kingaff

Troubleshooting Sudden Reboots/crashes

Seems there are a lot of posts on sudden reboots, so I thought I’d write up a troubleshooting post.

If anyone has any feedback or changes to make, let me know.

This is basically a “shotgun” approach, as sudden reboots and crashing can be caused by so many different things. Here are some of the basic things to try before posting and asking for help. If you have gone through all the below and you are still crashing, then make a new thread and be sure to include a detailed list of your computer’s specs, including:

CPU:
Motherboard:
Ram Speed, size and #sticks:
GPU:
PSU (do not leave this out):
OS:

Also include details of the crash. When does it typically happen? What EXACTLY happens?

And a history as well. At one point, did your system work? Have you recently installed some new hardware/software?

The main cause of system crashes is drivers. If you are ONLY able to boot up in safe mode, this is usually a sign that you have a driver issue.

1. Before going any further, update the following:
Video card drivers: NVIDIA Home or Advanced Micro Devices - Smarter Choice - Graphics and Media Processors
Motherboard chipset drivers: go to your motherboard maker’s website

Also take a look at how outdated your bios is, and consider an update. If there have been 3 or 4 revisions since the one you have currently installed, it is probably best to update.

2. Check your “Event Log”
As said above, the main cause of system crashes is outdated/corrupt drivers installed. It can be tricky to find which driver it is that is causing the crashes. If you are able to boot your computer up, right click on “my computer”, “manage”, “event viewer”. Here you will find 3 categories and a list of each. Your system reports any errors it encounters here. You want to search in all three categories and look for any details with a red symbol by them. Yellow are also significant, but red ones are the ones to research first.

To research them, the best way is to look inside the details provided. There is a link in there that tells you to “click for more details”. Do it. It’ll take you to the Microsoft site with some possible causes and solutions.

If the crashes seem related to a particular piece of hardware in your system, that is your hint. Update the drivers for that hardware and see if that fixes you up. For example, if you are getting constant error reports for your internet connection, try updating your NIC (network interface card) drivers.

Some other useful links in helping to figure out error codes:
http://www.microsoft.com/technet/sup..._advanced.aspx
EventID.Net

3. Auto reboot on error and reporting minidumps
Windows XP is set by default to reboot your computer when it hits a critical error. For most users I suppose this is a good thing, as no one likes to see the BSOD (blue screen of death). But in this case, we actually WANT to see the BSOD because it contains some useful information in diagnosing the problem. In order to force your computer to show you the BSOD, need to tell it to stop auto restarting on critical errors.

Right-click "My Computer", and then click "Properties".
Click the "Advanced" tab.
Under Startup and Recovery, click "Settings" to open the Startup and Recovery dialog box.
Clear the "Automatically restart" check box, and click OK the necessary number of times.
Restart your computer for the settings to take effect.

If you are suffering from BSOD, please see this thread here by Howard:
Before posting your minidumps, please read this. - TechSpot OpenBoards

4. Power supply issues
My personal favorite rant. People spend thousands on great parts for their computer, and then by a $30 power supply to power it all up. Cheap power supplies will crash your system, and these types of glitches are extremely hard to diagnose. They make your other parts act unstable, and it LOOKS like it is those other parts that are faulty. Or even worse, the cheap power supply actually does damage to other parts.

If you suspect your power supply, you can download software to check your voltages on the various rails. One such program is “speedfan”. You can find it here: SpeedFan 4.31 download - TechSpot Downloads

You want to check how many volts your PSU is putting out on the various “rails”. There is the 3.3v, the 5v and the 12v and the readings shouldn't deviate more than 5 or 6%. Note, this 5-6% rule only applies to the +3.3v, +5v and +12v rails (not the negative rails).

Another common problem is people updating their computers with newer parts, but not updating their power supplies to match these changes. Newer graphics cards in particular need a lot more power on the 12v rails. You should be up in the 24-30 amp range if you are running the latest graphics cards. You can tell how many amps your PSU is putting out on each rail by looking at the sticker that is attached to it (inside your case).

It is also possible that your PSU doesn't have enough Watts for the updates you have made. You can calculate what your system actually needs here: eXtreme Power Supply Calculator v2.0

(While we are here, if you have recently installed a new graphics card, many require their own power supply connection. This step is often forgotten, so if you are having problems running games, this may be your cause).

And a final note somewhat related: everyone should have a good surge protector. Well worth the investment!
Newegg.com - Once You Know, You Newegg

5. System temperatures
Most motherboards today have a feature built in which will automatically shut down your computer if it is reaching dangerously high temperatures in order to prevent damaging any parts. There are 5 main parts that could be overheating

CPU
Chipset
GPU
HDD
PSU

Download and install speedfan: SpeedFan 4.31 download - TechSpot Downloads to monitor your temperatures.

Your CPU and chipset temps shouldn’t be getting anywhere near the 65-70c mark when at full load. If your temp is at these levels when idle, you are running too hot.
Your GPU temp will depend on the card, but they usually run a lot hotter than CPUs so don’t be alarmed if it is showing a reading of 70c.

If you find that your system is in fact overheating, take a look at the fans. Are they caked in dust? Also check your case fans. Are they spinning? If all fans appear to be working, you may need to re-seat your CPU. Here is a well written article on how to do it: DIY: Installing a CPU and Heatsink

Graphics card overheats are fairly common. Usually the symptoms start with “artifacts” showing up on the screen, but sometimes it will actually shut down your system. To troubleshoot, you can run with your PCs case open, and position a house fan to blow directly on the graphics card and try gaming. If your crashes happen less frequently or stop completely, you have found the problem

Here is a list showing what the critical temperature is for most CPUs. You don't want your CPU within 10c of these temps:
The Heatsink Guide: Page not found

6. Undo Recent Changes to your system
Have you recently added any new software or hardware? You can always undo the system changes you have made, and tell your system to reset itself to a date where you knew everything works. Note: this will delete any new software/drivers you have installed after the date you choose to go back to.

Log on to Windows as Administrator.
Click Start, point to All Programs, point to Accessories, point to System Tools, and then click System Restore. System Restore starts.
On the Welcome to System Restore page, click Restore my computer to an earlier time (if it is not already selected), and then click Next.
On the Select a Restore Point page, click the most recent system checkpoint in the On this list, click a restore point list, and then click Next. A System Restore message may appear that lists configuration changes that System Restore will make. Click OK.
On the Confirm Restore Point Selection page, click Next. System Restore restores the previous Windows XP configuration, and then restarts the computer.
Log on to the computer as Administrator. The System Restore Restoration Complete page appears.
Click OK.

6. Faulty Ram / Dying ram

Another one of the most common causes of random crashes and/or system instability is your ram. Ram can easily become damaged from system spikes, or simply from basic wear and tear. And with ram it is not only physical damage that causes problems, but also compatibility issues.

Motherboards are very selective about what kinds of ram they work well with. And if you are working with more than one stick at a time (as many do), you run the risk of the different sticks of ram not working well with each other.

When in doubt, a good starting place is to run your system with just one stick of ram for a while and see if your problems lessen.

Visit your motherboard maker's website and make sure the ram you have is approved for your board, or at least meets the requirements for compatibility.

You can test your ram using memtest: Memtest86+ - Advanced Memory Diagnostic Tool
download it and run it on your system from DOS. Let it run for 6 or 7 passes (takes a few hours).

Other things to try based on Private Messages other users have sent me:
-if you have an APG card you can up the aperture in bios (advanced/chipset)
Mine was set at 64mb and I upped it to 128mb and I no longer get crashes etc (contributed by: A-thru-Z)

Cheers!

TBIRD

Thank you for this GREAT post.
TBIRD