acceptor dopant - an impurity element with only three electrons in its outer subshell. This leaves an extra space for an electron (a hole) in the crystal lattice that can easily accept an electron.

byte - a grouping of eight bits, or binary digits. On most computers, a byte is the smallest part of main memory that has an address.

energy gap - the region between the valence band and the conduction band in a semiconductor, which has no orbital states for electrons; in order to become part of an electric current, an electron must jump from the valence band to the conduction band, over the energy gap.

Java - a programming language specifically designed to write programs (applets) that can be downloaded over the WWW. Java is supposed to be independent of the type of computer on which it is executed.

monocrystalline – a property of certain materials (particularly silicon for this class) in which a uniform crystal structure exists throughout the material. This is essential for creating properly functioning integrated circuits.

MOS transistor - a simple form of transistor that employs the effects of an electric field under a "gate" to control whether the transistor is on or off.

search engine - a WWW server that allows you to seek html files that relate to a particular topic. They come in two flavors: spiders, that access various Web servers, look at their files for specified keywords, then place the URL of the file in a database indexed by the keyword; or indices, that group files that are suggested to them under various categories.

timesharing computer - a method of accessing a computer where a number of people using terminals all execute programs on a single machine. The computer gives each in turn a small part of the total available computation power, giving each the illusion that they are connected to a somewhat slower machine dedicated to them.

UniCode - a convention for representing text characters by 16-bit binary numbers (instead of the 8-bit numbers of the ASCII code); this allows a single code to consistently represent characters from many languages.

wafer - a thin, round disk of monocrystalline silicon that is the unit on which integrated circuits are created.

2. (20 points)

3. (5 points) Convert the characters below into ASCII code. Use hexadecimal values to display the result.

5. (5 points) What is the truth table for the following 2-input CMOS circuit? What logic function does the circuit implement?

6. Explain why floating point numbers in a computer may not represent a numeric quantity completely accurately.

There are really two issues here. One is that all floating point numbers in a computer have a fixed, finite number of bits to hold the magnitude of the value. For a single precision floating point number, the magnitude can be equivalent to 6 or 7 decimal digits (expressed, of course, as a binary number). If I try to enter 20 decimal digits, only the first 6 or 7 will actually be used. The rest are discarded, usually after rounding the last digit retained depending on the value of the digits discarded. So trying to use more digits of precision than the floating point number can hold leads to inaccuracies.

Also, there are finite decimal numbers that have no finite binary equivalent value (0.1 is an example), so even a finite decimal number may need more digits of precision than the floating point number can provide.

7. What is an advantage in using logarithms to graph values that vary widely in magnitude? What is a disadvantage of doing so?

The primary advantage is that use of logarithms allows us to graph widely varying quantities on a single graph. The primary disadvantage is that dramatic, exponential growth is hidden, giving a false impression of orderly, linear growth.

8. What are the components of a URL? How do they allow a browser to find material on the World Wide Web?

There are four components of a URL: the protocol to be used in the communication with the server, the name of the server, the directory path to the file to be accessed, and the name of the file to be accessed. The server name allows the browser to find a particular physical machine, the directory path tells the browser how to navigate through all the files available on that server, and the file name selects a single file.

9. What's the advantage in a computer-aided design system to specifying the dimensions of the design in terms of an arbitrary unit of measure (l )?

The main advantage is the design does not incorporate any absolute physical dimensions. This means it is possible to scale a single design to be produced using different fabrication processes, simply by specifying a physical value for l and producing a new set of masks. Primarily, that means a design can take advantage of advances in process technology.

10. Explain how doping a silicon wafer with phosphorus increases the ability to carry an electric current.

The phosphorous atom has five electrons in its outermost shell. When it substitutes for a silicon atom in the semiconductor crystal lattice, four of those electrons occupy the covalent (chemical) bonds of the four outer electrons of silicon. The fifth electron is only loosely bound to the atom, and can easily jump to the conduction band to take part in an electric current.

11. (13 points) We said that digital systems use codes (e.g., ASCII) to convert things such as text characters (or other finite, discrete collections of things), to numeric values inside the computer. But we said that for continuous (not discrete) events, such as sound waves or images, we had to use a sampling technique to convert the event to an internal representation the computer can manipulate.

What factors determine the accuracy of internal representation for a continuous event? What are the tradeoffs with those factors? What kinds of practical issues prevent us from creating a code to represent something like sound waves or images?

The factors that determine the accuracy are the sampling rate (how often the continuous event is measured) and the precision (how many bits are used to hold the measurement when it's taken). The more samples of the analog source that are taken in a given time (the sampling frequency), the more accurately changes in the source can be represented. Think of how many times the pitch of a song changes in a minute. One sample per minute wouldn't give a very good approximation to the original, whereas 250,000 samples per minute would be much closer. For accurate representation of the analog source, samples should be taken twice as often as the fastest change in the source (this is called the Nyquist rate).

Similarly, the precision (number of possible values that can be used to represent each sample) affects quality for a given sample rate. The more possible values (more bits per sample), the more closely each sample will approximate the actual source. Two possible values doesn't give a very close approximation, but 65,000 values gives a quite good approximation.

The fundamental problem with trying to create a "sound code" to define sounds created by real world sources is that there are an infinite number of possible sound "values." This means we would need an infinite number of values in the code, which is impossible. Note that even if it were possible, such a code would only define the possible sample values (the precision), not the sample rate (the frequency).