From: Fredmoreno@aol.com
Message-ID: <5ad06e3e.245653a4@aol.com>
Date: Mon, 26 Apr 1999 19:41:24 EDT
Subject: Cowl Flaps and cooling
To: lancair.list@olsusa.com
Reply-To: Fredmoreno@aol.com
Mime-Version: 1.0

          <<<<<<<<<<<<<<<<--->>>>>>>>>>>>>>>>
          <<  Lancair Builders' Mail List  >>
          <<<<<<<<<<<<<<<<--->>>>>>>>>>>>>>>> 
>>
	I read Earl Schroeder's note on cowl flaps and engine cooling, a 
favorite topic of mine and an area of some modest expertise.  You are correct 
in noting that many Russian radial engines and the radial engine on the 
Polish Wilga (perhaps others) use a "circular venetian blind" valve to meter 
the flow at the engine inlet.  It is effective, but not appropriate for 
application in speedy Lancairs.  The radial installations are high drag on 
high drag airframes and speed was not a major consideration (particularly on 
the Wilga).  However, operation in Siberia in winter is a major 
consideration, and hence the need for something to keep the engine warm when 
it is a bit chilly outside. 

	On most high performance experimental planes, designers optimize for 
speed and compromise everything else, including cooling to minimize cooling 
drag.  Few cool well in a sustained high power climb, frequently in spite of 
what owners might say. Others argue that "It climbs as well fast as slow so I 
climb fast with no penalty on climb rate" although this is a violation of the 
laws of physics. (Going slower makes less drag which frees up more power for 
climb, as long as you are above best L/G speed.) These comments are merely 
excuses for poor cooling.  Unfortunately there is lots of BS and little data 
on this topic. 

	I flew to the Lancair factory party some years ago alongside a 
beautiful 235, and he had to step climb to keep heads cool while we motored 
alongside in a TurboSkylane RG with our cowl flaps closed. "Cools fine in 
cruise" is simply not good enough.  It took us a long time to get him up to 
his optimum cruise altitude on a warm summer day.  This was a stock airplane, 
well built, so it suggests again that the cooling design is inadequate for 
hot climb, and optimized for cruise. 

	Experimental aircraft designers try to reduce drag by making the 
inlets especially small as this is the conventional wisdom concerning drag 
reduction.  It is wrong.  This has been verified in extensive NASA testing in 
the 1980s with a 6 cylinder Lycoming on a Piper Aztec.  It is better to have 
the inlets adequately sized or too large, and to control flow at the exits 
with well formed nozzles that squirt the air backward to recover some of the 
momentum lost when the flow was slowed to pass over the engine.  As long as 
the inlets have a radius on the lip to permit excess flow to spill over 
without inducing flow separation, the drag from a large inlet is negligible 
compared to the potential losses otherwise, or the penalty of a long step 
climb. 

	The question is then how to control the flow.  Cowl flaps that 
deflect the free stream flow outward (away from the fuselage) create suction 
when open helping augment the ram pressure obtained at the inlets when speed 
is low.  When closed, they should lie flush to the fuselage, and make a nice 
nozzle that smoothly accelerates the hot air in the cowl backward.  This 
gives adequate cooling flow over the broadest range of flight conditions.  

	If it is difficult or impossible to put cowl flaps at the 
conventional location on the LC2 planes, consider putting them in the side of 
the cowl opening outward, and closing fully for cruise, dumping the flow out 
the normal outlets.  The key is whether or not you can get them actuated 
easily with some push-pull cables.

	But before adding such complexity, I would recommend the following:
1) Enlarge the inlets to minimize the friction and pressure recovery losses 
that arise when too much air tries to rush through too small a hole.
2) Check your engine sealing to be sure that all the air going in the inlets 
goes where it should, and does not leak around the baffles, through holes, 
etc. A frequently leaky region is below and behind the spinner where 
depending on configuration air can come in the inlets, reverse behind the 
spinner, and then sneak downward around the crank flange.  I also believe 
that the rubber flap baffles we frequently use are ineffective, particularly 
with 4 cylinder engines.  The engines vibrate so much out by the heads that I 
suspect the flaps bounce on and off their seating surfaces letting some air 
through.  The heads wiggle up and down perhaps 0.050 every revolution, and I 
don't think the rubber is staying put under these conditions.  Only a theory, 
but look at the wear and polish at the mating surfaces in some installations, 
and you will begin to wonder also.   In any event, NASA found that in a brand 
new Aztec, supposedly well baffled, the flow leakage around the engine was 
equal to 50% of the cooling air going where it was supposed to be going.  In 
other words, the cowl had 150% of the required flow for cooling going through 
it, and a THIRD was leakage.  Obviously even "good" installations leak a 
BUNCH.  Start here looking for improvements.  I believe that a lot of the 
modifications to increase exit area are merely allowing the leakage flow to 
escape.  Keep it tight, then consider cutting more exit area in the cowl.
3) Consider making "doghouse" cover over the entire top of the engine that 
seals well to the inlets, and which connects to the inlets with smooth 
transitions (no SCAT hose).  This eliminates all rubber baffling.  See a new 
Piper inlet to see what I mean about good inlet design.  Perhaps we can 
convince Brent Regan (Brent, are you reading?) to send photos of his LIV 
doghouse so you can see what a good installation looks like.  It cools well, 
and allowed 320 knots at 27,000 feet (85% power) when racing without frying 
anything except the pilot's brain.

Hope this helps.  Remember: big inlets (properly configured) don't hurt, but 
small inlets certainly do.  And work especially hard to keep the leakage down 
to zip.  You may think your cowl is tight, but I bet it can still be improved 
- a lot more than you think.

Fred Moreno
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
LML homepage:   http://www.olsusa.com/Users/Mkaye/maillist.html