Hi Peter,

 I was looking at your web site for your new email address when
 I came across this.

 It might be a minor factor but Franklin figured out in the
 30's that if you apply the cold cooling air to the coolest part
 of the engine first you optimize the heat transfer rate.

 If you go over the exhaust pipes first you reduce the cooling
 because the cooling air gains heat and the rate of heat transfer
 is proportional to the temp. diff.

 I suspect this is why most AC have down draft cooling.

 Also you recover some cooling drag by squirting the hot air out
 in the right direction.

 Paul



 Resolution of the unequal-cooling problem (4/07)

 It seems perfectly obvious in retrospect, but it took me several years to recognize that the problem of the two hot cylinders was due to the baffling and not to some sort of mysterious dynamics in the cowling. I knew, rationally, that the large air inlet and the very large plenum beneath the engine ought to mean that the air there was essentially at a standstill and was pressurized to a reasonable fraction of the total pressure -- the diffusion losses would be the same as in any other cowling
 using a plenum -- and so the flow across the cylinders was most likely unaffected by air motion within the plenum. I would visualize the situation as resembling a large tank of water with a hole in the bottom, and the engine, inverted, blocking that hole. Nevertheless, I could not see how the baffling of cylinders #1 and #6 was so different from that of the other four that they would run as much as 80 deg. F. hotter, and so I kept thinking that some kind of mysterious blockages or eddies or
 other aerodynamic effects were responsible. The fact that a nose-high attitude in climb tended to reduce the temperature disparity encouraged me in that belief.

 The only blockage, as it turned out, was mental. I date its removal from April 22, 2007, when I was in Paso Robles at a meeting whose purpose was to plan some flight tests of some of Javier Arango's World War I airplanes. Always eager to pick a brain when one wanders within range, I described my cooling issues to Paul Robertson, who owns Aeronautical Testing Service in Arlington, Washington. He said that such things were not unusual, and that the solution always lay in tinkering with the
 baffling. By sweeping aside the whole matter of flow dynamics within the plenum, his remark, together, perhaps, with an article on baffling that happened to appear in Kitplanes magazine this month, made something click. It enabled me to perceive what should have been obvious all along: that what the two hot cylinders had in common was that their exhaust ports were not adjacent to another cylinder.

 I inspected the environments of the "cool" cylinders and observed that the neighboring cylinders' intake port geometry formed a converging duct around the unfinned outer surface of the exhaust port. I accordingly removed the baffling from the #6 cylinder (it's a little easier to remove than that of #1, so I attacked it first) and added some bits of sheet metal whose purpose was to imitate that duct -- that is, to contain and accelerate the air flowing over the surface of the exhaust port. A
 short test flight showed that this change had been quite effective; the #6 CHT had dropped by 50 deg. F or so. Similar modifications on the #1 (right rear) cylinder produced the same result.

 Research done on cylinder cooling back in the 1920s and 1930s showed that what was needed was to keep air flowing over every part of the cooling fins; air that became static or stalled did no good. Baffles therefore rest on the cooling fins -- air going by outside the fins does nothing -- and wrap around the downstream sides of cylinders, allowing air to escape through slots whose area is about the same size as the space between the fins. I had provided that kind of baffling on the downstream
 sides of my cylinders. But I had failed to appreciate the importance, for heat transport, of maintaining flow velocity over all parts of the upwind sides of the cylinders as well. The exhaust ports were in relatively stagnant, albeit cool, air; they needed to be enclosed in channels through which that air would travel at reasonably high velocity, pumped by the pressure difference between the two plenums.

Thanks for forwarding Peter's article.  He certainly has a way with words.

I certainly agree with your analysis about heat transfer.  However, I
have wondered if there are benefits to channeling cool air directly to
the area around the exhaust ports on a finicky aircraft engine - the
hottest part of the engine - in order to give the exhaust the best
chance of survival.  Instead of then routing this heated air over
other parts of the cylinder, maybe you could design the baffles/plenum
so that air coming off this portion of the head is segregated from
going over the portion of the cylinder that would normally be down
stream.

We can probably agree that it absolutely doesn't make sense to heat
the air used to cool the cylinders by running it over the exhaust
stacks first.  Though the stacks actually do cool the heads, they are
so hot that running 250deg vs 350deg air over them would make little
difference in how much cooling they do.  Running heat tape on them
generally raises CHT's.


Matt-


You got a good point there Matt. The heat in the exhaust valves is
the limiting factor. Did you know that some Mooney aircraft have oil
channels around the outside of the exhaust valve guides to directly
cool the guides/valves with oil?