots in WinXP can be caused by faulty hardware as often as software or driver conflicts these days. This post will try to highlight some of the common hardware problems, how to troubleshoot and possible solutions.
Quick Word about Software Conflicts:
With WinXP maturing as an OS and the assistance of WHQL, software developers are running into less and less conflicts with the XP OS in general. With that being said even mature 3rd party software goes thru revisions and in some cases a rush to market for competitive reasons and full compatibility testing was not carefully completed.
System Restore: By default WinXP has system restore turned on, in most cases this is how you would leave it. XP will automatically create a system restore point for you when it detects a significant system event like a new program install. But this does not always happen when you update a currently residing application. It is strongly recommended you create your Restore Point before you install or update any application. To do this: Start/All Programs/Accessories/System Tools/System Restore/Create Restore Point.
If you are suddenly experiencing BSOD’s/reboots/freezing the first question to ask yourself is , “What changed?” Did you install a new program or a revision to a program? At this point your best bet is to go back and uninstall what you just did.
For a list of common WinXP BSOD Codes go here:
http://www.geekstogo.com/forum/Windows-XP-Blue-Screen-of-Death-STOP-Codes-t43519.html
These codes are important and can help to direct your attention to the problem. A common problem with reboots/BSODs is a harddrive error or the beginnings of a harddrive failure. Sometimes a virus infection can cause a weak or aging HDD to fail or become inconsistent. A virus after all is just a piece of code which tells the computer to perform some function, if this function is to write blocks and blocks of code to the HDD, a weak HDD may not be able to keep up.
With that being said many HDD diagnostic programs, like Chkdsk or Fixdsk can put enough strain on a weak HDD to cause it's complete failure.
Other Common Hardware issues causing BSODs:
1) Failing or over taxed PSU: This is especially common after adding a new and fairly powerful graphics card or any new hardware actually.
A newer CPU like an Athlon 64/FX and the P4/P-D series uses in excess of 100Watts from the 12Volt rail alone from the PSU. A good GFX card can use 50Watts or more from the 5Volt rail.
Symptoms of an over taxed PSU are random BSODs, but especially under load like gaming or graphics editing or even watching a DVD.
Go here to check your PSU requirements:
PSUCalculator
ALL PSU are rated by maximum watts and this is NOT what they were intended to operate at continuously. You must take 20-30% off this for a good Brand Name PSU and 30-40% for a generic crappy. Also what the above link calculated you should add another 20-30% for safe operating and expansion headroom.
An Example calculation: Say the psucalculator gave you a wattage need of 250Watts. You buy a good Brand name Thermaltake. So you would buy a wattage of: 250X1.3= 325W minimum. Now add 30% for headroom and expansion (new stuff for the PC). 325X1.3=422.5, so you need 425W unit.
See how easy it is to under rate your PSU? Look for a PSU with lots of amps on the 12Volt output. This is what runs the CPU, motors (HDD, DVD/CD/FF) and fans. An AMD64 series or a 3.0+GHZ P4 will use 10 amps off the 12Volt rail alone. Here is a nice little list of PSU specs:
PSULIST
2) Overheating CPU/GPU or RAM: Check your temps in BIOS under PC Health or in Windows software like Everest/Motherboard Monitor etc. Ensure all fans are running and clean. Especially the HSF on the CPU and GPU. Ensure you have good thermal contact and have used thermal paste on the CPU before installing a new HSF. All else fails, take the side of the case off and put a big table fan blowing on the whole PC.
Temperature limits: AMD XP, Duron and Sempron (older core types) CPUs are around 85-90C. Athlon 64 it's a little lower at about 65C. FX series may go 5C higher. Newer Sempron cores (Paris/Palermo) 65-70C.
Pentium 2/3 < 1.1GHz range from 70-85C, but most fall in around 75C.
PIII Xeon 55-60C
P4 (1.3-2.5GHz) 65-75C (core dependent)
P-D 62-68C
For more of a complete list go here:CPUTempLimits
3) Bad or failing RAM: Remove, clean the copper fingers with an eraser and reseat the RAM. Run with one stick then the other. Run Memtest or Memtest86. Replace RAM.
Indicators for points 1) and 2): Running 3D games cause BSOD/Freezes. If the crash occurs almost immediately then 1) is most likely. If the game runs several minutes then it can be either 2) or 1) and 2). If you are running a 3D game and your FPS really starts to drop this is a strong indicator of point 2). This happens as the CPU or GPU automatically throttles back to prevent complete overheating and self destruction.
Indicators for point 3): Opening a new program or closing an active program. BSOD Stop code 02E.
If you are experiencing auto restarts/reboots: check the Start-up and Recovery options and "turn off auto restart on error". This will allow you to see the BSOD and the Stop code and can help to trouble-shoot the problem.
To do this: Right click on My Computer/Properties/Advanced tab/Startup and Recovery Settings in System Failure, uncheck Automatically Restart.
If this is already unchecked and the PC still reboots automatically this is a strong indicator of a overworking PSU. If the "Power Good" signal falls below 5V the CPU will automatically reset. This is the same effect as if you pressed the reset button on the front panel.
Minidumps: Often the only way to diagnose a BSOD problem is to analyze your minidumps. These are files Windows creates and saves to the hardware drive in the event of a system crash. Minidumps must be enabled on your PC first before Windows will generate these. To enable these:Right click on My Computer/Properties/Advanced tab/Startup and Recovery Settings. At the bottom in "Write debugging Information", scroll the window to Small memory dump (64K).
Advanced Techniques:
BSODs and Freezes are typical symptoms of an unstable OCed system. In order to correct for this it is practically mandatory for Ocers to increase the voltage on CPU, called Vcore, the RAM, called Vdimm and even the northbridge chipset.
You might say, “ But I haven’t Oced my system so this doesn’t pertain to me.”
My response or reasoning is this: The extra speed demanded of the CPU/RAM/Northbridge circuits in an Ocing attempt produces more heat in each of these components. This extra heat causes the internal semiconductor junction bandgap to shift to in energy and thus voltage. Without going into semiconductor transistor device theory, which is beyond the intended scope of this thread, suffice it to say the component (CPU, RAM etc) starts making “State” errors. These are 1’s and 0’s logic level errors. The binary logic that is the very “thing” that a PC is. By adding voltage to the Vcore or Vdimm we reinforce the junction bandgap voltage and 1 and 0 logic state errors are eliminated.
How does this affect me when I don’t OC my computer? Time and heat (i.e. aging) has a slow and accumulated effect of degrading the junction bandgap energy level. This occurs because the actual “dopant” atoms used to create the semiconductor junction in the first place begin to diffuse away from the junction location and penetrate deeper into the semiconductor lattice. The diffusion rate is driven by heat energy. In other words age is wearing out the proper operation of the CPU or RAM or northbridge itself!
To compensate for this we can increase the voltage on some of the key components in the PC, just like Ocers and create a stable system again.
WARNING: By adding voltage to these components we are also adding heat. Add heat we increase the aging rate. We need to add more voltage. You can see where this is going!
I recommend you add only the minimum voltage step available within your BIOS.
Start with Vdimm first. Reboot and see how it goes. If this does not help then add a little to Vcore and finally if required add a little to the northbridge chipset.
Hope this helps.
Cheers. :wave: