Motherboard Types and Features
Motherboard Types and Features
systems. The boards are often used with an Intel Atom processor and
are sometimes purchased as a motherboard-processor combo unit.
FlexATX Up to 9" x 7.5" Smaller version of MicroATX.
BTX Up to 12.8" wide The BTX boards can have up to seven expansion slots, are
designed for improved airfl ow, and can use an ATX power supply.
Form Factor Motherboard Size Description
MicroBTX Up to 10.4" wide Smaller version of BTX and can have up to four expansion slots.
PicoBTX Up to 8" wide Smaller than MicroBTX and can have up to two expansion slots.
NLX Up to 9" x 13.6" Used in low-end systems with a riser card.
PROCESSOR SOCKETS
Another important feature of a motherboard is the processor socket. This socket and the
chipset determine which processors a board can support. A socket for a personal computer
is designed to hold either an Intel processor or an AMD processor. Some older processors
were installed on the motherboard in a long narrow slot, but all processors sold today use
sockets. Now lets look at sockets for Intel and AMD processors.
Sockets and processors use different methods to make the contacts between them. Here is
a list of the more important methods:
A pin grid array (PGA) socket has holes aligned in uniform rows around the socket to
receive the pins on the bottom of the processor. Early Intel processors used PGA
sockets, but they caused problems because the small delicate pins on the processor
were easily bent as the processor was installed in the socket. Some newer Intel mobile
processors, including the Second Generation Core i3, Core i5, and Core i7 processors
use the PGA988 socket or the FCPGA988 socket in laptops.
A land grid array (LGA) socket has blunt protruding pins on the socket that connect
with lands or pads on the bottom of the processor. The fi rst LGA socket was the
LGA775 socket. It has 775 pins and is shown with the socket lever and top open in
Figure 4-6. Another LGA socket is the LGA1366 shown in Figure 4-7. LGA sockets
generally give better contacts than PGA sockets, and the processor doesnt have the
delicate pins so easily damaged during an installation. You learn how to use both
sockets in Chapt
Some sockets can handle a processor using a fl ip-chip land grid array (FCLGA)
processor package or a fl ip chip pin grid array (FCPGA) package. The chip is fl ipped
over so that the top of the chip is on the bottom and makes contact with the socket.
The LGA1155 socket has a fl ip chip version, which is called the FCLGA1155 socket.
The two sockets are not compatible.
A staggered pin grid array (SPGA) socket has pins staggered over the socket to squeeze
more pins into a small space.
A ball grid array (BGA) connection is not really a socket. The processor is soldered to
the motherboard, and the two are always purchased as a unit. For example, the little
Atom processors often use this technology with a Mini-ITX motherboard in low-end
computers or home theater systems.
When a processor is installed in a socket, extreme care must be taken to protect the
socket and the processor against ESD and from damage caused by bending the pins or
scratching the socket holes during the installation. Take care to not touch the bottom of the
processor or the pins or holes of the socket, which can leave fi nger oil on the gold plating of
the contact surfaces. This oil can later cause tarnishing and lead to a poor contact. So that
even force is applied when inserting the processor in the socket, all current processor
sockets
have one or two levers on the sides of the socket. These sockets are called zero insertion
force (ZIF) sockets, and this lever is used to lift the processor up and out of the socket.
Push the levers down and the processor moves into its pin or hole connectors with equal
force over the entire housing. Because the socket and processor are so delicate, know that
processors generally should not be removed or replaced repeatedly.
Table 4-3 lists the AMD sockets for desktop systems. AMD has chosen to use the PGA
socket architecture for its desktop processors. (Some of AMDs server processors use
Socket F, which is an LGA socket.) Figure 4-9 shows the AM2+ socket. The lever on the side
of the socket is lifted, and an Athlon 64 processor is about to be inserted. If you look
closely near the lower edge of the processor, you can see the small delicate pins that will
seat into the holes of the socket.
THE CHIPSET
A chipset is a set of chips on the motherboard that works closely with the processor to
collectively control the memory, buses on the motherboard, and some peripherals. The
chipset
must be compatible with the processor it serves. The major chipset manufacturers are Intel
(www.intel.com), AMD (www.amd.com), NVIDIA (www.nvidia.com), SiS (www.sis.com),
and VIA (www.via.com.tw).
Intel dominates the chipset market for several reasons: It knows more about its own Intel
processors than other manufacturers do, and it produces the chipsets most compatible with
the Intel family of processors.
INTEL CHIPSETS
Intel has produced far too many chipsets to list them here. To see a complete comparison
chart of all Intel chipsets, start at the Intel link ark.intel.com.
Here is a list of the more signifi cant chipset families by Intel:
North Bridge and South Bridge use a hub architecture. Beginning with the release in
2006 of the Intel i800 series of chipsets, a hub using the Accelerated Hub Architecture
is used to connect buses (see Figure 4-10). This hub has a fast and slow end, and each
end is a separate chip on the motherboard. The fast end of the hub, called the North
Bridge, contains the graphics and memory controller, and connects directly to the
processor by way of a 64-bit bus, called the Front Side Bus (FSB), system bus, or host
bus. The slower end of the hub, called the South Bridge, contains the I/O controller
hub (ICH). All I/O (input/output) devices, except video, connect to the hub by using
the slower South Bridge. Notice that in Figure 4-10, the primary PCI Express slot, the slot
designated for the video card, has direct access to the North Bridge, but other PCI
Express slots must access the processor by way of the slower South Bridge. On a
motherboard, when you see two major chip housings for the chipset, one is controlling
the North Bridge and the other is controlling the South Bridge (refer to Figure 4-2).
Other chipset manufacturers besides Intel also use the North Bridge and South Bridge
architecture for their chipsets.
Nehalem chipsets with the memory controller in the processor. The release of the
X58
chipset in 2008 was signifi cant because, with previous chipsets, the memory controller
was part of the North Bridge. But beginning with the X58, the memory controller was
contained in the processor housing. For example, in Figure 4-11, the Core i7 processor
contains the memory controller. Notice that memory connects directly to the processor
rather than to the North Bridge. Another signifi cant change is the 64-bit Front Side
Bus was replaced with a technology called the QuickPath Interconnect (QPI). The QPI
has 16 lanes for data packets and works similar to how PCI Express works. All Intel
chipsets since the X58 use QuickPath Interconnects. A motherboard using the X58
chipset is shown in Figure 4-12. The board comes with a fan that can be clipped to
the top of the North Bridge to help keep the chipset cool.
Nehalem chipsets, which Intel has begun to call the previous generation of chipsets,
support the Intel LGA1366 socket, the Core i7 processors, and PCI Express Version 2.
They can also support either SLI or CrossFire technologies. (SLI and CrossFire are two
competing technologies that allow for multiple video cards installed in one system.)
Sandy Bridge chipsets with the memory and graphics controller in the processor.
In
2011, Intel introduced its second-generation chipsets and sockets, which it code-named
Sandy Bridge technologies. Rather than using the traditional North Bridge and South
Bridge, only one chipset housing is needed, which houses the Platform Controller
Hub. The processor interfaces directly with the faster graphics PCI Express 2.0 bus as
well as with memory (see Figure 4-13). Therefore, both the memory controller and
graphics controller are contained within all Sandy Bridge processors. Sandy Bridge
processors, such as the Second Generation Core i7, use the LGA1155 or the LGA2011
socket, and Sandy Bridge motherboards use DDR3 memory. Sandy Bridge chipsets for
desktop computers include X79, P67, H67, Q65, Q67, and B65. The H67 chipset on
an Intel motherboard is shown in Figure 4-14 and earlier in Figure 4-1.
Ivy Bridge chipsets. Third-generation processors and chipsets by Intel, released in 2012
and codenamed Ivy Bridge, use less power, squeeze more transistors into a smaller
space, and perform better than earlier products. Ivy Bridge chipsets include B75, Q75,
Q77, H77, Z75, and Z77. Several Ivy Bridge processors use the LGA1155 socket for
backward compatibility with earlier motherboards. The Ivy Bridge chipset uses a single
Platform Controller Hub.
AMD CHIPSETS
AMD purchased ATI Technologies, a maker of chipsets and graphics processors (called a
graphics processor unit or GPU), in 2006, which increased AMD chipset and GPU offerings.
Signifi cant chipsets by AMD include the following:
The AMD A-series chipsets (code named Trinity) are designed to compete with Ivy
Bridge chipsets in the light notebook market.
The AMD 9-series chipset supports AMD CrossFireX technologies.
The AMD 9-series, 8-series, and 7-series chipsets are designed with the gamer,
hobbyist, and multimedia enthusiast in mind. They focus on good graphics capabilities
and support overclocking. The 9-series is the most current and supports 8-core
AMD processors.
The AMD 580X Crossfi re chipset supports ATI CrossFire.
The AMD 780V chipset is designed for business needs.
The AMD 740G and 690 chipsets are designed for low-end, inexpensive systems.
to component on the board. This system of pathways used for communication and the
protocol and methods used for transmission are collectively called the bus. (A protocol is
a set of rules and standards that any two entities use for communication.) The parts of the
bus that we are most familiar with are the lines of the bus that are used for data; these lines
are called the data bus. A bus can also carry electrical power (to power components on
the motherboard), control signals (to coordinate activity), and memory addresses (for one
program to tell another program where to fi nd data or instructions).
All data and instructions inside a computer exist in binary, which means there are only
two states: on and off. Binary data is put on a line of a bus by placing voltage on that line.
We can visualize that bits are traveling down the bus in parallel, but in reality, the voltage
placed on each line is not traveling; rather, it is all over the line. When one component at
one end of the line wants to write data to another component, the two components get
in sync for the write operation. Then, the fi rst component places voltage on several lines of
the bus, and the other component immediately reads the voltage on these lines. The CPU
or other devices interpret the voltage, or lack of voltage, on each line on the bus as binary
digits (0s or 1s).
The width of a data bus is called the data path size. Some buses have data paths that are
8, 16, 32, 64, 128, or more bits wide. For example, a bus that has eight wires, or lines, to
transmit data is called an 8-bit bus. Figure 4-16 shows an 8-bit bus between the CPU and
memory that is transmitting the letter A (binary 0100 0001). All bits of a byte are placed
on their lines of the bus at the same time: no voltage for binary zero and voltage for binary
one. For every eight bits of a bus, a bus might use a ninth bit for error checking. Adding a
check bit for each byte allows the component reading the data to verify that it is the same
data written to the bus.