Make your own free website on Tripod.com
BuildPC * CustomizeWindows * InstallOS * Internet * HardwareProblems * E-Mail * Home
PCWorkz

 

HARD DRIVE

In principle, a hard-disk drive is very similar to a floppy drive—a magnetic recording media is applied to a substrate material, which is then spun at a high rate of speed. Magnetic read/write heads in close proximity to the media can step rapidly across the spinning media to detect or create flux transistions, as required. When you look closely, however, you can see that there are some major physical differences between floppy and hard drives.
Booting a computer can take up to 30 seconds—often more. Some of this time is an artificial delay needed to initialize the hard drive. From the moment that power is applied to the hard drive, it can take anywhere from 7 to 10 seconds for the drive’s onboard controller to start and initialize the drive where it can be recognized by the system POST. This is known as the drive’s start time. Boot problems with a new hard drive are frequently caused by an insufficient delay at boot time. The BIOS attempts to check for the presence of a hard drive, which has not yet had time to initialize.

Modern hard drives are not simply “on” or “off.” They operate in any one of several modes, and each mode makes different power demands on the host system. This is particularly important because today’s PCs are becoming ever-more power conscious, so the ability to control drive power is an integral part of PC power-conservation systems. Typical hard drives operate in any of five different power modes:

SMART COMMAND SET Some of the newest hard drives use the Self-Monitoring Analysis and Reporting Technology (SMART) command set. SMART-compliant drives improve the data integrity and data availability of hard-disk drives by regularly checking for potential drive problems. In some cases, a SMART-compliant device will predict an impending failure with sufficient time to allow users to backup their data and replace the drive before data loss occurs. If you do or you don't have a SMART-compliant drive and you do recognize that you do have a problem then the examples below may be of service to you in your quest to remedy the particular problem you are experiencing.
Now it’s time to take a look at some problems and solutions. The important concept here is that a hard-drive problem does not necessarily mean a hard-drive failure. The failure of a sector or track does not automatically indicate physical head or platter damage—that is why software tools have been so successful. Even if one or more sectors are physically damaged, millions of sectors are on a hard drive. A few bad sectors do not render a drive faulty. One of the only times that a drive is truly irreparable is when physical media damage occurs on track 00, but software tools will help you to identify the scope of the problem.
Drive troubleshooting has the potential of destroying any data on the drive(s). Before attempting to troubleshoot hard-disk problems, be sure to back up as much of the drive as possible. If no backup is available, do not repartition or reformat the drive unless absolutely necessary, and all other possible alternatives have been exhausted.

Q&A

 

The hard drive is completely dead
The drive does not spin up, the drive light doesn’t illuminate during power-up, or you see an error message indicating that the drive is not found or ready. In most cases, you should suspect a power problem first. Be sure that the 4-pin power connector is inserted properly and completely. If the drive is being powered by a “Y-connector,” be sure any interim connections are secure. Use a voltmeter and measure the +5-V (pin 4) and +12-V (pin 1) levels. If either voltage (especially the +12-V supply) is unusually low or absent, replace the power supply. Also check your signal cable. See that the drive’s signal interface cable is connected securely at both the drive and controller ends. For IDE/EIDE drives, this is the 40-pin ribbon cable. If the cable is visibly worn or damaged, try a new cable.
The PC cannot use a hard drive that it can’t recognize, so enter the CMOS setup routine and see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all be correct—otherwise, POST will not recognize the drive. If you have an “auto-detect” option available, try that also. Remember to save your changes in CMOS and reboot the system.
If problems continue, the hard drive itself might be defective. Try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective, and should be replaced. If a known-good hard drive fails to operate, replace the drive controller board.

You see drive activity, but the computer will not boot from the hard drive
In most cases, there is a drive failure, boot-sector failure, or DOS/Windows file corruption. Check the signal cable first. Be sure that the drive’s signal interface cable is connected securely at both the drive and controller. If the cable is visibly worn or damaged, try a new one. You should check the CMOS setup next—see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all correct. Otherwise, POST will not recognize the drive. If it has an option to “auto-detect” the drive, try that as well.
The boot sector might also be defective. Boot from a floppy disk and try accessing the hard drive. If the hard drive is accessible, chances are that the boot files are missing or corrupt. Try a utility, such as DrivePro’s Drive Boot Fixer . You might also try running “FDISK /MBR,” which will rebuild the drive’s master boot record. Careful: the FDISK /MBR command might render the files on your drive inaccessible. Finally, you might have a problem with your drive-system hardware. If you cannot access the hard drive, run a diagnostic such as Windsor Technologies’ PC Technician. Test the drive and drive controller. If the controller responds, but the drive does not, try repartitioning and reformatting the hard drive. If the drive still doesn’t respond, replace the hard drive outright. If the controller doesn’t respond, replace the hard-drive controller.

One or more sub-directories appear lost or damaged
Both the root directory of a drive and its FAT contain references to sub-directories. If data in either the root directory or file allocation table is corrupt, one or more sub-directories might be inaccessible by the drive. Try repairing the drive’s directory structure. Use SCANDISK (with DOS 6.2 or later) to check the disk’s directory structure for problems.

Errors occur during drive reads or writes
Magnetic information does not last forever, and sector ID information can gradually degrade to a point where you encounter file errors. Start by checking for file structure problems on the drive. Use SCANDISK, to examine the drive and search for bad sectors. If a failed sector contains part of an .EXE or .COM file, that file is now corrupt and should be restored from a backup. If you cannot isolate file problems, you might need to consider a Low-Level (LL) format. This is an ideal solution because LL formatting rewrites sector ID information, but the sophistication of today’s drives makes LL formatting almost impossible. If the drive manufacturer provides a “drive preparation” utility, you should backup the drive, run the utility, FDISK, FORMAT, and restore the drive.

The hard drive was formatted accidentally
A high-level format does not actually destroy data, but rather it clears the file names and locations kept in the root directory and FAT—this prevents DOS from finding those files. You will need to recover those files. Use a utility which can reconstruct root directory and FAT data contained in a MIRROR file. This is not always a perfect process and you might not be able to recover all files.

A file has been deleted accidentally
Mis-typing or forgetting to add a drive specification can accidentally erase files from places you did not intend to erase. You can often recover those files if you act quickly. to restore the deleted file. This is not always a perfect process and you might not be able to recover every file.

The hard drive’s root directory is damaged
A faulty root directory can cripple the entire disk, rendering all sub-directories inaccessible. You might be able to recover the root directory structure. Use a utility, to reconstruct the damaged FATs and directories. You might also try other recovery utilities, such as DrivePro or ScanDisk. However, if you cannot recover the root directory reliably, you will need to reformat the drive, then restore its contents from a backup.

Hard drive performance appears to be slowing down over time
In virtually all cases, diminishing drive performance can be caused by file fragmentation. To a far lesser extent, you might be faced with a computer virus. Start the PC with a “clean” boot disk and be sure that no TSRs or drivers are being loaded. After a clean boot, run your anti-virus checker and be sure that there are no memory-resident or file-based viruses.
If the system checks clean for computer viruses, you should check for file fragmentation next. Start your defragmentation utility and check to see the percentage of file fragmentation. If it has more than 10% fragmentation, you should consider running the defragmentation utility after preparing Windows. Before defragmenting a drive, reboot the system normally, start Windows, access the Virtual memory controls for your version of Windows, and shut down virtual memory. Then leave Windows and boot the system “clean” again. Restart your defrag-mentation utility and proceed to defragment the disk. This process might take several minutes, depending on the size of your drive. Once defragmentation is complete, reboot the system normally, start Windows, access the Virtual memory controls for your version of Windows, and recreate a permanent swap file to support virtual memory. You should now notice a performance improvement.

The hard drive accesses correctly, but the drive light stays on continuously
A continuous LED indication is not necessarily a problem as long as the drive seems to be operating properly. Check the drive and drive controller for drive “light jumpers”—examine the drive itself for any jumper that might select Latched mode vs. Activity mode. If no such jumpers are on the drive, check the drive controller or motherboard. Set the jumper to Activity mode to see the drive light during access only. Next, consider the possibility of drive-light error messages. Some drive types (especially SCSI drives) use the drive-activity light to signal drive and controller errors. Check the drive and controller documents and see if any error is indicated by the light remaining on.

The hard drive is not accessible and the drive light stays on continuously
This usually indicates a reversed signal cable, which is most common when upgrading or replacing a drive system. In virtually all cases, one end of the signal cable is reversed. Be sure that both ends of the cable are installed properly (remember that the red or blue stripe on one side of the cable represents pin 1). If problems persist, replace the drive controller. It is rare for a fault in the drive controller to cause this type of problem, but if trouble persists, try a known-good drive controller board.

A “No fixed disk present” error message appears on the monitor
This kind of problem can occur during installation, or at any point in the PC’s working life. Check the power connector first, and be sure the 4-pin power connector is inserted properly and completely. If the drive is being powered by a Y-connector, be sure any interim connections are secure. Use a voltmeter and measure the +5-V (pin 4) and +12-V (pin 1) levels. If either voltage (especially the +12-V supply) is unusually low or absent, replace the power supply. Next, check the signal connector. Be sure that the drive’s signal cable is connected securely at both the drive and controller. If the cable is visibly worn or damaged, try a new one.
If problems persist, check the CMOS setup—enter the CMOS setup routine and see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all correct—otherwise, POST will not recognize the drive. You might also try “auto-detecting” the drive. Also check for hardware conflicts. Be sure that no other expansion devices in the system use the same IRQs or I/O addresses used by your drive controller. If so, change the resources used by the conflicting device. If your drive system uses a SCSI interface, be sure that the SCSI cable is terminated properly.
If problems continue, try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective. If problems persist with a known-good hard drive, replace the drive-controller board.

The drive spins up, but the system fails to recognize it
Your computer might flag this as a “Hard-disk error” or “Hard-disk controller failure” during system initialization. Start by checking the signal connector. Be sure that the interface signal cable is inserted properly and completely at the drive and controller. Try a new signal cable. Next, check any drive jumpers, and see that a primary (master) drive is configured as primary, and a secondary (slave) drive is configured as secondary. For SCSI drives, see that each drive has a unique ID setting and check that the SCSI bus is terminated properly. 2
Enter the CMOS setup routine and see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all correct—otherwise, POST will not recognize the drive. Try using the “auto-detect” feature if it is available. If the CMOS is configured properly, you should suspect a problem with the partition. Boot from a floppy disk and run FDISK to check the partitions on your hard drive. Be sure that there is at least one DOS partition. If the drive is to be your boot drive, the primary partition must be active and bootable. Repartition and reformat the drive, if necessary.
If problems persist, try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective. If a known-good hard drive fails to work as expected, replace the drive controller. If problems persist with a known-good floppy drive, replace the drive-controller board.

The IDE drive spins up when power is applied, then rapidly spins down again
The drive is defective, or it is not communicating properly with its host system. Check the power connector first. Be sure that the 4-pin power connector is inserted properly and completely into the drive. Always check the signal connector next, and see that the interface signal cable is inserted properly and completely at the drive and controller. Try a new signal cable.
Inspect the drive jumpers—the primary (master) drive should be configured as primary, and a secondary (slave) drive should be configured as secondary. For SCSI drives, see that each drive has a unique ID setting, and check that the SCSI bus is terminated properly. If problems persist, try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective.

A “Sector not found” error message appears on the monitor
This problem usually occurs after the drive has been in operation for quite some time, and is typically the result of a media failure. Fortunately, a bad sector will only affect one file. Try recovering the file. Use a utility and attempt to recover the damaged file. Notice that you might be unsuccessful, and have to restore the file from a backup later. Check the media itself. Use a disk utility, such as ScanDisk, to evaluate the drive, then locate and map out any bad sectors that are located on the drive.
If problems persist, perform a low-level format (if possible). Lost sectors often occur as drives age and sector ID information degrades. LL formatting restores the sector IDs, but LL formatting is performed at the factory for IDE/EIDE and SCSI drives. If an LL formatting utility is available for your particular drive (available right from the drive manufacturer), and ScanDisk reveals a large number of bad sectors, you might consider backing up the drive completely, running the LL utility, repartitioning, reformatting, then restoring the drive. Finally, if ScanDisk maps out bad sectors, you might need to restore those files from a backup.

A “1780 or 1781 ERROR” appears on the monitor
The classic 1780 error code indicates a “Hard disk 0 failure,” and the 1781 error code marks a “Hard disk 1 failure.” Start the PC with a “clean” boot disk and be sure that no TSRs or drivers are being loaded. If you haven’t done so already, run your anti-virus checker and be sure that there are no memory-resident or file-based viruses. Next, if you can access the hard drive once your system is booted, chances are that the boot files are missing or corrupt. Try a utility, such as DrivePro’s Drive Boot Fixer. Otherwise, you will need to repartition and reformat the disk, then restore disk files from a backup.
Check the hardware next—if you cannot access the hard drive, run a diagnostic such as Windsor Technologies’ PC Technician. Test the drive and drive controller. If the controller responds but the drive does not, try repartitioning and reformatting the hard drive. If the drive still doesn’t respond, replace the hard drive outright. If the controller doesn’t respond, replace the hard-drive controller.

A “1790 or 1791 ERROR” appears on the monitor
The classic 1790 error code indicates a “Hard Disk 0 Error,” although the 1791 error code marks a “Hard Disk 1 Error.” Check the signal connector first. Be sure that the interface signal cable is inserted properly and completely at the drive and controller. Try a new signal cable. There might also be a problem with the drive’s partition. Boot from a floppy disk and run FDISK to check the partitions on your hard drive. Be sure that there is at least one DOS partition. If the drive is to be your boot drive, the primary partition must be active and bootable. Repartition and reformat the drive, if necessary. If problems persist, replace the hard drive. If a known-good drive works as expected, your original drive is probably defective. If problems persist with a known-good floppy drive, replace the drive-controller board.

A “1701 ERROR” appears on the monitor
The 1701 error code indicates a hard-drive POST error—the drive did not pass its POST test. Check the power connector first, and be sure that the 4-pin power connector is inserted properly and completely. If the drive is being powered by a Y connector, be sure that any interim connections are secure. Use a voltmeter and measure the +5-V (pin 4) and +12-V (pin 1) levels. If either voltage (especially the +12-V supply) is unusually low or absent, replace the power supply.
Enter the CMOS setup routine and see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all correct; otherwise, POST will not recognize the drive. Try “auto-detecting” the drive. If problems persist, perform a low-level format (if possible). ST506/412 and ESDI drives might require LL formatting, but LL formatting is performed at the factory for IDE/EIDE and SCSI drives. If an LL-formatting utility is available for your particular drive (available right from the drive manufacturer), you might consider backing up the drive completely, running the LL utility, repartitioning, reformatting, then restoring the drive.

A “Bad or Missing Command Interpreter” error message appears
This is a typical error that appears when a drive is formatted in one DOS version, but loaded with another. Compatibility problems occur when you mix DOS versions. Start by booting the PC with a “clean” boot disk, and be sure no TSRs or drivers are being loaded. If you haven’t done so already, run your anti-virus checker and be sure that there are no memory-resident or file-based viruses. Finally, be sure that the drive is partitioned and formatted with the version of DOS that you intend to use. Also be sure to use FORMAT with the /S switch, or SYS C: to transfer system files to the drive.

An “Error reading drive C:” error message appears
Read errors in a hard drive typically indicate problems with the disk media, but might also indicate viruses or signaling problems. Check the signal connector first. Be sure that the interface signal cable is inserted properly and completely at the drive and controller. Try a new signal cable. Next, start the PC with a “clean” boot disk and be sure that no TSRs or drivers are being loaded. If you haven’t done so already, run your anti-virus checker and be sure that there are no memory-resident or file-based viruses.
Consider the drive’s orientation. If problems occur after remounting the drive in a different orientation, you might need to repartition and reformat the drive, or return it to its original orientation. Also check the media—use a utility, such as ScanDisk, to check for and map out any bad sectors. Once bad sectors are mapped out, you might need to restore some files from your backup. Try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective.

A “Track 0 not found” error message appears
A fault on track 00 can disable the entire drive because track 00 contains the drive’s File Allocation Table (FAT). This can be a serious error, which might require you to replace the drive. Before going too far with this type of problem, check the signal connector and see that the interface signal cable is inserted properly and completely at the drive and controller. Try a new signal cable. Boot from a floppy disk and run FDISK to check the partitions on your hard drive. Be sure that there is at least one DOS partition. If the drive is to be your boot drive, the primary partition must be active and bootable. Repartition and reformat the drive, if necessary. Try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective.

The FDISK procedure hangs up or fails to create or save partition record for the drive(s)
You might also see an error message, such as “Runtime error.” This type of problem often indicates a problem with track 00 on the drive. Before you do anything else, check the signal connector—be sure that the interface signal cables are inserted properly and completely at the drive and controller. Try some new signal cables. Enter the CMOS setup routine and see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all be appropriate. Check with the drive maker and see if there is an alternate “translation geometry” that you can enter instead. If the BIOS supports auto-detection, try “auto-detecting” the drive. Check your version of FDISK. The version of FDISK you are using must be the same as the DOS version on your boot diskette—older versions might not work. Next, run FDISK and see if any partitions are already on the drive. If so, you might need to erase any existing partitions, then create your new partition from scratch. Remember that erasing a partition will destroy any data already on the drive. Use a utility, such as DrivePro (from MicroHouse) or ScanDisk, to check the media for physical defects—especially at track 00. If the boot sector is physically damaged, you should replace the drive. Finally, check for emergency drive utilities. Some drive makers provide low-level preparation utilities, which can rewrite track 00. For example, Western Digital provides the WD_CLEAR.EXE utility. If problems still persist, replace the defective hard drive.

The high-level (DOS) format process takes too long
In almost all cases, long formats are the result of older DOS versions. Check your DOS version MS-DOS version 4.x tries to recover hard errors, which can consume quite a bit of extra time. You will probably see a number of “Attempting to recover allocation units” messages. Your best course is to upgrade the MS-DOS version to 6.22 (or MS-DOS 7.0 with Windows 95). Later versions of DOS abandon hard-error retries.

The IDE drive (<528MB) does not partition or format to full capacity
When relatively small hard drives do not realize their full capacity, the CMOS setup is usually at fault. The drive parameters entered into CMOS must specify the full capacity of the drive—using a geometry setup that is acceptable. If you use parameters that specify a smaller drive, any extra capacity will be ignored. If there are more than 1024 cylinders, you must use an alternate “translation geometry” to realize the full drive potential. The drive maker can provide you with the right translation geometry. Also check your DOS version—older versions of DOS use a partition limit of 32MB. Upgrade your older version of DOS to 6.22 (or MS-DOS 7.0 with Windows 95).

The EIDE drive (>528MB) does not partition or format to full capacity
This type of problem might also be caused by a CMOS setup error, but is almost always caused by poor system configuration. Check the CMOS setup for drive geometry—the drive parameters entered into CMOS must specify the full capacity of the drive. If you use parameters that specify a smaller drive, any extra capacity will be ignored. If there are more than 1024 cylinders, you must use an alternate “translation geometry” to realize the full drive potential. The drive maker can provide you with the right translation geometry. Also check the CMOS setup for LBA. EIDE drives need Logical Block Addressing to access over 528MB. Be sure that there is an entry such as “LBA mode” in CMOS. Otherwise, you might need to upgrade your motherboard BIOS to have full drive capacity. Check the drive controller. If you cannot upgrade an older motherboard BIOS, install an EIDE drive controller with its own controller BIOS—this will supplement the motherboard BIOS. Finally, check the drive-management software. If neither the motherboard or controller BIOS will support LBA mode, you will need to install drive-management software.

“Disk boot failure,” “non system disk,” or “No ROM basic— SYSTEM HALTED” error messages appear
There are several possible reasons for these errors. Start by checking the signal connector. Be sure that the interface signal cables are inserted properly and completely at the drive and controller. Try some new signal cables. Boot the PC with a “clean” boot disk and be sure that no TSRs or drivers are being loaded that interfere with drive operation. If you haven’t done so already, run your anti-virus checker and be sure that there are no memory-resident or file-based viruses. Next, enter the CMOS setup routine and see that all of the parameters entered for the drive are correct. Heads, cylinders, sectors per track, landing zone, and write precompensation must all be entered. Boot from a floppy disk and run FDISK to check the partitions on your hard drive. Be sure that there is at least one DOS partition. If the drive is to be your boot drive, the primary partition must be active and bootable. It is also possible that the hard drive itself is defective. Try a known-good hard drive. If a known-good drive works as expected, your original drive is probably defective. If problems persist with a known-good floppy drive, replace the drive controller.

A hard-drive controller is replaced, but during initialization, the system displays error messages, such as “Hard disk failure” or “Not a recognized drive type”
The PC might also lockup. Some drive controllers might be incompatible in some systems. Check with the controller manufacturer and see if there have been any reports of incompatibilities with your PC. If so, try a different drive-controller board.

The drive will work as a primary drive, but not as a secondary (or vice versa)
In most cases, the drive is simply jumpered incorrectly, but it might also have timing problems. Check the drive jumpers first. Be sure that the drive is jumpered properly as a primary (single drive), primary (dual drive), or secondary drive.
The drive-signal timing might also be off. Some IDE/EIDE drives do not work as primary or secondary drives with certain other drives in the system. Reverse the primary/secondary relationship. If the problem persists, try the drives separately. If the drives work individually, there is probably a timing problem, so try a different drive as the primary or secondary.

The hard drive is infected by a bootblock virus
You might detect the presence of a bootblock virus (a virus that infects the MBR) by running an anti-virus utility, or receiving a warning from the BIOS bootblock-protection feature. In every case, you should attempt to use the anti-virus utility to eradicate the virus. You might also remove a bootblock virus by using “FDISK /MBR” (although that could render the contents of your disk inaccessible). If you’re using drive overlay software, such as Disk Manager, you can usually rewrite the code through the “Maintenance Menu” within the Disk manager program itself.

An “Incorrect DOS version” error appears
You attempted to execute an external DOS command (i.e., FORMAT) using a version of the utility, which is not from the same DOS version as the COMMAND.COM file, which is currently running. Reboot with a corresponding version of COMMAND.COM, or get a version of the utility, which matches the current version of COMMAND.COM.