Although every personal computer uses the same essential sub-assemblies, each sub-assembly is designed a bit differently. This is especially true of the processing components contained on a motherboard - this is understandable given the tremendous speed at which PC components and technology is advancing. Unfortunately, variations in hardware make it difficult to use a single standard operating system. Instead of tailoring an operating system (and applications) to specific computers, a Basic Input/Output System (BIOS) is added on ROM ICs to provide an interface between the raw PC hardware and the standardized operating system - BIOS gives an OS access to a standard set of functions. As a result, every system uses a slightly different BIOS, but each BIOS contains the same set of functions which an OS can interface to. This chapter explains the internal workings of a typical BIOS, illustrates some means of identifying BIOS versions, and shows you the many features a modern BIOS must support.

Of course, BIOS is not limited solely to the motherboard, although most BIOS versions carry enough routines to support video and drive controller operations - in addition to other motherboard features. But, what happens when a new video card is developed that the system BIOS does not know how to work with, or an advanced drive controller board becomes available? A common practice in computer design is to include a BIOS ROM for major sub-systems such as video and drive control. One of the early steps of system initialization is to check for the presence of other valid BIOS ROMs located in upper memory (between 640KB and 1024KB). These are usually referred to as "expansion" or "adapter" BIOS. When another BIOS is located, it is also checksum-tested and used by the PC. In general, a PC may be fitted with up to five or more BIOS ROMs:

• System (motherboard) BIOS
• Video BIOS
• Drive controller BIOS
• Network adapter board BIOS
• SCSI adapter BIOS

The typical BIOS ROM occupies 128KB of space in the system's upper memory area (UMA) from E0000h to FFFFFh (within the PC's first MB of memory). Contrary to popular belief, BIOS is not a single program, but an arsenal of individual programs - most quite small. In general, BIOS contains three sections as shown in Fig. 9-1: the POST, the CMOS Setup routine, and the System Services routines. The particular section of BIOS code that is executed depends on the computer's state, and its activities at any given moment.

Although many novice technicians are aware that POST checks the system, few are aware that POST actually manages the entire system startup. The Power-On Self-Test handles virtually all of the initialization activities for a PC. POST performs a low-level diagnostic and reliability test of the main processing components - including ROM programs and system RAM. It tests the CPU, initializes the motherboard’s chipset, checks the 128 bytes of CMOS for system configuration data, and sets up an index of interrupt vectors for the CPU from 0000h to 02FFh. POST then sets up a BIOS Stack Area from 0300h to 03FFh, loads the BIOS Data Area (BDA) in low memory 0400h to 04FFh, detects any optional equipment (adapter BIOS ROMs) in the system, and proceeds to boot the operating system.

The hardware configuration for any given computer is maintained in a small amount of CMOS RAM, and a CMOS Setup routine is required in order for you to access the system’s configuration. Older i286 and i386 systems provided the CMOS Setup routine as a separate utility included with the system on a floppy disk. In most cases, the "Setup disk" was promptly misplaced or discarded. Starting with late-model i386 and later systems, the CMOS Setup routine has been integrated into the motherboard BIOS itself. The actual CMOS Setup program can vary tremendously between system manufacturers and motherboards, so there is no one standard for just what settings can be controlled, or where those entries are located. Check Chapter 12 for a detailed discussion of typical CMOS entries and their meanings.

The system services (also referred to as BIOS Services) are a set of individual functions that form the layer between hardware and the operating system. Services are called through the use of interrupts. An interrupt essentially causes the CPU to stop whatever it was working on and sends program control to another address in memory which usually starts a subroutine designed specifically to deal with the particular interrupt. When the interrupt handling routine is complete, the CPU's original state is restored, and control is returned to where the PC left off before the interrupt occurred. There are a wide range of interrupts that can attract the attention of a CPU, and interrupts can be produced from three major sources: the CPU itself, a hardware condition, and a software condition. Table 9-1 lists 56 standard interrupts found in a typical BIOS. Keep in mind that the BIOS used in your particular system may offer more or fewer functions depending on its vintage.

Interrupts produced by the CPU itself (known as processor interrupts) are often the result of an unusual, unexpected, or erroneous program result. For example, if a program tries to divide a number by zero, the CPU will generate INT 00h which causes a "Divide by zero" error message. There are five processor interrupts (00h to 04h).

The hardware interrupts are generated when a device is in need of the CPU's attention to perform a certain task. Hardware interrupts are invoked by asserting a logic level on a physical interrupt request (IRQ) line. The CPU suspends its activities and executes the interrupt handling routine. When the interrupt handler is finished, the CPU resumes normal operation. For example, each time a keyboard key is pressed, the keyboard buffer asserts a logic line corresponding to INT 09h (IRQ 1). This invokes a keyboard handling routine. PC/AT-compatible systems typically provide 16 hardware interrupts (IRQ 0 to IRQ 15) which correspond to INT 08h to 0Fh and 70h to 77h respectively.

Software interrupts are generated when a hardware device must be checked or manipulated by the PC. The print screen function is a prime example of a software interrupt. When the Print Screen button is pressed on the keyboard, an INT 05h is generated. The interrupt routine dumps the contents of its video character buffer to the printer port.

PC technology is constantly advancing in CPU’s, chipsets, memory, video, drives, and so on. As the hardware continues to advance, the BIOS must also advance to keep pace with the resources emerging on today’s systems. As a result, it is important for you to recognize the key features that are included in a modern BIOS. You do not need to understand the details of each feature right now, but you should at least recognize a "current" BIOS by reviewing its feature set. Table 9-2 lists the specifications for an Award Software EliteBIOS. The feature set for Mr. BIOS (Microid Research) is shown in Table 9-3. While many of the features listed in Tables 9-2 and 9-3 may seem a bit incidental, the core features of a modern BIOS can be broken down into several major areas:

• CPU support - BIOS should support a rich range of CPUs, preferably from various CPU makers like Intel, AMD, and Cyrix. Look for Pentium, Pentium MMX, Pentium Pro, and Pentium II support (though one BIOS will probably not support ALL these CPU families).
• Chipset support - the BIOS should support the latest chipset families (such as Intel’s 430TX chipset). Chipset support is critical because it is the chipset which allows motherboard designers to implement other features like USB and advanced memory.
• Memory support - the BIOS should be able to auto-size and support the most modern forms of memory (i.e. EDO and SDRAM). Memory error checking (parity and ECC) should also be supported.
• ACPI/APM support - the BIOS should be fully compliant with the Advanced Configuration and Power Interface (ACPI) specification (revision 1.0), and supports APM BIOS specifications through version 1.2. Power management is important for mobile systems, and is widely used in desktop/tower systems to reduce energy waste.
• Drive support - the BIOS must support large EIDE/Ultra-ATA hard drives (over 1024 cylinders) with very fast data transfer modes like PIO Modes 0-4 and Ultra DMA/33.
• PC 97 support - the BIOS should be compliant with the current Microsoft PC 97 BIOS requirements.
• I²O support - the BIOS may support I²O (Intelligent I/O) which allows the dynamic assignment of ports and resources for I/O devices in the PC.
• Boot versatility support - the BIOS should be able to boot from a number of different drives, and include the BIOS Boot Specification for Initial Program Load (IPL) devices. This currently supports booting from up to four IDE drives (including CD-ROM drives), SCSI drives, and network cards.
• Plug-and-Play support - the BIOS must detect and configure PnP devices during POST. The BIOS also communicates with Windows 95 to determine system resources. Support of Microsoft’s AML permits compatibility with PnP capability in future Windows operating systems (such as Windows 98/NT).
• PCI support - the BIOS must support Intel's Peripheral Component Interconnect (PCI) bus specification (version 2.1), including PCI-to-PCI and PCI-to-ISA bridging.
• USB support - the BIOS should support both Universal and Open HCI standards. It should maintain full core compatibility and providing Legacy support for USB hardware and multi-layered USB hubs.

Ordinarily, the BIOS forms a "transparent" layer between the operating system and PC hardware - you’re not even supposed to know the BIOS is there. In actual practice, however, the BIOS can be a root cause of many incompatibilities and problems in the PC, and you must often be able to determine as much information about the BIOS as possible when checking for known BIOS problems or upgrades. When a PC starts and the video system is initialized, a BIOS "banner" is typically the first item to appear. The BIOS banner usually lists the BIOS manufacturer and copyright date, then provides a ID string. There is a great deal of information encoded into most BIOS ID strings, and this part of the chapter will help you decode many of those cryptic ID codes.

Companion CD: If you simply need to extract the copyright date from your BIOS, try the BIOS.ZIP utility on the companion CD.

Older AMI BIOS - AMI BIOS prior to the Hi-Flex line often used a "4/4/6" pattern of BIOS information such as:

• A - BIOS options - particular options included with the BIOS: (D) diagnostics, (S) setup, (E) extended setup.
• BBB - Chipset or motherboard code - a 3-character code used to identify the corresponding chipset or motherboard.
• CCCC - Reference number - a manufacturer’s license reference number.
• DD - BIOS Date Month.
• EE - BIOS Date Day.
• FF - BIOS Date Year.
• GG - Keyboard controller version number - the BIOS and KBC must work together properly. This number indicates the KBC version designed to work with this BIOS.

AMI Hi-Flex line 1 - BIOS ID strings following this format are typically associated with AMI Hi-Flex BIOS. These types of AMI BIOS used several different BIOS ID strings at boot time, and the line below is typically shown first:

• A - Processor type - defines the vintage of CPU the BIOS is intended for: (0) 8088 or 8086, (1) 80286, (2) 80386, (3) 80386SX, (4) 80486, or (5) Pentium.
• B - Size of BIOS - indicates the BIOS ROM size: (0) 64KB ROM, or (1) 128KB ROM.
• CC - Major version number.
• DD - Minor version number.
• EEEEEE - Reference number - a number typically used only by the BIOS maker.
• F - Halt on POST error - details whether the POST will stop if an error is encountered or not: (0) no, (1) yes.
• G - Initialize CMOS in every boot - details whether the CMOS will be initialized at boot time: (0) no, (1) yes.
• H - Block pins 22 and 23 of the keyboard controller: (0) no, (1) yes.
• I - Mouse support in BIOS/keyboard controller: (0) no, (1) yes.
• J - Wait for <F1> if error found: (0) no, (1) yes.
• K - Display floppy error during POST: (0) no, (1) yes.
• L - Display video error during POST: (0) no, (1) yes.
• M - Display keyboard error during POST: (0) no, (1) yes.
• NN - BIOS Date Month.
• OO - BIOS Date Day.
• PP - BIOS Date Year.
• QQQQQQQQ - Chipset Identification/BIOS Name - this designation is intended to identify the chipset supported by this BIOS.
• R - Keyboard controller version number - the BIOS and KBC must work together properly. This number indicates the KBC version designed to work with this BIOS.

AMI Hi-Flex line 2 - BIOS ID strings following this format are typically associated with AMI Hi-Flex BIOS. These types of AMI BIOS used several different BIOS ID strings at boot time, and the line below is typically shown second:

• AA - KBC pin number - defines the pin number used for KBC clock switching
• B - KBC pin function - defines the function performed by the KBC clock switching pin: (H) switches clock to high speed, (L) switches clock to low speed.
• C - Switching enable - sets clock switching through chipset registers: (0) disable, (1) enable.
• DDDD - Port address to switch clock high.
• EE - Data value to switch clock high.
• FF - Mask value to switch clock high.
• GGGG - Port address to switch clock low.
• HH - Data value to switch clock low.
• II - Mask value to switch clock low.
• JJJ - Turbo pin - the pin number for Turbo Switch input.

AMI Hi-Flex line 3 - BIOS ID strings following this format are typically associated with AMI Hi-Flex BIOS. These types of AMI BIOS used several different BIOS ID strings at boot time, and the line below is typically shown third:

• AA - Keyboard controller pin - designates the lines used for cache control.
• B - Cache control - indicates whether high signal on cache control pin enables or disables cache: (H) enable, (L) disable.
• C - Controller signal - indicates if the high signal if used on the keyboard controller pin: (0) false, (1) true.
• DDD - Cache control through chipset registers: (0) cache control off, (1) cache control on.
• EE - Port address - to enable cache through special port.
• FF - Data value - to enable cache through special port.
• GGGG - Mask value - to enable cache through special port.
• HH - Port address - to disable cache through special port.
• II - Data value - to disable cache through special port
• JJ - Mask value - to disable cache through special port.
• K - Pin number - for resetting the 82335 memory controller.
• L - BIOS modified flag - incremented each time BIOS is modified.

The next step in understanding the BIOS is to recognize how it boots - the series of steps that take a PC from power-on to the point where it’s loading an operating system. Each BIOS is written a bit differently, and may have more or less steps than comparable BIOS versions. This part of the chapter looks at the boot sequences for AMI and Phoenix BIOS.

American Megatrends is renowned for their BIOS, PC diagnostics, and motherboards. AMI BIOS performs a fairly comprehensive suite of 24 steps in order to check and initialize the PC. The general AMI BIOS POST procedure is listed below:

1. Disable the NMI - BIOS disables the non-maskable interrupt line to the CPU. A failure here suggests a problem with the CMOS RAM IC or its associated circuitry.
1. Power-on delay - the system resets the soft and hard reset bits. A fault here indicates a problem with the keyboard controller IC or system clock generator IC.
1. Initialize chipsets - BIOS initializes any particular motherboard chipsets (such as the Intel or VIA chipsets) that may be present in the system. A problem here may be caused by the BIOS, the clock generator IC, or the chipset itself.
1. Reset determination - the system reads the reset bits in the keyboard controller to determine whether a hard or soft reset (cold or warm boot) is required. A failure here may be caused by the BIOS or keyboard controller IC.
1. BIOS ROM checksum - the system performs a checksum test of ROM contents and adds a factory preset value that should make the total equal to 00h If this total does not equal 00h, the BIOS ROM is defective.
1. Keyboard test - the system tests the keyboard controller. A fault here is likely the keyboard controller IC.
1. CMOS shutdown check - BIOS tests the shutdown byte in CMOS RAM, calculates the CMOS checksum, and updates the CMOS diagnostic byte. The system then initializes a small CMOS area in conventional memory and updates the date and time. A problem here is likely in the RTC/CMOS IC, or the CMOS backup battery.
1. Controller disable - BIOS now disables the DMA and IRQ controller ICs before proceeding. A fault at this point suggests trouble in the respective controller.
1. Disable video - BIOS disables the video controller IC. If this procedure fails, the trouble is probably in the video adapter board.
1. Detect memory - the system proceeds to check the amount of memory available. BIOS measures system memory in 64KB blocks. A problem here may be in the memory IC(s).
1. PIT test - BIOS tests the programmable interrupt timer (PIT) vital for memory refresh. A problem with the PIT test may reflect a fault in the PIT IC or in the RTC IC.
1. Check memory refresh - BIOS now uses the PIT to try refreshing memory. A failure indicates a problem with the PIT IC.
1. Check low address lines - the system checks the first 16 address lines controlling the first 64KB of RAM. A problem with this test typically means a fault in an address line.
1. Check low 64KB RAM - the system now checks the first 64KB of system RAM. This is vital since this area must hold information that is critical for system initialization. A problem here is usually the result of a bad RAM IC.
1. Initialize support ICs - BIOS proceeds to initialize the programmable interrupt timer (PIT), the programmable interrupt controller (PIC), and the direct memory access (DMA) ICs. A fault here would be located in one of those locations.
1. Load INT vector table - BIOS loads the system's interrupt vector table into the first 2KB of system RAM.
1. Check the keyboard controller (KBC) - BIOS reads the keyboard controller buffer at I/O port 60h. A problem here indicates a fault in the keyboard controller IC.
1. Video tests - the system checks for the type of video adapter in use, then tests and initializes the video memory and adapter. A problem with this test typically indicates a fault with the video memory or adapter respectively. After a successful video test, the video system will be operational.
1. Load the BDA - the system now loads the BIOS data area (BDA) into conventional memory.
1. Test memory - BIOS checks all memory below 1MB. A problem here is typically the fault of one or more RAM ICs, the keyboard controller IC, or a bad data line.
1. Check DMA registers - BIOS performs a register-level check of the DMA controller(s) using binary test patterns. A problem here is often due to a failure of the DMA IC(s).
1. Check the keyboard - the system performs a final check of the keyboard interface. An error at this point is usually the fault of the keyboard.
1. Perform high-level tests - this step involves a whole suite of tests which check such high-level devices as the floppy and hard disks, serial adapters, parallel adapters, mouse adapter, and so on. The number and complexity of these tests vary with the BIOS version. When an error occurs, an a corresponding text message will be displayed. If the system hardware does not match the setup shown in the CMOS setup, a corresponding error code will be displayed.
1. Load the OS - At this point, BIOS triggers INT 19h which is the routine that loads an operating system. An error here generally results in an error message such as "Non-system disk" .

Phoenix Technologies is one of the premier BIOS manufacturers for IBM-compatible PCs known for their extensive POST and versatility with OEMs. A typical Phoenix BIOS performs essentially the same steps as an AMI BIOS, but there are several variations as shown below:

1. Check the CPU - the registers and control lines of the CPU are checked. Any problems will usually be the result of a faulty CPU or clock IC.
1. Test CMOS RAM - The CMOS IC is tested. A fault is usually due to a failure of the RTC/CMOS IC.
1. BIOS ROM checksum - a checksum is performed on the BIOS ROM. If the calculated checksum does not match the factory-set value, an error is generated. A checksum problem is typically the result of a faulty BIOS ROM. Try replacing the BIOS ROM.
1. Test chipset(s) - the system checks any chipsets (such as the Intel or VIA chipsets) for proper operation with the BIOS. A problem here is typically due to a fault in the chipset. Replace the motherboard.
1. Test PIT - the programmable interrupt controller (PIT) is tested to ensure that all interrupt requests are handled properly. A problem here indicates that the PIT IC is defective.
1. Test DMA - the direct memory access (DMA) controller is tested next. A fault at this point is typically caused by the CPU, the DMA IC, or an address line problem.
1. Test base 64KB memory - BIOS checks the lowest 64KB of system RAM. A problem here is due to a fault in memory, or an address line problem.
1. Check serial and parallel ports - the system checks the presence of serial and parallel port hardware, and I/O data areas are assigned for any devices found.
1. Test PIC - The programmable interrupt controller (PIC) is tested to see that proper interrupt levels can be generated. A problem here is typically due to a fault in the PIC IC.
1. Check keyboard controller (KBC) - the keyboard controller IC is tested for proper operation. When a problem occurs, the keyboard controller is likely defective.
1. Verify CMOS data - data within the CMOS is checked for validity. If the extended area returns a failure, CMOS data has probably been setup incorrectly. However, continuous failures typically represent a faulty RTC/CMOS IC.
1. Verify video system - video RAM is tested, then the video controller is located, tested, and initialized. A fault is usually the result of a defective video controller. If the controller is located on an expansion board, try replacing the video board.
1. Test RTC - the real-time clock (RTC) is tested next, and each frequency output is verified. A problem here is usually due to a fault in the RTC, PIT, or system crystal.
1. Test CPU in protected-mode - the CPU is switched to protected mode and returned to POST at the point indicated in CMOS RAM offset 0Fh. When this step fails, the CPU, keyboard controller IC, CMOS IC, or address line(s) may be at fault.
1. Verify PIC 2 - counter #2 is tested on the PIC IC. If this test fails, the PIC IC is likely defective.
1. Check NMI - the NMI is checked to be sure it is active. A problem here often indicates trouble with the CMOS IC, but could also reflect problems in the BIOS ROM, PIC IC, or CPU.
1. Check the keyboard - the keyboard buffer and controller are checked
1. Check the mouse - BIOS initializes the mouse (if present) through the keyboard controller. A fault is usually caused in the mouse adapter circuit.
1. Check system RAM - all remaining system RAM is tested in 64KB blocks. Trouble usually means a defective memory IC.
1. Test disk controller - fixed and floppy disk controllers are checked using standard BIOS calls. Problems here are usually the result of defective controllers or faulty drives. If the controllers are installed on expansion boards, you can try replacing the respective expansion board.
1. Set shadow RAM areas - the system looks at CMOS to find which ROM(s) will be shadowed into RAM. Problems here are often due to a faulty adapter ROM, or problems in RAM.
1. Check extended ROMs - BIOS looks for signatures of 55AAh in memory which indicate the presence of additional ROMs. The system then performs a checksum test on each ROM. A problem with this step generally indicates trouble with the extended ROM or related adapter circuitry.
1. Test cache controller - the external cache controller IC is tested. A problem is usually due to a fault in the cache controller IC itself, or a defect in cache memory.
1. Test CPU cache - the internal cache present in the CPU is tested. A problem here is almost always due to a CPU fault.
1. Check hardware adapters - BIOS proceeds to check the high-level sub-systems such as the video system, floppy disk, hard disk, I/O adapters, serial ports, and parallel ports. Problems usually reflect a fault with the respective adapter, or an invalid CMOS setup.
1. Load the OS - At this point, BIOS triggers INT 19h which is the routine that loads an operating system. An error here generally results in an error message such as "Non-system disk" .

No matter how much time and effort are put into BIOS development, there are still many times when BIOS can come up short. Before you start troubleshooting, you should have an understanding of the places where BIOS is weakest.

As you might expect, no BIOS can possibly address every piece of hardware in the PC marketplace, or keep pace with the rapid advances of those devices that a BIOS does support. As a result, PC designers have devised a way to augment BIOS through the use of device drivers. Traditional CD-ROMs are an excellent example. There are a number of CD-ROM designs in use today - each CD-ROM and their corresponding adapter board use their own (often proprietary) circuitry to operate the drive and interface it to the PC bus. Neither the CD-ROM application, DOS, or BIOS are capable of identifying the drive or interface. To get around this, a low-level device driver is loaded into conventional memory from disk once the PC initializes. The low-level device driver translates a set of standard DOS calls into the instructions necessary to operate the adapter and drive. An extension of DOS (MSCDEX for MS-DOS-based systems) is also loaded into memory after the low-level driver. The DOS extension works seamlessly with MSDOS.SYS to provide applications with a standard set of software interrupt CD-ROM services. Generally speaking, device drivers all serve to supplement BIOS. Video, SCSI, and network adapters all make use of device drivers at some level.

NOTE: The newest BIOS versions DO support bootable CD-ROM drives which adhere to the "El Torito" standard.

The broad acceptance of "flash" memory allows BIOS to be reprogrammed "in-system" through the use of a downloadable program. There is no need to open the PC, or exchange BIOS ICs. This offers BIOS makers a great deal of versatility in the development of new BIOS, but it can also foster an attitude of laziness. Given the astounding speed at which new developments are proliferating, BIOS makers are under a great deal of pressure to create ever-more powerful and diverse BIOS. With traditional BIOS, programmers needed to create solid, well-tested code - because replacing thousands of BIOS ICs in the field is an expensive and cumbersome task. Now that BIOS can be quickly updated with relatively simple software, BIOS programmers can sometimes take the "release-it-and-patch-it-later" attitude. As a rule, BIOS code is still quite solid, but you should be aware that the potential for BIOS problems and oversights are now much higher than years past.

Another problem with BIOS ICs is their inherently slow speed. BIOS is typically recorded onto "flash" ROM ICs (older BIOS used conventional ROM ICs, or other programmable ROM ICs). These devices are necessary because BIOS data must be maintained even when power is removed. Unfortunately, permanent storage ICs such as these have hideously slow access times (150nS to 200nS) when compared to the fast RAM used in today's PCs (50nS to 70nS). When you consider that the services stored in a BIOS ROM are used almost continuously, it is easy to see that each delay is additive - the net result is an overall reduction in PC performance.

In order to overcome this limitation, it would be necessary to accelerate the access time of BIOS ROM. However, this is not too likely given the current state of semiconductor technology, so PC designers do the next best thing - ROM shadowing. The process of "shadowing" basically copies ROM contents from the BIOS IC into available RAM in the upper memory area. Once the copy is complete, the system will work from the copy rather than the original. This allows BIOS routines to take advantage of faster RAM. Not only system BIOS, but ALL BIOS can be shadowed. Video BIOS is particularly popular for shadowing. ROM shadowing can typically be turned on or off through the CMOS setup routine.

NOTE: Not all BIOS can be successfully shadowed. Shadowing problems can cause erratic system behavior and lockups. Whenever you encounter problems configuring a system, you should always try stabilizing the system by shutting down all shadowing options. You can restore shadowing options later and observe if system problems return.

In the race to wring every last clock-tick of performance from a PC, even the most elegantly-written BIOS is simply too slow for high-performance applications. If the application could work with PC hardware directly, system performance (especially disk and video sub-systems) could be substantially improved. Writing directly to hardware is hardly new - pre-IBM PCs relied on direct application control. The use of BIOS was included by IBM to ensure that variations in PC hardware would still remain compatible with operating system and application software. As it turns out, today's PC hardware functions are remarkably standardized (even though the actual components can vary dramatically). With this broad base of relatively standard features, software developers are reviving the direct control approach and ignoring the use of BIOS services in favor of routines written into the application. The trouble with this approach is that direct hardware control may not work on all system configurations, and any changes to the system hardware (i.e. upgrade or replacement parts) may cause the PC to malfunction when the particular application is executed.

As with all software-based products, BIOS code is subject to accidental errors or omissions (software bugs). When BIOS is developed, it is replicated by the thousands and purchased by motherboard manufacturers who incorporate the BIOS into their motherboards. If a bug is present in the BIOS, the system will typically lock-up or crash unexpectedly, or during a certain operation. Since the same BIOS may be used in several motherboards, the bug may not manifest itself in all cases. As one example, some users of AMI BIOS (dated 04/09/90 or earlier) reported problems with the keyboard controller when running Windows or OS/2. As you can imagine, BIOS bugs are particularly frustrating. If an application contains a bug, you can turn the application off. Unfortunately, you can not turn the BIOS off, so the only way to correct a bug in BIOS is to update the BIOS IC, "flash" the BIOS with an updated BIOS file, or replace the entire motherboard.

When investigating a customer complaint for a PC, you may wish to check with the BIOS manufacturer (through technical support, fax-back service, or CompuServe forum) and find if there have been any problems with the BIOS when used in the particular motherboard (i.e. a Phoenix BIOS in an Intel motherboard). If your symptoms match other symptoms that have been reported, a quick BIOS upgrade may save the day for your customer.

The turn of the century is approaching, and with it comes a perplexing problem with file dates. Traditional OS and applications use two-digit year designations (i.e. 1998 would be shown as "98"). The problem is that not all PC hardware or applications are suitable to move into the new century because they cannot properly handle the "roll-over" of year designations such as "00" for 2000, "01" for 2001, and so on. This is known as the "year 2000 problem".

As far as the PC hardware is concerned, Y2K problems are with the Real Time Clock (RTC), and its relation to the internal DOS clock device driver (CLOCK$) - which is actually a counter and not a real clock at all. You can verify if your system has an RTC problem by setting the date and time to 11:57pm (or 23:57) on December 31, 1999, and leaving the machine running to see what happen when it reaches 2000. DOS copes with the problem quite easily, but if you turn the power off and reboot, you might see a system date starting somewhere in 1980. The date 01/01/1980 is usually set if your CMOS contents are lost, and 01/04/1980 if an out-of-range date is encountered.

The reason for this discrepancy is the interaction between the RTC and DOS clocks. The RTC is part of the CMOS RAM chip that maintains the system BIOS settings, and is kept intact by a backup battery. Some of the older RTC chips cannot keep track of the centuries by themselves, so a byte is used in the CMOS to do the job instead. Also, the RTC timing components are "trimmed" at the factory to a certain tolerance (typically plus or minus 20 seconds a month) which will only be adhered to if the desired operating environment is maintained (i.e. temperature and humidity).

By contrast, the DOS device driver (CLOCK$) only interrogates the RTC (via the BIOS) when the machine starts, then proceeds to ignore it as long as the PC is running. The date supplied is converted to the number of days since January 1st 1980, and the number of seconds since midnight of the current day. The number of seconds since midnight is stored in the counter by the BIOS, and when DOS needs to read the clock, the BIOS is called to read the counter and the number of ticks is converted back to seconds. If the counter goes past midnight, it is reset to zero by the BIOS, and the first call after that is told that the day has advanced. As a result, if more than 24 hours has elapsed between calls, there is no way that DOS can tell which day it is.

NOTE: This is why there is often a time difference between your watch and your PC at the end of the day; the system clock has to compete for attention with other devices, and is often reprogrammed by games or other applications which use it for their own timing purposes. Being interrupt driven, the system clock’s accuracy depends on system activity.

As DOS operates between 1980 and 2099, it can figure out that 00 equates to 2000 (although DOS may have problems if the RTC specifically hands it a date of 1900 or any other incompatible date). In practice, the BIOS converts the date as well - some correct the time automatically at boot and supply DOS with 2000 instead of a hardware date of 1900. However, other BIOS cannot produce a date later than 1999 (i.e. Award BIOS 4.5G prior to November 1995 can only accept dates between 1994 and 1999).

Most new BIOS and RTC versions (released after 1996) are designed to deal with the year 2000 properly. But you will need to test older systems for rollover capability - especially as the year 2000 approaches. If your owner's manual doesn't mention this issue, and you can't call the computer manufacturer for clarification, take these steps:

• Set the time on your computer to three minutes before midnight (23:57:00), and the date to December 31, 1999.
• Turn your computer off and wait for five minutes.
• Turn your computer on, booting from a "clean" floppy disk (DOS only with no CONFIG.SYS or AUTOEXEC.BAT files). Check the date.
• If the date is correct, you’re all set.
• If the date is incorrect, try entering the correct date in CMOS Setup (remember to exit saving your changes), then reboot the system. If the new dates hold after adjusting the CMOS Setup, you’re all set.
• If the date is still incorrect after a reboot, or you find that you need to reset the system date every day, you’ll need to upgrade the system BIOS.
• If the date problems persist after upgrading the BIOS, you may need a new RTC (or upgrade the motherboard).

NOTE: Even if the BIOS and RTC are functioning properly for Y2K service, the operating system and applications must also be written to support dates after January 1, 2000.

You’ve got to be familiar with the myriad of error messages that a system can generate. Each time you start the PC, the Power-On Self Test (or POST) initiates a comprehensive series of tests to verify the computer’s hardware. Traditionally, the POST generates two types or error messages; beep codes, and POST codes. Beep codes are generated through the PC speaker before the video system has properly initialized. POST codes are single-byte hexadecimal characters written to I/O port 80h (or other I/O port) as each POST test is started. You can read the POST code using a POST reader card. By matching the beep code or POST code to your particular BIOS, you can determine the exact fault. See Chapter 19 for a comprehensive set of error codes.

The problem with beep codes and POST codes are their cryptic nature - you need a detailed code listing in order to match the code to the fault. However, current generations of BIOS and operating systems are starting to employ more "user-friendly" error messages. By displaying complete error messages (rather than simple codes), a great deal of guesswork is removed from the troubleshooting process. Remember that BIOS error messages are designed to enhance (rather than replace) beep and POST codes. You should also note that unlike beep codes and POST codes, many BIOS error messages are not fatal - that is, the system will continue to run after the error has been generated.

The following list outlines many of the most common BIOS error messages, explains their intended meaning, and suggests some corrective action. The list is a compilation developed from a variety of different systems, and presented in alphabetical order.

Symptom 9-1. 8042 Gate - A20 Error. There is a fault using gate A20 to access memory over 1MB. One or more SIMMs may be loose, or the keyboard controller may have failed. Check that each of the SIMMs are installed securely. Try replacing the keyboard controller, or replace the entire motherboard if necessary.

Symptom 9-2. BIOS ROM checksum error - System halted. The checksum of the BIOS code in the BIOS chip is incorrect - this is a fatal problem indicating the BIOS code may have become corrupt. You will need to replace the motherboard BIOS before the system will initialize.

Symptom 9-3. Cache memory bad, do not enable cache. POST has determined that your cache memory is defective. Do not attempt to enable the cache in your system. You should replace the cache RAM at your earliest opportunity. Until then, you may notice a decline in system performance.

Symptom 9-4. CMOS battery failed. The CMOS battery is no longer functional. You will need to replace the CMOS battery as soon as possible. If you haven’t yet lost CMOS contents, take a <PrintScreen> of each CMOS Setup page immediately to record the setup configuration.

Symptom 9-5. CMOS battery state low. The CMOS battery power is getting low. Make it a point to record your CMOS settings as soon as possible, then replace the CMOS battery promptly.

Symptom 9-6. CMOS checksum error - defaults loaded. CMOS RAM has become corrupt, so the CMOS checksum is incorrect. The system loads the default equipment configuration in an effort to ensure that the system can start. This error may have been caused by a weak battery. Check the CMOS backup battery and replace if necessary.

Symptom 9-7. CMOS display type mismatch. The video type indicated in CMOS RAM is not the one detected by the BIOS. Check your CMOS Setup and make sure the correct video type is selected (usually "VGA"). Remember to save your changes before exiting and rebooting.

Symptom 9-8. CMOS memory size mismatch. The amount of memory recorded in the CMOS Setup configuration does not match the memory detected by the POST. If you have added new memory, start your CMOS Setup and make the appropriate corrections (or simply "save changes and reboot" even though you change nothing). If you’ve made no changed to the system, try rebooting the computer. If the error appears again, some of your memory may have failed. Try a systematic replacement to locate a defective SIMM.

Symptom 9-9. CMOS system options not set. The values stored in CMOS RAM are either corrupt or nonexistent. Check your CMOS backup battery and replace it if necessary. Enter the CMOS Setup routine and reload any missing or corrupted entries. Remember to save your changes before exiting and rebooting.

Symptom 9-10. CPU at nnn. Displays the running speed of the CPU (where nnn is the speed in MHz). This is not an error, but a measurement. If the displayed speed is known to be different than the actual clock speed, you should check the motherboard’s clock settings and multipliers, or suspect an error in BIOS speed detection (you may need to update the BIOS).

Symptom 9-11. Data error. The diskette or hard-disk drive which you are accessing cannot read the data. One or more sectors on the disk(ette) may be corrupted. If you are using MS-DOS, run the CHKDSK or ScanDisk utility to check the file structure of the diskette or hard-disk drive.

Symptom 9-12. Decreasing available memory. An error has been detected in memory, and the "available memory" is being reduced below the point at which the fault was detected. Either a SIMM has failed, or one or more SIMMs may be improperly seated.

Symptom 9-13. Diskette drive 0 (or 1) seek failure. Your floppy drive was unable to seek to the desired track. A cable may be loose, or the CMOS Setup information may not match your actual floppy drive. Check your CMOS Setup, check your signal cable, and replace the floppy drive if necessary.

Symptom 9-14. Diskette read failure. The system was unable to read from a floppy disk. This is usually due to a dirty read/write heads, a loose signal cable, or a defective floppy disk. Try cleaning the read/write heads, try a different diskette, check the floppy signal cable, and replace the floppy drive if necessary.

Symptom 9-15. Diskette subsystem reset failed. The PC was unable to access the floppy drive system. The diskette drive controller may be faulty. Make sure that the drive controller is seated properly in its bus slot, and that all cables are attached securely. Try the drive controller in another slot, and replace the drive controller if necessary.

Symptom 9-16. Display switch is set incorrectly. Some motherboards provide a display switch that can be set to either monochrome or color. This message indicates the switch is set to a different setting than indicated in CMOS Setup. Determine which video setting is correct, and then either turn off the system and change the motherboard jumper, or enter CMOS Setup and change the video selection.

Symptom 9-17. DMA (or DMA #1 or DMA #2) error. A serious fault has occurred in the DMA controller system of your motherboard. In virtually all cases, the motherboard will have to be replaced (unless you can replace the DMA controller).

Symptom 9-18. DMA bus time-out. A device has driven the bus signal for more than 7.8 microseconds. This may be a random fault, but chances are that a device in the PC has failed. Try removing expansion devices first. Otherwise, replace the motherboard.

Symptom 9-19. Drive not ready. No diskette is in the drive. Make sure that the valid diskette is secure in the drive before continuing.

Symptom 9-20. Floppy disk(s) fail. The PC cannot find or initialize the floppy drive controller or the floppy drive itself. Make sure the drive controller is installed correctly (you might try a different expansion slot). If no floppy drives are installed, be sure the "Diskette Drive" entries in CMOS Setup are set to "none" or "not installed".

Symptom 9-21. Hard disk configuration error. The system could not initialize the hard drive in the expected fashion. This is often due to an incorrect configuration in the CMOS Setup. Make sure that the correct hard drive geometry is entered for the drive (or try "auto-detecting" the drive). If the problem persists, try replacing the hard drive.

Symptom 9-22. Hard disk controller failure. There is a problem with the hard drive system - either the hard drive or drive controller has failed. Check the drive controller first and make sure it’s seated properly in its bus slot. Try a different bus slot. Check that all the drive cables are secure. Make sure that the hard drive is spinning up. Try a new drive controller, and try a different hard drive if necessary.

Symptom 9-23. Hard disk(s) diagnosis fail. Your BIOS may run specific disk diagnostic routines. This type of message appears if one or more hard disks return an error when those diagnostics are run. In most cases, the drive itself is installed improperly, or is defective. Check the drive installation, and replace the drive if necessary.

Symptom 9-24. Hard disk failure. The hard drive failed initialization, and usually suggests that the drive has failed. Make sure that the drive signal cable is attached properly, and see that the drive spins up, then replace the hard drive if necessary.

Symptom 9-25. Hard disk drive read failure. The drive cannot read from the hard drive, and usually suggests that the drive has failed. Make sure that the drive signal cable is attached properly, and see that the drive spins up, then replace the hard drive if necessary.

Symptom 9-26. Incompatible Processor: CPU0 (or CPU1) is B0 step or below. You have installed an old version of a CPU which is not supported by the BIOS. In a single-microprocessor system, CPU0 refers to the system board microprocessor; in a dual-microprocessor system, it refers to the secondary microprocessor on the add-in card. The CPU1 message appears only on a dual-microprocessor system and always refers to the system board microprocessor. Replace the microprocessor with a current version of the microprocessor.

Symptom 9-27. Incompatible Processors: Cache sizes different. This message appears for a dual-microprocessor system if the CPUs use differing L2 cache sizes. Replace one of the microprocessors to make the L2 cache sizes match.

Symptom 9-28. Insert Bootable Media. The BIOS cannot find a bootable media. Insert a bootable floppy diskette or bootable CD, or switch to a known-good bootable drive.

Symptom 9-29. INTR #1 (or INTR #2) error. A serious fault has occurred with your interrupt controller on the motherboard. In virtually all cases, the motherboard will have to be replaced entirely.

Symptom 9-30. Invalid Boot Diskette. The BIOS can read the disk in floppy drive A:, but cannot boot the system from it. Use another known-good boot disk, or try booting from a different drive.

Symptom 9-31. Invalid configuration information - please run SETUP program. The system configuration information in your CMOS Setup does not match the hardware configuration detected by the POST. Enter the CMOS Setup program and correct the system configuration information. Remember to save your changes before exiting and rebooting.

Symptom 9-32. I/O Card Parity Error at xxxxx. An expansion card failed. If the address can be determined, it is displayed as xxxxx. If not, the message is I/O Card Parity Error ????. In either case, you’ll need to find and replace the defective expansion card.

Symptom 9-33. Keyboard clock line failure. BIOS has not detected the keyboard clock signal when testing the keyboard. Often, the keyboard connector is loose, or the keyboard is defective. Check the keyboard cable, and try another keyboard if necessary. If the problem persists, the keyboard controller may have failed. Try replacing the keyboard controller IC, or replace the entire motherboard.

Symptom 9-34. Keyboard controller failure. The keyboard controller on the motherboard is not responding as expected. Start by checking the keyboard connection, and try a different keyboard. If the problem persists, the keyboard controller may have failed. Try replacing the keyboard controller IC, or replace the entire motherboard.

Symptom 9-35. Keyboard data line failure. BIOS has not detected the keyboard data signal when testing the keyboard. Often, the keyboard connector is loose, or the keyboard is defective. Check the keyboard cable, and try another keyboard if necessary. If the problem persists, the keyboard controller may have failed. Try replacing the keyboard controller IC, or replace the entire motherboard.

Symptom 9-36. Keyboard error or no keyboard present. The system cannot initialize the keyboard. Make sure the keyboard is attached correctly, and see that no keys are pressed during POST. To purposely configure the system without a keyboard (i.e. if you’re setting up a server), you can configure the CMOS Setup to ignore the keyboard.

Symptom 9-37. Keyboard is locked out - unlock the key. If your system comes fitted with a key lock switch, make sure that the switch is set to the "unlocked" position. If there is no key lock switch (or the switch is set properly), one or more keys may be pressed or shorted on the keyboard. Try a new keyboard.

Symptom 9-38. Keyboard stuck key failure. In almost all cases, this is a keyboard problem. POST has determined that one or more keys on the keyboard are stuck. Make sure that nothing is resting on the keyboard, and see that no paper clips or staples may have fallen into the keyboard. Try a different keyboard.

Symptom 9-39. Memory address line failure at <address>, read <value> expecting <value>. An error has occurred in the address decoding circuitry used in memory. In many cases, one or more SIMMs may be improperly seated. Check that all SIMMs are installed correctly. If the problem continues, try systematic replacement to locate a defective SIMM. If you cannot find a defective SIMM, there is likely a problem elsewhere on the motherboard. Replace the motherboard.

Symptom 9-40. Memory data line failure at <address>, read <value> expecting <value>. An error has been encountered in memory. In virtually all cases, one or more SIMMs may be faulty or improperly seated. Make sure that every SIMM is seated correctly, and try a systematic replacement to locate a defective SIMM.

Symptom 9-41. Memory double word logic failure at <address>, read <value> expecting <value>. An error has been encountered in memory. In virtually all cases, one or more SIMMs may be faulty or improperly seated. Make sure that every SIMM is seated correctly, and try a systematic replacement to locate a defective SIMM.

Symptom 9-42. Memory odd/even logic failure at <address>, read <value> expecting <value>. An error has been encountered in memory. In virtually all cases, one or more SIMMs may be faulty or improperly seated. Make sure that every SIMM is seated correctly, and try a systematic replacement to locate a defective SIMM.

Symptom 9-43. Memory parity failure at <address>, read <value> expecting <value>. An error has been encountered in memory. In virtually all cases, one or more SIMMs may be faulty or improperly seated. Make sure that every SIMM is seated correctly, and try a systematic replacement to locate a defective SIMM.

Symptom 9-44. Memory write/read failure at <address>, read <value> expecting <value>. An error has been encountered in memory. In virtually all cases, one or more SIMMs may be faulty or improperly seated. Make sure that every SIMM is seated correctly, and try a systematic replacement to locate a defective SIMM.

Symptom 9-45. Memory size in CMOS invalid. The amount of memory recorded in the CMOS Setup configuration does not match the memory detected by the POST. If you have added new memory, start your CMOS Setup and make the appropriate corrections. If you’ve made no changed to the system, try rebooting the computer. If the error appears again, some of your memory may have failed. Try a systematic replacement to locate a defective SIMM.

Symptom 9-46. No boot device available. The computer cannot find a viable diskette or hard drive - typically because the drives have not been entered properly into CMOS. Enter the CMOS Setup program and configure the proper drive information. You should also verify that your diskette or hard drive has been prepared as bootable.

Symptom 9-47. No boot sector on hard-disk drive. The PC is refusing to boot from the hard drive. This is usually because the drive is not configured properly. Check the CMOS Setup and verify that the correct drive information has been entered (or select "auto-detect"). Also make sure to partition the drive with an active bootable partition, and format it as a bootable device. If the problem continues, try replacing the hard drive.

Symptom 9-48. No timer tick interrupt. The interrupt timer on the motherboard has failed. This is a fatal error which will probably require you to replace the motherboard.

Symptom 9-49. Non-system disk or disk error. The diskette in drive A: or (your hard drive) does not have a bootable operating system installed on it. If you’re booting from a floppy drive, make the diskette bootable. If you’re booting from a hard drive, make sure that the drive is partitioned and formatted for bootable operation.

Symptom 9-50. Not a boot diskette. There is no operating system on the diskette. Boot the computer with a diskette that contains an operating system.

Symptom 9-51. Off-board parity error. There is a parity error in memory installed in an expansion slot (i.e. a SIMM on the video adapter). The format is: OFF BOARD PARITY ERROR ADDR (HEX) = (XXXX) where XXXX is the hex address where the error occurred. Chances are that the memory installed at the error address has failed.

Symptom 9-52. On-board parity error. There is a parity error in memory installed on the motherboard in one of the SIMM slots. The format is: ON BOARD PARITY ERROR ADDR (HEX) = (XXXX) where XXXX is the hex address where the error occurred. Chances are that the memory installed at the error address has failed.

Symptom 9-53. Override enabled - defaults loaded. If the system cannot boot using the current CMOS configuration for any reason, the BIOS can override the current configuration using a set of defaults designed for the most stable, minimal-performance system operations. The CMOS may be ignored if the CMOS RAM checksum is wrong, or if a critical piece of CMOS information is missing which would otherwise cause a fatal error.

Symptom 9-54. Parity error. A parity error has occurred in system memory at an unknown address. Chances are that memory has failed. Try a systematic "check and replace" approach to isolate the replace the defective memory component.

Symptom 9-55. Plug-and-Play configuration error. The system has encountered a problem in trying to configure one or more expansion cards. Start the CMOS Setup routine and check that any PnP options have been set correctly. If there are any "configuration utilities" included with your particular system, try running those utilities to resolve any configuration issues.

Symptom 9-56. Press <TAB> to show POST screen. Some system OEMs (such as Acer) may replace the normal BIOS POST display with their own proprietary display - usually a graphic logo. When the BIOS displays this message the operator is able to switch between the OEM display and the default POST display. This can be helpful for troubleshooting purposes.

Symptom 9-57. Primary master hard disk fail. POST detects an error in the primary ("master") hard drive on the primary EIDE controller channel. Double-check the drive’s installation, jumpering, and cable connections. Otherwise, replace the drive outright.

Symptom 9-58. Primary slave hard disk fail. POST detects an error in the secondary ("slave") hard drive on the primary EIDE controller channel. Double-check the drive’s installation, jumpering, and cable connections. Otherwise, replace the drive outright.

Symptom 9-59. Resuming from disk. Award BIOS offers a save-to-disk feature for notebook computers. This message may appear when the operator re-starts the system after a save-to-disk shut-down. You will almost never find this type of message on a desktop or tower system.

Symptom 9-60. Secondary master hard disk fail. POST detects an error in the primary ("master") hard drive on the secondary IDE controller channel. Double-check the drive’s installation, jumpering, and cable connections. Otherwise, replace the drive outright.

Symptom 9-61. Secondary slave hard disk fail. POST detects an error in the secondary ("slave") hard drive on the secondary IDE controller channel. Double-check the drive’s installation, jumpering, and cable connections. Otherwise, replace the drive outright.

Symptom 9-62. Shutdown failure. There is a serious fault on the motherboard - usually associated with the CMOS RAM/RTC function. In most cases, you’ll need to replace the motherboard outright.

Symptom 9-63. Terminator/processor card not installed. This is an error that occurs with dual-CPU systems when neither a "terminator" card nor a secondary microprocessor card is installed in the secondary card connector. Make sure either a terminator card or a secondary microprocessor card is installed in the connector. Install the appropriate card and start the system again.

Symptom 9-64. Time of day clock stopped. The real-time clock (RTC) has stopped. The CMOS battery may be dead (or almost dead). Enter the CMOS Setup and correct the date and time. If the trouble continues, try replacing the CMOS backup battery.

Symptom 9-65. Time or date in CMOS is invalid. The time or date displayed in the CMOS Setup does not match the system clock. This can happen often under Windows 95, or other operating systems which can "desynchronize" system clock. Enter the CMOS Setup utility and correct the date and time. If the problem re-occurs, you may be able to determine a specific application that is causing the problem.

Symptom 9-66. Timer chip counter 2 failed. There is a serious fault on the motherboard - probably due to a failure of a programmable interrupt timer (or PIT). In most cases, you’ll need to replace the motherboard outright.

Symptom 9-67. Unexpected interrupt in protected-mode. An interrupt has occurred unexpectedly. Loose or poorly-inserted SIMMs can cause such a problem, so start by checking and reinstalling the SIMMs. A faulty keyboard controller can also result in interrupt problems. Try replacing the keyboard controller if possible, or replace the entire motherboard.

Symptom 9-68. Warning - Thermal Probes failed. This error is usually found in Pentium Pro systems with one or two thermal probes. At system start-up, the BIOS has detected that one or both of the thermal probes in the computer are not operational. You can continue to use the system, but be aware that the temperature probe(s) are disabled - a processor overheat condition will not shut down the system. You will probably have to replace the motherboard to correct this fault.

NOTE: The Pentium Pro has a built-in thermocouple that halts microprocessor operation if the CPU exceeds its rated temperature.

Symptom 9-69. Warning - Temperature is too high. During system start-up, the BIOS has detected that one or both microprocessors are overheated. This can happen if you try to restart the system too soon after a thermal shutdown. After displaying this message, the BIOS halts the processes and turns off the system. Let the system cool down before attempting to restart it.

Symptom 9-70. Bad PnP serial ID checksum. The serial ID checksum of a Plug and Play card is invalid. Try replacing the offending expansion card.

Symptom 9-71. Floppy disk controller resource conflict. The floppy disk controller has requested a resource that is already in use by another device. Try freeing the resources that are requested by the PnP system.

Symptom 9-72. NVRAM checksum error, NVRAM cleared. The extended system configuration data (ESCD) was reinitialized because of an NVRAM checksum error. Try rerunning the ISA Configuration Utility (ICU). If the problem persists, replace the NVRAM IC, or replace the motherboard.

Symptom 9-73. NVRAM cleared by jumper. The "Clear CMOS" jumper on the motherboard has been moved to the Clear position and the system has been initialized. CMOS RAM and ESCD have been cleared, and now must be reconfigured.

Symptom 9-74. NVRAM data invalid, NVRAM cleared. Invalid data has been found in the ESCD (which may mean that you have changed devices in the system). When this message is displayed, the BIOS has already rewritten the ESCD with current configuration data. Try rebooting the system.

Symptom 9-75. Parallel port resource conflict. The parallel port requested a resource that is already in use by another device. Try freeing the resources requested by the PnP system.

Symptom 9-76. PCI error log is full. More than 15 PCI conflict errors have been detected, and no additional PCI errors can be logged. Deal with the PCI errors already contained in the log in order to reduce the total number of errors.

Symptom 9-77. PCI I/O port conflict. Two devices requested the same I/O address, resulting in a

conflict. Try freeing the resources needed to allow both devices to be configured properly.

Symptom 9-78. PCI IRQ conflict. Two devices requested the same IRQ, resulting in a conflict. Try freeing the IRQs needed to allow both devices to be configured properly.

Symptom 9-79. PCI memory conflict. Two devices requested the same memory resource, resulting in a conflict. Try freeing the memory needed to allow both devices to be configured properly.

Symptom 9-80. Primary boot device not found. The designated primary boot device (hard disk drive, floppy disk drive, CD-ROM drive) could not be found. Check the installation and configuration of each boot device.

Symptom 9-81. Primary IDE controller resource conflict. The primary IDE controller has requested a resource that is already in use. Try freeing the resources that are needed to allow the IDE controller to operate.

Symptom 9-82. Primary input device not found. The designated primary input device such as the keyboard or mouse (or other device if input is redirected) could not be found. Check the installation and configuration of all your input devices. Make sure that the input devices are also enabled in CMOS Setup.

Symptom 9-83. Secondary IDE controller resource conflict. The secondary IDE controller has requested a resource that is already in use. Try freeing the resources that are needed to allow the IDE controller to operate.

Symptom 9-84. "Static device resource conflict" or "System board device resource conflict". A non-Plug and Play ISA card has requested a resource that is already in use. Try reconfiguring the ISA card to use other resources, or try freeing the resources needed by the ISA card.

This finishes up Chapter 9. Be sure to review the glossary and chapter questions on the accompanying CD. If you have access to the Internet, point your web browser to some of the contacts below:

American Megatrends: http://www.megatrends.com

Award BIOS: http://www.award.com

Hardware IC newsgroup: comp.sys.ibm.pc.hardware.chips

IBM SurePath BIOS page: http://www.surepath.ibm.com/

MicroFirmware: http://www.firmware.com/catalog2.htm

Microid Research (Mr. BIOS): http://www.mrbios.com/

SystemSoft: http://www.systemsoft.com

Unicore: http://www.unicore.com/

Year 2000 reference: http://www.year2000.com (or http://www.sbhs.com/y2k)