Michael wrote:
Chung wrote:
Michael wrote:
Can someone recommend a satisfactory test unit to measure 'peak'
frequency coming from an amplifier.
What exactly do you mean by "peak" frequency?
I am testing for ultrasonic frequencies (greater than 100khz) coming
from SACD's through my amplifier. I want to know what the highest
frequency that is reproduced.

That's not "peak frequency" you're trying to measure, it's the
bandwidth.

Normally, if you were measuring the trhoughput bandwidth, you'd
simply have a source, such as an oscillator, that can produce at
least the frequency range you expect, along with a meter, such as
a wide voltmeter, to measure it. Pick a frequencyb in the middle to
set your reference point, note the amplitude on the meter at that
point,
and then tune the oscillator to a high enough frequency that the meter
drops by some amount, say, 3 dB or 20 dB or whatever you decide
defines the band limit. That's your highest frequency.

However, because you're looking for the highest frequencies from
an arbitrary signal, it's more complicated. You have no control over
the content of the signal, so you need to use something capable
of doing broadband (100 kHz or greater) spectrum analysis. These
days that means, generally, FFT or similar analysis. TO get that sort
of capability in a stand-alone instrument is not at all cheap, and 100
kHz is and above is well above the range of pretty much all PC-based
measurement tools (even though some cards may advertise a 192 kHz
sample rate, that still only gives you a maximum high frequency limit
less than 96 kHz, and the real bandwidth may be substantially less.
I was once asked to evaluate a 96 kHz samplking system that was to
be used for gunshot transient analysis, and at 96 kHz, it still only
had
a 20 kHz bandwidth).

Now, are you looking for spurious signals above 100 kHz, such as
clock noise and such, or are you looking for real signals. If the
latter,
you're going to spend a huge amount of money and effort to achieve
a null result. If the former, you're going tro spend a huge amount of
money and effort to possibly achieve a result that neither you, the
manufacturer of the equipment, or the whole SACD world in general
really don't want to know.