Refreshing Your Memory Basics






Refreshing Your Memory Basics

You can look at memory in the following ways:

  • As the physical chips and modules that mount on the motherboard or in one of a PC’s interfaces

  • As the scratch pad space that holds data and instructions before and after their use by the processor

For the two A+ exams, you need to know something about both of these considerations, including the form factors, purposes, and uses of the various types of memory. You also need to know something about memory configuration and a bit more about the troubleshooting processes. (Troubleshooting in general is one of the major focuses of the A+ exams.)

Memory systems are complex and offer a depth of material — the kind of stuff engineers get misty-eyed over. For such a physically small thing, memory is a large subject. But not here! For example, I give no lengthy diatribes on why parity memory is better than nonparity memory. I give you just the facts, along with a small amount of background information to help you understand the exam’s questions — quite a trick in itself.

Differentiating RAM from ROM

 Remember  The PC has the following two basic types of memory:

  • RAM (random-access memory): Holds the instructions and data in use by the operating system and software applications before and after the data is passed to the CPU. RAM is volatile; a steady power source is required to hold its contents. Should the power stream be broken, anything that is stored in RAM is erased. No power, no data — pretty straightforward.

  • ROM (read-only memory): Is nonvolatile and retains its contents through even the darkest power outage. For more information on ROM, see Chapter 4.

RAM is by far the faster of the two types of memory. In fact, RAM is often used to shadow the BIOS ROM to improve its performance during the boot process. If you’re asked what copying the contents of the BIOS ROM into RAM for faster access to the instructions by the CPU is called — the answer is shadowing.

 Tip  When volatile memory loses its power source, it also loses its contents. Most types of RAM, especially DRAM, are volatile, and that’s why when you turn off the PC’s power (or trip over the power cord), everything in RAM is lost. Nonvolatile memory, like ROM and other types, does not lose its contents when the power source is lost.

Random-access memory (RAM)

RAM is the primary memory of the PC and is generally installed directly on the motherboard in a variety of package types, which I deal with later. Many types of RAM exist, and you need to know about the following ones for the A+ Hardware Technology exam:

  • DRAM (Dynamic Random-Access Memory)

  • EDO RAM (Extended Data Output RAM)

  • RIMM (Rambus In-line Memory Module)

  • SRAM (Static RAM)

In addition to the types of RAM that are used in primary memory, you should also know about the following types of RAM that are used on video cards:

  • VRAM (Video RAM)

  • WRAM (Windows Accelerator Card RAM)

You need a solid understanding of RAM, its packaging forms, and its technologies for the exam. Many questions assume that you know the types of RAM, their characteristics, and where they are installed. For example, you may be asked which is the fastest of the RAM types (SRAM), or how DIMMs are installed (in vertical sockets on the motherboard).

Dynamic RAM (DRAM)

DRAM (pronounced “dee-ram,” but not in the same context as “dee-bears” or “dee-bulls,” although it could apply to Marshal Faulk) is the RAM that computer types talk about. DRAM has, until recently, always been the most common type of memory in PCs.

Compared to other forms of integrated circuits, DRAM isn’t a complex circuit and, as a result, is not expensive. However, its design also requires that it be refreshed regularly or it loses its contents. This need for constant refreshing gives DRAM its dynamic tag.

 Instant Answer  DRAM must be refreshed every 2 milliseconds. A special refresh logic circuit reads and then rewrites the contents of each DRAM address, regardless of whether it’s in use.

DRAM is also the slowest of the memories, clocking in with access speeds around 50 nanoseconds (ns) or higher (higher, in this case, means slower). Older DRAM had an access speed of about 120 ns.

DRAM technologies

 Remember  DRAM comes in a variety of popular technologies. The following lists the characteristics of each of the DRAM technologies:

  • Extended Data Out (EDO): This is the most common type of DRAM. It’s common in most Pentium and later PCs, except those with memory buses over 75 MHz.

  • Fast Page Mode (FPM): This type of DRAM is occasionally called non-EDO RAM. It’s generally compatible with motherboards with memory buses with speeds under 66 MHz.

  • Burst Extended Data Out (BEDO): This DRAM is EDO memory with pipelining technology added for faster access times. BEDO memory allows much higher bus speeds than EDO memory.

  • Synchronous DRAM (SDRAM): Like its SRAM cousins (see the section, “Static RAM (SRAM),” later in this chapter), SDRAM is tied to the system clock and reads or writes memory in burst mode.

  • Rambus DRAM (RDRAM): Rambus is a proprietary DRAM technology developed by Rambus, Inc. (www.rambus.com) that has memory speeds of up to 3.2 Gbps. RDRAM comes on a module that is similar to a DIMM, called a RIMM (Rambus In-line Memory Module).

  • Synchronous Link DRAM (SLDRAM): This is an enhanced version of SDRAM memory that uses a multiplexed bus to transfer data to and from the chips rather than fixed pin settings. SLDRAM has transfer rates as high as 3 Gbps. Unlike RDRAM, this is an open technology.

  • Video RAM (VRAM): This specialized DRAM is used on video cards and not for main memory. VRAM applies dual porting, which means that it can be written to and read from at the same time. This allows the processor and the refresh circuitry for the monitor to access VRAM at the same time. Another type of video RAM is Windows RAM (WRAM), also called Windows Accelerator RAM, which has essentially the same properties as VRAM.

You will definitely see DRAM technology questions on the A+ exams. Familiarize yourself with the general descriptions that I provide here, which should be enough for FPM, EDO, Burst EDO, RDRAM, and SDRAM. VRAM and WRAM are covered in more detail in Chapter 12.

DRAM packaging

DRAM memory comes in the following package forms:

  • DIP (Dual In-line Packaging)

  • SIMM (Single In-line Memory Module)

  • DIMM (Dual In-line Memory Module)/RIMM

SIMM and DIMM packages are like mini-expansion boards that have either surface-mounted SOJ (Small Outline J-lead) or TSOP (Thin, Small Outline Package) DRAM soldered on one (SIMM) or two (DIMM) sides of a circuit card. Figure shows the basic forms of a DIP and a memory module (a SIMM is shown).

Click To expand
Figure: The two basic package forms of DRAM: a memory module (SIMM) and a DIP chip.

SIMMing right along

The memory standard on middle-aged PCs (486s and early Pentiums) is the SIMM (see Figure). The edge connector on a SIMM has either 30 pins or 72 pins. A SIMM’s memory capacity ranges from 1MB to 16MB in either a one-sided or two-sided style, with chips soldered to one or two sides of the board.

Click To expand
Figure: A typical SIMM.

Because the Pentium processor uses a 64-bit path to memory, 32-bit SIMMs must be installed in pairs. Each SIMM bank has two sockets and both sockets must be filled before another bank receives a SIMM.

Moving up to DIMM

The Dual In-line Memory Module (DIMM) is the memory standard on most newer and larger PCs. Because its 64-bit memory matches the 64-bit data path of the Pentium processor, you need to install only one DIMM at a time. In comparison to the SIMM, a DIMM has 168 contact pins as opposed to the 30 and 72 pins of the SIMM. A DIMM looks just like a SIMM, except that it’s slightly larger, has memory chips on both sides, and has about twice as many contacts on its edge connector.

DIMMs come in different voltages — 3.3V and 5.0V — and as buffered or unbuffered, which yields four possible combinations.

A smaller version of the DIMM, the 100-pin Small Outline DIMM (SODIMM), which is used primarily in laptop computers, supports PC100 SDRAM. An even smaller DIMM package, the 144-pin MicroDIMM, which is used in subnotebook portable computers, supports PC100 SDRAM.

On the RIMM

Where SIMM and DIMM are generic names for a type of memory module, RIMM is a trademarked name for the Rambus memory module. A RIMM looks like a DIMM, but the RIMM has 184 pins on its edge connector. RIMMs transfer data in 16-bit chunks. A RIMM is packaged inside an aluminum sheath called a heat spreader. The heat spreader covers the entire assembly to protect against overheating. Figure shows the parts of a RIMM.

Click To expand
Figure: RIMMs are packaged inside of a heat spreader.

A smaller version of the RIMM, the SORIMM (Small Outline RIMM), is similar to the SODIMM, with the exception of the Rambus technology.

For more information on memory than your brain can possibly hold, check out Kingston Technologies’ “Ultimate Memory Guide” at www.kingston.com/tools/umg.

Installing DRAM

The installation procedures for DIP memory and SIMMs and DIMMs are quite different. As a PC repairperson, you have probably had an occasion to install a SIMM or DIMM. However, you may not have had to install a DIP memory chip. You need to have some idea of the processes that are used to install memory in a PC.

 Remember  You do have your ESD (electrostatic discharge) protection on, don’t you? You should take steps to protect the PC — and especially memory chips or modules. It doesn’t take much of an ESD to severely damage the memory that you’re installing. Always perform the following steps before installing memory in the PC (and for any other operation inside the case):

  1. Turn off the PC and disconnect the AC power cord.

  2. Follow the instructions in the PC’s owner’s manual on how to locate the memory expansion sockets.

  3. Before touching anything inside the case or opening the memory’s package, make sure that you first touch an unpainted, grounded metal object, such as a chassis wall or support, to discharge any static electricity that’s stored on your body or clothing.

  4. Handle memory modules carefully. Don’t bend or flex them, and always grasp them by the edges.

Putting in the SIMMs

To install a SIMM, align it with the socket at about a 45-degree angle. As the SIMM is lifted to a vertical position and seated in the socket, the clamping clips on the ends of the socket grab the module and hold it in place. To remove a SIMM, release the clamping clips, push the module to a 45-degree angle, and then lift out the module.

Dropping in a DIMM

A DIMM is installed by inserting the module into an available memory socket. A DIMM is keyed to match the socket, so it only goes in one way. After the DIMM is aligned to the socket slot, firmly press down on the module until it seats in the socket slot and locks into place with a snap.

Nearly all DIMM sockets have ejector tabs. To remove a DIMM, press down on the ejector tabs, and the module should pop up and out of the socket slot. Carefully lift the module out of the socket.

Putting in a RIMM

A RIMM is installed in special RIMM connectors. Check the motherboard’s documentation to see whether a pair of RIMM connectors exists on the board. Most likely, they are the connectors that you couldn’t identify. Both slots of a RIMM connector set must be occupied by two RIMMs or a single RIMM and what is called a C-RIMM (Continuity RIMM). A C-RIMM doesn’t contain memory; it’s only a pass-through module that completes the memory channel.

You install a RIMM almost exactly as you would a DIMM — press down until the locking clips snap onto the module. To remove the module, press the ejector tabs outward; this pops the module out of the socket.

Hot-swapping flash memory

Notebook and palmtop computers can have flash memory added with a PC card. A PC card looks like a credit card (see Figure) and slips into a slot that’s usually located on the side of the notebook’s base. One feature of PC cards is hot-swapping, which allows you to remove and replace the card while the system is running. Chapter 14 covers portable systems and PC cards in more detail.

Click To expand
Figure: A PC card flash memory module.

The Personal Computer Memory Card International Association (PCMCIA) is the standards authority on PC cards. This organization has developed three primary standards for PC cards that define the cards’ size and use. PC cards are 85.6 millimeters (mm) by 54 mm (or about 3.4 inches by 2.1 inches) in size. The thickness of the card is set by the three card type standards. See Chapter 14 for more information, but briefly the three standard PC card types are as follows:

  • Type I: Up to 3.3 mm thick and used primarily for adding memory

  • Type II: Up to 5.5 mm thick and commonly used for I/O devices, such as data/fax modems, network interface cards, and mass storage devices

  • Type III: Up to 10.5 mm thick and used for rotating mass storage drives

Static RAM (SRAM)

SRAM is static, which means that it can retain its data using a very low voltage electrical charge and doesn’t need to be refreshed. As long as it has a power stream, SRAM holds its charge and contents. SRAM also has very fast access times, in the range of 15 to 20 ns (comparing to the 50 to 120 ns of DRAM). SRAM in a DIP package is physically two pins longer than DRAM in the same packaging, and because it’s a more complex technology, it costs a lot more. Figure contrasts SRAM to DRAM in terms of their general characteristics.

Figure: DRAM versus SRAM

DRAM

SRAM

Slow and must be constantly refreshed

Fast and doesn’t require refreshing

Simple

Complex

Inexpensive

Expensive

Physically small

Physically large

SRAM is packaged as either a single DIP chip, on a COAST (Cache On A Stick) module in a variety of increments, or embedded into other chips, such as processors or controllers.

SRAM is available as either synchronous or asynchronous. Synchronous SRAM uses the system clock to coordinate its signals with the CPU, and asynchronous doesn’t.

Caching in on SRAM

SRAM is most commonly used on a PC as cache memory. The type of SRAM and the amount that a motherboard supports vary with the motherboard. However, because of its greater cost and bigger physical size, SRAM is most typically used for the Level 2 (outside the processing chip) cache memory that is mounted directly on the motherboard.

Memory caching enables the CPU to work more efficiently. Cache memory stores data and instructions and then fetches them in anticipation of what the CPU will want next. When cache memory is correct, which is about 90 percent of the time, the CPU can get what it needs from the much-faster SRAM instead of the slower DRAM. It is now common for a system to include from 128K to 512K of SRAM cache on the motherboard. Chapter 3 includes information on how caching works.

Comparing DRAM to SRAM

In spite of its constant thirst for power, DRAM works better for PC main memory because it’s cheaper and needs less space. SRAM costs too much and takes too much space to be used for main memory. On the other hand, because of its speed and the fact that not much memory is needed, SRAM is perfect for cache memory.



 Python   SQL   Java   php   Perl 
 game development   web development   internet   *nix   graphics   hardware 
 telecommunications   C++ 
 Flash   Active Directory   Windows