■ The analog and digital boards. The digital
board has the RCM4120 inserted.
tle “headroom,” but not a whole lot. The
resistor values I show on the schematic are
based on the following two examples.
frequency sine wave input. This procedure allows you to
adjust the input signal amplitude and will store both the
processed and unprocessed amplitudes. I suggest 0 dBm
or 1 VRMS — see the discussion that follows. The
calibration function displays the value of the input signal
you are using for calibration so that you can monitor it
while adjusting it. This yields the best calibration for the
Note: The “processed” amplitude is the output value
of the FFT algorithm.
1) Suppose you want to monitor the
output of a 100W power amplifier that is
driving an eight ohm speaker. The output
voltage at full power is therefore: V =
sqrt(100* 8) = 28. 3 VRMS. In order to
measure this signal, you need to divide it
by 28. A set of “reasonable” resistors is
then: R1 = 2.7K and R2 = 100 ohms. I
have designed the circuit board so that R2
is an SMD and R1 is a thru-hole part so it
can be 1/2 watt. By not activating K1, the
attenuator will be used giving the overall
circuit a “gain” of 1/28. There is a
disadvantage to this circuit — it gives the audio circuit an
input resistance of 2.8K. If this is not acceptable, you can
increase the resistor values — but not too much or the
ratio will be affected by R5 and R6.
2) Now, suppose you want to monitor a signal that is 100
mVRMS. All you have to do is multiply it by 10. Using the
values shown in the schematic: With IC3 turned off, the
gain is (1K+9K)/1K that is a voltage gain of 10 ( 20 db);
with IC3 turned on, R11 is shorted out that would yield
a gain of 1 (0 db).
The maximum signal presented to the ADC must be
less than 5V peak-to-peak, about 1.7 VRMS. I suggest that
you use 0.7 VRMS as a standard because it is equivalent
to 2.0V peak-to-peak, as well as 0 dBm into 500Ω and
makes a nice reference value. You could also use 1 VRMS
since this would be 2.828V peak-to-peak. Both leave a lit-
The program has a feature that causes it to read
calibration values from its Flash memory when it first
starts. I have set up the program so that if it detects that
no calibration values have been previously stored, it will
use default values that should be very close — assuming
the circuit is built with the
values in the schematics.
The circuit has enough gain
for a typical microphone if you
use the suggested transformer.
Most mics have an output
voltage of about - 60 dbm (0.7
mV into 500Ω). The transformer
has a gain of about 12 db, so
with the available gain of 50 db
that should be adequate for
most mics. Note that there is
also voltage available for mics
that need phantom power. This
should be applied to the primary of the transformer as
shown in Figure 7.
One of the requirements of
the FFT is that you must sample
the input signal at least twice as
■ FIGURE 6.
The LCD and