What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Help with measuring cowl pressure differential

Roger all that Gary...good work as well.


Here's a thought for adding the radius to a 6 or 7. Make a small lap joint inside the radius (full width), and have that sit on top of the flange at the firewall. The lap joint would contain the nutplates, and the radius itself would then be flush with the flange itself (and thus the belly of the plane).

The only issue I see with a either lap joint or the radius on top of the flange (Gary's method), is working around the rivets in the flange. I gues one could drill them out and reinstall after radius installation, or drill holes in the lap joint or radius to fit over the shop heads. How'd you do it Gary?

Bob, I worked around the flange rivets with holes in the fairing.

Nothing at the top of the fairing. It is just bent "tangent" to the firewall and fits snugly with just two 8/32 screws holding it against the firewall.

I am also not happy with the smoothness of the curve going out onto the belly and plan to rework that with an arrangement similar to what you have described. I have a small bump bluff body on the belly (only 1/2" high) but the curve does not match the radius of my firewall fairing exactly.

But ......I am encouraged by the very streamlined flow in this area indicated by oil drop tracks. Lottsa air going through here!!
 
Thoughts from anyone on the lap joint? Overthinking it? :eek: Cheers, Bob

Roll a 2"~3" radius on the end of a 0.019 stainless sheet which extends back under the belly a foot or so. Attach with blind rivets....belly burn through fire protection, and of course, no lap.
 
Roll a 2"~3" radius on the end of a 0.019 stainless sheet which extends back under the belly a foot or so. Attach with blind rivets....belly burn through fire protection, and of course, no lap.

Ah-ha moment! Thx! ;)

Cheers,
Bob
 
OK, sorry it took an extra week, but here are the photos of the "uglied up" aluminum exit curve thing (whatever you want to call it). Note that on the photo from the bottom you can see the line of RTV across the ege. I had that covered with aluminum tape, but it had problems sticking really well, and kept curling up...possible from air curving inwards from the side of the cowl exit. Anyway, it's fine with just the RTV, so far.

Also, the bottom picture is for Dan...I installed pass-through ports on my firewall today for pressure testing. No more running the lines through the heater duct...too hot for that. I'm hoping to have everything wrapped up tomorrow morning so I can fly a pressure test or 2.

imag0177vm.jpg
[/URL][/IMG]

imag0178m.jpg
[/URL][/IMG]

 
Sonny,

Be interesting to see if you see any speed gain, or improvement in cooling. It is going to take more modification than that to outrun my plane haha:D
 
Sonny,

Be interesting to see if you see any speed gain, or improvement in cooling. It is going to take more modification than that to outrun my plane haha:D

Hey Steve! Yeah, I don't know, brother...I just waxed my airplane...it just FEELS faster! I'm not expecting a speed gain, but I really am hoping for a cooler running engine.

Oh, and get this...just finished my annual condition inspection and I found that my throttle cable clamp was slightly pushed out of position. I wasn't maxed out on the movement of the throttle! I think I was missing a few RPM at the top end! I really don't think anything will change, though...I'll keep you posted.

Looking forward to that Reno trip!
 
I had that covered with aluminum tape, but it had problems sticking really well, and kept curling up...

I've been using a lot of Shurtape AF100. It's rated to 260F, necessary in the exhaust area. The acrylic adhesive is just right, good stick but peels without a problem on paint. Only thing I've noticed is that the surface must be perfectly oil free or adhesion falls off a cliff. I wipe with MEK and a clean rag prior to taping anything.

Cowl extension for area reduction, pitot-static on left, shielded temperature probe on right.

2ew37dx.jpg


..I installed pass-through ports on my firewall today for pressure testing.

Bad to the bone.
 
I wanted to update this thread in case anyone didn't see my "original" thread about this...but, I posted a couple pictures of my EFIS with the results in the other thread here.

The jist of it is that the ramp probably netted a decrease of 3 or 4 degrees (F)...coupled with blocking off the heater scat opening in the baffling, it was about 7 degrees cooler overall.
 
James Cowl RV-10

All, I posted this over on the other forum before running across this thread... Great thread by the way!

We have been flying with the James cowl and plenum for a few years now. *As some know we went through an extensive exercise in order to run the airplane near lean which was solved by adding the turbo injectors.

We really never had much of a cooling issue, but lately I think due to higher ambient temperatures we have had to limit extended climbs. *So it is time to make some modifications to provide more margin. *I found that there was a change to the inlet diameter of the James cowl. *A quick e-mail to Will and he sent out larger inlet rings at no charge with apologies for having to install them.

So now what - we didn't want to install the rings without a good baseline of our current condition so we started on a path to gather data. *I am disappointed in myself that I didn't collect this data during the whole turbo injector experiment.

We got a manometer that measured from -10" to +10" of water. *We started by plumbing in four lines, first to the plenum chamber above the engine, second to the cowl chamber below the engine (near the sump), third to the pressurized injector manifold (which is air plumbed from the intake air box), fourth to the static pressure system.

Last night we gathered some readings during climb at 105kts then in cruise at 8000 PA *and 4000 PA.*

The values listed respectively were:
Plenum to Cowl: 2.5", 4", 4"
Injectors to Cowl: 7", 9.5", 10"
Plenum to Static: 4.5", 6.5", 7.5"
Injectors to Static: 9", >10", >10"
Cowl to Static: 2", 3", 3"

I know taking the values to the cowl and to static might have been redundant, but it is an experiment after all.

So now for my question, I am not sure what to make of the "Cowl to Static" reading. *There was an article in the September 96 issue of Sport Aviation by Jimmy Tubs titled "Engine Cooling Problems" where he says "A significant pressure drop indicates that the bottom cowling is damming up the air causing lazy flow over the cylinders and oil cooler". *So what does a significant pressure drop mean? *We obviously do not have enough differential pressure over the cylinders, I have found numbers for the 320 and 360 and was told at one time that over 6" differential pressure is needed, but I want to attack the right end of the system. *No point in opening up the inlet if the outlet may be to blame. *The good news is I have the data, the bad news is I am struggling with the end interpretation of where to go.

By the way, the data gather absolutely proves that the turbo injector setup is more than doing its job! *In fact I think it is safe to say the we're exceeding the pressure on the nozzles of the standard setup.

Thanks in advance - Jason
 
Plenum to Cowl: 2.5", 4", 4"
Plenum to Static: 4.5", 6.5", 7.5"
Cowl to Static: 2", 3", 3"

The "plenum to static" value less the "cowl to static" should equal "plenum to cowl". It's within a half inch, so there is measurement consistency.

Your upper plenum pressures appear quite low at first glance, but you're flying test altitudes different from the standard routine we're using, and more important, you did not tell us TAS. So, not enough data to make a judgement.

Would you be interested in the installation of a standard piccolo tube setup, and flying a standard test routine? Your results would then be directly comparable.

Rough comparison; a data set from 1-29-12, for my RV-8, at 3500 PA. TAS is by the NTPS 3-leg method.

2enlq9i.jpg


Your "4000 PA at cruise" is the same velocity and density ballpark as my 3500PA at the two faster speeds...but your pressures are Upper 7.5", Lower 3", and a DeltaP of 4".

Mr.Tubs is certainly correct in that you must have the necessary delta to drive air through the fins. Pressure difference is how we make air move. However, it is only necessary to reduce lower cowl pressure (un-dam the flow, to use the Tubs view, i.e. open the outlet) when you are unable to generate good upper cowl pressures.
 
Last edited:
Some Additional Info

Dan thanks for the reply! Some more data to help out here, our 4000' data was at 156kts TAS and the 8000' data was gathered at 158kts TAS. TAS was read off the GRT box, not as a three leg run - so take it for what it is worth.

We are using a Dwyer Magnehelic gauge with a range of -10 to +10 inches of water. Our piccolo tubes are rather crude; I just punched .040 holes in 1/4" OD plastic tubing at random over a 4" length. I am connecting the gauge in the cockpit to the piccolo tubes using plastic tubing with a 5/32" OD. The plastic tubing is the same stuff that is used for pitot static installations, but obviously a different size. In the 10 it is difficult to route through the heater vents, so I am instead running the lines up through the top of the cowl where it meets the firewall then running the lines through the side vent of course all help in place with some speed tape.

The top piccolo tube is located at the engine spine, the lower piccolo tube is tied to the fuel line the runs between the servo and spider. The injector pressure is taken by disconnecting one of the cylinders and connecting the line directly to the line from the turbo nozzle manifold.

I would be happy to document a 'standard test' to contribute meaningful data, as long as we can live with the smaller lines running to the cockpit.
One of the things I have been trying to understand is what we want the value of the lower cowl pressure to be relative to static. We are seeing 2" to 3" is that a good number, should that number be higher or lower, should it be negative? I think that would tell me which end needs to be attacked right? My intuition tells me that I want it to be as low as possible since it is directly opposing the flow of air from the upper to the lower deck. However, I wonder if that number can ever get to zero since in theory we should have a vacuum sucking the air out of the exit. Great - now my brain hurts again!

Thanks again for the help and meaningful input!

Jason Kreidler ? 4 Partner Build ? Sheboygan Falls, WI
RV-10 - N44YH - #40617
Flying ~450 hours
 
...our 4000' data was at 156kts TAS

Ok, that confirms being in the same ballpark. Your slightly higher test altitude shouldn't lower available dynamic pressure more than 1.5%, maybe 2%, yet you have 7.5" plenum pressure at 156 knots and I'd have over 12". The maximum available dynamic pressure (Q) at this velocity and altitude would be a little over 14". Measuring the delta between upper plenum and aircraft static tells you what percentage of that available Q is being converting to increased plenum static pressure.

We are using a Dwyer Magnehelic gauge with a range of -10 to +10 inches of water. Our piccolo tubes are rather crude; I just punched .040 holes in 1/4" OD plastic tubing at random over a 4" length.....The top piccolo tube is located at the engine spine, the lower piccolo tube is tied to the fuel line the runs between the servo and spider.

You'll need a manometer with more range if you get your cowl working right. Switching to the "standard" piccolo tube setup which Sonny, Ken, and I are using would be nice.

The injector pressure is taken by disconnecting one of the cylinders and connecting the line directly to the line from the turbo nozzle manifold.

Can't comment. Perhaps a new thread for the injector discussion?

One of the things I have been trying to understand is what we want the value of the lower cowl pressure to be relative to static. We are seeing 2" to 3" is that a good number, should that number be higher or lower, should it be negative?

As high as possible while still maintaining the necessary delta between upper and lower.

I think that would tell me which end needs to be attacked right?

It's quite clear which end needs work. 7.5 / 14 = 53%...a rather miserable conversion of available Q to upper plenum pressure.

My intuition tells me that I want it to be as low as possible since it is directly opposing the flow of air from the upper to the lower deck.

Only if you can't improve the upper plenum pressure.

However, I wonder if that number can ever get to zero since in theory we should have a vacuum sucking the air out of the exit.

In theory we want an exit jet with a velocity as close to freestream as we can get. That requires pressure.

Try this example....if you have 7 in the upper plenum and lose 4 across the fins you have 3 available to push air out the exit. If you have 14 in the upper plenum and lose the same 4 across the fins you have 10 available to push air out the exit. Same cooling, but the second example has far less cooling drag.

Lets suppose you can't increase upper plenum pressure....you're stuck with only 7". Opening the exit area will indeed decrease lower cowl pressure. Let's say it is now down to 2". The result will will a 5" delta across the fins, so you get more mass flow. Cooling capacity increases, but the decreased lower cowl pressure slows exit velocity. You get a double whammy as punishment for taking the lazy man's path (a big exit). Cooling drag is mass x loss of momentum, and you increased both.
 
Last edited:
Wow has this discussion gone off the deep end. Still a couple questions come to mind. If we're measuring differential pressure, it really doesn't matter if we're using total pressure or static pressure, differential is differential, but how do we verify the accuracy of those readings under the cowl? If you use single point probes, proper orientation of the pick up tube is critical. To accurately measure total pressure the probe must face directly into a laminar airstream. If we're measuring static, the probe must be exactly perpendicular to the laminar airstream. Inside a cowling there is the potential for very high airflows, much turbulence, as well as areas of stagnation, and most importantly airflow striation. And this is keeping it simple. The use of a picollo tube, (multiple samples in a linear area) might help some with accuracy, but you might also find that the orientation of this tool will yield widely varying results. The most accurate method my simple mind can come up with to approach any relative accuracy would be to install some sort of vel grid device over the inlet and outlet of the area in question. I'd speculate it's already been done, who knows, but in reality, this is really not such an easy project to get right, numbers is numbers though, and I guess even if they're not accurate, what's it really matter anyhow?
 
I am starting to get it!

You're making good progress at educating me!

Meter range - I will search for the same instrument the rest of the team is using and try to pick one up. I would be really pumped to need range above 10" - it means I am making progress towards the solution.

Piccolo tubes - I can either make a set from your plans or if you would be so kind to supply I will send you my information.

I am worried that the 'fix' will not be as simple as just increasing the inlet diameter but will also include a significant increase in plenum volume. But we are a bit away from knowing for sure what the fix is until we have good data gathered.

Thanks, Jason
 
Wow has this discussion gone off the deep end.

Would you care to elaborate?

how do we verify the accuracy of those readings under the cowl?

At the moment we can't claim global accuracy between Jason's airplane and the airplanes in the test group. However, we have taken some pains to standardize the large area piccolo installation on all the test group airplanes. We even use the same manometer model. Jason is welcome to make the exact same installation. Identical installs and test flight plans go a long way toward making the data comparable within that fleet.

BTW, CR3405 explores several measurement methods and relative accuracy.
 
One of the more accurate methods for measuring a ducted airflow is a duct traverse. You take velocity pressure readings at equidistant points in the duct where the airflow is likely to be laminar, average the Vp readings, then calculate the total airflow. If you simply pick a spot, take one Vp reading at one point somewhere in the duct, you'll essentially have nothing reliable. You'll have varying results if you take a reading of the airflow along the walls of the duct because of friction losses, read air moving around corners, flowing around obstacles, entering or leaving a transition, turbulent air, etc., worthless data. True, velocity and pressure are inversely proportional, but how do you take reliable readings in turbulence? That's why you take readings at equidistant points in a clean straight piece of ductwork, and average the readings. The one positive thing to look at is that in a constant state of flow in a sealed duct, total and static pressure in the duct will remain relatively stable while the quantity of air is moving through the duct, but the challenge is to find the spot that is undisturbed that reliably represents true conditions. That brings us back to measuring differential pressures under a cowl. Unless you can find a way to reliably measure the airflow before it is disturbed, as well as measure it relatively undisturbed as it's being discharged, you'll have an issue with reliable readings. You could try installing pick ups at exactly the same positions under each of the cowls, but unlike laboratory conditions, exact placement would be near impossible because no two installations are exact duplicates. Garbage in, skewed results. There is a possibility? One might try a couple automotive style mass flow sensors (hot wire annometers), but once again? Who knows. We'd be trying to measure a huge amount of airflow with a device that was not designed for the job. I do know that trying to install single point pick ups, especially if you're tryting to duplicate installations for multiple data acquisition is a little bit of futility if you're looking for any kind of accuracy. The piccolo route might be as accurate as needed? You would need multiple locations for any real accuracy, and be very precise in the duplicate installs.
 
Pressure recovery

Dan

Have you noticed any change in plenum pressure as you modified the outlet? Thinking about one of the objectives of testing, determining pressure recovery characteristics of various cowl configs, it would seem that how much air is bled out through the outlet would be a factor. That is, seal the outlet (and spinner) and almost any inlet would have great pressure recovery.

Conversely, if the bottom cowl is full of big holes, you might not ever see decent pressure recovery. At least not without a maw like a gray whale.
 
Wow has this discussion gone off the deep end. Still a couple questions come to mind. If we're measuring differential pressure, it really doesn't matter if we're using total pressure or static pressure, differential is differential, but how do we verify the accuracy of those readings under the cowl? If you use single point probes, proper orientation of the pick up tube is critical. To accurately measure total pressure the probe must face directly into a laminar airstream. If we're measuring static, the probe must be exactly perpendicular to the laminar airstream. Inside a cowling there is the potential for very high airflows, much turbulence, as well as areas of stagnation, and most importantly airflow striation. And this is keeping it simple. The use of a picollo tube, (multiple samples in a linear area) might help some with accuracy, but you might also find that the orientation of this tool will yield widely varying results. The most accurate method my simple mind can come up with to approach any relative accuracy would be to install some sort of vel grid device over the inlet and outlet of the area in question. I'd speculate it's already been done, who knows, but in reality, this is really not such an easy project to get right, numbers is numbers though, and I guess even if they're not accurate, what's it really matter anyhow?


If we were trying to meet a pressure recovery spec or some other critical desgin criteria I would agree having ABSOLUTE accuracy would be important. Our testing (Dan, Sonny and I) was concerned more about consistent and relative data. We each used the same measuring device, picollo tubes (Dan fabricated for us) and located them in the same locations. Our standardized testing methodiligy resulted in consistent results from test-to-test providing credibility to the "relativistic" nature of our testing. Once we were convinced we could develop repeatable data throughout our various flight regimes (airspeed mainly) we investigated configuration changes (mostly exit area) and the relative affect of those changes to the pressures in question.

Within the scope of our budget (both $ and time) it was very informative and I'd stand-by our data as being as credible as required to understand, in general, what is going on. Could it be better - you bet and even with $2 million and two years of testing we could still have questions needing answers... This isn't NASA and we are not solving problems to the 12th decimal place... but you are welcome to do so and share your data with us!! :)
 
Just curious...has anyone considered making the ramp out of fiberglass? Seems like that might be easier given the curves and interference.

Scott
 
Scott, aluminum would be easier if the goal is a simple radius on the bottom of the firewall. Jury is still out regarding the worth of a more elaborate converging exit duct. Building one in aluminum requires some hammerforming and a stretcher, so glass may be a worthwhile option.
 
Back
Top