What's new
Van's Air Force

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

Boundary Layer Separation

DanH

Legacy Member
Mentor
Just one of those fun little details spotted during my annual. Up under the wing with the panels off, and noticed the paint on the pitot mast was looking scruffy. No proiblem...remove it, sand it, repaint it.

This is a shop made streamlined mast, an epoxy/micro airfoil cast around a standard 1/4" Vans tube This thicker glass/epoxy sheath slides up to cover the nut.

Looked closer back at the shop. Tell me if I'm wrong, but it looks like the paint was popping off right where the boundary layer was separating.
.
 

Attachments

  • Pitot Mast Right.jpg
    Pitot Mast Right.jpg
    120.5 KB · Views: 475
Trip

Looks like you need a strip of tape the trip the boundary layer a little early. Could be evidence of a separation bubble?
 
Since it appears to have the separation further forward toward the top where the shape is fuller, I would assume separation at higher speeds. Time to streamline it a bit.
I would think vortex generators would be going the wrong direction.

Did you change the shape when you repainted?
 
Boundary layer

If you are having separation, I think the shape is wrong and you are getting to much turbulence, which causes a decrease in speed. I don’t have a clue, it’s to cold in Texas to think.
 
I think I'd agree with John S. Worth noting is that the Reynold's Number there is relatively low, even at the speeds you fly. But for this one, we need Steve.

Dave
 
Hmmm, this is a puzzle.

From what I can see in the pictures, it really doesn't look like a shape that would separate.

A laminar bubble wouldn't normally have enough energy to "popcorn" the paint off - and I wouldn't really expect separation to either.
If it was predisposed to popcorn because of very low adhesion, then I suppose a laminar bubble might do it. The surface doesn't really look smooth enough to support a laminar boundary layer in the first place, but at low Reynolds no. perhaps it is laminar in the front half.

Dan,

Can you take a picture from an end-on view so I can get a better idea of the cross section shape?
 
My very rough calculation, estimating chordwise dimensions from DanH's finger, has Rn at the damage varying from about 6x10^4 near stall to 2x10^5 near top speed. Even if the surface was smoother than it looks, these are probably a bit too high for conventional laminar separation bubbles.

If the coating's peel strength was a problem overall, I think it would be more likely to be damaged nearer the leading edge of the body, especially if high sideslip variations were routinely flown.
 
I suspect a non-aerodynamic reason for the paint blistering pattern. Maybe skin oil from holding it by that corner?
 
Looks like you need a strip of tape the trip the boundary layer a little early. Could be evidence of a separation bubble?

second that

Golf balls, which operate in the same Reynolds regime need some help too. Invigorate the bl early. Either make little things thin or stimulate to avoid bl separation. Thx Dr. Horner
 
Extend discussion

So if we extend this discussion to other parts of the RV, do the gear leg fairings have the potential to have BL problems, or is the reynolds number too high?

Just thinking out loud that maybe we need to trip the BL on the gear leg fairings?
 
A quick analysis of a similar mast configuration

My first thought was to agree with Steve's opinion on a non-aerodynamic cause.

I did a quick model of a similar mast attached to a stub wing with flaps down, one at 4 degrees alpha, 0 degrees sideslip, and one 4 degrees alpha and 10 degrees sideslip.

The model...
config.jpg

First image is right side at zero sideslip

probea0right.jpg

Left and right view at the 10 degrees sideslip.

probea10right.jpg

probea10left.jpg

In summary, the suction forces on the mast are pretty benign.... maybe the wear is actually due to some abrasives carried in the flow during takeoff/landing? Mr. Horton flies his aircraft quite a lot I'm guessing....

Just my 2 cents...
 
Hmmm, this is a puzzle.

Yes, I found it curious. Look at how the streamlines ended at the paint loss.

I suppose it could be a "chicken or egg" thing, where laminar flow transitions at a paint failure which happened for some other reason. However, I'm not known to contaminate a surface I'm about to paint. The lower mast was painted at the same time, assembled; one would think any contamination would show up on both components. The finishing methods are the same as every other part on the airplane, the only exception being it was painted with a rattle can.

Can you take a picture from an end-on view so I can get a better idea of the cross section shape?

Freshly repainted below. The mast itself is quite slim (1.75" chord, 0.3125" thick), but the upper sheath starts at 2" x 0.4375" and transitions to 3" x 1.125" where it meets the wing. I was trying to keep in under 3 to 1.

Back when I built it: https://vansairforce.net/community/showthread.php?t=68242
.
 

Attachments

  • Mast1.jpg
    Mast1.jpg
    68.1 KB · Views: 78
  • Mast2.jpg
    Mast2.jpg
    53.4 KB · Views: 82
  • Mast3.jpg
    Mast3.jpg
    89.1 KB · Views: 79
  • Mast4.jpg
    Mast4.jpg
    63.3 KB · Views: 82
Maybe it's not separating, just transitioning from laminar to turbulent.

Yes, and I may have mis-used the terms.

....it's not a heated pitot tube is it?

Nope, cold, epoxy / glass.

Since it appears to have the separation further forward toward the top where the shape is fuller, I would assume separation at higher speeds. Time to streamline it a bit.
Did you change the shape when you repainted?

Certainly the thickest section. Can't change it without a redesign, need internal clearance for the B nut.

Dan, how much rain has that pitot mount seen?

I'm not known to be bashful about a little water, but....

If the coating's peel strength was a problem overall, I think it would be more likely to be damaged nearer the leading edge of the body...

My thinking too.

Ultrasonic damage from the 50 cal.

Nope, the paint comes off the other guy's airplane...;)
 
A strip of zig zag tape at around 20-25% chord might get a fresh-enough turbulent boundary layer to hold on whereas a natural transition might not happen before a laminar separation, as others have suggested.

Even if the flow is separated, it is still less drag than the bare pitot tube and B nut hanging out there.

That upper sleeve is pretty thick. And it is fairly convex, with max thickness pretty far back and keeping its thickness pretty far aft so that the trailing edge closure angle is very steep. Take a look at NACA 4-digit sections in Abbott and VonDoenhoff. e.g NACA 0024. pg 327 in my Dover edition.

Not suggesting you remake it, just suggestions for the future. Unless....

You could scab on some additional material on the back half to increase the chord length, keep the max thickness the same, and shift the max thickness forward a little. Just more of your microballoon casting added to the back to extend the airfoil shape.


All that said, it is remarkable that the unsteady flow at 200mph could damage paint like that. Transonic, sure. But at our speeds, there isn't really enough energy in the flow. I wonder if there were trapped air bubbles from pinholes or something that made the paint fragile?
 
Dan,

Is it possible you had some moisture from wet sanding not completely dried out from the outer layer of shaping compound/filler?
It can take weeks to fully dry out.
 
You could scab on some additional material on the back half to increase the chord length, keep the max thickness the same, and shift the max thickness forward a little. Just more of your microballoon casting added to the back to extend the airfoil shape.

Thanks Steve, I'll put it on the list. And I see what you mean about the profile.

All that said, it is remarkable that the unsteady flow at 200mph could damage paint like that. Transonic, sure. But at our speeds, there isn't really enough energy in the flow. I wonder if there were trapped air bubbles from pinholes or something that made the paint fragile?

I dunno. It just struck me as interesting, and generally speaking, if the smart people around here also find it interesting, I learn something new.
.
 

Attachments

  • IMG_20230131_143212097.jpg
    IMG_20230131_143212097.jpg
    55.4 KB · Views: 99
Thanks Steve, I'll put it on the list. And I see what you mean about the profile.



I dunno. It just struck me as interesting, and generally speaking, if the smart people around here also find it interesting, I learn something new.
.

Learning something new……. me too. I have the same 1/4” pitot tube on my RV6 and I have been thinking about incorporating something like this:

https://antisplataero.com/products/universal-streamline-fairing-14

- along with a small shop made small fairing for the B-nut, but wondering what the benefit is. It might look a little better, but it’s under the wing, so not that visible. Does it gain me one knot? I doubt it. Maybe if I applied the same mini fairing to all three of my antennas, that would give me a knot or more cumulatively….. something to experiment with. No big gains here. The positives have to outweigh the negatives, but all of these are cheap.
 
Thanks Steve, I'll put it on the list. And I see what you mean about the profile.



I dunno. It just struck me as interesting, and generally speaking, if the smart people around here also find it interesting, I learn something new.
.

Another way to do it would be to make a short bump fairing for the B nut like what people put over their fuel drains, then have the thinner blade section cover the full length of tube.
 
Another thing you could try for drag reduction is to just shorten the vertical part of the probe. It only needs to be clear of the wing boundary layer, which is fairly thin at that point under any flight condition where you're likely to be concerned about airspeed.

The Van's tube-type total pressure probe is fairly insensitive to angle of attack, and is unlikely to be affected significantly by wing proximity. To reduce its AoA sensitivity further, you could modify the probe tip using something like the example below, from one of Gracey's many NACA publications on the subject.
 

Attachments

  • NACA-TN-2331-TubeA7.png
    NACA-TN-2331-TubeA7.png
    96.6 KB · Views: 75
Note the exemplar probe shown is 1.00" diameter. That's a lot of cross-section to incorporate into a drag-reduction effort.
 
If a fella wanted to have some fun, how would he visualize the surface flow experimentally? Oil dots, powder, yarn, what?
 
Learning something new……. me too. I have the same 1/4” pitot tube on my RV6 and I have been thinking about incorporating something like this:

https://antisplataero.com/products/universal-streamline-fairing-14

...

Note the old film near the end of the video on the antisplataero.com website.

https://youtu.be/mX74jP3bKw0?t=378

So counter intuitive to me that adding all that streamline area (friction) to a round tube would reduce overall drag. Yet that old MIT film clearly shows the difference. 9.3 times! Amazing!

Edit: So what is the drag of 5" of 1/4" tube and how many HP does it take at 200 mph?

Finn
 
Last edited:
So cutting the vertical section of the pitot in half would cut the drag in half. How long would this section of the pitot need to be in order to function properly?
 
Note the old film near the end of the video on the antisplataero.com website.

https://youtu.be/mX74jP3bKw0?t=378

So counter intuitive to me that adding all that streamline area (friction) to a round tube would reduce overall drag. Yet that old MIT film clearly shows the difference. 9.3 times! Amazing!

Edit: So what is the drag of 5" of 1/4" tube and how many HP does it take at 200 mph?

Finn

Wow, that video is eye opening. Guess I’ve got some streamlining to do when my bird is home from the paint shop.
 
The video illustrates a round object 90 degrees to the air flow, most antennas are swept back where the round element changes to more of an elliptical shape to the airflow, significantly reducing the drag.
 
If a fella wanted to have some fun, how would he visualize the surface flow experimentally? Oil dots, powder, yarn, what?

Seems like I remember Dave Anders writing that he used coloring crayons dissolved in lacquer thinner or some such solvent on his landing gear struts to make sure they were inline with the airflow. That is, he adjusted the angle until airflow around both sides met at the trailing edge without separation along the chord.

Perhaps solvent (not oil!!!) dissolved crayon would help to visualize the airflow you are interested in. Or perhaps Dave Anders has already solved the problem you have raised! (P.S. I do not think Dave Anders would put something on his beautiful and fast RV-4 that is difficult to remove!)

And perhaps my memory is like the quotation from Col. Kittinger's funeral, "Old age is when you remember more clearly the things that never happened!"
 
Last edited:
If a fella wanted to have some fun, how would he visualize the surface flow experimentally? Oil dots, powder, yarn, what?

Way back in the olden days I used oil dots in my aero work on F-5000 cars. Not the best but we did make progress.
 
If a fella wanted to have some fun, how would he visualize the surface flow experimentally? Oil dots, powder, yarn, what?

There are several variations of oil flow techniques that can work well, or be frustrating.

If the object is white, I love to use dirty crankcase oil. The suspended carbon particles are super small and the oil flows smoothly.

For dark objects, you want a white-pigmented oil. Many will get powered tempera paint from the art store and dissolve that in 30 wt oil. But it doesn't dissolve very well, and the particle size is too large. A few drops of oleic acid will act as an emulsifier and help somewhat. Once mixed the best you can, strain it through a coffee filter. What then looks like white paint will work ok, but when it flows it leaves 'grit' behind that may or may not influence the flow.
I have seen others obtain titanium dioxide pigment that is made for paints, and mix that into the oil. It is finer than the tempera paint power, dissolves pretty well.

I would NOT try crayons melted into oil. crayons are mostly wax, and the mixture will probably cool and stick and be very hard to clean off.

In any case, slather the oil on, go fly a circuit, then land and jump out and take pictures.

Another variation we use a lot in wind tunnels (where you can turn the lights out) is to use oil with a fluorescent dye in it. Conveniently Aeroshell 100 fits the bill nicely. So a black object works well, and you might get good results at night. Slather it on, go fly, taxi back to an assistant with a hand-held black light and a camera.
 
So counter intuitive to me that adding all that streamline area (friction) to a round tube would reduce overall drag. Yet that old MIT film clearly shows the difference. 9.3 times! Amazing!
Finn

My dear old friend and colleague, RT Jones, used to have a plaque on the wall behind his desk. On it was a 1" segment of 1/8" diameter wire, and next to it a 1" thick slice from an airfoil. It was a NACA 0010 airfoil section, with a 1" thickness, and a 10" chord. The plaque noted that the drag of the wire and the airfoil were the same!
 
My dear old friend and colleague, RT Jones, used to have a plaque on the wall behind his desk. On it was a 1" segment of 1/8" diameter wire, and next to it a 1" thick slice from an airfoil. It was a NACA 0010 airfoil section, with a 1" thickness, and a 10" chord. The plaque noted that the drag of the wire and the airfoil were the same!

Steve could you give me an idea of how many HP it takes to drag a 5" segment of 1/4" tube through the air at 200mph?

Finn
 
Note the exemplar probe shown is 1.00" diameter. That's a lot of cross-section to incorporate into a drag-reduction effort.

It's the shape that's important, not the diameter. I've used similar shapes on probes as small as 2mm dia, with similar effect.
 
So cutting the vertical section of the pitot in half would cut the drag in half. How long would this section of the pitot need to be in order to function properly?

Cutting it in half would be OK, except maybe in inverted stalls :). This would reduce power required for level flight by very roughly 1/4 HP at max speed (if my conversions between SI and US customary units are correct). Note this is only applicable to the simple tube pitot probe, not to anything with static ports, AoA ports, or more complex shapes.
 
Another way to do it would be to make a short bump fairing for the B nut like what people put over their fuel drains, then have the thinner blade section cover the full length of tube.

Something like this, on the right?
.
 

Attachments

  • Pitot.jpg
    Pitot.jpg
    24 KB · Views: 107
A test exercise

This discussion is all about learning, so I thought I would add a bit to my previous post...


First, a caveat... the results I am presenting are from a simulation that is inexact... but, I am kind of comparing apples to apples, so the relative results tell us something...

I analyzed a few probe mast configurations, using different cross-sections, aspect ratios and Reynolds numbers.


1: A simple tube
2: Blunt airfoil mast representing what one might make to be robust in the scale we need to make these masts
3: A symmetrical laminar section

I ran each model at different Reynolds numbers to 'scale' the model to the appropriate conditions the probe would see on our RVs.... For instance a 1.5 inch chord probe at 150 knots is running at an Re of about 196000. If it was a full scale wing sized object... it would be about 4500000.

Here is an image of each model at Re 200000. Three models have an aspect ratio of 14.5 which would represent a mast of chord ~1.0 inch by ~ 14.5 inches. ( A really long probe) and the other three at AR 7 which would be closer to a real probe to see what difference the AR has on the drag.

The Tube (First image AR 14.5, second AR 7 )
tubeRe200000.jpg tubear7.jpg

This model had a 3D Cd of 0.0464 at Re 200000
Cd of 0.0341 at Re 200000 AR 7
Cd of 0.0189 at Re 4500000


The Blunt foil (again at AR 14.5 and 7)
bluntRe200000.jpg bluntar7.jpg
This model had a 3D Cd of 0.0279 at Re 200000
Cd of 0.0278 at Re 200000 AR7
Cd of 0.0181 at Re 4500000


The symmetric foil
symRe200000.jpg symAr7.jpg
This model had a 3D Cd of 0.0130 at Re 200000
Cd of 0.0133 at Re 200000 AR 7
Cd of 0.0086 at Re 4500000


If you look closely at the cylinder, you see a large 'red' zone at the trailing edge . This indicates stagnated flow over a large area so I REALLY doubt the absolute value of the tube drag number since the boundary layer code was not able to accurately model the separation region. (I am working on this)
Its probably better to fall back on first principles and grab the flat plate Cd
from a bible like Hoerner's Fluid Dynamic Drag instead..... Also, the drag numbers are nonsensical at the higher Re of 4500000. The boundary layer code just cant cut it on the cylinder.

Hey Wait a minute! A tube is going to have a much larger aspect ratio in the real world. So a 1/4 inch by 7 inch tube will have an AR of 28, and a reduction in Re to 50000....

What does that look like?

tubAr28.jpg

Well, the Cd at 50000 is about 0.0727 (don't believe it) but its still trending the correct way..... higher at lower Re., and the the boundary layer is still messed up but you get the idea.


In relative terms though, you can see the effect of cross section on transition to turbulent (The dark line running along the span) and the relative increase in Cd between shapes.

In real world numbers, the drag of your mast might be for the symmetrical mast:
Drag in LBF = 0.5*Cd*Rho*V*V*S / G
Drag = 0.5 *0.0133 * .0752 LB/FT3 * 230 F/S* 230 F/S *0.0486 FT^2 / 32.2
Drag = 0.0399 lbf Symmetric foil.

From Hoerner's drag book, a cylinder at Re50000 has a Cd of about 1.2, so the drag of a 0.25 inch x 7 inch tube would be 0.9 LBF or 22 times the symmetrical shape.....

In summary, streamlined is better. Aspect ratio doesn't make a lot of difference.... probably because we are generating very little lift force.... I wish we could have been able to model the tube better...

And...that's not a lot of drag force...but it all adds up..

Hope this was as much fun to read as it was to prepare...

Chris
 
Last edited:
Back
Top