Memory Basics

408.995.0900

Fax 408.995.0933

www.huffcomm.com

Basic, Basic Memory Basics

While there are myriad different types of memory chips, the majority of those manufactured today have a simple, basic premise:

They are organized as a bunch of rows which can be counted (from the bottom). In many of today's memories, the number of rows climbs into the millions. The bottom most row is row 0, the next is row 1, the next is row 2, etc. Each of these rows is identified by its number. This number is also called the Address and so, by calling out the number of some row, we have Addressed the location.

(Note: EEs often start counting from zero. Then when we reach seven, we know we've got eight. It's easier for us somehow.)

Most memories also share other traits (Fig. A):
• An Address Bus a signal path through which a row number is communicated to the memory. A Bus simply refers to a group of one or more conductors which route together and conduct similar information such as an address or a control signal.
• A Control Bus a signal path upon which control signals, such as Read or Write commands, are sent to the memory.
• Some small number of columns, typically 1, 4, 8 or 16. The number of columns defines the "width" of the memory. If the width is 8 columns, as in a 2-Mbit x 8 memory, then each row contains 8 separate storage locations. In this case, any row number specifies 8 bits of information, one bit in each of the columns of the row specified.
• A Data Bus a bus that has the same number of conductors as the memory has columns. The Data Bus conducts data TO the memory (for storage) during a write operation and conducts data FROM the memory during a read operation.

Most memories also respond to two basic types of control signals: Read and Write (excepting Read-Only memories which usually Š don't write).

A read instruction is accompanied by an address (row number). In response to the read instruction, the memory locates the row addressed and then returns the information (data) which is stored in that row.

A write instruction is accompanied by an address AND data. The memory locates the addressed row and then stores the data in the appropriate place.

Figure A

Now you have it. That's Memory Basics. From here, we simply start to look at enhancements to these basic principals and soon we'll know most everything that's important about memory chips.

Single Port/Multi-Port Memory Devices
The basic structure described above is sometimes referred to as Single-Port memory based on the notion that the device has a single channel (address, data, and control signals) through which the memory can be accessed (Fig. B). There are however, applications in which it is beneficial to have multiple ports through which to access the memory.

Figure B

For example: Say there was a central contacts file or address book on one computer where everyone in your office could look up and update addresses and phone numbers for the customers and vendors who are important to your operation. This approach would have the advantage of having a single version of all contact information which everyone could use. If one person learned of a customer phone number change, he could update the system and anyone subsequently seeking that information would find the correct phone number. This would work great except for the fact that everyone would have to line up at one computer to get to the information.

Another approach might be to put a copy of the central contacts file or address book on each computer in the office. This would start out great since everyone could get to the information independently and through their own computer. But clearly, the system would go down hill fast since updated information wouldn't be available to anyone except those using the computer upon which the update was made. Everyone else would be frustrated by looking up old information and then, redundantly, going through the process of updating their files. Pretty soon, no one would be using the contacts system, complaining that it has bad information.

The ideal solution is to have a central storage center where the current information is stored and updated but which is accessible from each of the computers in the office. (No lines, no bad data o.k. so let's pretend it's a perfect world.)

This is the premise of the Multi-Port memory. A Multi-Port memory has two (Dual-Port) or more channels through which independent "users" can simultaneously access the information (Fig. C).

Figure C

One type of Dual-Port memory in use today is the Video RAM (VRAM,) which has two ports: One Parallel Port and one Serial Port (Fig. D).

The Parallel Port has separate "parallel" conductive paths for each of the buses (control, address and data). This has the advantage of providing high-speed data transfers.

The Serial Port uses a single conductor to transfer all of the information (control, address and data signals) by sending each in sequence one bit at a time. This "serial" method is similar to that used by a telegraph or those lights used to blink information between ships sailing on the high seas at night.

Using the parallel port, the system processor can quickly load up the video information to be displayed on the monitor and then is free to attend to other tasks, while the video processor transforms the digital information into the video display. The video controller, meanwhile, uses the serial port. This enables it to work at its own pace gathering data as needed from the same storage center through the Serial Access Manager (SAM) block of the VRAM.

Figure D

"Basic, Basic Memory Basics" first appeared on EBNONLINE.COM on September 6, 1999.