In a message dated 10/5/2002 8:16:10 AM Pacific Daylight Time,
glcasey@adelphia.net writes:

<< A properly designed inlet doesn't reduce efficiency by much if it
 is too large as the pressure in front of it will just push air around the
 sides and if they are generously radiused there will be very little
 associated drag.  The Lancair cowl is a good example of this.  One of the
 fastest piston engined planes built at the time was the Corsair and look at
 the huge cooling inlet.  The cowling was radiused to provide for a smooth
 flow around the outside and that is what mattered.  The prop hub was then in
 a relatively stagnant area and thus didn't need a spinner. >>

Whoaaaaa, let's be VERY careful of the assumption that stagnation does not
create drag.

First the theoretical side.  Take a flat plate of plywood and force it to fly
normal to the air.  Half of the drag can be quantified by separation of the
airflow on the back side of the plywood.  The other half (the so-called
"pressure drag") is caused by the stagnation of the air forward of the
plywood.  In this area, air at a given p and v has to give up that v and
create p.  

On to practical drag reduction.  On an aircraft, oil streak and tuft testing
can reveal areas of stagnation and separation.  Properly designed aerodynamic
improvements can clean these areas up and enable attached air flow aft and
decrease stagnation forward.  But how much drag?  How do we quantify the
improvement and how does that relate to the theory?

On three radically different configurations we have quantified this.  In
finding the separation area and curing it, we were able to account to less
than 0.5% of the total drag by simply looking forward from the tail at areas
of separation (picture this as the reverse perspective of "frontal area") and
calculating the square footage of flat plate drag equivalent as half of the
square feet of separated cross sectional area;  then subtracting the wetted
area drag added by the fairings considering the boundary layer to be fully
turbulent at the test Rn.  Worked perfectly.

On to stagnation.  Other areas exhibited stagnated flow and in some cases
reversed flow.  Theses areas were quantified as equivalent flat plate drag
area by looking at their cross section from the front and dividing by 2.  
Fairings were added and the wetted area drag of the fairings subtracted from
the calculated drag improvement.  Partially laminar boundary layers existed
in some areas and we had to also add in some extra drag equivalent on the
baseline to account for reversed flow (even worse than stagnation, this is
when the air actually turns 180 degrees and goes forward on the aircraft).  
After all this, the drag reduction from reduction of stagnation was still
greater than we could directly account for.  What else happened?

Turns out that since stagnation disrupts airflow forward on the airframe, it
is capable of really messing things up downstream as well.  Some downstream
separation, vortex flows, and disruptions of spanwise flow creating
discontinuities in the spanwise lift distribution (higher induced drag) that
were all improved by reducing stagnation, accounted for the extra drag
reduction.

Be very wary of oversize inlets.  On two aircraft that I am now working on,
they create other problems that have increased the total airframe drag by
over 10%.  And this is with radii and fillets that anyone would consider
generous.  As for big cowlings and little spinners making no difference, I
would point to the large number of Unlimited air racers that have achieved
large drag reductions through the use of highly oversized spinners vs. their
stock configurations.

Eric Ahlstrom