[William Sommerwerck]

As friendly professor Peter Schickele once said... "Truth is truth. You can't
have opinions about truth."

I'm accused of having obsessions no one else cares about. (I'm willing to
confess to an excessive fondness for gerbils.) But I do care about the truth,
and make no apologies for it.

Part of the reason for this obsession is that, as an under-employed technical
writer and editor, I see technically incompetent people with little or no
ability to write or edit /being paid/ to do so. * Is it any wonder I blow my
stack over errors that a competent editor would have caught?

Anyhow...

Here are the excerpts from Frenzel's "Electronics Explained" that disturbed
me. I believe in encouraging people to think for themselves, so instead of
explaining what's wrong, I invite people to say what /they/ think is wrong
(including "important stuff" that's been left out). I might prompt a little,
but I prefer that people figure out these things on their own.

The purpose of this exercise is not to start arguments, but to get people
thinking about what they understand or don't understand.

(Frenzel is a master of tautological writing. He says the same thing over and
over in a slightly different way each time, rather than expressing it once,
simply. He would have had a great career as a paid-by-the-word pulp-fiction
writer.)

----------------------------------------------
p89

Recall that there are two basic types of electronic signals -- analog and
digital. A digital signal is one that varies in discrete steps. Unlike an
analog signal, which varies continuously, a digital signal has two levels or
states. The signal switches or changes abruptly from one state to the other.

Figure 5.1 shows a DC digital signal that switches between two known levels
such as zero volts or close to it (0.1 volts) or 0V and +3.3V. The positive
voltage can be anything between about 1 volt and 12 volts with 3.3 and 5 being
the most common.

[Those who can get p89 to appear might ask themselves whether the figure
actually shows a digital signal. How do you know it's digital and not analog?
Can a pulsed signal be analog? Can a sine wave be digital?]

Digital signals with two discrete levels are also referred to as binary
signals. Binary means two -- two states or two discrete levels of voltage.

Humans use the decimal number system that represents quantities with digits 0
through 9. However, digital equipment and computers do not. Internally,
digital equipment processes binary data.
----------------------------------------------
p118ff

Analog-to-Digital Conversion

The process of converting an analog signal into a digital one is called
analog-to-digital conversion, and is performed by an analog-to-digital
converter (ADC). The process, also referred to as sampling, is illustrated in
Figure 5-26. The ADC looks at the analog input and periodically takes a sample
of the voltage at that instant, captures it, then converts it into a
proportional binary number. We say that we are digitizing the signal. The
sample points are shown by the dots on the analog curve. The binary value of
the sample is shown to the right of the curve. The conversion process actually
results in a sequence of binary numbers that represent the analog waveform.
These values are usually stored in a RAM or transmitted to other circuits as
shown in Figure 5-27. Note the symbol for an ADC.

Digital-to-Analog Conversion

To recover the original signal, we put the data sequence previously captured
by the ADC into a digital-to-analog converter (DAC) (see Figure 5.28). The
output is a version of the analog signal. The DAC output is not a perfect
reproduction, but just an approximation. This is shown in more detail in
Figure 5.29. Each binary input results in a constant voltage output from the
DAC during the sample period. The result is a stepped approximation to the
original signal. The rate at which the binary data is [sic] sent to the DAC
must be the same as the sampling interval to recover the original frequency
information in the signal.

Resolution and Sampling Interval

The key to good data conversion is to use greater resolution and faster
sampling rates. Resolution refers t the number of bits used in the data
conversion. In Figure 5.27, only 4 bits are used, so the resolution is poor.
The voltage range is only divided into 16 intervals, meaning that amplitude
variations at [sic] less than 0.625 volt are missed. This problem can be
corrected by using more bits. ADCs are available in many bit sizes. The most
common are 8, 10,, and 12 bits, but 14 and 16 bits are available. Some methods
of ADC produce resolutions of 20 to 26 bits. The result is a finer conversion
of amplitude detail. As an example, if the 0- to 10-volt range in Figure 5.27
was [sic] a 12-bit ADC, the individual smallest voltage increment that can be
detected is 10/2^12 = 10/4096 = 2.44mV instead of the 0.625 volt[s] in the
figure.

Another critical specification is sampling rate. To retain all the frequency
detail in a signal, the sampling rate must be at least twice the highest
frequency in the signal. This called the Nyquist criterion. For example, when
digitizing music with a frequency range of 20 Hz to 20 kHz, the sampling rate
must be at least double the 20-kHz frequency. In most systems, a rate of 44.1
kHz or 48 kHz is used.

----------------------------------------------

For those who haven't fallen asleep (or died), here's part of my review of
another bad book, "Signals and Systems Using MATLAB".

I wrote "...on p9, he confuses "analog" with "continuous time" -- which is NOT
correct. One can have analog data that are discrete-time."

One reader responded "...it is impossible to have an analog signal that is
discrete time. An analog signal can be represented by a discrete time signal
very well, but an analog signal IS continuous, where a discrete time signal is
defined as being discontinuous. Since it is impossible to have a continuous
signal be equal to a discontinuous signal, your argument is invalid."

This confusion of time and amplitude remains common. (Note also the confusion
between "signal" and "data".) You might want to chew on it a bit. (ar-ar)


Thanks for your interest. Have at it!


* Several years ago I complained to a Wiley editor about a miserable book on
the history of radio. He was a very nice person -- but admitted he had almost
no technical knowledge. So what made him think he could edit technical books?
He'd applied for and taken a job he was unqualified for. I think I have a
right to be outraged.

"'We already know the answers -- we just haven't asked the right questions."
-- Edwin Land