Page 83 - The Art and Science of Analog Circuit Design

Signal Conditioning in Oscilloscopes and the Spirit of Invention

Get excited. Enjoy the thrilling early hours of a new idea. Stay up all
night, lose sleep, and neglect your responsibilities. Freely explore tan-
gents to your new idea. Digress fearlessly and entertain the absurd.
Invent in the morning or whenever you are most energetic. Save your
"real" work for when you are tired.
Master the fundamentals of your field. The most original and creative
engineers I have known have an astonishing command of undergraduate-
level engineering. Invention in technology almost always stems from the
novel application of elementary principles. Mastery of fundamentals al-
lows you to consider, discard, and develop numerous ideas quickly, accu-
rately, and fairly. I believe so much in this concept that I have begun
taking undergraduate classes over again and paying very careful attention.
Honestly evaluate the utility of your new idea at the right time: late
enough not to cut off explorations of alternatives and wild notions, but
early enough that your creativity doesn't go stale. In this stage you must
ask the hardest questions: "Is this new thing useful to anyone else? Ex-
actly where and how is it useful? Is it really a better solution or just a
clever configuration of parts?" Even if you discover that your creation
has no apparent utility, savor the fun you had exploring it and be thankful
that you don't have the very hard work of developing it.
Creativity is not a competitive process. It is sad that we engineers are
so inculcated with the competitive approach that we use it even privately.
You must suspend this internal competition because almost all of your
new ideas will fail. This is a fact, but it doesn't detract a bit from the fun
of inventing.
Now it's time to get on to a very old and interesting analog design
problem where there is still a great deal of room for invention.

Requirements for Signal Conditioning
in Oscilioscopes

Most of my tenure as an electrical engineer has been spent designing
analog subsystems of digital oscilloscopes. A digital oscilloscope is a
rather pure and wholesome microcosm of signal processing and measure-
ment, but at the signal inputs the instrument meets the inhospitable real
world. The input signal-conditioning electronics, sometimes referred
to as the "front-end" of the instrument, includes the attenuators, high-
impedance buffer, and pre-amplifier. Figure 7-1 depicts a typical front-
end and is annotated with some of the performance requirements.
The combination of requirements makes the design of an oscilloscope
front-end very difficult. The front-end of a 500MHz oscilloscope devel-
ops nearly IGHz of bandwidth and must have a very clean step response.
It operates at this bandwidth with a IMQ, input resistance! No significant
resonances are allowed out to 5GHz or so (where everything wants to
resonate). Because we must maintain high input resistance and low ca-
pacitance, transmission lines (the usual method of handling microwave