[Bob Axsom]

I read Christopher Zavatson's article "Cooling Drag - Making the Air Work for You" in the brand new EAA Sport Aviation (December 2007). I have exchanged e-mails with Christopher, Paul Lipps, John Huft and Tom Martin about the subject and I am a little reserved about the subject. The one thing I have not experimented with is defuser tubes in the cooling air inlet path (well the hard plenum is another item that I am too dubious about to try so far). I have new wheel pant development on my plate for this winter but I will have more time before racing season starts next year. I want to put something in the queue that has a high probability of increasing the speed of the airplane - no other parameter is worth my time. OK here's a request for information:

Given no prop extension and using Van's cowl has anyone that is already seeing 200+ mph added defuser tubes and seen an increase in speed?

If so can you share any details of configuration and speed increase?

Tom Martin mentioned that they should be at least two inches long - I think I could fit that in but the right side will be tight.

I have already done a lot of experimenting and testing so please respond with real first hand data and experience only.

Bob Axsom

 

[dbuds2]

Difuser angles

I seem to remember a total divergence angle of 10 degrees or less is recommended for maximum pressure recovery. We were running some diffuser case and ejector rigs at work and if the expansion angle get larger than 10 degrees, there are increased losses.

You'd just have to figure out how fit a 10 deg expanding flow path to the inlet area and/or exit area.

 

[rv969wf]

Yep they work!!!!

They do work if designed correctly. I have 2 5/8" round inlet diffusers that are nearly 6-8" long that cool my engine with no problem, I do have an external source for the oil cooler and I'm seeing 325F CHT's in 95F -100F air. Winter time I change out the the cowl inlets to 2 1/8" to get CHT's up.

Yes they work if the inlets are made correctly along with a host of other cooling issues that need to be followed up on and corrected.

 

[gmcjetpilot]

Sorry, I slept at Holiday Inn Express Last Night

I'm reserved about the subject.....I have not experimented with is diffusion tubes......hard plenum is another item that I am too dubious about to try so far. I have already done a lot of experimenting and testing so please respond with real first hand data and experience only. Bob Axsom

I love the threads that start with:
I ONLY want smart people who know what they are talking about with certified data to reply.
I'm not smart but do stay at hotels a lot.

Bob, sorry I only have second hand DATA from Dave Anders. The data shows SPEED increase by going from the stock cowl to a new cowl/sealed rings/sealed plenum. His RV4 goes around the 260 mph mark. The "diffuser" is only part of a system. Rings, sealed diffuser (fancy word for duct) & plenum make up a total package.

Dave Anders RV-4 gained at least 6 mph by conservative estimate (closer to 8 mph) by going from stock to a sealed round inlet cowl w/ sealed diffuser & plenum. Dave's speed increased from 235 MPH to 248 MPH when he made several mods, including the cowl, inlet rings & plenum. He estimates 6 mph gain was from drag reduction, the remaining 7 mph gain was from increases in HP & prop efficiency (Custom Hartzell - thinner & new twist - 0.78 to 0.805). Gaining 6 mph @ 248 mph range is like gaining 12 mph in the 200 mph range. Dave's other AERO mods was a "firewall fairing" and new gear leg fairings. Even if the cowl represented half the drag reduction (being very conservative), that matches the 6-8 mph gain I expect and is typical. Bob you can't deny the gain, can you? What difference did Dave's new gear leg fairings & firewall fairing contributed? I don't know, it's the problem of multi mods made at the same time. I talked to Dave; he's very pro cowl/ring/duct/plenum mod.​

You're right, a RV, a 'tractor airplane', short or long cowl (just 1-9/16" longer), there's not much room to develop a perfect "diffuser", which would be real long, like 4 or 5 feet long to get perfect/ideal/optimal diffuser wall divergence (angle). Just because you can't make it perfect, does not mean significant gains are not possible. Also curved surfaces help, along w/ avoiding sharp corners. The devils in the details. Vans stock seal between cowl and baffle is kind of iffy at best. A round duct at LEAST starts with a bit of "good" flow (at the ring), which is better than nothing. Just sealing the duct from all leaks alone is golden. It just turns out round is easier to seal and has no corner. Oval would work, to make it more stock looking. Some folks just hate the look of round inlets for aesthetics only.

Van's cowl is NOT bad, just approx 5-8 mph worth of drag, verses a typical "tight" system. Bob, this has been proven over & over. What kind of cowl do you see on the RV winners at races you go to?

I don't get why you're sceptical but I noticed. You resist considering this cowl/plenum mod but play with tiny little internal baffles and consider fancy pneumatic wheel pant seals? You make your own wing tips but don't want to glass up new cowl inlets and plenum?

I've encouraged you to go w/ a Holy Cowl-Sam James-LoPresti-NASA-Mississippi State style cowl for what, 2 years? It's OK, I'm use to being ignored . I've been involved in RV's since 1985. I know you'll get measurable speed gains w/ the cowl ring/ diffuser/ plenum & gain very little w/ more internal cowl baffles & wheel pant seals. I can't change your mind, but your're throwing away known gains for small and unsure gains. I understand the resistance of tearing out an existing cowl/baffle set-up. It's a lot of work. If PhD Aero Professors, LoPresti & Dave Anders think its good.......may be they can persuade you?


******
From Dave Anders Notes:
Propeller - Inlet - Cowl relationship
1) propeller thrust distribution for inlet location.
2) boundary layer thickness at back edge of spinner and shank
of the propeller.
3) differential pressure across inlet.
4) reverse flow on standard inlets.
5) prop extension to decrease flat plate area.

Inlet (stagnation inlet)
1) shape - round = smallest surface area/volume
easiest to provide best seal to plenum.
2) size - calculated from mass flow required for BTU from
engine chart.
Mass flow = area X velocity X plenum efficiency
Oversized inlets >velocity resulting in >pressure thus >cowl drag around the inlets.
3) round lip edge to decrease inlet stall and reduce inlet drag

Plenum design
1) uses Bernoulli’s Law: volume = >velocity = >pressure
2) Greater *p across the cylinders carries off more BTU.
3) diffuser can diverge 5 deg/side ± (air flow remains attached due to >Velocity and >pressure)
4) A = area of inlet (calculated by mass flow requirement)
A = area at the rear of the effect diffuser
A = area over the cylinder and head, A area = A area
5) slowing the free stream velocity 10 - 40% >Cooling drag losses.
6) there should be no sharp bends or edges to minimize flow separation.
(My note: there's less leakage, no doubt)

Outlet - starts under the cylinders
1) area as small as possible to re-accelerate the air to free
stream condition to >drag.
2) Convergence zone leads to exit fuselage/firewall airfoil
rounded gentle curved sides should be ~ 35 deg.
3) shape - NACA studies
- straight sided rectangular shallow angle converging ramp
- ramp as parallel to free stream flow as possible at end
- width to depth ratio of ramp
a) 7 to 1 for cooling systems
b) 1 to 1 for exhausts (augmenter)
4) inlet to outlet ratio: what’s recommended, what works & why.
a) stock Van’s: RV4 ratio ~ 39sq” X 60sq” = 150%
RV6 ratio ~ 44sq” X 56sq” = 127%

(Dave's notes go on; he describes what he feels is a better 76% ratio, eg smaller outlet than inlet.
Obviously low speed climb condition is critical & req's more area than high speed level cond.)
******


Without debate, the hard DATA:

No leaks +
Smaller inlets +
Better shaped inlets +
Better location inlets
= less cooling drag​

More esoteric but true:

Better flowing round inlets (nozzles) with airfoil profile
(not laminar but not a swirling reverse flow mess) +
Better area transitions (which could be better still **)
= less cooling drag​

** Now my opinion, no data sorry, I'm convinced two separate L/R plenums are better than one big dog house plenum (Sam James). Area transition is larger than needed. Data? RV buddies have great luck with two separate plenums. You just don't need air over the middle of the case. Drag reduction is unknown, it may be small but cools nicely, so at least there's no loss & it weighs less. Another area I'd improve, leakage around the spinner. Min spinner/cowl gap is a common approach, but a seal behind the gap could reduce leakage into the lower cowl, which reduces efficiency. If air flow is out of the spinner gap it causes external plume-interference drag. Down side? Sealing is hard, and we're talking nitpick small gains, unlike the known large gain of a cowl/ring/diffuser/plenum system. Gaining 6 or 15 mph is like finding 10 or 15 HP with out burning more gas!

Bob, when you race, what kind of cowl do the winners have? I'm not sure what convincing you need but the DATA is out there. It has been shown over & over. All the best. I can see a Bob Axsom special home rolled cowl/plenum in your future? If you don't do it fine; it just gives me an advantage over you at the races.

PS Evel Knievel Passed, I remember those jumps as a kid> I had the Evel Knievel Lunch Box. Remember the lunch boxes with hot/cold thermos w/ glass liners that always broke. Wish I had it today, probably worth a few bucks. My Mom threw away all my Hot Wheels!

 

[Bob Axsom]

Read all

Good input so far.

I have made extensive baffle additions in the lower cowl to improve the flow after the cooling air leaves the cylinders. I won't give the details because there are so many pieces in the implementation that it would be distractive. What worked for me was:

A five surface baffling below the engine:
- One tapered baffle curving down from the full width of rear of the engine and plenum to the width of the outlet 3/16" below the bottom of the airplane
- Two baffles sloping in from the sides of the lower cowl to the rear baffle to direct the air in to the width of the outlet at the maximum angle (with respect to the firewall) allowed by the left and right cylinder stagger (the left side angle is less than the right)
- Two baffles extending out from the valve covers on the left and right side to the lower cowl below the upper/lower cowl separation.

This separates the cowl interior into three independent chambers the plenum, the area below the engine and the finally the rest of the inner cowl. The rest of the inner cowl is the area above the horizontal baffles outboard of the plenum and the area between the first three baffles and the firewall. The gain was on the order of 4 kts above the baseline speed of 170.67kts at 6,000 ft density altitude with everything set to the maximum. From memory the number is 174.7kts.

Additional large baffles originating at the front of the engine to turn the air toward the outlet avoiding the top of the filter air box and the nose gear web structure did not improve the speed. When I extended the these to the lower lip of the cowl outlet in a smoothly curved fashion the speed was reduced. I believe this was due to the acceleration of the exit air between the two opposing curved baffle surfaces followed by a divergence of the exiting air into the external air below the cowl. I think there may be something to be gained there but it requires something different than I have tried and the nose gear may make any further gains there impossible.

I performed carefully incremented experiments with inlet area reduction and I found a strong positive relation between inlet area and cylinder head temperature but after applying the national test pilot school spread sheet to the GPS test data I found that my aircraft speed was insensitive to inlet area changes and shape changes within the range of my experiments.

Every baffle is sealed with rubber and/or RTV in my installation except for the blast tube holes, the heater hole and the oil cooler in the rear of the plenum. When I block these off the holes other than the oil cooler for racing there is a small speed increase that cannot be ignored (every fraction of a knot matters). I also appear to get a small increase from the 3+ inch exhaust pipe turn down tip cut off. Cooling is not a problem in my airplane. The new short race wingtip configuration I developed for my airplane gave me another 3kt speed increase (this kind of speed change is major).

If someone has implemented defuser tubes in the inlet like RV969WF and has independent test data showing the speed gain or loss from that modification alone, it would be influential in my winter time work planning.

Bob Axsom

 

[rvbuilder2002]

I also enjoyed the article, but...

I was disappointed that there wasn't before and after test data.
That is the type of info that tells the real story of what is happening.

 

[elippse]

Please! Please! When you give the speed increase from a mod, don't just say you got 5 mph more, but say what the before and after speeds were so the percent increase can be calculated. The percent increases from each mod, expressed as a decimal, (3.3% = 1.033) can be multiplied together to give the overall increase which can be expected. One thing of note. I have noticed in my testing of RVs that the low 4.8:1 aspect ratio of the -6 causes much increased induced loss at higher density altitudes with its lower power/speed. Lower density, lower speed, higher CL, higher induced loss! As the plane's drag is decreased from mods, and the speed increases, the greater dynamic pressure will allow the wing to operate at a lower CL and decrease this loss, thus giving slightly more speed than would be expected just from the mod's percentage. For more speed from wing mods - down low - decrease area; up high, increase span with slashed tips that give more span/area!

 

[pierre smith]

Alan, some info...

They do work if designed correctly. ......

.......Yes they work if the inlets are made correctly along with a host of other cooling issues that need to be followed up on and corrected.

Alan, if you don't mind divulging, what taper ratio are your diffusers? The small 2 1/8" get wider as the go to the rear...then what is that diameter?

Thanks

 

[Sticky1]

Pictures

Can all-u-all post pictures. I have been baffled by the cooling within my bird. My oil temps are high. The CHTs & Egts are somewhat normal. the info contained herein is very helpfull however I am not intune with all the jargon.

Should the inlet tubes on an IO-540 be 8" in dia. and extend to the cyl 1 & 2?? I was also thinking about spliting the top cowling into 2 sections with downdraft cooling.

I am still baffeled........This is truly a black art..!!

 

[Sticky1]

???

What is the best placement in relation the the prop...

Currently 1.5" behind the prop is where the tube starts. How far out from the prop root should it start..?

 

[groucho]

Please forgive my lack of data on this, but...

One of the things I haven't seen done in an RV is something I did on my VariEze. I redid my around-the-cylinder baffling with RTV & fiberglass cloth. It makes for a near perfect seal around the cylinders. Less leakage means better cooling & less drag. Very messy to do though.

I did these concurrent with an engine rebuild (9.5:1, ported, new exhaust pipes, new prop extention, new prop, new spinner, cowling smoothing), so it's extremely hard to determine the actual benefit of the baffling. BUT cooling drag is a HUGE on all of our planes. IMHO, I think this type of cooling airflow direction is worth doing if you're looking for every last knot.

FWIW, I did a few other aero mods to my VariEze besides the mods listed above. When I bought my EZ, it'd do 156mph TAS at 8500MSL (crappy!). When I sold it, it would do 205mph at 8500MSL. The engine stuff listed above was responsible for 25mph worth of gain (180 to 205mph). No kidding.

Check out the late 90's era digital pictures (these are full resolution ):

From underneath the engine...

 

[rv969wf]

Go to Walmart

Alan, if you don't mind divulging, what taper ratio are your diffusers? The small 2 1/8" get wider as the go to the rear...then what is that diameter?

Thanks

Here is a picture of my high dollar (.75cent) mold from "Wallmart Spruce".... I used this CocaCola plastic cup as my mold to make my cowl inlet diffusers, then reshaped them later to fit the the cowl. When I'm using the 2 1/8" inlets or the larger 2 5/8" inlets, both of them open up to a final size of around 4 1/2" that blend into the plenum.

Here is another example on a RENO Racer with rectangle inlets.

Here is a picture of the cowl inlets on "Endeavor" that won at RENO 07.

Groucho is dead on correct with making baffles on top of the cylinders as most of the winning or really fast planes at RENO had this modification done as shown below just like how Brian (Groucho) did his plane.

 

[zav6a]

Why two?

Seeing how little opening Alan is able to get by with, how about eliminating one of the openings altogether? Based on everything I've read here and in Chris's article, the upper chamber seems to provide relatively uniform pressure distribution if pressure recovery is done well.

Open area at 2.5 inch diameter ring is is 4.9 in2. At 3.5, it goes to 9.6 in2 and at 4.5 inches, it is a whopping 15.9 in2.

The phenomenon of air flowing in one opening and out the other would certainly be eliminated.

Would look a bit odd but might be worth a knot or two!

 

[Bob Axsom]

Good Question

With the single plenum that is worth looking into. I may test it by making a plug for the inlet in front of cylinder #2 because of the greater gap between the prop and the inlet. It certainly would stress test the uniform pressure idea. I suspect the uniform pressure is a bit of an idealistic myth and the direct flow plays a bigger part in cooling than it is given credit for. The worst case is sitting still so a ground test should reveal high CHTs on #2 and #4 if there is a problem. After that test is passed a flight test in our rural area would be possible without risk to others. It is not high on my list but it is an idea not to be thrown away without serious consideration. Several airplanes use single cooling air inlets but they are usually updraft with equal distribution of direct inlet air.

Bob Axsom

 

[rv969wf]

One or two inlets.

Seeing how little opening Alan is able to get by with, how about eliminating one of the openings altogether?

When I was doing testing along time back, I did block of each inlet one at at time and take test flights.

This is what I found out on my RV with the settup I'm currently using. My left cowl inlet #2 cyl is more efficient than the right inlet #1 cyl at cruise speed. My thoughts are this: the #2 cylinder is spaced further back than #1 cylinder and this allows a longer diffuser which in my opinion is more efficient. Now another issue that I found out, in climbout at say around 120-140 MPH, the right #1 cyl cowl inlet provided the best cooling, maybe from prop rotation, prop wash, I'm guessing. NOTE: my prop is within 1/8" of the cowl inlet at high pitch. How do I know this? because when I made my first set of diffuser inlets, the prop came in contact with the carbon fiber and ground it down, this is with a constat speed prop ok, then I later clearanced things. No it did not hurt the prop.

So which ever inlet I blocked, that front cylinder CHT's gained around 30F degrees roughly over the other side that was open.

After trying this and that, I left both cowl inlets the same size to balance out between climb and cruise and I also made 3-D shaped block offs in front of 1&2 cyl fins that are angled upwards towards the plenum with spacing behind them. Hope some of this makes since.

Something else that I found out that works is an Anti-Reversion lip at the exit point of the diffuser to help with the reversion of the air so the air does not try to go back into the inlets. Finding the right cowl outlet and inlet size that work together is a key in my opinion. Manometer testing and temp probes placed in many areas of the cowling will help out in this area.

Note: Not all RV's have the same cowlings, plenums, nor shapes of inlets or prop extensions, so some of this might not apply or work, but these are RV's with an EXP rating so play around with caution.

 

[Bob Axsom]

Outstanding input Alan

I will not try it since you have already done it and the results are what they are. You can't buy that kind of first hand experimental test result anymore. My test plug results have something in common with yours. I extended them beyond the rear of the spinner to get as close to the rear of the prop in the max pitch condition as I could. Then I added a layer of fiberglass to eliminate a designed in valley in the leading edge of the plugs to see if that would help the speed any. Sometime after the test flight I noticed a spot in the flat black paint at the trailing edge of the prop blade. I looked at the plugs which had long since been removed. The left one was clean but the right one had a light (as in hard to see) scuff on the outboard leading edge. The farther out from the aircraft centerline you go, the farther back the trailing edge of the prop extends at high pitch because of the shape of the blade. It is not enough assume the prop pitch can rotate all the way to trailing edge contact with the surface of the spinner bulkhead and add a safety margin of a 1/2" when you extend a plug laterally 1.75 inches into the cooling air inlet. You have to determine the center of pitch rotation, the maximum pitch angle, the distance of the blade trailing edge from the center of pitch rotation at the location consistent with the outboard end of the plug and make sure you have good clearance between the prop and the plug in worst case interference conditions.

Bob Axsom

 

[Bill Wightman]

Cowl Flaps, anyone?

Great stuff here!

Anyone know of an RV'er who has installed a cowl flap(s)?

Cooling drag is a very big hit on our airplanes. Its analysis rests in knowing inlet-outlet flow momentum change. Its a classic control volume analysis, with the trick in knowing the heat added and pressure losses across the cylinders. This is really hard to nail down.

BUT, a cowl flap would control mass flow through the whole system (ie: the whole inlet - outlet system). Additional gains from cowl flaps can include parasite drag cleanup by reducing profile discontinuities on the cowl outlet (fairing down the exhaust stair-step by closing down the flap).

Back in 1994, I did a full-blown cowling design analysis for a new Part23 project. I still have the spread sheets and some of the data used to do that cowl. We built it and flew it. I had pitot style inlets sized at 4-1/2 inches diameter each (memory here), which worked just fine in flight. The outlet area was fixed at around 60 in^2 even though I was pushing for a cowl flap (marketing squashed the idea). These sizes are large compared with what you might see on an experimental. I know I could have gone smaller - alot smaller - but had too many constraints from our management and the FARs.

Running the same analysis on my RV4, I roughly expected a 10mph gain was possible with a good cowl flap, the rest being left alone. Mass flow could be controlled and parasite drag reduced with a good flap. But I sold the airplane before I did any mods like that.

My main point in all of this is to drive the momentum losses downward by controlling the OUTLET area, not the inlet area. Wondering why I haven't seen anyone try a cowl flap. I'd be willing to assist in the design/analysis of such an installation if someone would go ahead with it.

But, thanks to everyone for so much excellent information here!

 

[SvingenB]

Wondering why I haven't seen anyone try a cowl flap. I'd be willing to assist in the design/analysis of such an installation if someone would go ahead with it.

I'm not sure I quite get this regarding loss/gain in top speed, but is the idea to use as little air as possible to do as much cooling as possible? Could this cowl flap be used to control engine temp? if so this would be interesting because here the temp changes from +25 in the summer to -25 (deg C) in the winter, and RV-4s are known to be running cold, too cold in the winter.

 

[Bill Wightman]

Cowl flap usage

SvingenB: The idea is to control how much air flows through the cowl, and across the engine. We want just enough airflow to cool the engine - allowing excess cooling flow lets the engine run cooler, but also adds drag.

Reducing the inlet area is fine, and will yield results. But with fixed inlet/outlet areas, we're still required to run at lease enough area to cool to cool the engine under worst case operation, like slow speed high power climbs on a hot day.

Then when you level off, airspeed picks up and outside temps are likely cooler, and most of us will reduce power. Now our fixed geometries are allowing more cooling air than we need, and thus more drag. If we had a cowl flap, we'd just close it down to control the mass flow over the engine for cruise operation. Lots of certified airplanes do this, and it was a well proven tactic on the warbirds of WWII.

The cowl flap gives several advantages:

- Control of how much airflow we allow over the motor: reduction in cooling drag due to reduced mass flow with flap closed.
- A degree of control over CHT and oil temps not present with fixed areas.
- Reduction in parasite drag with flap closed.

Disadvantages:

- Added cost and complexity to the airplane.
- Increased weight.
- Another system to possibly fail and give us grief in flight.

But if its speed you're after, then IMHO, you just have to control the mass flow through the system. And it has to be reduced to just what's needed for what ever phase of flight you're in for the least drag and best performance. That requires variable geometry, since operational considerations vary so greatly for an airplane.

My .02

 

[Bob Axsom]

Thought a lot about it

The obvious benefit of a cowl flap cannot be ignored as you work the cooling drag problem. However, I have laid under the plane visualized every kind of hinge point motion direction and control mechanism I could imagine for an "A" model like mine and I have not come up with anything that has enough positive potential to justify the implementation abuse to the airplane. With conventional gear the way is MUCH clearer and the implementation possibilities much greater. I have gone into this much deeper than I want to get into here but on an "A" model I think the outlet will have to be relocated to implement a beneficial cowl flap (or flaps). If the center flow path were closed and outlets placed on the bottom outboard of the nose gear that would probably work in the normal fashion to throttle the air mass flow.

Bob Axsom

 

[flyeyes]

(snip). However, I have laid under the plane visualized every kind of hinge point motion direction and control mechanism I could imagine for an "A" model like mine and I have not come up with anything (snip)
I think the outlet will have to be relocated to implement a beneficial cowl flap (or flaps).
Bob Axsom

Bob--this isn't a priority for me since my temps are good )I've actually got some leeway to start closing down inlets which will probably be a "next winter" project) but I've spent some time thinking about it.

If you size the normal center outlet for the cruise condition, why not put cowl flaps outboard of the center tunnel? One or two small flaps just in front of the firewall could give extra outlet area for climb, and close completely for cruise. This especially makes sense to me since my oil cooler is on the right side of my firewall and air leaving the cooler already has to make a turn to get out of the middle outlet.

 

[Bob Axsom]

We are drifting here

The question of this thread is very specific so I ask that the thread focus on the non-standard tubes connecting the air inlets to the plenum and their measured effect on airplane speed exclusive of all of the other components in the engine cooling air system. No technofluff just good hard facts if there are any - which I am beginning to doubt.

Bob Axsom

 

[Tom Martin]

I currently have a cowl flap installed on my EVO1. The results have been mixed and for the most part dissapointing. The flap is about 6" wide and closes the area between the two exhaust pipes. The neutral position is the old cowl line. When I open the flap, lower from neutral, the temperatures go down as well as the airspeed, which you would expect. However when I pull the flap closed from the neutral position the temperatures go up but the speed actually goes DOWN. Not much but definately not as fast. The one good thing about this mod is that you do not have to do multiple tests you can see the results directly on the airspeed idicator as you change the position of the flap.
The interesting thing about the position of the cowl flap is that it affects the trim of the airplane. With the plane trimmed for level flight, if I close the flap from the neutral position, the nose goes down. Returning the flap to the neutral position will raise the nose to trimmed flight again. For me this means that somehow I am creating more drag, by turbulence perhaps, than I am saving from reduced airflow. The turbulence can be in two places, at the actual cowl flap area or backed up air spilling out the inlets.
For winter use the flap will be benificial and as a side benefit when you are descending pulling the flap closed helps to reduce shock cooling.
I would not install a cowl flap again. On the F1, in my opinion, if you have the inlets operating correctly, and the exit area cleaned up, the stock cowling is pretty effiecent in the stock form.
Years ago I had tried a cowl flap on my RV4 with very similiar results. I thought I was smarter now but apparently not!

 

[rv969wf]

I agree

Tom, your post is correct as I found out the same thing as I tried a cowl flap which was a little different than yours on my RV-6 several years ago and it did make temp changes, but when the cowl flap was opened or closed it caused my RV to pitch up or down and I had to make small adjustments with the elevator trim, yep it was very pronoused and the speed change was nill to nothing and made no changes in airpspeed that I could see, so I went back to other thoughts and ended up with what I've got now. Thinking here, but I might be wrong, the air might have been building up inside the plenum when the cowl flap was closed and caused a reversion at the inlets like your talking about and made the airplane speed change nill to none or worse. I did all this along time ago and did not have photos or tuft work till later years after trying a few other things so I'm kinda out in the wind with what was going on a few years ago with testing things on the cowl flap idea, but later on a few years passed and I did do some tuft work with the inlets and did see reversion at the inlets and that's when I found the sweet spot and came up with the diffusers / size that I have at present that work.

I don't know what it is with these VANs RV's or Rockets, but the cowl flap idea I feel is maybe more for the spam cams at lower airspeeds. If I had the time I'd like to design a means to control the air inlet area between winter and summer and not at the cowl outlet and leave that area fixed.

Anyone have any thoughts on this???

 

[gmcjetpilot]

Bob is fighting "The Force"

The question of this thread is very specific so I ask that the thread focus on the non-standard tubes connecting the air inlets to the plenum and their measured effect on airplane speed exclusive of all of the other components in the engine cooling air system. No technofluff just good hard facts if there are any - which I am beginning to doubt. Bob Axsom

Bob just do it, your throwing away 5-8mph. The article said 5 kts, which is about right. He even got 20F-30F lower CHT's. What more do you want? You say you don't want technofluff but ignore the results. I know you're fighting the idea. What else do you need to know? It works. If you want evidence, take the article to heart and the data from Dave Anders I posted. This is the most popular mod to RV's for a reason. Nothing wrong with the original RV cowl baffle design, but it can be improved on. You can use your existing cowl and existing baffles. You're handy with fiberglass. Glass in some inlet rings, make a top for the baffle (metal or fiberglass), make a transition from the soft duct to plenum (like the article shows) and with two soft ducts/tubes & clamps you're in business. Simple right. You can always write a big check to James Aircraft, but many have modified the stock cowl and baffles with success. There is a right way and a wrong way to roll your own.

"focus on the non-standard tubes connecting the air inlets to the plenum and their measured effect on airplane speed exclusive of all of the other components in the engine cooling air system"

You know I am going to be honest with you Bob. It makes no sense to look at it that way, at least to me. Its the total package from the lip of the cowl inlet > ring throat > soft duct > into the plenum, all of which is considered the diffuser. You can't pick one component and say what affect does that have. The soft duct is a necessary evil to connect the inlet rings and plenum. Remember high velocity air need careful attention. The tube is part of the effort to flow high velocity air smoothly. It's not perfect because there are are some transition lips, but it's leak free. There is no way you can say that about a stock Van cowl to baffle interface.

"hard facts if there are any - which I am beginning to doubt."

I'm speechless, yes its true. I don't have anything to say.



*********************
Technofluff nerd stuff warning - Do no read below unless you are using "The Force"
*********************

The article mentioned stagnation (air momentum = zero) but skipped the external drag that happens. Big inlets means stagnation outside the cowl. When the air stops externally before going into the inlet, that air than spills out and mixes back in with the fast moving external air going around the airplane. That's not efficient and it's drag; stopping and re-accelerating air is drag.

You're cutting inlet area down by at least 45% and still getting the same cooling! IT'S LESS DRAG. However to get that you have to be very efficient (careful) with the (high velocity) air going in the cooling system (diffuser). You need smooth non-turbulent air flow down stream. The "non-standard tube" is part of that effort. The only way to do that is with careful ducting (diffusing if you will). Bigger low velocity inlets or the stock Van's inlets are fine and less critical, but they have more drag (like 30-50% more). Last and very simple to understand, absolutly and positively NO LEAKS in the high pressure part of the cooling system is a must for efficiency. You can't get better than a round tube clamped to round flanges. As well elimination of the soft baffle seals is goodness. Bob, don't fight the force, use the force, the stock soft seal design leaks at 200 mph in ways you can't see. Also a side benefit of not using the cowl to seal the top of the cooling plenum is reduced stress and maintenance on the cowl. Bottom line, How good? About 5-8 mph. I'm done.

 

[Bob Axsom]

I do not agree or disagree

Deep Breath - I have extensively modified and fine tuned the cooling system in my RV-6A and for my efforts I gained 4kts over the baseline speed of 170.67kts. Then I made new wing tips and gained 3 more knots over the demonstrated best case cooling system speed on my airplane. My current demonstrated 6,000 ft density altitude speed with everything trimmed and set for maximum speed is 177.8 kts (3-way with GPS tracks and speeds input to the NTPS spread sheet). At these speeds, on this airplane, in this configuration, I am confident that I have reached a point where further increases in true airspeed based on cooling system changes will be exceedingly small and will require very precise and correct changes. I cannot afford to expend my resources on this kind of pursuit. The idea that I can arbitrarily put a separate lid on my plenum and gain TAS with no other changes is not believable. The possibility of gaining speed by reducing plenum leakage alone is not believable. The possibility of reducing the air mass needed to cool the engine by reducing plenum leakage is believable. Expanding on that idea, tuning the inlet to take in the reduced amount of air mass to achieve the cooling of the engine has the potential for increasing the TAS if it is processed properly and external aerodynamic changes made possible by the smaller inlet requirement are incorporated. I have achieved cooling efficiency already that does not require the amount of air mass which is entering the plenum but incremental efforts to reduce the air mass flow in the system by reshaping, relocating and reducing the cross section of inlets has only been effective in reducing the air mass flow and increasing the CHTs incrementally up to the 400 degree limit in direct relation to the inlet cross sectional area without an increase in TAS. To my satisfaction the system is not performance bound by plenum leaks in its current configuration. I think there is potential for improvement at the front end by more effective closure of the front of the plenum and sealed defuser tube inputs but I'm not sure - which is why I asked the initial question. I observe that out of over 1,000 observations of this thread, no one has come forward with a test of the isolated TAS effect of sealed defuser tube input to the plenum. If you look at the round inlets to the New 230+ kt Mooneys and Columbias lack of attention to this possibility is hard to ignore. I may try the defuser tubes anyway but I will be thrilled if the TAS goes up to 178kts (0.2kt increase). I hope you can appreciate that after a couple of years of work and significant speed improvement on my airplane I am not asking for generic cooling system design philosophy but specific TAS results isolated to one system element - inlet defuser tubes. I have received some good input here from all of you and I see it as a medium risk low gain potential experiment that I may or may not conduct.

Bob Axsom

 

[Andy_RR]

I sat down and read NASA CR3405 on the weekend since I was bored and my empenage kit is still several thousand miles from where I was/am.

If you haven't already, I recommend you read this report in detail from start to finish, because many of the points raised here are covered and there is at least one conclusion opposite to what George has postulated.

If I may summarise my understanding of it is this:

- cooling flow leakage past (not through) the engine is the major player
- leakage at the front of the engine is a large part of it.
- throttling the cooling flow downstream is the way to go
- circular cooling air entries are generally the way to go

Interestingly, large, low velocity ratio inlets were more effective at dynamic pressure recovery than the smaller, high velocity ratio designs as we see on the James cowl etc. I think this has more to do with the ability to diffuse high velocity air over a relatively short distance.

Another interesting point was that the stagnation point of the cowl was not axial with the propellor, but below by some distance. The airflow past the cooling inlets has a significant vertical velocity component which is not reflected in many of the cowling designs I've seen. It is on Mr LoPresti's PA-28 cowl though - I've no idea how effective it would be, but it could help with packaging the diffuser along the top of the cowling.

They did mention that the smaller circular inlets were more prone to stalling on the bottom edge which could dramatically reduce the pressure recovery that was possible. Maybe this is what Mr LoPresti is attempting to avoid?

I also wonder why the edges of these sleek circular entries are always sharp edged. I understood that, kind of like a re-entrant nozzle, the edge should be as rounded as possible - kind of like the large entry nozzles that are put on turbofans when they're tested on ground-based test-rigs.

Cheers,

A

 

[Nuisance]

Ah, the early morning musings of the sleep deprived...

What we are talking about here are small differences. Van already made all the big gains in efficiency with his excellent design. All we can do is make little bitty changes and get little bitty gains.

Bob, you are asking for irrefutable prove that you will make substantial gains before you do any work. I'm sorry, it doesn't work that way. That is why we are called experimenters.

I can't give everyone before and after numbers on my cooling plenum. I built it that way from the start, because I believed in the science. Bob used the word "believe" several times in his last post. To me, math and science is fundamental, and if it makes sense to me, it is kinda like religion. That is why I ended up being an engineer I guess.

One of my planned winter projects is to refine the airflow from the inlets into the plenum. It will be a lot of work including repainting the front of my cowl. It might not gain me anything. That is the game.

I have played this game many times in the 5 years I have been flying my airplane. During those years I have wasted time and money on things that didn't work, but I have gained speed as well.

I think this is a lot of fun, but then I am weird.

John

 

[Tom Martin]

Bob
I aggree with your thoughts on the standard inlets. I feel I have gone as far as I can with my rocket cowling to get the cooling that I reguire for my type of flying. It is my opinion that the main advantage of round inlets is not so much that they are round but that they are farther forward. This gives a much greater distance between the opening and the front of the cylinder in which to create a better diffuser. A nicely made rectangular opening would be just as easy to seal as a round opening. Mathematically I believe that a round opening gives the greatest area in relation to perimter so in that regard it is more optimal. However most round inlet installations have what I consider to be a draggy area between the inlet and the spinner. I look for direct bug splats to define draggy areas. I wonder why the round inlets could not be moved closer to the spinner. This would reduce this draggy area and probably reduce the total wetted area of the front of the cowl.
However, if you are looking for the most bang for your time spent, I do not think you will see huge increases from this work as you have already shown some significant improvements from internal sealing and baffling. Claims of huge speed increases from round inlets are only valid if the internal workings of the cowl are also done and you are already there.
I think there are more gains to be made, for your time, if you continue with your wheel pant, gear faring, experiments.

 

[pierre smith]

Camera aperture

..... If I had the time I'd like to design a means to control the air inlet area between winter and summer and not at the cowl outlet and leave that area fixed.

Anyone have any thoughts on this???

Alan,
A 35 MM camera aperture adjustment comes to mind. If you could cut your inlet inserts in half from left to right or top to bottom, then hinge them on one side and with a cockpit controlled vernier cable to the opposite side, adjust them smaller or larger. Doing this in flight would allow you to see airspeed and oil/cht changes, no?

The Bearcats and Corsairs had these "fan" type cowl exit air flaps that overlapped each other but also, the firewall had a generous radius to ensure laminar air flow out.

Regards,

 

[born2fly]

stock to ???

It seems like an easy mod to change from the leaky stock rectangular intakes to a sealed oval. This could even be a plug-in, at least for a while. Might even look good. Has anyone done this - or disqualified it? If it's useful, how far inside the plenum would the diffusers project?

G.

 

[Bob Axsom]

That is the question essence

Tom Martin tells me they need to be at least 2" long. My thought is there needs to be a web across the rear of the existing inlet to contain the blowback across the front of the plenum with a sheet rubber seal across the opening for the defuser tube that has an opening smaller than the defuser tube to seal it when it is plugged in from the front. I plan to use the four inner inlet screws to mount the plug/defuser tubes in the inlet. This is how I mounted the non-defuser tube test plugs when I conducted my inlet reduction tests. I have seen no hard data that supports these tubes as a value added part of the system.

Bob Axsom

 

[petehowell]

More cooling drag information

I realize this is fast glass, but this guy Chris Z., really seems to know his stuff. I enjoyed reading his insights.

http://www.n91cz.com/cooling/webCowlrep.htm

main site here

Just realized this might be old news here - sorry

 

[chaskuss]

I sat down and read NASA CR3405 on the weekend since I was bored and my empenage kit is still several thousand miles from where I was/am.

If you haven't already, I recommend you read this report in detail from start to finish, because many of the points raised here are covered and there is at least one conclusion opposite to what George has postulated.

If I may summarise my understanding of it is this:

- cooling flow leakage past (not through) the engine is the major player
- leakage at the front of the engine is a large part of it.
- throttling the cooling flow downstream is the way to go
- circular cooling air entries are generally the way to go

Interestingly, large, low velocity ratio inlets were more effective at dynamic pressure recovery than the smaller, high velocity ratio designs as we see on the James cowl etc. I think this has more to do with the ability to diffuse high velocity air over a relatively short distance.

Another interesting point was that the stagnation point of the cowl was not axial with the propellor, but below by some distance. The airflow past the cooling inlets has a significant vertical velocity component which is not reflected in many of the cowling designs I've seen. It is on Mr LoPresti's PA-28 cowl though - I've no idea how effective it would be, but it could help with packaging the diffuser along the top of the cowling.

They did mention that the smaller circular inlets were more prone to stalling on the bottom edge which could dramatically reduce the pressure recovery that was possible. Maybe this is what Mr LoPresti is attempting to avoid?

I also wonder why the edges of these sleek circular entries are always sharp edged. I understood that, kind of like a re-entrant nozzle, the edge should be as rounded as possible - kind of like the large entry nozzles that are put on turbofans when they're tested on ground-based test-rigs.

Cheers,

A

Andy
Thanks for the link to the NASA report. Do I have to slog through all 1,000+ pages of it? Can you suggest a smaller section that has the relevent data to this discussion?
Charlie Kuss

 

[dsb]

Andy
Thanks for the link to the NASA report. Do I have to slog through all 1,000+ pages of it? Can you suggest a smaller section that has the relevent data to this discussion?
Charlie Kuss

It's 152 pages... Only 65 are 'reading' pages, the rest are graphs and pictures...

The high lites according to Dave:

P 26... 55% of the entering air bypasses the engine...

P 27... With a 'dog house' upper plenum, the same pressure drop across the engine was achieved with 38% less flow entering the inlets...

P 28... 8% leakage around the baffles where they meet the cylinders...

P 29... Installing a front engine baffle to remove cowl/spinner gap losses netted 9%...

P 42... Inlet lip contours should be well-rounded with large radii to minimize inlet stall...

P 57... Cowl Exit Study: Poor inlet design and leaky baffles can be compensated for by subsequently increasing the exit area to achieve the required cooling, at a considerable drag cost...

P 65... Conclusion #7: The aerodynamic behavior of the cowl inlets are a major factor in the effectiveness of the cooling installation. There is an obvious need for basic inlet design guidance...

HTH,
Dave

 

[zkvii]

Help....

P 42... Inlet lip contours should be well-rounded with large radii to minimize inlet stall...

I've done the sealing / hard top plenum thing and it is working well. I'm now working on inlets and have delta pressure sensor measurments running etc. I have the data to show the inlet shape makes a difference - what I'm trying to work out is which of the A-Series ratios we should be using for RV speeds. ( If anything I would like to work on the lower portion / best climb type speed (80-100kts) as once into the cruise speeds the deltas are over double.)

I think the Vi/Vo ratio is the first step and I'm assuming we take Vo around the 90 kts mark - but what is Vi for the cowl and therefore which shape should we be aiming at? And which series does that equal.

(I know Chris Z has commented that he had a 'high inlet velocity' but don't know the actual numbers)

Carl

 

[Bob Axsom]

Another option

It's 152 pages... Only 65 are 'reading' pages, the rest are graphs and pictures...

The high lites according to Dave:

P 57... Cowl Exit Study: Poor inlet design and leaky baffles can be compensated for by subsequently increasing the exit area to achieve the required cooling, at a considerable drag cost...

HTH,
Dave

It is not necessary to increase the exit area if you improve the efficiency of the flow in the lower cowl and with a reduction in drag if it is properly done.

Bob Axsom

 

[tin man]

Cooling air needs to spend a bit of time on the cooling fins to absorb the heat.
A proper exit to inlet ratio is necessary to maximize the effeciency of any cooling modification. In some casese enlarging the exit area may actually have a negative affect on drag and cooling efficiency. Lots of pieces to this puzzle.
Tin Man

 

[dsb]

It is not necessary to increase the exit area if you improve the efficiency of the flow in the lower cowl and with a reduction in drag if it is properly done.Bob Axsom

Just to clarify... I am not saying any of this... Just quoting NASA CR3405...

Dave

 

[Andy_RR]

From my reading of CR3405 and my limited understanding of fluid dynamics, 'smoothing out' the flow in the large volumes above and below the engine is not the "low-hanging fruit". I think of it more in terms of energy management. Dynamic pressure is 1/2rho.v^2 so you are better off targeting the flow management of the higher velocity air where the square term is having most effect - i.e. around the entry and exit points.

Also, CR3405 suggests that exit throttling is preferable to inlet throttling, so large, low velocity ratio inlets are not a bad thing, are less prone to stalling at the edge and have the effect of extend the effective length of your diffuser into the upper plenum volume.

A good reentrant nozzle arrangment around the exit will be my next requirement, where the low velocity air in the lower plenum needs to be accelerated back up to the free stream velocity as efficiently as possible. Van's rounded edge ramp is the first step, but I think it can be made more efficient at the edges. I'll scratch my head about this some more when I begin the fuselage build of my RV-8.

The rest of it appears to be ensuring that the effective orifice size of the engine cooling system is no larger than absolutely necessary - i.e. no leakage path past the engine - everything must go through the fins etc.

That's enough of my theory - can't wait to try it in practice on my bird.

A

 

[Bob Axsom]

That's the spirit

Everything is reversable (at some cost) so if you want to go fast, use the best available information NACA/NASA and other sources have documented, concieve how you want to try it in your installation (that they never saw or tested) implement it and test the results. Be careful to test the same way EVERY TIME using a repeatable and valid test method. I gained 4 knots on my RV-6A with cooling drag mods alone and others have done much better I'm sure.

Bob Axsom

 

[rv6ejguy]

Andy,

Would you be interested in instrumenting your cowling setup for pressure and velocity at key points? It would be interesting to compare data to my liquid cooled installation.

 

[Andy_RR]

experimental? - you bet!

Ross, I'll be very keen to do some studies as soon as I'm finished with the rivetting - it's a while off though!

 

[rv6ejguy]

Ross, I'll be very keen to do some studies as soon as I'm finished with the rivetting - it's a while off though!

Cool, I may be done the RV10 about the same time you are done your -8. In the meantime I'll try to get my gauges mounted again in the 6A and get some new numbers.

 

[dsb]

One part of this equation that I haven't read anything on (If you have a reference, please let me know...) is how this works from a thermal perspective. I have heard that a P-51 actually has negative drag through the cooling duct, supposedly from the heat transfer of the radiator. In the case of an air cooled horizontally opposed engine, it would seem that if the interior of the lower pressure plenum were constrained to an appropriate size then the thermal transfer would be sufficient to re-accelerate the air to negate (at least partially...) the drag from that air re-entering the exterior stream.

Dave

 

[Bob Axsom]

I've heard that before

I have heard that before from good sources. One in particular is John Huft who has a very fast RV-8. He won the RV Blue class in the AirVenture Cup Race last year at 235 mph.

Bob Axsom

 

[rv6ejguy]

One part of this equation that I haven't read anything on (If you have a reference, please let me know...) is how this works from a thermal perspective. I have heard that a P-51 actually has negative drag through the cooling duct, supposedly from the heat transfer of the radiator. In the case of an air cooled horizontally opposed engine, it would seem that if the interior of the lower pressure plenum were constrained to an appropriate size then the thermal transfer would be sufficient to re-accelerate the air to negate (at least partially...) the drag from that air re-entering the exterior stream.

Dave

Later calculations showed that the stock P51 was a pretty good system but it did not produce zero drag or thrust at the typical speeds or coolant temperatures run. Various sources disagree. Some more conservative ones say the system offset about the equivalent of 150hp worth of drag compared to more conventional under wing radiators.

The higher Delta T of air cooled installations offers more possibility to recover higher pressures at the exit due to increased expansion but this is probably offset by the poor geometry and high pressure losses internally on something like the typical RV installation. Exit air trying to find its way out past obstructions and thudding into the firewall can't be that great from a drag point of view.

I think you'd need sealed bottom plenums on a Lycoming with smooth, decreasing volume ducts to twin exit nozzles- maybe boosted with exhaust augmenters. Lots of work to build which is why you don't see much of this being done. Would be cool to build such a system and instrument in flight.

 

[dsb]

Later calculations showed that the stock P51 was a pretty good system but it did not produce zero drag or thrust at the typical speeds or coolant temperatures run. Various sources disagree. Some more conservative ones say the system offset about the equivalent of 150hp worth of drag compared to more conventional under wing radiators.

The higher Delta T of air cooled installations offers more possibility to recover higher pressures at the exit due to increased expansion but this is probably offset by the poor geometry and high pressure losses internally on something like the typical RV installation. Exit air trying to find its way out past obstructions and thudding into the firewall can't be that great from a drag point of view.

I think you'd need sealed bottom plenums on a Lycoming with smooth, decreasing volume ducts to twin exit nozzles- maybe boosted with exhaust augmenters. Lots of work to build which is why you don't see much of this being done. Would be cool to build such a system and instrument in flight.

Something like this?

http://www.ez.org/feature/F0502-1/F0502-8.jpg

 

[rv6ejguy]

Yes, that's what I was thinking. Now to fit that in an RV cowling...

 

[frankh]

Wow

Something like this?

http://www.ez.org/feature/F0502-1/F0502-8.jpg

Does that thing even have exhaust pipes?...Or does the exhaust simply dump into the lower plenum..I gues it could as its a pusher, assuming the lower plenum don't melt!