Cad Guidebook: A Basic Manual for Understanding and Improving Computer-Aided Design (11 page)

The precision of the image is controlled by the monitor’s resolution. The
resolution is an indication of how finely or coarsely the image can be presented,
and it is based on the number of tiny regions on the screen that can be discreetly
activated or “painted.” These tiny regions are called pixels (or picture elements).
If one looks very closely at most monitors the pixels can be seen. Obviously, if a
monitor can display more pixels, then it can show a better or more precise
picture.

32 Chapter 2

The resolution for a monitor is usually expressed in terms of the number of
pixels in the horizontal and vertical directions. For instance, a monitor may be
said to have 1024 by 768 resolution. This means that there are 1024 pixels in a
single “row” across the screen in the horizontal direction, and there are 768 pixels
in a single “column” of the screen in the vertical direction. With a monitor able to
display a resolution such as 1280 by 1024, the total number of pixels exceeds one
million; this may be referred to as a megapixel system.

Keep in mind, however, that the monitor does not generate these pixels, it
merely displays them. It is the graphics adapter that creates them, but there is a
limit to the ability of a monitor to display these pixels. Therefore, a graphics
adapter that can produce 1600 by 1200 resolution would not be used to full capa-
bility with a monitor that is only designed to handle 800 by 600.

The monitor actually generates the display image by a method of horizon-
tal scanning. The electrical signals from the graphics adapter are translated into
the control of a beam that excites special chemicals inside the monitor to make
the colors appear. This beam proceeds as a horizontal line across the screen, and
this line then may scan from the top of the screen to the bottom. The repeating of
this beam produces the refresh rate. The rate at which this refreshing occurs (such
as a number of times a second) can also have an effect on the performance of the
monitor. Another ramification of the scanning behavior of the monitor is that
computer graphics data from a computer program is often referenced from the
top of the screen, instead from the bottom (which may be more natural in a math-
ematical sense where the 0,0 point is at the lower left corner). In other words, the
location of a pixel at specific X and Y coordinates may have the Y measured
from the top of the screen with higher numbers proceeding DOWN from the top;
the X coordinate usually then proceeds from left to right, as usual.

2.7.2 Graphics Adapters

As mentioned previously, the graphics adapter is the device which really creates
the image to be displayed on the monitor. It is, therefore, probably the most im-
portant peripheral for a computer running CAD software. The performance and
capability of the graphics adapter could easily have the greatest impact on a de-
signer’s productivity. They can be quite complicated devices, but a basic under-
standing of its functioning should prove quite valuable in optimizing
performance.

Graphics adapters are electronic components on a circuit board that can be
installed into a computer. Although this circuitry could be self-contained on the
“motherboard” of the computer system, usually computer manufacturers permit a
variety of graphics adapters for their computer systems (at a wide variety of
costs). Therefore the graphics adapter is left as a removable circuit board. Also,
the graphics adapters have often been in a cycle of improvement that is indepen-

Computer Hardware Basics 33

FIGURE
2.6

Simplistic schematic of graphics systems and process.

dent of that for complete computer systems, and leaving them as an expansion or
option allows the user to upgrade the graphics adapter without replacing the en-
tire computer. The graphics adapter may also be referred to as the graphics card
since these expansion circuit boards are often referred to as expansion cards. Or,
the graphics adapter may be referred to as the video card, but this may be mis-
leading since video also refers to analog-type signals like from a videotape or
television.

Although the graphics adapter is just an expansion, it could have almost as
many components as a complete computer. The graphics adapter may have its
own type of CPU that makes calculations and manipulates data in its own mem-
ory. This processor may be called a coprocessor to distinguish it from the main
processor, or it may be called the GPU (Graphics Processor Unit). The presence
of the co-processor has a very large impact on performance. The graphics adapter
will also have its own graphics memory systems similar to the complete com-
puter (or a portion of the main memory designated as the graphics memory). In
addition, the graphics adapter will have various programs running on it to create,
process, and transmit the image to the monitor. The amount of this programming
can vary widely depending on the coprocessor and memory.

The basic operation of the graphics adapter is pretty simple. As mentioned
above, there is a graphics memory set aside by the design of the computer system
(the graphics memory is a set of memory chips of the appropriate size and speed).
First, the main computer program (such as the CAD software) determines what
image is needed on the screen. Second, data is created for that image in the

34 Chapter 2

graphics memory. Third, the graphics adapter takes what has been put into the
graphics memory and translates it into an image on the screen as quickly as pos-
sible. The graphics adapter is constantly scanning the data in the graphics mem-
ory and then creating the signal to scan down the monitor as it refreshing. Since
there is a sort of “shuttling” of the data between the computer system and the
monitor, and often there is electronic circuitry to temporarily hold or “buffer” the
data as it is being shuttled, the terminology of “graphics buffer” or “picture
buffer” or even “frame buffer” is often used to refer to the kind of data held by the
graphics memory and various memory chips on the graphics adapter itself. The
details of these processes is beyond the scope of this work, but the key concept is
that there is a special graphics memory area and the graphics adapter is
constantly trying to get the data that is symbolized in that memory sent to the
monitor. Figure 2.6 shows a simple schematic.

The most misleading statement in the previous paragraph is the second step
of the process—creating the data that is put into the graphics memory. In some
cases, the CAD software will actually use the main CPU to figure out how to
translate the concept of an analog geometric entity (such as a circle) into a set of
digital data that can reside in the graphics memory. Obviously, the monitor with a
limited number of pixels can not show a perfect circle; it needs to be broken
down into a set of pixels that approximates the circle. If the CAD software itself
creates this approximation and then sends the data to the graphics memory, then
the graphics adapter does not perform this second step, and only has to do the
third step (send the image data to the monitor).

However, in other cases, the CAD software will not determine the approxi-
mation of the geometry and send the data to the graphics memory. In this case,
the CAD software “calls on” the power of the graphics adapter (which would
probably have a coprocessor in this case) to make the approximation and create
the data in the graphics memory. In this situation, the graphics adapter is doing 2
jobs—creating the data in the graphics memory and sending the image to the
monitor. Obviously, this will only work if the programming running on the
graphics adapter is compatible with the way the CAD software needs the data to
be created. The compatibility issues raised by graphics adapters are generally ad-
dressed by the use of industry standards: agreements between computer and
electronic device manufacturers to have components behave in certain, predeter-
mined fashions based on some sort of language or protocol. The standards may
arise from industry groups such as the IEEE or EIA. Or, they may be considered
de facto standards which means that one manufacturer’s design is so pervasive,
all other manufacturer’s change to match it accordingly. A very early example of
a de facto standard for graphics was TCS or PLOT10 for Tektronix®
4014 termi-
nals. A more recent standard would be OpenGL®.

At first, the task of creating digital approximation of geometric entities cre-
ated by CAD software may seem simple. However, this task can be complicated,

Computer Hardware Basics 35

such as when the CAD program is relying on the graphics adapter to create
shaded 3-D models. In this case, the CAD program may be completely relieved
from interacting with the user for the manipulation of the 3-D model. Once the
CAD program defines the desired 3-D surfaces to the graphics adapter, the user
may manipulate and “slice through” the 3-D model by interacting only with the
graphics adapter. In this case, the coprocessor may be doing millions of compli-
cated computations per second.

Beyond the presence and capability of the coprocessor, probably the next
most important characteristic of a graphics adapter is the amount and configura-
tion of the graphics memory. As mentioned earlier, the graphics memory may be
simply a section or region of the computer system’s main memory. In this case,
there will be a somewhat fixed amount of graphics memory to consider since it is
an integral part of the computer design. In other cases, the computer may allow
for variation or expansion of the graphics memory. In general, more graphics
memory (or Graphics RAM, Video RAM, etc.) indicates a better graphics adapter
system.

The graphics memory has enough addresses or locations for storing data
within it so that each pixel on the monitor gets a certain predetermined amount of
information. If a graphics adapter operates at a resolution of 1280 by 1024, then
there are 1,310,720 pixels (1280 times 1024). So, at least these many addresses
are needed (or about 1.2 megabytes) in the graphics memory. Of course, each
byte can contain any pattern of 8 ones and zeros. These 8 bits can then be used to
define specific characteristics of a given pixel. The most obvious characteristic is
color.

Looking at an 8-bit per pixel system, then, the emission of colored light is
based on 3 primary colors (red, green, and blue), so 2 of the 8 bits can represent
the amount of red, 2 of the 8 bits can represent the amount of green, and 2 of the
8 bits can represent the amount of blue (ignoring the other 2 bits). Any of those 2
bit combinations can have 4 different values (00, 01, 10, 11), so there are 4 levels
or shades of each individual color possible for this fictitious graphics adapter.
Therefore, the total number of possible shades (combinations of the red, green,
and blue) is 4 times 4 times 4 or a total of 64.

Note that all the pixels have some sort of definition in their corresponding
location in graphics memory, even if it looks like the screen is blank. If the graph-
ics adapter is functioning and blank areas of the screen are black, then it just
means that those parts of the Graphics Memory is set to all 0s (no red, no green,
no blue/black) by whatever program is running. As mentioned previously, it is the
job of the CAD software to get this graphics memory loaded with the right data
(on its own or by calling on the coprocessor), and then it is the graphics adapter
that looks at the pattern of bits and then sends the appropriate signal to the moni-
tor so that it displays the image desired. This process of turning bits (or digital
data) into a “real” electrical signal (or analog) is called a digital-to-analog con-

36 Chapter 2

version, and the reverse is called analog-to-digital. This process is performed by
“D to A” converters (DACs). This is an important function of the graphics
adapter, but it is performed behind the scenes.

Although the resolution in the example of graphics memory is pretty good
(about 1 megapixel), the number of colors is not very good. Of course, at one
time 64 colors was acceptable for 2-D CAD, but more likely at least 64,000 col-
ors would be needed to show realistic 3-D models for a CAD system (curved sur-
faces need to show a very gradual change in the color to give the illusion of
depth). The way to improve the available colors (or the palette) is to increase the
amount of graphics memory so that each pixel has a greater depth of information.