Alan Brown, an English gentleman who is very knowlegable in aerodynamics gave a great presentation on the relationship of "wing chord, tail moment, and aerodynamic center" at the Salinas Valley Modelers club meeting last night. Upon finishing his "Lesson" he responded to a member asking about the less obvious causes of stall, and commented on the effect of a flat stabilizer w/ a square leading edge often found on ARF aircraft. His main point was that it caused the stab to stall sooner because of seperation from the un-aerodynamic L.E. This would lead to "Pitch Up", and aircraft stall.

While there are other factors that may play a larger role in an aircraft's tendency to stall such as CG or wing loading, it got me thinking about the importance of covering all the bases in my goal to fly safer and more accurately. I am presently assembling a Seagull PC-9, and will build a symmetrical airfoil stab to replace the original flat style stab.

Any feedback and experience regading this matter from Ya'll is appreciated.

Sincerely,

Bill

Bear in mind that any airfoil when it stalls does not loose all it's lift. In fact most don't even loose lift but continue to increase lift despite being stalled. What does happen in all cases is that the drag from the stall increases a LOT and it's that drag that slows the plane down so far that it suddenly loose lift due to the loss of airspeed.

But a flat plate may well be one of those that suffers loss of lift when it stalls.

Just a note...

A stalled horizontal stabilizer results in a rapid pitch DOWN of the airplane...not UP.

Recovery from a tailplane stall is the exact opposite of the technique used to recover from a stalled main wing. Moreover, a tail stall is perhaps one of the least likely things you would ever experience in modeling, unless you fly in icing conditions which we don't usually do.

As for your PC-9 stab? Aside from the aesthetic appeal of a nicely shaped stabilizer, I wouldn't expect to see miraculous improvements in the models performance. Some improvements perhaps, but not dramatic.

Oh, and let's not forget that a flat stab and elevator combo isn't really a flat airfoil any more when the elevator deflects. It ends up being a cambered airfoil with a really bad shape...

Thanx for the responses!

Note: I was incorrect in stating that the PC-9's stab had a "Square" leading edge. It is rounded.

I considered that the amount of effect a flat stab with square L.E. has on model aircraft would be small, but hey, it's all about obsessing, right?

I'm still not sure that the stab stalling before the wing wouldn't produce pitch up. They are both creating lift while in a positive AOA. If I'm all wet, feel free to tear me a new one.

Cheers,

Bill

In a conventional aircraft layout (stab behind wing, engine up front...) the horizontal stab's function in level flight is to push down on the tail and not to produce "lift." This is assuming that the aircraft's center of gravity is forward of the neutral point on the wing. If the center on gravity is behind the neutral point, not only will the stab need to lift the tail, but the pilot will also likely need a diaper.

In a conventional aircraft layout (stab behind wing, engine up front...) the horizontal stab's function in level flight is to push down on the tail and not to produce "lift." This is assuming that the aircraft's center of gravity is forward of the neutral point on the wing. If the center on gravity is behind the neutral point, not only will the stab need to lift the tail, but the pilot will also likely need a diaper.

Absolutely, but the situation I'm referring to is not ( AOA ) level flight. My whole thinking about this is probably moot however, as I've made the classic mistake of picking one component and trying to link one reaction to it, when there are numerous variables that will either add to or counter the effect. Re-reading the previous replies made me realize this. For instance, BMatthew's point that a stalled surface doesn't immediatly stop providing lift, but does greatly increase drag would give creadance to Race City's notion that the result would be pitch down, etc.

In the words of Gilda Radner, " Well... That's different. Nevermind".

Bill

Thre are a lot of full scale and models with flat stabs -- other than needing bracing to keep from going the wrong way when the elevators are deflected -- they work just fine
for years I felt the nice curved stabs were superior - and for being "clean" yes they are -however- as far as how the plane flies-- I will bet money that is a carefully done comparison test - you can't find any differences --in models such as sport models or aerobatic stuff -

A stabilizer with an airfoil profile may be more aesthetically pleasing, and may help a little when you have a highly-deflected control surface, and does give a slightly-softer feel around neutral. As far as the vast majority of models are concerned, though, the "plank" style of control surface works just fine, and is MUCH easier to construct. The main consideration is to make sure the total area and the amount of fixed versus movable surface are correct for the job the airplane's intended to do.

Thanx again for the replies, as this has been educational for me.

One thing I should point out is that Alan Brown's main complaint with the flat constructed stabs was not that it was flat as much as that if the leading edge wasn't rounded, but square, that it would cause seperation at an earlier angle of AOA, possibly causing the tail to stall before the wing. I'm sorry I didn't make a better point of this earlier. As I discovered on my PC-9, the stab's L.E. is rounded, and I've decided to use it after all ( The realization that I'd have to buy 3 rolls of Oracote for the stab might have played a roll as well.. Ya Think? )

Cheers,

Bill

Good subject - good responses - very good thread!

paul

Flat stabs even with rounded leading edges suffer from shallow angle of attack flow separation anyway. There's no doubt that the square edge would make it a bit worse but by how much is something that you'd need to learn in a wind tunnel.

Well, this is not the first time I have seen this topic here. I'm a poor student so constructing adequate models and investing in the flow visualization and measurement equipment is out, but I do have access to a computational fluid dynamics (CFD) package called FLUENT. This is an extremely powerful tool for fluidic design as it allows for not only flow visualization but also measurement of forces/coefficients acting in any direction you would like.

The setup:
I tested a "square" stab with a length of 5" and a thickness of 0.25". This seems to be fairly standard for most model stabs. The model velocity was set to 60 mph at standard temperature and pressure. I varied the angle of attack from 0 to 11 degrees and mapped the associated drag coefficients. The attached images are at angles of attack of 0, 5, 7, 9, and 11 degrees from left to right. The same setup was then used for the same stab with full rounding of the LE and TE.

Results:
The flow separation on the square stab starts at or before 5 degrees and becomes much more severe as the angle of attack is increased. (visible by the lighter blue near the leading edge). Also, as expected the square stab has vortices that are created and shed off the TE (I would suppose that future work would be placed on analyzing the frequency of the shedding and seeing if it is in the audible range which would be pretty cool!). The round leading edge has a huge impact on the separation progression as well as the shed TE vortices. They are actually not visible in these shots because they are extraordinarily small and occur primarily in the boundary layer. I have also included a graph of the drag coefficient vs. the angle of attack. The trendlines are not totally accurate as more data is needed as well as real model validation. As is evident from the graphs the rounded leading edge produces a drag coefficient that is 50% smaller than the square edge.

M, could you repeat the square leading edge with the rounded trailing edge....

Er... isn't is usually the other way around Paul? Rounded LE and square TE?

Either way that's an amazing outcome. I sure would not have thought that it would be so dramatic a difference for the rounding of the LE.

-Paul
Sure, here it is. As expected the drag is lower than the purely square stab, but the the separation is still present at very low angles of attack. All in all I think this has made the case for at least rounding all surfaces for the best preformance.


I have included some other images generated by FLUENT. These are pictures of our SAE AeroDesign aircraft and of particular interest is the ability to predict the position and magnitude of the wing tip vorticies. We used this data to design our winglets and were able to test it prior to committing to a prototype. Fun stuff. As I am still only scratching the surface of what this program is capable of doing I welcome all opportunities to do CFD analysis for anyone who might have an interesting problem or is just after some baseline drag data. I have done things such as characterized turbulator placement by observing the exact location of the separation point at various angles of attack. This program does have 3-D capability, but the meshes are significantly harder to generate and the processing time goes up exponentially!! (the SAE aircraft took 178,000 iterations over almost three days[X(] ) So if there are any other flow mysteries or challenges I welcome you to let me know!!!

-Matt

Whoops, here are the other pictures. The rocket nozzle is one of the designs we are working on for the Air Force and NASA. It is about 1.2mm wide at the exit plane and will be placed on some of the new micro-satellites being developed.

There is a fair chance that the modeling shown is more appropriate for our models than not. It's been accepted that the flow we experience in our models at their low Reynold's numbers is more correctly described with fluid dynamics than by other systems.

There has been a running battle (war?) within these forums about not tapering our TEs and about squaring them off in order to cure flutter problems. Seldom does that war deviate into discussing what extra drag having a square TE will cause. It'd be interesting to see that test with a rounded LE and square TE to compare with the round-round. It'd be interesting because since the war revolves around the TE, it'd be interesting to see the flows from a rounded TE versus a square TE when the flow to each is relatively uninterrupted (not pre-screwed-up so to speak).

mhm21,
Those are really kewl pictures. It's an excellent thing your sharing them with us. I hate to ask a favor, but......

Is it possible to describe your test profile as having a straight taper to the TE? Most of our stab/elevator and stab/rudder planforms have a flatplate stab with a rounded LE and with the elev/rudder surface having a straight taper to the TE. Usually the TE isn't "sharp" (It's thickness isn't 0.). It's usually a square TE, but is most often less than half the thickness of the stab and the LE of the elev/rudder.

If not, then thanks for sharing and for answering the extra request above. It's been informative.

.
There was a method in my madness...
The old "trailing edges MUST be square" myth.
Rounded shure looks good, dunit?

Just got a chance to get back to this thread.. Wow, I didn't know I'd stir something up.

Thank you Matt for taking the time to run your tests and share such great info with all of us. Even with the limited input data, I think you have shown us that there is some benificial info in the results.

I like the replies regarding trailing edge. I never considered it's affects, or knew there were different schools of thought on it's design. That made me take a look at the PC-9's stab again ( it's elevator's chord is slightly tapered, but finishes in a flat trailing edge ) and a closer inspection of the whole kit in general. The more I look, the more I realize just how many subtle things I am oblivious to, or take for granted.

Cool stuff!

Bill

Hey Matt, if it's not too much trouble how about running these again but this time with the more "standard" rounded leading edge and square trailing edge? Something that I'm not sure you can model with the software is the tendency to flutter. One reason often given for squaring off the trailing edge is that rounded ones tend to develop a stronger flutter while square ones have more drag but also a highly reduced flutter energy situation.

This is an extremely interesting thread thanks to your computational fluid vews. VERY interesting. And a hearty THANKS! for taking the time to run these tests!

In reading the chart --
it looks like the differences from AOA are more linear in the square edged example.
Most of my flat plate square edged stuff flies at under 25 mph so I would suspect variences would be lesser in any squared edge / rounded TE etc..

Hey Dick, do you just leave your leading edges square?

I'm putting together my first Depron flat foamie 3D model and like the detail nut I am I rounded the LE. Sure hope I didn't ruin the sudden stop charactaristics....

Yes - at the speeds involved -I can find no detectable differences in sharpening/rounding etc.
The real differences are in making the thing reasonably rigid . It is a real eye opener on "what really makes a difference" when one can fly a model right in front -up close and personal whilst watching things bend or twist.
The wings which simply flex-cause no problem as long as it is a straight up/down flex -
The other thing is proving how much the CG can be shifted at these low wing loadings . The slower one flies the more the cg can be shifted . this also changes how much control input is needed.
If you approach foamies with an open mind (as you will) and simply look to see what really matters on super light , high powered , slow speed craft - you may find some nice surprises. At this point we must have done at least a hundred of these and still find them very intersting
Our latest is a full contoured fuselage bipe ( stock kit ) which tho a bit faster , works very well as an outdoor flyer - stable smooth very aerobatic no bad tendencies and the wings ? flat squared off foam on all surfaces