On Apr 10, 4:21 am, "ilam" wrote:
I am in a process of understanding the importance of S-Parameter at
high frequencies, please answer my following questions:

1. Why Measuring a wave (voltage or current) is more easier than
volage and current? ( Why measuring S-Parameter is more easier than
other parameters?)

2. How the voltage and current waves are measured?

Measuring voltage and current are difficult at high frequncies, that
is why S-Parameter is more useful at high frequncies. Is the same
thing is true for simulation?

Regards,
Ilam

Hewlett-Packard were one of the early proponents of using S-parameters
for calculations as well as for measurements. HP had a vested
interest, in that they were one of the first companies to make
accurate instruments for measuring high frequencies, up into the
microwaves. Part of that effort was customer education, and there are
some good applications notes from HP about S-parameters, and about RF
measurements in general. You should be able to find many of them on
the internet, with a bit of searching. Some of them deal with
accuracy, and in them you should find answers to your questions.
You'll find a great deal more than just those answers.

Measuring voltage or current at audio frequencies is easy enough. But
how do you measure a current at 10GHz without disturbing and changing
it? How do you accurately measure a voltage at that frequency, when
the tiniest of capacitances will change the reading? But if instead
you want to measure the standing wave ratio on an accurately-made
piece of transmission line, that can be done relatively easily, and
especially you can determine when the standing wave ratio goes to
zero: when there is no variation in voltage along the length of the
line. Then too, you can build a bridge circuit to enable measurements
at a particular impedance. 50 ohms is convenient, though it's
somewhat arbitrary. If you pick too high or too low an impedance, you
run into the same difficulties as if you try to measure a voltage or a
current, but at impedances between perhaps 20 ohms and 100 ohms, in a
coaxial-line environment, you can build loads--resistors--that
accurately terminate the line, and you can build a Wheatstone bridge
circuit, essentially, that will tell you if you have a load that's the
impedance of the bridge arms. The bridge is the most sensitive at
balance, when all the arms are the same impedance.

Also note that while it's very difficult to measure current or voltage
accurately at high frequencies, if you can build a resistive load,
then measuring power accurately is possible. When you are first
starting, you can measure power by determining the temperature rise of
a load resistor, which you made by trimming things carefully so that
there were no standing waves on the precision line feeding that
resistor. You can calibrate that quite accurately by feeding it DC
power, to determine the temperature rise per watt. If you learn to
make very tiny resistors and measure very smal temperature
differences, you can make a thermal bridge that can detect power at a
low level accurately.

Let us know if you have trouble finding the HP applications
notes...look for S-parameter notes, and look for notes on making
accurate measurements with (vector) network analyzers. Also, do a
search for information on the history of RF and microwave
measurements.

For simulation, it is simply convenient to work in the same domain as
you make measurements. As you note, there are other parameter sets
that give you equivalent information, and with modern computers, there
is effectively no loss of accuracy if you deal in any other set of
parameters used for 2-port (or N-port) networks. But do note, please,
that these are linear parameters, and they won't give you information
about nonlinearities, either ones you wanted as in frequency mixers or
ones you didn't want, that cause unwanted distortion.

Cheers,
Tom