Jon Nicoll wrote:

> I'm getting interested in the idea of making an system to measure the
> volume of air breathed by a subject.

  Three methods in common use:

 1} Propeller drives optical chopper.  Pony spirometer and others.

 2} Pressure differential measured across restriction.  Restriction
consists of multiple planes of mesh for laminar flow.  Welch-Allyn use
this with, I think, a 68HC05.

 3} Hot wire, either single or double (records direction and self-
compensates).  Various manufacturers.

> I thought I'd ask for previous experience on this august forum...

  Whoa!  It's December ;-)

> I'm somewhat familiar with ways of measuring air speed - am I right in
> guessing that the sensible way to measure breathed air would be to
> measure the air speed across a known area and integrate etc?

  Integrate flow, yes.

> - relatively low air speed. Also you need to keep the pressure
> 'normal', ie I don't want to have to breathe through a narrow tube

  Sensors have to be accurately scaled for the job.

> - keeping the flow of air uniform across any such cross-section, so
> that the speed measurement is a true indication of the volume passing.
> Methods of keeping flow laminar? or just calibrate it for inacuracies
> somehow...

  (Multiple) mesh planes.

> - discriminating between breathing in and breathing out. Ideally I'd
> like to be able to do this automatically (two sensors, detect which
> changes first to give the direction?)

  Only the single hot wire gives no direction indication.

> - otherwise just a valve to make breathed out air go an alternate
> route.

  Forget it; too complex.

> - doing the integrating etc.

  Use a microprocessor.
--
  Cheers,
        Paul B.