[Richard Crowley]

"Randy Yates" wrote ...
FYI, the reason for my question is that I am attempting to review the
design of the front-end of a heart rate monitoring device. This design
is based on the SSM2167 (e.g., from National). My thinking is that this
is the wrong part for the job since the sensor is fed to the preamp via
a run of cable perhaps 4 feet in length, and there could therefore be
lots of common-mode noise.

It does sound like a less than optimal design. OTOH, if they are only
looking for cardiac sounds, they could make the case that everything
of interest is *below* the power mains frequency (50/60Hz) and can
be simply low-pass filtered. So the kinds of things we do for wideband
audio are maybe not necessary for that special case?

Also, I believe the AGC action is, in this stage, not appropriate.

I agree from an audio perspective. But maybe the full waveform is not
what they are looking for? Maybe only the cardiac rate is desired here
and preserving the exact waveform is not important? It would presumably
be different if they were running a full EKG.

This appears to be an electret type of microphone in that DC power is
fed to the device, so the poster who discussed how to feed phantom power
and simultaneously receive the input differentially was right on the
money.

There are some clever (and simple/elegant) circuits for using commodity
electret elements, feeding power, AND providing a fully balanced cable
run from the capsule to the preamp, etc.

When I went in for a sleep study, they taped a (disposable) sensor on
my finger with a SMD red LED over my fingernail, and what appears
to be a small photo-resistive element on the opposite side of my finger.
They said that they read both cardiac rate AND blood oxygenation
from the sensor. I kept it as a souvenir. It is possible that there is also
an IR LED and/or sensor in there, but I haven't put it under the micro-
scope yet.