<<But the problem I have is that the hole in the carb throat
is about the size of the scoop entrance. doesn't the air have to
speed back up again to get through the carb?  So the pressure falls
and we are back where we started.>>

Just in case Fred says it different I'll take a stab at this.  Very
excellent dissertation by Fred, by the way.
Yes, the air will have to speed up again to go into the throttle body, so in
theory the best thing to do would be to make a gradual transition from the
inlet (sized according to the airspeed of the aircraft) and the throttle
body (sized by the manufacturer so you have no control over that).  First
problem is the air filter and you need a large area filter with the air
velocity through the element uniform across the total area.  This forces you
to slow the airflow down by gradually increasing the area.  Another
consideration is the flow losses in the plumbing and these losses will be
less with a low velocity air flow.  Speeding up the airflow is much easier
to do efficiently than slowing it down - a good bell-shaped inlet will do a
very good job in a short distance.  The reason is that the pressure drops as
the velocity increases, inherently maintaining attached flow to the surface.
To decelerate the air the pressure increases as it travels and hence will
look for any opportunity to create eddies and locally reversed flow,
destroying the efficiency of the process.  Fred can probably explain it
better.

Incidentally, has anyone used the cowl just in front of the windshield for
an air inlet?  This would seem to be a good source of low velocity, high
pressure air.  The only problem I can see is that in high angle of attack
situations the pressure might not be so high.  Ram air capture for an inlet
system would still be effective at high speed, however.  I realize that most
of the engines we are working with have the engine inlet facing the front so
this might not be convenient.

Gary Casey
ES