"Rich Sherman" wrote

Can you guys help me to create a test bench setup in order to
perform
measurements for
amplifier frequency response (-3dB bandwidth).

In other words what test equipment do I need to test and generate
a
Amplitude vs. Frequency
plot of a given amplifier. Yet it sounds expensive to me.

Does one 'sweep' an amplifier from 20Hz to 20KHz to obtain an
amplitude
response curve with an Audio Sweep Generator?

Or does one test the amplifier at selected frequencies and then
generate a table of data points of Frequency vs. Amplitude?

Maybe a PC solution exists for such an interface on the market
today.

I would like to plot the frequency response of my home-brew amps
and also rebuilt classics.

If it's a manual thing then some ideas on how to go about
calibrating the
input levels
and measuring the gain and output levels would be helpful.

My test gear at the moment is an EICO Audio Sine Wave Signal
Generator, an
40 MHz
Dual Trace Oscilloscope and also a Stabilock Model 4032 RF
Communications
Analyzer.
Plus the usual Volt-Ohm meters, power test resistors and small
coaxial BNC
Cables, etc.

You can use the dual trace scope for a quick and easy plot. Use one
channel to monitor the generator, and one to monitor the output of
the amp. Set the generator level at 1V (or whatever) at 1k and the
scope to show this at a convenient near-full-screen vertical scale.
Set other channel to appear exactly the same, so it looks like one
trace.

Then scan through the frequency range with the generator, noting the
proportional difference in amplitude between the traces. It will be
obvious where you need to note many points, and where just a few
will be enough. You should need to go well beyond the audio range to
find the top end limit.

You can also note the phase shift, but soon you will notice that
this can be entirely predicted from the amplitude variation so you
won't need to bother.

This method takes into account variations in generator output, and
it is simple to follow what you are doing. It assumes some frequency
standard...either a frequency-accurate generator or scope, or a
separate counter.

Alternatively, you can simply use the generator and a voltmeter. A
true rms meter is preferable but not really necessary as long as
there isn't lots of variation in distortion level.

Of course you can use a frequency scan output if you have one from
the generator, but not so easy to measure.

cheers, Ian