<<This exposes a common error in entrance diffuser design. While it is
certainly desirable to have a radiused transition to prevent "tripping" the
flow, continuing an expanding radius beyond the optimum angle throughout the
length of the diffuser is counter productive.>>

Rob makes a very good point and I've seen this rule "violated" a number of
times.  As the air slows and rises in pressure remember that the pressure
forces are trying to make the air go back upstream.  This force is countered
by the momentum of the air, but that is getting less and less as the
velocity reduces.  Eventually the air flow reverses at the boundary layer
and travels upstream, "tripping" the flow and making it separate from the
wall.  The trick is to expand the area until just BEFORE this happens and
then ending the diffuser in a sharp edge.  This inhibits the air from
crawling upstream at the boundary layer.  A diffuser that shaped like a
trumpet may look pretty, but it will almost certainly be less efficient than
one shaped like a cone with a sharp trailing edge.  This works for boat
propellers, venturis, ailerons, race cars and a host of other gizmos.

Gary Casey