The second point is that PS noise
can cause oscillation. When digital signals switch, they can cause the power
supply voltage to drop. This can change
the internal bias of the amplifiers which
may cause them to draw more power
and/or shift the output voltage.
Since op-amps use negative feedback, any change in the output will be
“corrected” by the amplifier, which
causes an output shift and can cause
additional power supply variations.
Unfortunately, the op-amp cannot
respond instantaneously, so this correction is delayed to some degree. At some
point, if the correction if large enough
and of the proper polarity, positive feedback will occur which can cause ringing
or full-blown oscillation. To reduce PS
noise, three isolated voltages are used.
The + 9 volts are used for the electret microphones. This is done for two
reasons. The first is to isolate them
from any other power supply. Since
their output is directly related to the
positive supply, any noise here will be
directly coupled into the output. Note
that this + 9 volts is well bypassed and
isolated from the other two supplies.
The second reason is that the + 9 volt
supply will provide a larger signal than
using a + 5 volt supply.
There are two + 5 volt supplies:
one for the analog circuit and the other
for the digital circuit. Again, this is to
reduce noise coupling. A very important point is that there is one and only
one connection between the analog
ground (AGND) and the digital ground
(DGND). More than one ground interconnection can cause ground loops
between the analog and digital circuits.
Ground loops couple noise and allow
differential currents to flow. Not good.
The single ground connection should
be at the power supplies.
I used low-power, three-terminal
regulators. You can use the high-power
version if it’s more convenient. I also
added LEDs as a visual indicator that
the power supplies were working. I
find this useful, but it’s not necessary.
Since I had high-intensity LEDs on
hand, I used 5.1K current limiting resistors for R9 and R10. If you have less
bright LEDs, you may want to change
the resistors to 1K or even 510 ohms.
I chose the LMC6482 op-amp
because I had it on hand, as well. The
CMRR (Common Mode Rejection
Ratio) and PSRR (Power Supply
Rejection Ratio) are good for the
LMC6482 (at 82 dB). Most single supply amplifiers should work reasonably
well. If you have noise problems with
your amp, these are the characteristics
to compare. The input impedance of
your amplifier should be in the
megohms, at least (the LMC6482 has
> 10 teraohm input impedance).
Both amplifiers are the same.
Keep the leads short and close to the
amplifier. If you stand the gain
resistors — R7 and R8, R4, and R5 —
on end like I did (see Photo 2), place
the body of the resistor at the input
pin of the op-amp. This eliminates the
small amount of noise that might
come in from the 1/2 inch of exposed
resistor lead. (Every little bit helps!)
Ordinary point-to-point wiring was
used, but care was taken to keep the
lead dress as short as possible.
The variable resistors (R2, R6) are
used to set the sensitivity of the
amplifiers. The full gain of 1,000 may
not be needed. The input capacitor
(C1, C2) is used to block the DC that
is present at the electret microphone.
The microphone resistor (R1, R5) is
used to supply power and limit the
current to the microphone.
A number of different microphones were tried with various levels
of success. The best all-around
performance came from an ordinary
electret microphone. I used ones I
already had, which are apparently now
obsolete. They are about 0.27” high
and 0.39” in diameter. This appears to
be similar to Jameco 320178CK, but
anything with good low frequency
response (about 20 Hz) should work
fine. I attached the microphones to a
one inch square piece of 0.25” thick
Pulse Speed Timer
■ PHOTO 2. The circuit board layout is
straightforward. Point-to-point wiring
was used but care was taken to keep
leads short and to separate the analog
section from the digital section.
plexiglas (see Photo 3). I drilled a hole
slightly larger than the diameter of the
microphone, almost all the way through
the plastic so that the microphone
would have a solid backing. I drilled a
smaller hole the rest of the way through
for the wires.
The microphone was set in place
with silicone adhesive. Note that
about 0.2” of the microphone
protrudes from the face of the plastic.
This projection is important because it
presses the microphone into the skin.
You must use shielded cable for the
wires and keep the length to a maximum of 24”. Also be sure to observe
the polarity of the microphone leads.
The center conductor of the shielded
cable goes to the positive terminal.
■ PHOTO 3. The microphones were
cemented into plexiglas holders about
an inch square with silicone adhesive.
The protrusion of the microphones
provides better pick-up of pulse and
April 2007 41