Gary:

> 1.  Pressure recovery (converting the high velocity free-stream to low
> velocity higher pressure air with minimum energy loss) is critical and it
> looks like the Lancair cowl does this ahead of the entrance as the cowl
> openings are larger than they could be.  This is probably a good, efficient
> approach.  However, once the air enters the cowl, no attention at all was
> given to pressure recovery.  Should there be?  I think so.  The idea is to
> get maximum pressure above the engine.

Max pressure above the engine is a help, but not the primary concern! Any
increase in the DeltaP from top to bottom is helpful, but as it turns out,
not the major factor.  The cowl opening on most airplanes, as it turns out,
are larger than they should be!  More air comes OUT of the front of the
cowls TOWARD the prop than you can imagine.  There's more than enough air
going in.

> 2.  Eliminate all wasted air flow across the engine, primarily by
> eliminating any leakage paths.  This is the most obvious point and one on
> which everyone seems to agree.

This is also helpful, but believe it or not, if the engine is well baffled
by design, you can put a 1" hole in the back plate and not significantly
affect cooling.  Yes, I know that seems counter-intuitive and it is, but
it's a true statement--by data collection.   It is much more important to
get the air that is available to go to the right places.  THAT is the issue.

> 3.  Control of air flow patterns inside the cowl is important, but here
> again it looks like Lancair (and most other airframe builders) have given no
> thought at all to this.  Their assumption must have been that once the air
> is inside the cowl it will just magically get to where it needs to go.  The
> air at this point is at a relatively low velocity so this as some validity,
> but I'm not convinced.

BINGO.  You're on the right road.  Until you put 36 probes on an engine (six
CHT probes per cylinder) you can't tell where that air goes in a dynamic
situation.  It is CONFOUNDING.

> 4.  Exactly how the air gets around the cylinders has to be critical and
> this seems to be done mostly by the engine-manufacturer-provided baffles,
> but I'm not convinced these are optimized either.  The most important
> airflow around the engine has to be across the fins between the valves and
> I've never seen any directed effort at maximizing this flow path.  My
> judgment is that the temperature around the cylinder head is nowhere near
> constant, but how does one improve upon this?

BINGO, again.  Explained above.  You're properly identifying the problems
conceptually.

> 5.  And finally, the airflow in the lower cowl has to be managed.  In most
> systems I see zero effort in this regard.  Even Lancair, although they went
> to the trouble to provide a curved, converging flow path at the top of the
> cowl outlets, put sharp edges at the sides, which accounts for more than
> half the entrance periphery.  As I see it, the objective of the first four
> points above is to cool the engine while allowing the HIGHEST lower cowl
> pressure to be maintained.  This pressure can then used as the energy source
> to accelerate the air back up to somewhere near free-stream velocity, but it
> can only do that if the exit nozzles are efficient.
>
> I would appreciate any discussion on these points.

Gary, you are on the right path and have properly identified the set of
problems.  The issue is identifying the solutions which requires tufting an
engine, putting a video camera under the cowl and going flying while
watching the airflow patterns; then hooking up 36 probe CHTs
circumferentially and MEASURING the cooling around each cylinder.  Only then
will you know where to add airflow and where to reduce airflow.  This is no
trivial endeavor.

Your astute observations have you on the right track.  Now, all it takes is
a ton of time and money to measure the effects you properly suspect are at
work (you will have to hand make the thermocouples for the circumferential
measurements and create an instrument which will record the data).

Now, if you think that's difficult, you oughta try to explain why all of the
weird things you have to do to make it work right, work right... and explain
that to someone who is unfamiliar with the issues--oh, yeah, and try to do
that over a keyboard in less than three hours!

Walter Atkinson