Hi Lee,
Rise time is
actually a proxy for bandwidth, and I think this is
the real parameter we should be measuring....
You're right about bandwidth being what we really want to know. The
problem is in measuring it. It may seem obvious that we just look at it
with a spectrum analyzer. If you think about exactly how a spectrum
analyzer works you'll discover there are problems.
A spectrum analyzer is basically a receiver with a sweeping Local
Oscillator, a narrow IF filter and an AM detector. The width of the IF
filter determines the frequency resolution of the measurement we will
get from the spectrum analyzer. When we use a narrow filter for high
resolution the LO sweeps slowly. The spectrum analyzer's AM detector
produces a DC level corresponding to the signal level passing through
the narrow IF filter, and the input frequency being measured depends on
the frequency that the LO is at, at any particular time.
In order to measure the bandwidth occupied by a pulsed signal using a
spectrum analyzer we have to integrate over a lot of spectrum analyzer
LO sweeps. We could just send a string of dits and let it integrate for
a minute or two, and probably get a good indication of the bandwidth
occupied by a string of dits from that radio. Then what about rigs that
have a different envelope shape on the first dit than on subsequent
dits? You could use a 555 timer circuit to key the rig with delays so
that every dit is a "first dit".
You can also manually do the spectrum analysis using a receiver with a
narrow filter and an accurate S meter. The frequency resolution is
limited by IF filters available. You might get narrower resolution using
an audio filter and measuring the audio power output as you move the
receiver to different frequencies. You would need the AGC to be OFF.
The signal level would have to be low enough that no part of your system
is going into compression or clipping, and yet high enough that the CW
signal sidebands are out of the noise floor of the measuring system.
There are a lot of things that can go wrong to mess up this measurement.
Getting an accurate measurement of occupied bandwidth of a pulsed signal
using a spectrum analyzer or a receiver is not as trivial as it may
seem. That is one of the reasons for measuring the rise time and doing
the math instead.
DE N6KB
|