1. The average time interval between a storage peripheral (usually a
disk drive or semiconductor memory) receiving a request to read or write
a certain location and returning the value read or completing the write.
2. A measurement of time in nanoseconds (ns) used to indicate the speed
of memory. Access time is a cycle that begins the moment the CPU sends
a request to memory and ends the moment the CPU receives the data it
requested. Memory modules complete this process in as fast as 6ns for
PC-133 MHz memory, while older modules can take up to 80ns or more.
Ambyx test system
Also called 'Ambyx Oven.' A burn-in and test system, developed by Micron,
that performs burn-in and many functional tests under high-stress conditions
to ensure long-term quality and reliability of our parts.
A machine that measures critical dimensions of designated areas on the
die at different process levels through the use of a SEM (Scanning Electron
Application specific processor
Highly integrated logic chip designed for specific applications to work
alongside a microprocessor (e.g., a math co-processor, graphics processor,
artificial intelligence processor, LAN processor, digital signal processor).
These chips offload some of the specialized number crunching from the
The area of the RAM that stores the bits. The array consists of rows
and columns, with a cell at each intersection that can store a bit.
The large rectangular section in the center of the die where the memory
(American Standard Code for Information Interchange) A method of encoding
text as binary values. The ASCII system requires nearly 256 combinations
of 8-bit binary numbers to support every possible keystroke from the
A process in a multitasking system whose execution can proceed independently,
"in the background.
Describes a type of L2 cache that is not in synch with the system clock.
Asynchronous cache is slower than its synchronous counterpart, but is
capable of delivering information to the CPU at a rate 10 percent faster
than standard DRAM. Asynchronous cache was first used to boost memory
performance in 386 systems and is still widely used today.
A Synchronous DRAM feature that allows the memory chip's circuitry to
close a page automatically at the end of a burst.
The capacity to move data on an electronic line such as a bus or a channel.
In short, the amount of data moved relative to a specific time frame.
It is expressed in bits, bytes, or Hertz (cycles) per second. Essentially,
a measure of the capacity of data that can be moved between two points
in a given period of time.
A slot or group of slots that must be populated with modules of like
capacity and fulfill the data width requirement of the CPU.
A method of diagramming memory configurations. The bank schema system
consists of rows and columns that represent memory sockets on a system:
rows indicate independent sockets and columns represent banks of sockets.
A printed circuit board (PCB) that does not have any components on it.
Burst EDO - A variant on EDO DRAM in which read or write cycles are
batched in bursts of four. Burst EDO bus speeds will range from 40MHz
to 66MHz, well above the 33MHz bus speeds that can be accomplished using
Fast Page Mode or EDO DRAM.
Ball Grid Array - a square package with solder balls on the underside
for mounting. Use of BGA allows die package size to be reduced by allowing
more surface area for attachment. Smaller packaging allows more components
to be mounted on a module making greater densities available. The smaller
package improves heat dissipation improving performance. See CSP and
A method of encoding numbers as a series of bits. The binary number
system, also referred to as base 2, uses combinations of only two digits
- 1 and 0.
Basic Input Output System - often referred to as CMOS, the BIOS provides
an interface for a computer's hardware and software. The BIOS configuration
determines how your hardware is accessed.
Short for Binary Digit, the smallest unit of data that can be processed
or stored by a computer. A bit can have a value of either 1 or 0. Bits
make up 'computer' language the same way letters of an alphabet make
up human languages. Different combinations of different bits form 'words'
and 'sentences' (actually signals) that a computer understands. Before
these words and sentences can be transmitted from the CPU to memory,
or vice versa, they must be broken down into 8-bit segments called bytes.
Older computers were designed to handle only 8-bit data segments, while
newer models have progressed to 64-bit segments. This larger bit width
capacity generally means better and faster computer performance.
A physical unit of information in a logical record; block size is usually
expressed in bytes.
A circuit or system drawing concerned with major functions and interconnections
Square metallic pads on the die where the ball bond is attached. The
bond pad is used to find acceptable eye points.
This is when there is so much memory the chipset needs assistance to
deal with the large loading introduced by the large amounts of memory.
A buffer isolates the memory from the controller to minimize the load
the chipset sees. This means adding logic, particularly drivers, to
a SIMM or DIMM to increase the output current. Buffering is used to
overcome signal attenuation due to capacitive loading. Modules that
are "buffered" usually have small buffer chips mounted on
The process of exercising an integrated circuit at elevated voltage
and temperature. This process accelerates failure normally seen as "infant
mortality" in a chip. (Those chips that would fail early during
actual usage will fail during burn-in. Those that pass have a life expectancy
much greater than that required for normal usage.)
Bursting is a rapid data-transfer technique that automatically generates
a block of data (a series of consecutive addresses) every time the processor
requests a single address. The assumption is that the next data-address
the processor will request will be sequential to the previous one. Bursting
can be applied both to read operations (from memory) and write operations
The central communication avenue in a PCs system board. It normally
consists of a set of parallel wires or signal traces that connect the
CPU, the memory, all input/output devices, and peripherals and allows
data to be transferred from one system component to another. Busses
come in a variety of bit widths and speeds. To prevent data bottlenecks,
the components attached to a bus must operate at close to the same speed
as the bus.
A single transaction occurring between the system memory and the CPU.
A unit of information made up of 8 bits. The byte is the fundamental
unit of computer processing; almost all aspects of a computer's performance
and specifications are measured in bytes or multiples of bytes such
as kilobytes (~1,000 bytes) or megabytes (~1 million bytes), or gigabytes
(~ 1 billion bytes)
A small fast area memory holding recently accessed data, designed to
speed up subsequent access to the same data. Typically used between
a processor and main memory.
The circuit in control of the interface between the CPU, cache and DRAM
Cache RAM is high-speed memory (usually SRAM) which is dedicated to
storing frequently requested data. If the CPU needs data, it will check
in the high-speed cache memory first before looking in the slower main
memory. Cache memory may be three to five times faster than system DRAM.
Most computers have two separate memory caches; L1 cache, located on
the CPU, and L2 cache, located between the CPU and DRAM. L1 cache is
faster than L2, and is the first place the CPU looks for its data. If
data is not found in L1 cache, the search continues with the L2 cache,
and then on to DRAM.
The property of a circuit element that allows it store an electrical
A type of memory typically used in laptop and notebook computers. Credit
card memory features a small for factor and is named for its similarity
to the size of credit card.
(Column Address Select/or Strobe)--A control pin on a DRAM used to and
activate a column address. The column selected on a DRAM is determined
by the data present at the address pins when CAS becomes active.
CAS-RAS (CBR) (CAS before RAS)
CAS before RAS. Column Address Strobe Before Row Address Strobe. A fast
refresh technique in which the DRAM keeps track of the next row it needs
to refresh, thus simplifying what a system would have to do to refresh
When a device that was initially good now fails to function under any
Extra data bits provided by a DRAM module to support ECC function. For
a 4-byte bus, 7 or 8 check bits are needed to implement ECC, resulting
in a total bus width of 39 or 40 bits. On an 8-byte bus, 8 additional
bits are required, resulting in a bus width of 72 bits.
A detail test pattern designed to exercise each individual cell in the
memory and find possible shorts between adjacent columns and data buses
Complementary High-density Metal Oxide Semiconductor.
Complex Instruction Set Computing. This design logic is usually associated
with microprocessors. CISC chips use instructions, or commands, that
involve several steps in one.
The number of pulses emitted from a computer's clock in one second;
it determines the rate at which logical or arithmetic gating is performed
in a synchronous computer. An electrical current that alternates between
high and low voltages. The speed of the clock is measured in Megahertz
The rate at which a computer's internal system clock operates. The
clock is used to synchronize operations between the components within
Complementary Metal Oxide Semiconductor. A process that uses both N-
and P-channel devices in a complimentary fashion to achieve small geometries
and low power consumption. On a PC CMOS generally refers to the BIOS
information stored on a CMOS chip.
Cache On A Stick. Coast modules were used to upgrade a motherboard's
L2 cache and Tag memory on some socket 7 and older motherboards.
Chip On Board. A system in which semiconductor dice are mounted directly
on a PC board and connected with bonded wires or solder bumps. The dice
are usually mechanically protected with epoxy.
Part of the memory array. A bit can be stored where a column and a row
Compact Flash Memory
A fast, postage stamp size RAM that is removable. The CF Card weighs
half an ounce, with roughly one-fourth the volume and one-half the thickness
of a PCMCIA Type II Card. The CF Card fits into a CF PC Card Adapter
making it compatible with a standard PCMCIA Type II slot on any notebook
or desktop computer. This allows the easy transfer of stored digital
information from the CF Card to a computer or printer. Currently, the
most readily available application for the CompactFlash Card is the
digital still camera.
One of the major units in a computer that interprets and carries out
the instructions in a program.
With respect to semi-conductor packages, the condition of leads in a
package having all elements, or all elements in a seating plane, between
two parallel planes.
(Central Processing Unit)--The chip in a computer that has primary responsibility
for interpreting commands and running programs. The CPU is the most
vital component of a computer system. The speed of the CPU has a significant
impact on overall system performance, but the CPU doesn't act alone.
If slower memory is paired with a fast processor, the processor will
be forced to wait for the memory to respond. When the speed mismatch
is extreme, the user will see numerous memory errors and even complete
Continuity RIMMs are used to fill all unused RIMM sockets in a system.
CRIMMs do not use any active components, and are used to continue the
channel so that the signal can be properly terminated at the motherboard.
Chip Scale Package. CSP is a type of BGA in which the package is roughly
the size of the die. CSP is also known as mBGA or micro-BGA.
In resonant circuits, the decay of oscillations due to the resistance
in the circuit.
The signal line that carries the data read from the RAM (Random Access
A marking on all PCB and DRAM components indicating the manufacturing
date of the product.
(Double Data Rate) or SDRAM II--The next generation of the current SDRAM.
DDR finds its foundations on the same design core of SDRAM, yet adds
advances to enhance its speed capabilities. As a result, DDR allows
data to be sent on both the rising and falling edges of clock cycles
in a data burst, delivering twice the bandwidth of standard SDRAMS.
DDR essentially doubles the memory speed from SDRAMs without increasing
the clock frequency.
An individual rectangular pattern on a wafer that contains circuitry
to perform a specific function. The internal circuitry is made of thousands
of tiny electronic parts. 'Die' refers to a semiconductor component
or part that has not yet been packaged (also known as 'IC' (Integrated
Circuit) or 'chip').
Die pick-up tool
The bondhead tool on the machine that picks up the die from the precisor
and places it on the leadframe.
The physical measurements of the die.
A material that conducts no current when it has voltage applied to it.
Two dielectrics used in semiconductor processing are silicon dioxide
and silicon nitride.
A layer of deposited oxide used to isolate metal 1 from metal 2 on double-level
metal processes. This must be done in such a way to prevent hillock
formation on level 1.
The intermingling of molecules of two or more substances. When high
temperature processes are done in diffusion tubes, the high temperature
accelerates diffusion. Typical diffusion furnace temperature is 950
degrees Centigrade, or 1742 degrees Fahrenheit.
Dual Inline Memory Module. A printed circuit board with gold or tin/lead
contacts and memory devices. A DIMM is similar to a SIMM, but with this
primary difference: unlike the leads on either side of a SIMM, which
are "tied together" electrically, the leads on either side
of a DIMM are electrically independent, ie actually separate circuits
which allows for wider and faster data transfer.
(Dual In-line Package) A form of DRAM component packaging. DIPs can
be installed either in sockets or permanently soldered into holes extending
into the surface of the printed circuit board. The DIP package was extremely
popular when it was common for memory to be installed directly on a
Data mask signal used by SDRAMs to provide byte masking during write
operations. There is one DQM signal for every 8 bits of data width.
A computer memory address that is included as part of the instruction.
Direct memory access
A computer feature that allows peripheral systems to access the memory
for both read and write operations without affecting the state of the
computer's central processor.
Systems using intelligent input/output controllers and direct - memory
- access control to free the CPU of the details of block transfers.
The introduction of an element that alters the conductivity of a semiconductor.
Adding boron to silicon will create a P-type (more positive) material,
while adding phosphorus or arsenic to silicon will create N-type (more
(Dynamic Random Access Memory) DRAM is the most common type of memory
and is "dynamic" because in order for the memory chip to retain
data, it must be refreshed constantly ( a pulse of current through all
of the memory cells every few milliseconds). If the cell is not refreshed,
the data is lost. DRAM temporarily stores data in a cell composed of
a capacitor and a transistor. Each cell contains a specified number
of bits. These cells are accessed by row addresses and column addresses.
(See also RAM and SRAM.)
(Direct Rambus DRAM) A totally new RAM architecture, complete with
bus mastering (the Rambus Channel Master) and a new pathway (the Rambus
Channel) between memory devices (the Rambus Channel Slaves). A single
Rambus Channel has the potential to reach 500MBp/s to 800Mb/s in burst
mode; a 20-fold increase over DRAM.
A printed circuit board that sends signals from the interface board
of the oven to the DUT board and back to the interface board. Each oven
slot has a corresponding driver board located in the back of the oven.
The process of preparing product for shipment in moisture vapor barrier
bags. This process includes tubed or reeled product and a clay desiccant,
and an HIC (Humidity Indicator Card), vacuum-sealed in a moisture vapor
Device Under Test. It is used interchangeably with UUT (Unit Under Test).
Type of RAM (Random Access Memory). To keep data in the D(ynamic)RAM
memory, this data needs to be 'refreshed' (recharged). The electric
charge fades out of a DRAM like air seeps out of a balloon. Because
of this change, it is called Dynamic.
Electrically Erasable PROM.
Electrically Alterable Read-Only Memory.
Error Correction Code. A method used to check the integrity of data
stored in memory . ECC memory improves data integrity by detecting errors
in memory and is more advanced than parity because it can detect both
multiple-bit errors and single-bit errors (parity only detects single-bit
errors). ECC is typically found in high-end PCs and file servers where
data integrity is key. An ECC scheme capability is partially determined
by the sophistication of the "systematic code" employed. The
systematic code is like a reference table that the memory system uses
to determine whether or not the memory has returned the correct data.
Every time data is stored in memory, this code is responsible for the
generation of check bits which are stored along with the data. When
the contents of a memory location is referenced, the ECC memory logic
uses the check bit information and the data itself to generate a series
of "syndrome bits". If these syndrome bits are all zeros,
then the data is valid and operation continues. If any bits are ones,
then the data has an error and the ECC memory logic isolates the errors
and reports them in the operating system. In the case of a correctable
error, the ECC memory scheme can detect single and double bit errors
and correct single bit errors.
A memory feature that allows for faster back to back accesses.
EDO Parity RAM
EDO Parity RAM offers the high performance of EDO memory and has built-in
parity which greatly improves reliability. Ideal for high-end PCs and
entry-level servers, EDO Parity modules are compatible with any system
that accepts a standard 72-pin EDO module and are rapidly becoming the
new standard on high-end systems.
(Extended Data Out) EDO RAM is similar to FPM memory, a form of DRAM
technology that shortens the read cycle between memory and CPU. but
provides improved performance by keeping available data longer in memory.
It eliminates much of the wait time by allowing the processor to access
data during the refresh cycle. In other words, the computer can load
data as it is searching for new information. EDO memory is generally
10 to 20% faster than FPM memory. A computer must support EDO memory
in order to notice an increase in performance.
(Enhanced Dynamic Random Access Memory)--a form of DRAM that boosts
performance by placing a small complement of static RAM (SRAM) in each
DRAM chip and using the SRAM as a cache. Also known as cached DRAM,
Electrically Erasable Programmable Logic Device. A CMOS PLD made by
using EEPROM technology. It can be erased and reprogrammed.
Electrically Erasable, Programmable, Read-Only Memory chip. EEPROMs
differ from DRAMs in that the memory stays in even if electrical power
is lost. Also, the memory can be erased and reprogrammed.
Electrostatic discharge (ESD)
The dissipation of electricity. ESD can easily destroy semiconductor
End Of Buffer.
Electrically Erasable, PROgrammable, Read-Only Memory chip. EEPROMs
differ from DRAMs in that the memory stays in even if electrical power
is lost. Also, the memory can be erased and reprogrammed.
(ECC on SIMM) A data-integrity checking technology designed by IBM that
features ECC data-integrity checking built onto a SIMM.
Erasable Programmable Read Only Memory
The process of applying a cured-plastic protective housing to components.
A mold compound. An Assembly step.
A process using a chemical bath (wet etch) or a plasma (dry etch) that
removes unwanted substances from the wafer surface.
A local area network allowing several computers to transfer data on
a communications cable.
Even parity and odd parity are two different parity protocols used to
check the integrity of data stored in memory. A memory manufacturer
can use either protocol in a memory product. Even parity adds an additional
bit to every byte of data to make the total number of 1's in the segment
even. When the byte passes to the CPU, the parity circuit checks the
byte to be sure it is still even. If it is, the data is considered to
be valid and the parity bit is removed from the byte. If instead it
registers as odd, it is considered to be invalid and a parity error
Description of the rate at which parts fail, usually expressed as percent
Unlike odd and even parity, fake parity is not capable of detecting
an invalid data bit. It was designed to artificially 'satisfy' a parity-enabled
computer without actually checking the data for errors. Fake parity
attaches a bit to each byte of data just like odd and even parity protocols.
The difference is that fake parity simply adds the correct parity bit
as the data is sent to the CPU instead of attaching it before the data
is stored to memory, and recalculating it before the byte passes to
Material that fails various tests within the component manufacturing
Fine BGA is a ball grid array package with a fine pitch ball arrangement
on the underside of the package (larger than CSP).
Failures In Time.
Front Side Bus is the data channel connecting the processor, chipset,
DRAM, and AGP socket. FSB is described in terms of its width in bits
and it's speed in MHz.
In computing: A status bit that causes some indication of the state
or condition of the processing unit.
Flash memory is a non-volatile memory device that retains its data when
the power is removed. The device is similar to EPROM with the exception
that it can be electrically erased, whereas an EPROM must be exposed
to ultra-violet light to erase. Flash memory does not need a constant
power supply to retain its data and it offers extremely fast access
times, low power consumption, and relative immunity to severe shock
or vibration. These qualities combined with its compact size, make it
perfect for portable devices like scanners digital cameras, cell phones,
pagers, hand-helds and printers. Flash chips have a lifespan limited
to 100,000 write cycles, which means flash will never replace main memory
A Teflon Polyurethane wafer holder used to transport individual wafers.
Flatpacks can be stacked to carry and protect several wafers at a time.
A flat, rectangular IC package type with leads sticking out from the
sides of the package.
A circuit with two stable states that can be changed from one to the
other. Flip-flops are the storage element in most of the SRAMs.
Pertaining to the condition of a device or circuit that is neither grounded
nor connected to any potential. (Potential is voltage course or current
In Silicon Gate MOS technology: a gate that is not directly connected
to the rest of the circuit. Used in EEPROMs.
Fast Page Mode - A common DRAM data-access scheme. Accessing DRAM is
similar to finding information in a book. First, you turn to a particular
page, then you select information from the page. Fast-page mode enables
the CPU to access new data in half the normal access time, as long as
it is on the same page as the previous request. This feature is used
to support faster sequential access to DRAM by allowing any number of
accesses to the currently open row to be made after supplying the row
address just once.
A device or system that can change the frequency of an alternating current,
whether or not it changes the voltage or phase.
Approximately 1 billion bits: 1 bit x 1,024 (that is, 1,073,741,824
bits) Or exactly 2^30 bits.
A unit of measurement approximately equal to 1024 megabytes. Computer
components process data in bytes or multiples of bytes such as kilobytes
(~1,000 bytes), megabytes (~1 million bytes), and gigabytes (~ 1 billion
The wire used to make a physical connection from the device to the leadframe.
Die that fail functionality testing. These failures have a visual defect
99 percent time, such as poly or metal bridging, missing geometries
or layers, particles or contaminates.
A structure, attached to or part of a semiconductor device that serves
the purpose of dissipating heat to the surrounding environment; usually
metallic. Some packages serve as heat sinks.
Hyper Page Mode also known as EDO.
High Temperature Operating Life.
Connection to a CPU that is configured or programmed to provide data
path between the CPU and external devices such as a keyboard, display,
or reader; it may be an input port or an output port, or it may be bi-directional.
Integrated Circuit. A tiny complex of electronic components and their
connections that is produced in or on a small slice of material (as
In Circuit Emulator.
Identification Detect. Pins present on DIMMs to provide information
to the system using the module.
Standard set by IEEE (Institute of Electrical and Electronics Engineers)
for communication between pieces of electronic apparatus.
Used to describe the occurrence of premature failures at a higher than
A burn-in process whereby electrical functionality of the parts is continuously
or periodically monitored and recorded under various voltages, temperatures,
and refresh conditions during the burn-in process. This continuous or
periodic monitoring of the functionality of each IC allows intelligent
decisions to be made.
A PCB that interfaces between the computer and an interface board.
Current x Resistance = Voltage. Also an abbreviation for Infrared.
An amorphous, doped polysilicon used as an underlying layer for the
HSG poly to increase conductivity.
International Standards Organization.
Joint Electron Device Engineering Council - the group that establishes
the industry standards for memory operation, features and packaging.
Back To Top To Top
Notches on a memory module that prevent it from being installed incorrectly
or into an incompatible system.
Kilobit Approximately one thousand bits: 1 bit x 210 (that is,
A unit of measurement approximately equal to 1024 bytes. Computer components
process data in bytes or multiples of bytes such as kilobytes (~1,000
bytes), megabytes (~1 million bytes), and gigabytes (~ 1 billion bytes).
Level 1 cache. A small cache integrated in processor that provides a
small working space for quick access to the most recently used data.
A specialized memory unit that enhances DRAM performance by providing
the CPU with data at speeds ten times faster than DRAM. The L2 cache
is comprised of Static RAM (SRAM), a high-speed RAM that does not need
to be refreshed to retain its data. Most computers have two different
memory caches; L1 cache, located on the CPU, and L2 cache, located between
the CPU and DRAM. L1 cache is faster than L2, and is the first place
the CPU looks for its data. If data is not found in L1 cache, the search
continues to the L2 cache and then to DRAM. In early processors, the
L2 cache was not integrated into the processor but rather built into
the motherboard, and was in some cases upgradeable. See COAST.
Process which uses a YAG (Yittrium-aluminum-Garnet) laser to melt the
silicon in a dot matrix to form wafer scribe numbers.
Circuit element that stores a given value on its output until told to
store a different value.
An undesired phenomenon in an integrated circuit whereby a circuit locks
in a certain state and will not change.
The effective input voltage at which a flip-flop changes states.
The metal extensions from an IC package or discrete component that connects
the component to the PCB. The leg or contact point of the component
that is either physically soldered to a PC board or placed within a
socket for connection.
A metal structure that is part of the device. The die is attached to
Leads or Legs: The official name for the metal 'feet' on an IC. Also
called 'pins.' The part of the lead assembly that is formed after a
portion of the lead frame is cut away. The part's connection to the
Undesirable conductive paths in components, subsystems, and systems;
also the current through such paths.
Accelerated testing of electronic components to establish their field
A circuit that produces a voltage output approximately proportional
to the input voltage, generally over a limited range of voltage frequency.
Power supply design in which the voltage is held constant by dissipating
50% of the input voltage times and output current as a margin.
A method of selecting memory or input/output devices that dedicates
one address line per chip selection; results in overlapping or noncontiguous
memory; used because it is the cheapest method of selection.
A pin in the mold which locates the leadframe in the correct position
on the mold for processing.
An integrated circuit which provides a fixed set of output signals according
to the signals present at the input.
Several individual device functions on an integrated circuit chip.
Detail test pattern designed to check for decoder and cell interaction
Amount of memory equal to 1 bit x 1,0242 or 1,048,576 bits of information.
Amount of memory equal to 1,048,576 bytes of information. (Abbreviated
A Term commonly used to refer to computer system's random access memory
(see also RAM). The term memory has also been used to refer to all types
of electronic data storage (see storage). A computer system's memory
is crucial to its operation; without memory, a computer could not read
programs or retain data. Memory stores data electronically in memory
cells contained in chips. The two most common types of memory chips
are DRAM and SRAM.
A logical unit of memory in a computer, the size of which is determined
by the computer's CPU. For example, a 32-bit CPU calls for memory banks
that provide 32 bits of information at a time.
The amount of memory in an IC and how it is accessed. Also, a code on
the lot traveler used to indicate the IC's memory configuration (e.g.,
1M1 = 1 Meg x 1, 4M4 = 4 Meg x 4, etc.).
The logic chip used to handle the I/O (input/output) of data going to
and from memory. See chipset.
Minimum amount of time required for a memory to complete a cycle such
as read, write, read/write, or read/modify/write.
- Cache Data SRAM: quick-access chip.
- DRAM dynamic random access memory.
- SDRAM synchronous dynamic random access memory.
- DDR SDRAM double data rate dynamic random access memory.
- SLDRAM synchronous link dynamic random access memory.
- RDRAM (also DRDRAM) Rambus dynamic random access memory.
- EPROM: erasable, programmable, read-only memory.
- PROM: programmable, read-only memory.
- RAM: random access memory.
- ROM: read-only memory (permanent memory that cannot be changed).
- SRAM: static random access memory.
Megahertz is a measurement of clock cycles in millions of cycles per
A unit of measure equivalent to one-millionth of a meter; synonymous
Millions of Instructions Per Second. This measurement is generally used
when describing the speed of computer systems.
Metal Nitride Oxide Semiconductor. The technology used for EAROMs (Electrically
Alterable ROMs); not to be confused with NMOS.
Moisture vapor barrier bag
A vacuum-sealed bag designed to keep the moisture out so that the parts
inside will not be damaged.
Contained on one chip or substrate, as a microprocessor system including
not only the logic but also memory or input/output circuits.
Metal-Oxide-Semiconductor. Layers used to create a semiconductor circuit.
A thin insulating layer of oxide is deposited on the surface of the
wafer. Then a highly conductive layer of tungsten silicide is placed
over the top of the oxide dielectric.
Device in which current flow occurs in a single channel of P- or N-type
material and is controlled by an insulated electrode on the surface
of the channel region.
The set of chemical and metallurgical steps used to make MOS Large Scale
Metal Oxide Semiconductor Transistor.
Motherboard (also known as Mainboard)
Also known as logic board, main board, or system board; your computer's
main electronics board, which in most cases either contains all CPU,
memory, and I/0 functions, or has expansion slots that support them.
Mean Time Between Failures.
Mean Time To Failure.
Memory Unit. Usually a printed circuit board assembly populated with
memory chips that stores a certain quantity of memory. Intel term for
one of the types of cards in a memory system card set.
A computer logic circuit that produces an output which is the inverse
of that of an AND circuit.
Literal: One-billionth (10 to the -9). Diffusion: A tool used to measure
the thickness of a film on a wafer.
One billionth of a meter.
One billionth of a second. Memory data access times are measured in
nanoseconds. For example, memory access times for typical 30- and 72-pin
SIMM modules range from 60 to 100 nanoseconds.
Charge caused by the presence of electrons, not their absence.
A unit of force in the meter-kilogram-second system needed to accelerate
a mass of one kilogram one meter per second per second.
Usually 4 bits or half a byte.
N-channel Metal Oxide Semiconductor. This pertains to MOS devices constructed
on a P-type substrate in which electrons flow between N-type source
and drain contacts. NMOS devices are typically two to three times faster
than PMOS devices.
A term created by Apple Computer, Inc. that describes a memory module
which uses 16-Mbit technology. For a given capacity, a noncomposite
module will have fewer chips than a composite module.
Types of memory that retain their contents when power is turned off.
ROMs, PROMs, EPROMs and flash memory are examples. Sometimes the term
refers to memory that is inherently volatile, but maintains its content
because it is connected to a battery at all times, such as CMOS memory
and to storage systems, such as hard disks.
Nanosecond (ns). One billionth of a second; used to measure the speed
of the parts (e.g., -07 nanoseconds).
Non-Volatile Random Access Memory.
Even parity and odd parity are two different parity protocols used to
check the integrity of data stored in memory. A memory manufacturer
can use either protocol in a memory product. Odd parity adds an additional
bit to every byte of data to make the total number of 1s odd. When the
byte passes to the CPU, the parity circuit checks the byte to be sure
it is still odd. If it is, the data is considered to be valid and the
parity bit is stripped from the byte. If instead it registers as even,
it is considered to be invalid and a parity error is generated.
Output-Enable. On a part, where data-in and data-out are shared on the
same pins, the OE must be triggered to request output data.
A unit of measure of electrical resistance.
A circuit interruption that results in an incomplete path for the current
flow. (e.g., an open wire which opens the path of the current).
Software controlling the overall operation of a multipurpose computer
system, including such tasks as memory allocation, input and output
distribution, interrupt processing, and job scheduling.
An electronic circuit which amplifies 'linear' (also called analog)
A conductive bonding island located strategically on circuit chips for
inter-connecting circuit elements or for bringing connections from circuit
leads to the outside.
The number of bits that can be accessed from one row address.
Mode in which if RAS is kept low and the DRAM is given a column-address
without being given a new row-address, the chip will remember which
row it was on the last time and automatically stay on that row. It is
like saying that all the bits along one row are all on the same 'page,'
and the part will assume the same page is intended until a different
page is specified.
Programmable Array Logic. A device that can be programmed to do certain
logic functions. Then a fuse inside of the device can be blown so the
programmed information can never be changed. Sometimes called a PLD
(Programmable Logic Device) Language.
A test of the DUT that checks for pin leakage, the amount of current
it draws, opens, and shorts.
A test that measures values, also called a dynamic test as opposed to
a functional or Go/No-Go test.
A series of voltage and current tests performed on all products in Probe.
The test checks for variations in the fabrication process. Test results
are used by engineers to modify or correct processes.
A quality control method that checks the integrity of data stored in
a computer's memory. Parity works by adding an extra bit of data to
each byte to make the total number of 1's either odd or even An error
is detected if the parity circuit determines that this number has changed,
indicating that some of the data may have been lost or otherwise corrupted.
Two different parity protocols exist, even parity and odd parity. Parity
protocols are capable of detecting single bit errors only. To enable
multiple-bit error detection, manufacturers must use a more advanced
form of error checking called Error Correcting Code (ECC). See also
A bit added to a group of bits to detect the presence of an error.
The portion of a language translator (compiler or assembler) which determines
the logical structure of the program being completed.
A device incapable of current gain or switching such as a resistor or
A circuit element without an energy source such as a capacitor or resistor.
Intel's PC100 specification defines the requirements for SDRAM used
on 100 MHz FSB motherboards. Around the middle of 1998, Intel introduced
the BX chip set to their motherboard designs. One element in this new
architecture will include an increase in the PC main memory bus speed
(Host bus) from 66 to 100 MHz, called PC 100. To match the 100MHz bus
speed, 100MHz SDRAM modules is the required memory technology for this
new chip set.
The PC133 specification details the requirements for SDRAM used on 133MHz
FSB motherboards. PC133 SDRAM can be used on 100MHz FSB motherboards
but will not yield a performance advantage over PC100 memory at 100MHz.
(Printed Circuit Board) A component made up of layers of copper and
fiberglass; the surface of a PCB features a pattern of copper lines,
or "traces," that provide electrical connections for chips
and other components that mount on the surface of the PCB. Examples:
motherboard, SIMM, credit card memory, and so on.
(Personal Computer Memory Card International Association) A standard
that allows interchangeability of various computing components on the
same connector. The PCMCIA standard is designed to support input/ output
devices, including memory, fax/modem, SCSI, and networking products.
Presence Detect. Indicator pins on SIMMs and DIMMs that provide information
to the system using the module.
Pin Grid Array.
The metal extensions from an IC package or discrete component that connects
the component to the PCB.
Pin one hole
The hole located on the 'pin one' side of the leadframe.
An indentation or mark on the top of the part that indicates where the
first lead of the die inside is located.
Pipeline Burst Cache
A type of synchronous cache that uses two techniques to minimise processor
wait states - a burst mode that pre-fetches memory contents before they
are requested, and pipelining so that one memory value can be accessed
in the cache at the same time that another memory value is accessed
Programmable Logic Array. An array of logic elements that can be programmed
to perform a specific logic function. It can be as simple as a gate
or as complex as a ROM and can be programmed (often by mask programming)
so that a given input combination produces a known output function.
Programmable Logic Devices. Devices with 10-100 times higher level of
integration than a TTL; called programmable because they can be customized
in software rather than in hardware.
P-channel Metal Oxide Semiconductor. This pertains to MOS devices constructed
on an N-type silicon substrate in which holes flow between source and
Protective covering over the die; also called Die Coat
Poly-crystalline layer of silicon used for the silicon gate contact
in silicon gate MOS devices; also used for interconnections between
A PCB with components.
To turn the system's power OFF.
To turn the system's power ON.
Plastic Quad Flat Pack. A square, flat package with gullwing leads located
around all four sides of the package.
Cache that is closest to the processor: typically located inside the
CPU chip. Can be implemented either as a unified cache or as separate
sections for instructions and data. Also referred to as Level 1 cache
or internal cache.
Wire used to make electrical contact with a pad on a die; usually made
of either beryllium copper, tungsten, or palladium. The diameter of
the probe shank is 10 mils, the diameter of the standard probe tip is
1.5 mils, and the length is 7 or 14 mils.
A fiberglass card (P.C. Board) that has a hole in the center in which
there are pins that are aligned and placed on pads located on the die.
As the pins on the probe card are placed on die pads, the probe card
tests and sorts die for functionality.
Memory that is custom-designed for a specific computer.
A device or method used to keep the output voltage of a device at a
high level; often a resistor network connected to a positive supply
Printed Wiring Board; board upon which there are layers of printed circuits
where DRAMs can be attached with solder so that memory can be accessed.
Quad flat pack
QFP: A flat, rectangular, integrated circuit with its leads projecting
from all four sides of the package without radius.
Random Access Memory - A type of memory that can be written to and read
from in a nonlinear (random) manner. When an application is opened.
it is transferred from the hard drive to RAM where it is more readily
accessible. RAM enhances system performance because it can process requests
from the CPU more quickly than the hard drive. The kind of RAM used
in main memory on most computers is Dynamic RAM (DRAM). DRAM stores
data as electronic signals. These signals must be constantly refreshed
to keep them from dissipating. The more RAM your computer has, the more
data it can store at one time and subsequently the more efficiently
your computer will operate. The data held in RAM is lost when the computer
is turned off. The term random derives from the fact that the CPU can
retrieve data from any individual location, or address, within RAM.
A period of time in the oven when the temperature goes up.
Random failure region
The portion of the bathtub curve that represents the useful portion
of device life.
The difference between the smallest and largest values in a set of data.
This is the simplest measure of variation.
Row-Address-Strobe: the signal that tells the DRAM to accept the given
address as a row-address. Used with CAS and a column-address to select
a bit within the DRAM.
Rambus DRAM is an evolutionary type of DRAM that uses a 16-18 bit data
path and is designed to operate with FSB speed of 800 MHz producing
a burst transfer rate of 1.6 gigahertz. Rambus DRAM technology is a
system-wide, chip-to-chip interface design that allows data to pass
through a simplified bus. Rambus uses a unique RSL (Rambus Signaling
Logic) technology. Rambus is available in two flavors: RDRAM and Concurrent
RDRAM. The third line extension, Direct RDRAM, was developed in stages
and went into production in 1999. In late 1996, Rambus agreed to a development
and license contract with Intel that lead to Intel's i820 and i840 chip
sets supporting Rambus memory being released in 1999.
The amount of time required for the output data to become valid once
the read and address inputs have been enabled; generally called access
A mode of operation used in core memory systems.
A generic term for Random Access Memories.
An electrical process used to maintain data stored in DRAM. The process
of refreshing electrical cells on a DRAM component is similar to that
of recharging batteries. Different DRAM components call for different
This is a count of the number of rows (in thousands) refreshed at a
time in a refresh cycle. Common refresh rates are 1K, 2K, and 4K. DRAM
stores data as a series of electron charges in individual cells. This
data must be constantly recharged or 'refreshed' to keep the data from
dissipating. The refresh rate refers to the size of the data that must
be recharged, and is typically expressed in kilobytes (~1,000 bytes).
Two common refresh rates are 2K and 4K, with 2K being the faster rate.
Registers delay memory information for one clock cycle to ensure all
communication from the chipset is collected by the clock edge, providing
a controlled delay on heavily loaded memories.
An identifier that indicates the position of a memory location in a
computer routine relative to the base address as opposed to the memory
location's absolute address.
The amount of time data must remain stable after a device or circuit
has been clocked; also called 'hold time.'
A material that prevents etching or plating of the area it covers; also
Rambus Inline Memory Modules used for RAMBUS DRAM. A form of chip packaging
that is similar to DIMMs using Direct Rambus DRAM memory subsystems.
Reduced Instruction Set Computing. The design methodology is usually
associated with microprocessors. RISC chips use simpler instructions,
or commands, than CISC chips. However, they need to use more steps to
perform many functions that CISC chips perform in one step. SPARC and
MIPS chips are based on RISC.
The amount of time required for a signal level change to increase from
ten percent to ninety percent of its final specified value.
Returned Material Authorization; required if a customer desires to return
products. Also refers to parts that have been returned from a customer.
Read Only Memory - A form of random access memory that can only be read
from, not written to. Most systems use ROM to store the instructions
a computer needs during the startup process.
Part of the RAM array; a bit can be stored where a column and a row
Describes how many rows are on a wafer map in the X direction. (X =
left to right. Y = top to bottom).
An acronym to represent RS/1 Programming Language. RPLs are functions
or procedures written in RS/1's built in programming language. These
RPLs are used to automate tasks and analyses.
A marking on a wafer that identifies the wafer and the lot it came from.
The scribe is located on the front of the wafer, opposite the major
Synchronous Dynamic Random-Access Memory. A DRAM technology that uses
a clock to synchronize signal input and output on a memory chip. The
clock is coordinated with the CPU clock so the timing of the memory
chips and the timing of the CPU are "in synch." The synchronization
eliminates time delays and allows for fast consecutive read and write
capability, thereby increasing the overall performance of the computer.
SDRAM has two separate memory banks that operate simultaneously, while
one bank prepares for access, the other is being accessed. SDRAM allows
the CPU to access memory approximately 25 percent faster than EDO memory
because it is controlled by the system clock. SDRAM can only be used
in computers designed for it and cannot be mixed with any other type
of memory. SDRAM can operate at 100MHz, 133Mhz and features a burst
mode that allows it to address blocks of information instead of small
Cache that is second closest to the processor; typically located on
the system board. Also referred to as Level 2 cache and external cache.
A memory technology that enables DRAM to refresh on its own-independent
of the CPU or external refresh circuitry. This technology is built into
the DRAM chip itself and reduces power consumption dramatically. It
is commonly used in notebook and laptop computers.
Soft Error Rate. An error caused by temporary disruption of memory cell.
Serial Presence Detect (SPD)
Serial Presence Detect. An enhanced presence detect that uses an EEPROM
to store modules timings, configuration, and the manufacturer's data.
An element, such as silicon, that has intermediate in electrical conductivity
between conductors and insulators, which conduction takes place by means
of holes and electrons.
The sense amp acts as a distributor of current on the die.
A device or circuit capable of sensing very low voltages and amplifying
them to some higher voltage level.
Syncronous Graphics RAM. A single port DRAM designed for graphics hardware
that require high speed throughput such as 3-D rendering and full-motion
A reduction in die (chip) size. A reduction in the size of the circuit
design resulting in smaller die sizes that increases the number of possible
die per wafer.
Single In-line Module. Same as SIP except with a connector edge instead
Single In-line Memory Module - a high-density DRAM package alternative
consisting of several components connected to a single printed circuit
board. A small PCB designed to mount in a socket on a larger PCB providing
a large memory upgrade in a small space. This board provides the connection
between multiple memory chips and the computer system. SIMMS come in
various pin configurations, the most common type being: 30 pin and 72
pin. A 30 pin SIMM has a row of 30 tin or gold pins long the bottom
of the module which determine the amount of data the module can handle.
These pins connect to only one memory chip as opposed to DIMMs which
can connect to multiple chips.
An interconnect component mounted on the system board, or motherboard,
designed to hold a single SIMM.
Single In line Package. A component or module that has one row of leads
along one side. Many resistors come in SIP form.
Synchronous Link Dynamic Random Access Memory. SLDRAM is the next evolution
of SDRAM using a multiplexed command bus allowing fewer pins to increase
bandwidth and allow higher FSB speeds.
Small Outline Dual Inline Memory Module. Smaller and thinner than standard
DIMMs, SODIMMs are typically used in laptop computers and mobile computing
devices. An enhanced version of a standard DIMM. The small outline DIMM
is about half the length of a typical 72-pin SIMM. SO DIMMs come in
a variety of pin sizes and can be installed either singly to support
32-bit systems, or in pairs to support 64-bit systems.
An error caused by a temporary disruption of the memory cell.
Soft error fail
A part with a temporary, single-bit failure during the soft error test.
Small Outline Rambus Inline Memory Module. SORIMMs have a smaller profile
that standard RIMMs and are used in laptop computers and systems that
have strict size requirements.
Small Outline J-lead package. A rectangular package with leads sticking
out of the side of the package. The leads are formed in a J-bend profile,
bending underneath and towards the bottom of the package.
Statistical Process Control. The use of statistics to determine uniformity
around a target value.
SPD (Serial Presence Detect)
An enhanced presence detect that uses an EEPROM to store manufacturer
The time it takes the RAM to put information into its memory or get
information out of its memory. It is measured from the time that an
address and proper control signals are given, until the information
is stored or placed in the device's output(s).
Our coding for the speed that the stored information in the part can
be retrieved by a computer. For DRAMs, a -5 is 50 nanoseconds, a -6
is 60 nanoseconds, a -7 is 70 nanoseconds, etc. For SRAMs, a -10 is
10 nanoseconds, etc.
An operation found in coat and develop programs that stands for spin
The sudden drastic portion of a pulse that significantly exceeds its
average amplitude. Standard deviation A measure of variation for a particular
process or product characteristic. This is often abbreviated as 'STD
DEV' or 'STD'.
Standard Readability Assessment.
(Static Random Access Memory) An integrated circuit similar to a DRAM
(Dynamic Random Access Memory) with the exception that the memory does
not need to be refreshed. SRAM is faster and more expensive than DRAM
and is generally used for speed-critical areas of the computer such
as cache memory.
Smallest or lowest address that a memory system will respond to.
Unlike volatile memory, static memory retains its contents even when
the main current is turned off. The trickle of electricity from a battery
is enough to refresh it.
A medium designed to hold data, such as Floppy Diskette, Harrd Disk
Drive or CD-ROM.
An input that allows parallel data to be entered a synchronously.
The actual structural material on which semiconductor devices are fabricated,
whether passive or active. The term applies to any supportive material,
such as the materials used in the fabrication of printed circuits.
A J-leaded or Gullwing package or BGA that can be mounted directly on
the surface of P.C. Boards (as opposed to through-hole packages).
The menu in BIOS containing user-configurable options for a PC's hardware
A kind of L2 cache that is synchronized with the CPU. This eliminates
the lag time created while the CPU waits for cache memory to fulfill
its requests. Synchronous cache is typically 3 to 5 percent faster than
asynchronous cache, and is a full 20 percent faster than standard DRAM.
See also Asynchronous Cache.
The subset of the CPU address bits used to compare the tag bits of the
cache directory to the main memory address being accessed. Tag memory
acts as an index for the information stored in L2 cache.
Cache is physically divided into two sections. The Tag RAM section stores
the Tag address of the location of the data in cache. This section is
smaller than the Data RAM section, which stores the actual data or instruction.
To Be Determined. Used on quotes in reference to shipping dates.
A measurement of the speed or how fast the part is; it is the time it
takes to get a bit of information out of the part after CAS comes down.
Thermal Coefficient of Expansion. A constant that describes the changes
in linear dimensions with respect to temperature for a device or material.
A code for TSOP (see TSOP). A package-type code.
Thin Small Outline Package. It is thinner and slightly smaller than
an SOJ and with gullwing-shaped leads. A thin, rectangular package with
leads sticking out the sides of the package. TSOP DRAM mounts directly
on the surface of the printed circuit board. The advantage of the TSOP
package is that it is one-third the thickness of an SOJ package. TSOP
components are commonly used in small outline DIMM and credit card memory
An electrical process every product goes through which tests the parts
for parametric, speed, and functional failures.
A semiconductor device that uses a stream of charge carriers to produce
active electronic effects.
The process of creating a printed circuit on a printed circuit board
This number (x8, x9, x32, x36, x64, x72, x80) refers to the bit depth
of a module, or to the size of the data path used to access the memory.
This is where the chipset controller deals directly with the memory.
There is nothing between the chipset and the memory as they communicate.
Micron (or micrometer). A unit of length equal to one millionth of a
Microsecond: One millionth of a second.
Collector Common Voltage.
Memory that loses its contents when the power is turned off. A computer's
main memory, made up of dynamic RAM or static RAM chips, loses its content
immediately upon loss of power. Contrast ROM, which is non-volatile
Video RAM. DRAM with an on-board serial register/serial access memory
designed for video applications. VRAM has two separate data ports. One
is dedicated to updating the image on the screen while the other one
is used for changing the image data stored in memory. This "dual-ported"
design gives higher performance than DRAM which cannot read and write
simultaneously but is more expensive.
Vss is the abbreviation for the ground on a connection. (Like the ground
wire on a battery.)
A condition, transaction, or event that changes or may be changed as
a result of processing additional data through a system.
Write-Enable; WE must be pulsed low when data is written to the chip.
Time expended from the moment data is entered for storage to the time
it is actually stored.
An addressing mode in which the address is given as an unsigned binary
number that specifies one of the memory locations between 0 and 256(decimal).