Diagnosing the most common memory problems
Memory sizing error during POST (initial system boot)
This is one of the more common memory problems, occurring mainly in older computers such as the 486, earlier 60-120MHz Pentiums and even some of the Pentium Pro vintage that use EDO memory. Unfortunately though, we have seen a few instances of this occurring in new PC’s as well.
The 486 and early Pentium types
When this “memory sizing” error occurs in older computers of the 486 and early Pentium types, it involves systems using SIMM modules and the addition of more modules, or the replacement of old modules with new but with an increase in the total amount of memory. This increase in memory causes a problem with the BIOS on the motherboard as the BIOS recalls the old amount of memory and must be adjusted to reflect the new amount as well as any memory parameter changes. To adjust the values in the BIOS settings, boot the system into the BIOS menu (you should have the manual for you motherboard to do this), make the adjustments, and then save the settings. If you need more definitive help, visit our Technical Support area, in the menu select Motherboard and BIOS Issues and then select Clear CMOS or Passwords.Pentium (150-233MHz) through Pentium II (233-300MHz)
Having a “memory sizing” error in Pentium (150-233MHz), Pentium Pro (150-200MHz) and Pentium II (233-300MHz) based computers is extremely rare. Not impossible, but rare! This happens every once in a while on some of the early TX, HX, GX motherboards. Obviously, the first place to check would be the motherboards BIOS settings. The next place to look would be the modules themselves. Check and verify that all modules are alike and that the parameters of the memory modules are supported by the motherboard, especially issues such as parity versus non-parity and ECC (error correcting code) support.Later Pentium II through Pentium III and IV
Later Pentiums, those motherboards with the GX, LX, BX and 800 series chipsets (with auto recognition), are not normally susceptible to this problem. On those motherboards that permit the user to modify all BIOS settings, it is possible to see a memory sizing error, but in only those situations where the user has the ability to set the BIOS to a predetermined memory amount and auto-recognition has been disabled. Later machines are more apt to see parity errors, ECC errors and poor performance due to a mismatch of memory types, such as those experienced with poor quality memory modules.
Troubleshooting Memory Failures without Memory Testing Equipment
This section, and all that follow from this point forward, presume the reader has, at the least, a general understanding of computer hardware and operating systems. Although you should be a trained technician to perform some of these tasks, a basic understanding of hardware and operating systems will allow you to perform some of the various diagnostic procedures that will be outlined.
We will do our best to describe the entire process for you, however it is beyond the scope of a Web site troubleshooting guide to provide for all the possible situations you might face during a troubleshooting session searching for system memory failures. If, after following the diagnostic guidelines and methods outlines here, you require further assistance with either memory or non-memory related failures, please keep in mind that we do provide diagnostic services, repairs and upgrades. Of course you should always consult any operating and troubleshooting manuals provided by your computer manufacturer. If your computer is still under warranty, you may want to contact your manufacturers support system. If you still find that you do need assistance, please send us an e-mail and we will do our best to help you.
The following list represents the typical system problems normally associated with memory failures. Of course, there are other types of issues and failures, other than memory, that will cause these same problems. You will need to work your way through each issue below and pursue each possibility under that issue in order to root out the actual cause of the problem.
Issues and Possible Causes:
You computer will not boot-up even in DOS.
There are several possible causes for this to occur, aside from a memory problem. Here are some of the possibilities you may want to check.
- Computer power supply is failing or has failed.
- Motherboard failure, including the on-board voltage regulator.
- The processor has overheated due to a fan failure or a prior overheat condition has damaged he processor.
- Video card failure.
- Lightening strike or power spike damage to your telephone line has damaged the modem, which is causing the ISA or PCI bus to fail.
- Total physical floppy drive, hard drive or CD-ROM drive failure is causing a system electrical short.
- Memory module failure.
Initial Testing Procedure: Disconnect everything from the motherboard except the power supply connection, processor and memory and try to start the system and listen for beep tones. Count the beep tones and then refer to your owners manual. Click on these links for a list of AMI Beep Codes or Award Beep Codes and their meanings. If there is a memory error, you will receive the appropriate beep tones. If the only beep tone you receive is for a lack of a video card, shut the power off, insert your video card and repeat the process. Add back one component at a time until you discover the cause of the problem. If you have a memory error, then try swapping out the memory modules or run the software diagnostics found here.
Your system refuses to allow HIMEM.SYS to load from your Config.sys file.
There are several possible causes for this problem:
- This error may be caused by a fault in the computer’s external cache controller or by bad or mismatched memory chips on the external cache (generally 64, 128, or 256K cache SRAM).
- Incorrect Processor speed setting on the motherboard or within the BIOS settings.
- The Processor may be failing.
- Mismatched or bad system RAM.
Initial Testing Procedure: Disable the external cache on the motherboard. If disabling the external cache does not correct this problem, there may be a problem with the motherboard at the current processor speed, A20 handler problems, or bad or mismatched main system RAM chips.
If your computer displays the “ERROR: Unable to control A20 line!” message when you start your computer, HIMEM.SYS did not load properly and therefore cannot make extended memory or the high memory area (HMA) available. If HIMEM.SYS is not loaded correctly, programs such as EMM386.EXE and Microsoft Windows 3.0 and 3.1 are unable to use extended memory, and Microsoft MS-DOS cannot load into the High Memory Area. This problem occurs if HIMEM.SYS incorrectly identifies your machine type, or if the /CPUCLOCK and /MACHINE switches need to be added to the DEVICE command for HIMEM.SYS.
Your computer starts to boot-up but then hangs, or boots into the operating system and then upon starting a large application, the system reboots itself or hangs.
There are several possible causes for this problem:
- Your power supply is inadequate for the load (number of devices) connected to it, is overheating or the voltage to the motherboard is inconsistent.
- The processor may be overheating.
- One or more chips in the upper memory areas of your memory modules are bad or are going bad.
Initial Testing Procedure: Click here and verify that your power supply is adequate for the number of attached devices. If it isn’t, replace it with one that is. We can send you a new power supply via Federal Express. Check your processor and make sure that it and the attached heat sink and fan are clean and the fan is working. Like the Power supply, we can have a new fan to you over night. Lastly, swap your memory modules between slots, e.g., swap module 0 with module 1 or 2 and then start the system. If you need new modules, just let us know the make and model of the computer or the motherboard as well as the total amount of memory and we can furnish these as well.
Click here to Continue
Click here to go to the Performance Center Home Page