Once you find a fault there are a number of things that can be done to determine the cause of the fault that should enable you to fix the problem. The information that should be collected and the actions to take will depend on the nature of the problem. The main types of faults are:
• | System restart or crash (Blue Screen Of Death) or system instability and |
• | BurnInTest reported errors. |
Finding the cause of a fault generally relies on finding information about the fault and performing a series of tests that help to narrow down the possible causes to, for example, a specific hardware component or device driver.
Collecting information to help narrow down the cause of a fault
General
BurnInTest reporting will provide information about the test failure. A description of the BurnInTest reported errors are shown in the Common error messages section. Generally the system component that has a problem is identified by the test that reports the error.
More detail can be logged about errors or activity with Level 1 and Level 2 trace files respectively (see Test reporting). Level 1 activity trace logging mainly outputs additional information regarding error conditions. Level 2 trace logging outputs Level 1 content as well as a significant amount of activity reporting that helps with the understanding of what the tests are doing in quite some detail (these files are generally very large). These reports can be selected to be in text or HTML format, however text files are generally easier to analyze.
Windows logs information for certain system and application events. This can include information such as device drivers failing and reloading. These events are included in the BurnInTest level 2 trace file so that they can be seen directly in the context of the test that is occurring. This can help explain some errors.
System restart or crash (Blue Screen Of Death)
When a serious error occurs in Windows, a device driver or hardware the system can crash. Windows can be configured to either restart on a crash or show a diagnostics screen (better known as the Blue Screen of Death, or BSOD). If the system crashes during testing, it is suggested that:
• | Windows be configured to show the BSOD, rather than automatically restart. E.g. In Windows 7 this can be configured in Control Panel->System->Advanced system settings->Startup and Recovery Settings->Uncheck System failure, Automatically restart. Often the BSOD will provide valuable information about the crash, such as the name of the device driver that crashed. If this is identified, then you can take actions such as updating the device driver that failed, or replacing or uninstalling the device hardware that failed. |
• | Check the Windows System event logs for errors around the time of the failure. |
Testing to narrow down the cause of a problem
If a fault occurs and it is not obvious which system component has caused the failure, you should try an narrow down the potential causes. Examples are:
- Test with another similar system (if you have one)
- Reduce the BurnInTest test set until you have the minimum test set that provokes the problem.
Some suggestions for common faults
What follows is some hints on how to go about finding the cause of a particular system instability. (i.e. The system locks up, you get the windows blue screen, etc..). We don’t want to try and explain the steps involved in each of these processes, they are just points that may warrant future investigation.
• | Check you don’t have any viruses. |
• | Check the hard drive for errors using the built in Windows error checker. (Right click on the drive letter in Windows explorer, select properties then tools / Error checking ) |
• | Check that free space is available on the hard disk(s) for the windows swap file. |
• | Don’t run all the BurnInTest tests at once when looking for a fault. Run just the 2D graphics, then run just the 3D graphics, then just the disk, etc.. This will allow the problem to be isolated to one area. |
• | If BurnInTest is reporting strange errors, try turning the 3D test off. Bad DirectX 3D video drivers can cause a lot of strange problems. |
• | Know that faulty RAM can show up in many different ways. eg. Disk I/O error or a system crash. |
• | Boot Windows up in Safe Mode and see if the system is more stable. |
• | If you suspect faulty hardware to be the source of the problem, and you know what you’re doing, pull out all the "optional hardware", e.g. LAN cards, I/O cards and see if the system is more stable. |
• | Once again, if you know what you’re doing, start swapping out components of the system to see if the fault can be localized. Obviously you’ll need some spare hardware to do this. |
• | Have a look through the issues in the precautions section of the online help file. |
• | Check that the power supply that you have is large enough to power all the components installed in your system. A power supply typically supplies power at several different voltages (e.g. 3.3V, 5V, 12V). You should check that there is adequate wattage available at each of these voltages. PC power supplies are typically in the 250 - 500 Watt range. (Avoid power supplies that are 250 watts or less ) |
• | Spikes in the main power supply can cause problems. If you suspect that you have dirty mains power consider getting a power conditioner / UPS / surge protector. |
• | Check that internal cards and cables are not loose and don't have dirty contacts. |
• | Ensure that your computer is operating at a reasonable temperature. Check that fans are operating correctly and heat sinks are making good contact. A Pentium III without it's heat sink and fan will quickly reach over 100 degrees Celsius. With a heat sink and fan the temperature is maintained at a more reasonable 45 degrees Celsius |
• | If you’re really stuck you may want to try a reinstallation of Windows on a reformatted disk. Think carefully about this option before you attempt it, there are lots of good reasons why you don’t want to reformat your hard disk. |
• | Make sure you have got the most up to date software drivers for your hardware. Drivers are a never-ending source of problems. |
• | Check that you haven’t ended up with an overclocked CPU and don’t know about it. |
• | Check that your BIOS settings are correct. Note down the original settings before you change anything however. You might also want to check that you have the latest BIOS that is available from your motherboard manufacturer. |
• | Check that you haven’t purchased the cheapest and nastiest hardware in the hope of saving a couple of dollars (or pounds, francs, etc.). Often it may not be the cheap hardware that causes problems but the quality and support of the software drivers that come with the hardware that are a problem. Don’t shop on price alone, check out the support and product reviews. |
General precautions for testing
Have a look through the issues in the precautions section of this help file.
|