[bret]

Question for the engineer type, on the diffusers inside surface the air is expanding and slowing down and changing directions, what is the ideal surface for the inside walls? My question is should we leave it a little rough so the airflow will stick to it? like a vortex generator ontop a wing? or slick like glass?

 

[Marc Bourget]

The problem with diffusers for this purpose is you rarely have enough "room" to maintain cross-section (approx. 7%) expansion - which maintains attached flow. Some surface treatment may maintain attached flow, but it would take a lot of development work.

I understand from NASA reports that a properly designed inlet can be oversized with little or no drag penalty. The ability to achieve this should shift the focus from inlet to outlet (rarely done).

Much more potential to reduce cooling drag on the exit side.

mjb

 

[BillL]

Well, if you are going to make a buck, make it smooth. I think the shape, and expansion schedule is more important than the surface finish. I think just to make it nice to look at from the ground.

For the SJ-Long, the left side has plenty if length for proper expansion, the r/s not as good, it is on the theoretical edge of separation.

I will take some data on performance, but not until I am flying.

Having said that, there are methods to get good expansion in a short distance, but the wetted area drag is elevated with concentric cones. Data should indicate the need.

 

[rv7charlie]

Hi Bret,

Here's my take, from many hours of reading up on ducts, and asking aero engineers, while trying to design ducts for my liquid cooled engine. Higher authorities are welcome to correct me if I'm wrong.

I think that the answer you're looking for is, smooth is better.

Vortex generators get stagnant air moving faster, which shrinks the 'boundary layer' of dead, more or less motionless air near the wing surface (where the bugs are sitting on your windshield while driving). They make the air 'turbulent', but I guess you could say it's organized turbulence, rather than random movement that may be in all the wrong directions if you're trying to generate lift.

On a wing, you don't make the wing surface rough; that would be the equivalent of a layer of frost. So you wouldn't intentionally make the inside of the duct rough, either.

Since there's not much of a real pressure recovery 'diffuser' (term used in most of the inlet duct papers) going on inside a cowl (most cowls have 'external' pressure recovery, in front of the inlets), the surface probably isn't that critical. But if you're going for every last ounce of efficiency, then go smooth.

FWIW,

Charlie

 

[rv6ejguy]

Since the flow is turbulent to begin with here and also being disturbed regularly by the prop blade passage, surface finish of the plenum walls won't make any measurable difference to cooling effectiveness.

Seal every leak and concentrate on the exit geometry if you want the best cooling for the least drag. Dan Horton has covered this extensively with lots of in-flight testing if you do a search here.

 

[BillL]

Since there's not much of a real pressure recovery 'diffuser' (term used in most of the inlet duct papers) going on inside a cowl (most cowls have 'external' pressure recovery, in front of the inlets), the surface probably isn't that critical. But if you're going for every last ounce of efficiency, then go smooth.

This does not sound right - - do you have some references that show there is no pressure recovery from the inlet aft?

Chris Zavatson's report on his Lancair had a relatively high velocity across 100-200kts, like 1.0 ratio to free field. The pressure recovery varied, but the minimum was .6 based on 100% free field velocity recovery. About the best from the CR3045 report. It seems his diffusers were effective, but they were quite short. The SJ-long has more distance available, so should be able to do well.

 

[rv7charlie]

Not what I said.

The stuff I've read (and been told by aero engineers) is that *most* pressure recovery in cowls for air cooled flat engines happens in front of the inlet. The Sam James (actually Will James, IIRC) cowls probably do have a bit more inside the inlets than a standard Van's cowl, but they're still nowhere near an optimum diffuser. Ross (rv6ejguy) can till you how difficult it is to build a perfect diffuser, even without the constraints of a flat engine.

The above doesn't mean the recovery isn't there; it just means that it happens mostly in front of the inlets. Picture a pitot tube (*and* its static source) sitting a fraction of an inch in front of the inlet. If external diffusion is happening, the airspeed will be lower than the a/c's airspeed. If the diffusion is happening inside the cowl, then the airspeed in front of the inlet will be roughly the same as the a/c's airspeed.

I repeat; this stuff is just what I read on the internet. plus what aero guys have told me. The above are my words; I hope I haven't gotten it too far wrong.

Charlie

 

[Marc Bourget]

For the sake of accurate speech/description, how can you achieve pressure recovery external to the cowl? IMO, you can see a pressure build-up, but not "recovery"

Now a chicken vs. egg question, since there is "no wind in the air", is the inlet air speeding its way into the cowl, or vice-versa? I agree at most E/AB velocities, the air is incompressible, but the energy associated with accelerating the air into/through the inlet, then slowing it down for pressure recovery seems to be ignored.

 

[rv6ejguy]

Cooling is easy, cooling with low drag is harder. The flight test data gathered by Dan Horton showed how difficult it is to recover cooling air momentum in typical opposed air cooled engine installations even with extensive mods.

Highest delta between upper plenum and lower cowling gives you best cooling but may not result in lowest drag. Dan's results and mine clearly show that you need variable geometry exits to to get the best overall compromise between cooling and drag throughout all phases of flight.

The typical air cooled opposed engine installation is far from ideal as has been previously mentioned here. You have short distances to diffuse the air and turn it efficiently without separation, you have very non-optimal shapes/ volumes from inlet to plenum and fin exit to cowling exit, not to mention, often multiple obstructions in the exit path.

 

[bret]

OK, smooth it is.....I have no more fingerprints on my fingers from all this sanding stuff.......

 

[BillL]

OK, smooth it is.....I have no more fingerprints on my fingers from all this sanding stuff.......

Yeah - After sanding for a day, I have to enter the iPhone password. No print.

Nice looking part! And quick too.

 

[rv7charlie]

For the sake of accurate speech/description, how can you achieve pressure recovery external to the cowl? IMO, you can see a pressure build-up, but not "recovery"

Now a chicken vs. egg question, since there is "no wind in the air", is the inlet air speeding its way into the cowl, or vice-versa? I agree at most E/AB velocities, the air is incompressible, but the energy associated with accelerating the air into/through the inlet, then slowing it down for pressure recovery seems to be ignored.

I'm not an aero guy; all I know is what I was told at a Holiday Inn (by aero guys) a while back. But Google can be your friend, assuming extra work.
https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=cooling+inlet+external+diffusion+

You'll get additional results if you vary the search terms slightly.

Consider a pitot tube. Where does the pressure increase happen due to increasing airspeed?

I do wish a real aero guy would jump in & correct my poor descriptions.

Charlie

edit: Here's one place to start:
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19980214918.pdf
And a 'scholarly search':
https://scholar.google.com/scholar?q=cooling+inlet+external+diffusion&hl=en&as_sdt=0&as_vis=1&oi=scholart&sa=X&ved=0ahUKEwj29OzjnOzMAhUC3GMKHZX-Aj4QgQMIGjAA

(I didn't say it would be easy. At least for me it's not...)

 

[DanH]

I do wish a real aero guy would jump in & correct my poor descriptions.

There's nothing wrong with your descriptions...but hey, I'm not a real aero guy.

edit: Here's one place to start:
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19980214918.pdf

Yep, that's one of the good ones. Figs 5 and 6 are gold. Also note that although not technically correct, "pressure recovery" is apparently an accepted term.

To the OP's question, duct quality (both shape and finish) is important at high velocity ratios, and a non-issue a very low velocity ratios. Waxing poetic....

Make the inlet big
and the outlet small
and you don't need
duct walls at all