to: bob@bobthecomputerguy.com (Robert Harris)
subject: push me, shove you...

 Back to inducer/ejector stuff!

 Assume we have a single cylinder engine operating on a normal four stroke
cycle.  Airflow through the intake port, particularly at low RPM, is
intermittent in highly discrete pulses.  The air column has to start and
stop each time the valve opens, though there's the air spring effect.

 If we could keep the air column moving more of the time, it wouldn't
have to accelerate from zero every time the valve opened.  This would
reduce pumping losses.  You couldn't really call it a supercharge effect,
though pressure ratio across the intake valve would be greater.

 Assuming a relatively low restriction exhaust system, the exhaust port
spends more time under vacuum than under pressure.  The exhaust event
partially overlaps the intake event, but there's still a useful amount of
vacuum there.

 Run a passage tangentially connecting the intake and exhaust ports.
Near the inlet side, place a reed valve.  You could steal the reed
valve from a Mazda or Chrysler exhaust aspirator system; they're already
designed to work at exhaust temperature.  

                    |
             |      |      |
             |      V      |
             |             |\
             |             | +  reed valve
             |             | |\
             |             |\| \
             |             | \  \
             |    *****    |  \  \  
             |  *       *  |   \  \ 
             | *         * |    \  \
              *   INTAKE  *      \  \
               *         *        |  |
                *       *         |  |
                  *****           |  |
                  *****           |  |  bypass passage
                *       *         |  |
               *         *        |  |
              *  EXHAUST  *       |  |
             | *         * |      |  |
             |  *       *  |      |  |
             |    *****    |     /  /
             |             |    /  /
             |             |   /  /
             |      |      |  /  /
             |      |      | /  /
             |      |      |/  /
            |       |       ------|
            |       |         |   |
            |       |         |   |
            |       V         |   |
            |                 V   |
            |                     |

           stepped exhaust port (optional)

 Power stroke (down):  piston is forced down under pressure
 Exhaust stroke (up):  most pressure is bled off near BDC; pressure in the
                       exhaust pipe is below atmospheric during much of the
                       exhaust stroke
                       ---when exhaust pressure drops below inlet pressure,
                          the reed valve opens---
                       ---intake air column is accelerated and is pulled
                          through the connecting passage to the exhaust
                          port---
 Intake stroke(down):  intake valve opens with inlet charge *already*
                       moving; no losses involved with accelerating the
                       intake charge 
                       ---the exhaust pressure could still be below the
                          intake pressure, depending on exhaust system
                          restriction.  Reed stiffness could be adjusted
                          to reduce air inadvertently bypassed around to
                          the exhaust side---

 We're not trying to *add* anything, as in creating boost - we're just
reducing some of the flow losses in the intake tract.  The effect would
be most apparent at high load and low RPM; when heavily throttled or at
high RPM it would probably do very little.  The net effect would be a
boost in near-WOT torque at low RPM.