Community hub
0 subscribers8 pages, 0 posts
Recent from talks
All channels
Be the first to start a discussion here.
Be the first to start a discussion here.
Be the first to start a discussion here.
Be the first to start a discussion here.
Contribute something
Welcome to the community hub built to collect knowledge and have discussions related to Nonvolatile BIOS memory.
Nothing was collected or created yet.
Nonvolatile BIOS memory
View on Wikipediafrom Wikipedia
 | This article may need to be rewritten to comply with Wikipedia's quality standards. (May 2021) |
Nonvolatile BIOS memory refers to a small memory on PC motherboards that is used to store BIOS settings. It is traditionally called CMOS RAM because it uses a volatile, low-power complementary metal–oxide–semiconductor (CMOS) SRAM (such as the Motorola MC146818[1] or similar) powered by a small battery when system and standby power is off.[2] It is referred to as non-volatile memory or NVRAM because, after the system loses power, it does retain state by virtue of the CMOS battery. When the battery fails, BIOS settings are reset to their defaults. The battery can also be used to power a real time clock (RTC) and the RTC, NVRAM and battery may be integrated into a single component. The name CMOS memory comes from the technology used to make the memory, which is easier to say than NVRAM.[3]
The CMOS RAM and the real-time clock have been integrated as a part of the southbridge chipset and they may not be standalone chips on modern motherboards.[4][5] In turn, the southbridge has been integrated into a single Platform Controller Hub. Alternatively BIOS settings may be stored in the computer's Super I/O chip.[6]
The chipset built-in NVRAM capacity is typically 256 bytes.[4] For this reason, later BIOS implementations may use a small portion of BIOS flash ROM as NVRAM, to store BIOS setup and hardware configuration data.[7]
Today's UEFI motherboards use NVRAM to store configuration data (NVRAM is a portion of the UEFI flash ROM), but by many OEMs' design, the UEFI settings are still lost if the CMOS battery fails.[8][9]
CMOS battery
[edit]
CR-2032 button cell, the most common CMOS battery
.jpg)
VARTA-manufactured Ni–Cd barrel battery that has leaked onto the mainboard of a synthesizer and damaged legs and traces
The memory battery (aka motherboard, CMOS, real-time clock (RTC), clock battery)[2][10] is generally a CR2032 lithium coin cell. This cell battery has an estimated life of three years when power supply unit (PSU) is unplugged or when the PSU power switch is turned off.[11] This battery type, unlike the lithium-ion battery, is not rechargeable and trying to do so may result in an explosion. Motherboards have circuitry preventing batteries from being charged and discharged when a motherboard is powered on. Other common battery cell types can last significantly longer or shorter periods, such as the smaller CR2016 which will generally last about 40% less time than CR2032. Higher temperatures and longer power-off time will shorten battery cell life. When replacing the battery cell, the system time and CMOS BIOS settings may revert to default values. Unwanted BIOS reset may be avoided by replacing the battery cell with the PSU power switch turned on and plugged into an electric wall socket. On ATX motherboards, the PSU will supply 5V standby power to the motherboard to keep CMOS memory energized while the system is off.
Some computer designs have used non-button cell batteries, such as the cylindrical "1/2 AA" used in the Power Mac G4 as well as some older IBM PC compatibles, or a 3-cell nickel–cadmium (Ni–Cd) CMOS battery that looks like a "barrel" (common in Amiga and older IBM PC compatibles), which serves the same purpose. These motherboards often have a four pin straight header, with pin 2 missing, for connecting to an external 3.6v battery, such as the Tadiran TL-5242/W, when their soldered-on batteries run out. Ni–Cd batteries have a tendency to leak devastatingly after a period of disuse, damaging components and traces on the circuit board near the battery.[12]
See also
[edit]References
[edit]- ^ Mueller, Scott (2004). Upgrading and Repairing PCS. Que. ISBN 978-0-7897-2974-3.
- ^ a b Fisher, Tim. "What is CMOS? (CMOS & CMOS Battery Definition)". About.com. Archived from the original on 9 August 2016. Retrieved 2 September 2015.
- ^ Mueller, Scott (2003). Upgrading and Repairing PCS. Que. ISBN 978-0-7897-2745-9.
- ^ a b "Intel 100 Series Chipset Family PCH Datasheet". Intel. August 2015. Retrieved 16 April 2016.
- ^ "82430FX PCISET Data Sheet" (PDF). Intel. Retrieved 23 January 2023.
- ^ Mueller, Scott (2004). Upgrading and Repairing PCS. Que. ISBN 978-0-7897-2974-3.
- ^ "AMIBIOS8 Flash Update & BIOS Recovery Methods" (PDF). Archived from the original (PDF) on 2009-11-22.
- ^ "UEFI NVRAM - OSDev Wiki". wiki.osdev.org. Retrieved 11 September 2020.
- ^ Yao, Jiewen; Zimmer, Vincent (2020). "Firmware Resiliency: Detection". Building Secure Firmware. pp. 115–162. doi:10.1007/978-1-4842-6106-4_4. ISBN 978-1-4842-6105-7. S2CID 242541772.
- ^ Ask a question » Replacing the battery of your motherboard (CMOS battery)
- ^ "CMOS Battery on the Intel NUC". Intel. 24 August 2020.
- ^ Williams, Al (July 8, 2018). "Amiga 2000 Emergency Repair". Hackaday. Archived from the original on July 8, 2018.
External links
[edit]Nonvolatile BIOS memory
View on Grokipediafrom Grokipedia
Overview
Definition and Purpose
Nonvolatile BIOS memory, often referred to as CMOS RAM, is a type of battery-backed static random access memory (SRAM) located on computer motherboards that retains BIOS or UEFI firmware settings without requiring continuous power supply.[3] This small memory component stores critical system configuration data, including the real-time clock (date and time), boot device priority, and hardware parameters such as disk drive types and enabled peripherals.[7] In modern personal computers, it typically provides 256 to 512 bytes of storage capacity, sufficient for these parameters while minimizing hardware footprint. The primary purpose of nonvolatile BIOS memory is to enable the firmware to automatically recall and apply user-customized settings during each system boot, eliminating the need for manual reconfiguration after power cycles or shutdowns.[3] This contrasts sharply with standard volatile RAM, which discards all data when power is removed, ensuring reliable and consistent system initialization.[7] By preserving these configurations, it supports seamless operation of the boot process and hardware detection, contributing to overall system stability and user convenience. Key characteristics of nonvolatile BIOS memory include its extremely low power consumption—often in the microwatt range—sustained by a dedicated onboard battery, and its physical integration directly on the motherboard adjacent to the BIOS firmware chip for efficient data access.[3] Historically, this technology was introduced with the IBM PC/AT in 1984 via the MC146818 real-time clock chip, which included 64 bytes of CMOS RAM, of which approximately 50 bytes were allocated for BIOS setup information; subsequent systems expanded this to 128 bytes to accommodate additional features.[8]Role in System Boot Process
Upon power-up, the BIOS or UEFI firmware initiates the boot sequence by reading configuration data from nonvolatile memory, such as CMOS RAM, to initialize essential hardware components including the CPU, memory, and peripherals, set the system clock and date, and determine the boot device order for locating and loading the operating system loader.[9][10] This integration ensures that the system adheres to previously configured parameters without requiring reconfiguration each time, allowing for a streamlined transition from power-on to OS handover.[11] During the Power-On Self-Test (POST) phase, the firmware performs an integrity check on the nonvolatile memory contents via a checksum validation process to confirm data reliability.[12][13] If the checksum fails due to corruption—often from power loss or hardware issues—the system may automatically load safe default settings, enter a recovery mode, or prompt the user to access the setup utility by pressing a designated key such as Delete during startup.[14] The verified settings from nonvolatile memory are then applied to configure system behavior, such as enabling or disabling onboard devices like USB ports to manage legacy support or security, or adjusting CPU multipliers to optimize performance based on stored overclocking values.[15][16] For error handling in invalid memory scenarios, the firmware may emit diagnostic beep codes to alert users without display output; for example, implementations from manufacturers like AMI or Award use patterns such as multiple short beeps to signal configuration errors, directing troubleshooting toward memory reset or replacement.[13]Hardware Components
CMOS RAM Chip
The CMOS RAM chip employed in nonvolatile BIOS memory is a static random-access memory (SRAM) fabricated using complementary metal-oxide-semiconductor (CMOS) technology, which utilizes pairs of p-type and n-type transistors in a bistable configuration to maintain data states with minimal power consumption.[17] This architecture enables low static power dissipation, making it suitable for battery-backed operation in personal computers. In x86 systems, the chip is typically addressed through I/O ports 70h (for register selection) and 71h (for data read/write), allowing byte-level access to its contents.[18] The chip's capacity is generally 64 bytes in early implementations, with the first 14 bytes dedicated to real-time clock (RTC) functions such as timekeeping, alarm settings, and control registers, while the remaining 50 bytes store BIOS configuration data like boot order and hardware parameters.[17] Later variants expanded this to 128 bytes total, maintaining the RTC allocation but providing 114 bytes of additional space for extended settings.[19] Organized as a byte-addressable array, the memory supports direct interfacing; in modern motherboards, it is often connected via the Low Pin Count (LPC) bus or System Management Bus (SMBus) through the southbridge chipset for efficient low-bandwidth communication.[20] In legacy systems, the CMOS RAM chip is integrated as a dedicated RTC-plus-RAM module, typically in a 24-pin dual in-line package (DIP) that can be socketed for replacement or soldered directly onto the motherboard near the southbridge for thermal and electrical proximity.[17] However, since the 2010s, RTC and CMOS RAM functionality have been integrated directly into the motherboard chipset, such as Intel's Platform Controller Hub (PCH), eliminating discrete chips while retaining battery-backed SRAM.[21] Representative examples include the Motorola MC146818, an early CMOS RTC with integrated 50-byte user RAM, and the Dallas Semiconductor DS12887, a compatible upgrade featuring 114 bytes of nonvolatile RAM and an internal lithium backup cell.[19] Under battery backup, the chip exhibits high power efficiency, with typical leakage currents below 1 μA—such as 0.5 μA at 3.0 V and 25°C for the DS12887—allowing data retention for several years on a standard coin-cell battery.[19] This low standby current is achieved through CMOS transistor gating that isolates the RAM array when unpowered by the main supply.[17]Backup Power Sources
The primary backup power source for nonvolatile BIOS memory is a coin-cell lithium battery, most commonly the CR2032 type, which operates at 3 volts and provides a capacity of approximately 220 milliampere-hours (mAh). This battery connects in parallel to the CMOS RAM to deliver continuous standby power during periods when the main power supply unit (PSU) is disconnected or powered off, thereby preserving critical BIOS data such as system configuration, date, and time settings without relying on the primary system power.[22][23] The battery maintains nonvolatility by supplying the essential Vcc voltage (typically 3V) to the RAM cells, counteracting inherent charge leakage in CMOS circuitry that would otherwise lead to data corruption within hours or days. Lithium-based coin cells feature a minimal self-discharge rate of 1-3% per year at room temperature, enabling reliable long-term energy retention with negligible loss over extended periods. In practice, the battery's operational lifespan in a motherboard setup ranges from 5 to 10 years, depending on the low quiescent current draw of the CMOS components (often in the microampere range) and external factors like ambient temperature and humidity, which can accelerate degradation if elevated.[24][25] For circuit integration, the battery is wired directly to the CMOS power rail via a Schottky or similar diode to block reverse current, ensuring the battery does not discharge when the PSU's 5V standby rail (active even in powered-off states) takes over supply duties through an OR-ing configuration. This diode-protected setup prevents unintended charging of the non-rechargeable lithium cell and maintains efficient power handover. In server-grade motherboards, a baseboard management controller (BMC) typically monitors the battery voltage in real time, triggering alerts or logs for low thresholds (e.g., below 2.5V) to facilitate proactive maintenance and avoid settings loss.[26][27] While batteries dominate consumer applications, alternative sources like supercapacitors appear in select embedded systems, where their high capacitance (often 0.1-1 farad) supports brief retention intervals with faster recharge from transient power events, though limited energy density makes them unsuitable for prolonged standby in standard PCs. ACPI-defined standby power rails, such as the persistent 5VSB line, occasionally supplement battery operation by providing auxiliary power during low-power states, but they remain uncommon as primary backups in fully off configurations due to dependency on PSU presence.[28]Operation and Configuration
Data Storage Mechanism
The data storage mechanism in nonvolatile BIOS memory primarily involves writing user-configured settings to CMOS RAM during the BIOS setup process. Users access the setup utility by pressing a key like F2 at boot, navigate to modify parameters such as boot order or hardware detection, and save changes, prompting the BIOS firmware to update the memory. This write operation uses I/O port 70h to specify the target byte address (ranging from 00h to 3Fh in the original 64-byte CMOS space (as in the IBM PC AT), or up to 7Fh in later 128-byte implementations) followed by outputting the new data value to port 71h. In certain BIOS variants, such as those from AMI, higher-level write access and checksum updates are handled through interrupt 15h with vendor-specific functions like AX=DA20h subfunctions for setting bytes and recalculating integrity checks.[29][30] Reading data from nonvolatile BIOS memory follows a parallel port-based procedure, essential for loading settings during the power-on self-test (POST) and ongoing system operation. The BIOS outputs the desired address to port 70h, then inputs the corresponding byte from port 71h into the AL register for processing. This direct hardware access ensures low-latency retrieval of configuration flags and real-time clock (RTC) values. RTC synchronization relies on BCD-encoded data across dedicated offsets—such as 00h for seconds (00-59 in BCD), 02h for minutes (00-59 in BCD), and 04h for hours (00-23 in 24-hour format or 01-12 in 12-hour with AM/PM flag)—allowing the 32.768 kHz crystal oscillator to maintain accurate timekeeping even without main power.[29][8] Checksum validation protects against data corruption in CMOS RAM by computing a stored integrity value against the current contents during each boot or access. Typically implemented as a 2-byte sum-check at offsets 2Eh (low byte) and 2Fh (high byte), it covers configuration bytes from 10h to 2Dh; the checksum equals the two's complement of their byte-wise sum, ensuring the total sums to zero modulo 256 per byte. Some implementations use a full 16-bit sum or CRC-16 for broader coverage, including up to 7Dh in extended CMOS. Upon mismatch—often due to battery failure or electrical noise—the BIOS detects the error via interrupt 1Ah or POST routines, clears affected memory regions, and reverts to factory default settings to maintain system stability.[29][8] Data within nonvolatile BIOS memory is organized as compact binary structures to optimize the limited 64-128 byte space, prioritizing essential flags over verbose formats. Settings appear as bit-level flags in allocated bytes; for instance, byte 10h specifies the types of floppy drives A and B via 4-bit nibbles, with values like 24h indicating a 1.2 MB 5.25-inch A: drive paired with a 1.44 MB 3.5-inch B: drive. RTC bytes use BCD for decimal readability, while higher offsets (e.g., 40h-7Fh) include reserved regions for vendor extensions, such as OEM-specific hardware options or diagnostic flags, without disrupting core compatibility.[29][8]Reset and Default Settings
Nonvolatile BIOS memory can be reset to factory defaults through hardware or software methods, ensuring the system reverts to safe, pre-configured parameters when custom settings cause instability.[31] Hardware resets typically involve using a jumper on the motherboard, such as the CLR_CMOS pins, to short the battery line to ground and discharge the CMOS RAM.[4] This process requires powering off the system, moving the jumper cap from its default position (usually pins 1-2) to the clear position (pins 2-3) for 5-10 seconds, then returning it to the default and powering on.[32] Some modern motherboards feature a dedicated CLR_CMOS button near the battery that achieves the same discharge when pressed for 5-10 seconds with power disconnected.[33] These methods effectively clear all stored settings by removing the backup power, forcing the memory to lose its data.[32] Software resets provide a non-invasive alternative, accessible via the BIOS setup menu or operating system tools. In the BIOS interface, selecting options like "Load Optimized Defaults" restores the manufacturer's recommended configuration while preserving basic functionality.[31] Third-party utilities such as CMOS De-Animator allow users to invalidate the CMOS checksum from within the OS, prompting a reset to defaults upon reboot without hardware intervention.[34] On Linux systems, commands targeting the NVRAM device, such as loading the nvram module and writing zeros to /dev/nvram, can clear the memory contents and trigger a reset.[35] Default settings in nonvolatile BIOS memory consist of pre-programmed values designed for reliable operation, including automatic boot device detection to prioritize available drives.[36] These configurations vary by original equipment manufacturer (OEM); for instance, Award BIOS defaults often emphasize stable overclocking profiles, while Phoenix BIOS prioritizes compatibility with older hardware peripherals.[37] Security features in BIOS setups, such as administrator passwords, restrict access to the setup menu and prevent unauthorized software-based resets or modifications.[38] However, performing a hardware reset clears all stored passwords, including admin ones, effectively removing protection until reconfigured. For example, in older Phoenix Technologies BIOS systems used in devices like Samsung and Acer laptops from the 2000s and 2010s, removing the CMOS battery clears passwords such as the Supervisor Password or Setup Password that block access to the BIOS menu or boot process.[39][40] This underscores the importance of documenting passwords, as CMOS clearing erases them without recovery options beyond the reset itself.[41]Historical Development
Origins in Early PCs
The nonvolatile BIOS memory was developed in conjunction with the IBM Personal Computer released in 1981, primarily to address the challenges of storing adapter configurations in an era of rapidly evolving and often incompatible hardware options for personal computers. This initial design incorporated a 64-byte CMOS RAM structure, drawing inspiration from nonvolatile storage mechanisms in earlier minicomputer systems that required persistent parameter retention for multi-user and operational stability. However, the original IBM PC relied on physical DIP switches and jumpers for these configurations, as integrated nonvolatile memory was not yet implemented on the system board.[42] A key milestone occurred with the IBM PC XT in 1983, which for the first time integrated the Motorola MC146818 real-time clock (RTC) chip featuring embedded RAM, backed by a battery to maintain data persistence in consumer personal computers. This innovation addressed early limitations of volatile memory by ensuring that critical system data, such as time and basic configurations, survived power cycles without manual reset. The battery backup was particularly valuable in multi-user environments, where clock drift from repeated power-ons could disrupt scheduling and logging accuracy.[43] Early implementations faced significant challenges due to limited storage space in the CMOS RAM, leading many systems to continue using DIP switches as a primary or supplementary configuration method despite the availability of nonvolatile options. IBM's design for nonvolatile BIOS memory profoundly influenced the Industry Standard Architecture (ISA) bus specifications, standardizing access ports for CMOS data that became ubiquitous in compatible systems. The first widespread adoption came with the 1984 IBM PC AT, which utilized the 64-byte CMOS to enable more comprehensive storage of BIOS parameters like disk types and memory sizing directly on the motherboard. Capacity remained at 64 bytes through the 1980s, with initial expansions to 128 bytes appearing in systems like the IBM PS/2 starting in 1987.[44]Evolution Through the 1990s and 2000s
During the 1990s, the evolution of PC bus architectures, particularly the adoption of Extended Industry Standard Architecture (EISA) and Peripheral Component Interconnect (PCI) buses, drove significant expansions in nonvolatile BIOS memory capacity to handle the growing array of hardware configuration options. Early CMOS RAM implementations, limited to 64 bytes in the 1980s, increased to 128 bytes by the early 1990s and reached 256 bytes as standard by the mid-decade to support EISA's extended configuration space and PCI's plug-and-play capabilities, which required storing additional parameters for device detection and resource allocation.[45][46] Major BIOS vendors like American Megatrends Inc. (AMI) and Award Software responded to these demands by integrating Plug and Play (PnP) support, first standardized with Windows 95 in 1995, which further complicated setup storage needs in nonvolatile memory. AMI's Hi-Flex AMIBIOS, updated in 1993 for EISA and PCI compatibility, and its later versions like AMIBIOS 6, natively incorporated PnP enumeration, necessitating expanded memory for dynamic device profiles and IRQ assignments. Award BIOS similarly evolved, adding PnP features that increased setting complexity by managing legacy ISA conflicts alongside new bus standards.[47][45] In the 2000s, nonvolatile BIOS memory continued to scale, with 512 bytes becoming the de facto standard by 2005 to accommodate advanced features like larger hard drive support and integrated peripherals on chipsets such as Intel's 915 series. The integration of the Advanced Configuration and Power Interface (ACPI), introduced in 1996 but widely adopted in the early 2000s, added power management flags to CMOS storage, allowing operating systems to control sleep states (S1-S5) and device power rails directly from BIOS-persisted configurations. Additionally, a shift toward serial interfaces like I²C simplified access to RTC chips and extended NVRAM, as seen in devices like Texas Instruments' bq32002, enabling more efficient communication with super I/O controllers on motherboards.[48][49][50] Reliability enhancements during this period included software-based write-protect bits in BIOS firmware to safeguard critical settings against corruption and multi-checksum algorithms that computed separate integrity values for RTC data, system parameters, and extended configuration blocks, alerting users via POST errors if discrepancies arose. The Year 2000 (Y2K) issue prompted widespread BIOS updates to correctly manage the RTC's century byte (register 0x32 in the MC146818-compatible chips), preventing date rollovers from resetting to 1900 and ensuring proper leap year calculations post-1999. Vendor-specific advancements, such as Phoenix Technologies' BIOS 4.0 release in 2001 (with updates extending into 2002), introduced secure modes including two-level password controls such as Supervisor Password or Setup Password that blocked access to the BIOS menu or boot process, boot sector protection, and virus check reminders stored in protected NVRAM regions; these features were commonly implemented in laptops from manufacturers like Samsung and Acer during the 2000s and 2010s, where such passwords could be reset by removing the CMOS battery to clear the nonvolatile memory. Intel's platform specifications around 2006, aligned with early EFI developments, recommended a minimum 2KB NVRAM capacity to support future-proofing for ACPI tables and boot variables.[51][52][53][6][54][55]Modern Implementations and Alternatives
Integration with UEFI
The Unified Extensible Firmware Interface (UEFI), first specified by Intel in 2005 and subsequently evolved under the UEFI Forum, preserves core principles of nonvolatile BIOS memory by incorporating NVRAM to store EFI variables that persist across power cycles and resets.[56] These variables serve as key-value pairs for essential system configurations, such as boot settings and hardware parameters, with implementations providing varying NVRAM capacities depending on the hardware, often in the range of 64-256 KB. Access to this storage is facilitated through the EFI variables API, which includes runtime services like GetVariable(), SetVariable(), GetNextVariableName(), and QueryVariableInfo() to manage data securely even after the OS loader takes control.[57] In UEFI's hybrid model, legacy BIOS compatibility is achieved via the Compatibility Support Module (CSM), which emulates traditional BIOS operations—including reads from CMOS memory—for older operating systems like Windows XP that lack native UEFI support.[58] Meanwhile, native UEFI environments leverage SPI-connected NVRAM for efficient, scalable storage, decoupling variable management from the limitations of battery-backed CMOS RAM.[59] A representative application is the storage of boot manager entries, such as Boot#### variables, which allow the firmware to dynamically select and load bootable files without relying on fixed legacy structures.[57] UEFI further enhances nonvolatile storage through fault-tolerant mechanisms, where variables are maintained in dedicated NVRAM partitions designed for atomic updates to prevent corruption during power failures or incomplete writes.[57] This approach enables dynamic allocation of space, exceeding the rigid byte constraints of traditional CMOS (typically 256 bytes), by using protocols that reclaim and redistribute storage as needed for growing configuration demands.[60]Shift to Flash-Based Storage
In the 2010s, motherboard manufacturers increasingly integrated SPI NOR flash memory to store both BIOS/UEFI firmware and nonvolatile settings, transitioning away from battery-backed CMOS RAM to eliminate the need for maintenance-prone power sources. Chips such as the Winbond W25Q series, with capacities starting at 8 MB (e.g., W25Q64FV at 64 Mbit), provide ample space for firmware code alongside dedicated regions for configuration data like boot order and hardware parameters.[61] This approach leverages the inherent nonvolatility of NOR flash, which retains data without external power, offering typical retention periods of up to 20 years under standard operating conditions.[62] Key advantages of this shift include reduced hardware complexity and improved reliability, as flash memory avoids battery degradation and leakage issues while supporting infrequent write operations to settings without significant endurance concerns—NOR flash typically endures 100,000 program/erase cycles per sector, far exceeding BIOS update frequency.[63] In laptops and ARM-based devices, such as those incorporating Apple's T2 security chip, integrated flash handles firmware and secure settings storage, enabling features like encrypted boot without discrete batteries.[64] Similarly, embedded ARM systems benefit from always-on flash for persistent configuration in power-constrained environments. Implementation involves partitioning the SPI NOR flash into sections for firmware, variables, and runtime data, often with built-in wear-leveling algorithms to distribute writes evenly across sectors and prevent localized degradation.[65] UEFI variables, which encompass BIOS settings, are managed through EFI runtime services like GetVariable and SetVariable, allowing OS-level access to flash-stored data post-boot without emulating legacy CMOS behavior.[66] Since around 2020, flash-based storage has become increasingly common in consumer PCs and especially in servers, with major vendors like Dell implementing it on Intel platforms to reduce reliance on CMOS batteries for NVRAM preservation in certain models—though batteries persist in many systems for compatibility—evident in 14th-generation PowerEdge servers as of mid-2024, where BIOS firmware version 2.22.2 or later maintains settings in flash even after battery removal.[67] This evolution aligns with broader industry trends toward solid-state nonvolatility, enhancing longevity and simplifying design in Intel's post-2018 platforms.[68]Common Issues and Maintenance
Battery Failure Symptoms
When the backup battery for nonvolatile BIOS memory begins to fail, one of the most common initial symptoms is the system clock resetting to an incorrect date and time upon each boot, as the battery can no longer maintain the real-time clock (RTC) in the absence of main power.[23] This occurs because the RTC relies on the battery to preserve time data when the computer is powered off. Similarly, customized BIOS settings, such as boot order or overclocking parameters, may revert to factory defaults after shutdown, requiring manual reconfiguration each time the system starts.[22] Users may also encounter persistent POST (Power-On Self-Test) errors during startup, including messages like "CMOS checksum error" or "CMOS battery failure/low," which indicate that the integrity check on stored BIOS data has failed due to insufficient power retention.[23] These errors typically halt the boot process, with the BIOS displaying a prompt such as "CMOS Checksum Bad - Press F1 to continue" or "Press F2 to Setup" (specific messages and keys vary by manufacturer, BIOS version, and model, commonly including F1 to continue/resume/load defaults, F2 to enter setup, or others like Del or F10). Pressing the prompted key to continue may load default values and allow the operating system to boot, or entering the BIOS setup and exiting (with or without saving changes) clears the temporary halt, permitting OS loading. However, the prompt recurs on subsequent boots until the underlying cause is resolved, with a failing CMOS battery being the most common cause of checksum-related errors. Other hardware or configuration issues can trigger similar POST halts and prompts, such as CPU fan not detected, missing hardware enabled in BIOS (e.g., a legacy floppy drive configured but absent), USB overcurrent, or other anomalies.[5][69] These errors arise when voltage drops prevent reliable storage of configuration data in the CMOS RAM. In some systems, diagnostic indicators such as motherboard LEDs or beep codes signal the issue; for example, certain Dell models display an amber LED flashing three times followed by one white flash to denote low CMOS battery voltage.[70] To diagnose battery degradation, technicians often use a multimeter to measure the battery voltage directly on the motherboard; readings below 3 V typically confirm failure, as this falls short of the nominal 3 V required for stable operation.[71] System-wide effects can extend beyond BIOS issues, potentially leading to operating system boot failures if critical time-dependent configurations, such as secure boot keys, are repeatedly lost and require re-enablement.[72] Battery failure in nonvolatile BIOS memory is prevalent after 5-7 years of use, though lifespan varies from 3-10 years depending on battery quality and environmental factors.[73] High temperatures accelerate lithium self-discharge in these batteries, roughly doubling the rate for every 10°C increase above room temperature, which shortens effective life in warm operating environments.[74]Troubleshooting and Replacement
To diagnose issues with nonvolatile BIOS memory, begin by entering the BIOS setup utility during system boot—typically by pressing keys such as Delete, F2, or F10, depending on the motherboard manufacturer—and modify a non-critical setting, such as the system date or time.[4] Power off the computer completely, unplug the power cord, and wait at least five minutes to ensure any residual power dissipates; then, power on and re-enter the BIOS to verify if the modified setting was retained.[75] If the setting reverts to defaults, this indicates a failure in the nonvolatile storage mechanism, often linked to power supply problems in the CMOS RAM.[4] For further diagnostics, employ hardware monitoring tools where supported by the motherboard. Software like HWiNFO can display the CMOS battery voltage in its sensor readings if the system exposes this data via the embedded controller or Super I/O chip, allowing real-time assessment of voltage levels (ideally around 3V). The replacement procedure for the CMOS battery, commonly a CR2032 coin cell, involves the following steps to safely restore nonvolatile memory functionality. First, power off the computer, unplug the power supply unit from the wall outlet, and press the power button several times to discharge any residual electricity.[23] Open the computer case, locate the silver coin-cell battery on the motherboard (often near the PCIe slots or I/O panel), and gently remove it by pressing the retaining clip or lifting the holder—avoid using metal tools to prevent short circuits.[75] Leave the battery out for at least five minutes (or use the motherboard's CMOS clear jumper if available, by moving it from pins 1-2 to 2-3 for 10 seconds) to fully reset the memory.[4] Insert a new CR2032 battery with the positive side facing up, ensuring proper orientation, then close the case, reconnect power, and boot into the BIOS to reconfigure settings, as they will have reset to defaults.[23] For advanced fixes on faulty BIOS chips, particularly in cases of corruption or hardware failure, reflowing the solder joints on the EEPROM chip using a hot air rework station can temporarily restore connections if oxidation or cold joints are suspected, though this requires precision to avoid damaging nearby components and is best performed by technicians with soldering experience. More reliably, use a hardware programmer such as the CH341A to extract, backup, and reflash the BIOS firmware directly onto the chip after desoldering it from the SOIC-8 socket, following the manufacturer's pinout and voltage specifications (typically 3.3V).[76] Firmware updates from the motherboard vendor can also patch checksum calculation bugs in older boards, resolving persistent errors by reloading optimized NVRAM handling routines—download the latest version via USB or within the existing BIOS interface and apply it cautiously to prevent bricking.[77] For non-responsive laptops where standard troubleshooting fails, a deep BIOS reset can be performed by accessing the motherboard and temporarily removing the CMOS battery. This procedure involves powering off the laptop, disconnecting the AC adapter and removing the main battery if possible, opening the bottom casing to access the motherboard, locating and removing the small round coin-cell CMOS battery, waiting 10-15 minutes to allow residual power to discharge and clear the settings, then reinstalling the battery and powering on the system using only the charger.[78][32] Users should consult model-specific disassembly instructions, such as those available in official manuals or video tutorials on platforms like YouTube, and proceed only if experienced, as improper handling may void the warranty or cause damage; professional assistance is recommended for those without technical expertise.[78][32] Preventive measures for nonvolatile BIOS memory in server environments include scheduling annual inspections to visually check the CMOS battery for swelling, leakage, or corrosion, and measuring its voltage with a multimeter (should exceed 2.7V under load) to preempt failures.[79] Review system event logs quarterly for warnings related to RTC or NVRAM integrity, and proactively replace the battery every 5-7 years based on usage.[80] For high-uptime systems demanding uninterrupted timekeeping, integrate external real-time clock (RTC) modules, such as those based on the DS3231 or NXP PCF85263 chips, which use dedicated batteries or supercapacitors for redundancy and connect via I2C to the motherboard, ensuring clock accuracy even during power events.[81]References
- https://wiki.gentoo.org/wiki/CMOS_BIOS_Memory
- https://wiki.softhistory.org/wiki/AMIBIOS_6