Impact Stab Analysis

[Howard Rush]

Summary

I used XFoil to calculate characteristics of this peculiar airfoil for seven elevator positions.   The data show that the boundary layer transition jumps abruptly from the stab LE to the crack.  A limited test shows that Reynolds number effect is puny. 


Method

I used XFoil via Profili Pro.  I eyeballed a stab drawing to pick something close to the mean aerodynamic chord of the original Impact stabilizer and elevator.  I picked a place where chord was 7 inches, which is close enough.  I used Bill Lee's Reynolds number calculator to find the Reynolds number for a 5.3-second lap time on 70-foot lines on a sea level standard day.  XFoil rounded it off to 307K.  I assumed incompressible flow (Mach = 0) and picked 6 for the XFoil Ncrit value on the advice of real aerodynamicists. I calculated my own airfoil shapes with deflected elevator, rather than using Profili's flap function.   XFoil barfed when I tried to replicate the actual flap hinge geometry on wing airfoils, so I prophylactically faired the stab to the elevator with straight lines between tangents to the stab TE and elevator LE or to the point where the LE or TE curve transitioned to a straight line.  It appears that XFoil or Profili went on to do additional smoothing.  This smoothing may have perverted the airfoil shape some.  I entered one point every three degrees around the stab LE and TE and the elevator LE, but no points on the straight line segments between these curves.  The spline function may not have felt sufficiently constrained by these straight segments.  I did not close the 1/32" elevator TE thickness on the files I entered into Profili.  XFoil gave warnings that it did not converge, but went ahead to calculate most conditions.

I have difficulty communicating with Profili, as I do with people.  Sometimes it refuses to plot some conditions, often for cause.  Many of these refusals seem to be for no aerodynamic reason.  The 25-degree elevator plots, for example, are truncated.  The plots for elevator deflection up to 10 degrees look pretty complete, but those for steeper deflections are weirder.  These are ugly airfoils, so I guess we shouldn't expect much.


Results

Igor Burger and Frank Williams have written about movement of the laminar-turbulent boundary layer transition as a cause for stunt planes "hunting".  I was interested in boundary layer transition flying level right side up and upside down and in trying to figure out why the Impact's peculiar configuration of zero stab incidence, a flat stabilizer, and elevator downrig works so well at preventing hunting.  In particular, I was interesting in extrapolating its nice characteristics to electric stunt planes.  The first attachment shows the airfoil and the elevator positions I tested.  Next is lift vs. angle of attack.  Behold that the stab probably spends most of its time at negative alpha.  Next is moment coefficient (of the stab only, referenced to stab area and chord).  Then is transition point.  Xtr is the fraction of distance from the LE of the stab to the TE of the elevator.  The hinge line is at .6. 

Some folks were speculating that a low-aspect-ratio stab would benefit from having a higher Reynolds number than a longer, skinnier stab.  I ran the five-degree-elevator (neutral for the Impact) case at three different Reynolds numbers.  Lift coefficient and transition plots are shown.  Looks like you wouldn't pick a stab chord based on Reynolds number. 

These data might be useful if we knew at what alpha or lift coefficient the stab operates in various tricks.  Figure that out and let me know.

 

[Dave_Trible]

Howard,  thank you for the interesting work.  The visual I get is that the elevator may be sort of swimming around a few degrees in the wash of the stab and needs to deflect to a place where it is presenting a surface more cleanly to the airflow.  What do you think Profili would do with a stab and elevator of equal thickness and little elevator taper until at least the last bit.  It would appear to provide pressure with nearly no added deflection. Do you think the required droop is something of a function of the difference in stab and elevator thickness?  I am a little surprised that landing gear drag doesn't offer enough forward pitching moment that induced moment is needed.

Dave

[Serge_Krauss]

Howard, that's really interesting! Thanks for doing the work. Is your angle of attack taken from a chord from the leading edge of the stabilizer to the trailing edge of the elevator or just for the stabilizer? Also what causes the stabilizer lift to jump suddenly or vary some at 20- and 25 degree deflections - would that be felt at the handle or seen?

I think I remember Igor mentioning changes in optimal stabilizer/elevator thicknesses for different chords at really low RN. TR 586 doesn't really show what he saw. I do remember his discussing sharper leading edges for the thinnest sections, but I don't know whether he settled on that. I think he did some computer modeling, but don't have time to search. Even though you were concerned with Impact qualities, If you venture further afield, it would be interesting to see if results are any different for a section like the one shown below, which operates at RN's closer to 210K. It's tapered with flat top and bottom except for the leading edge and slightly convex thinning just before the hinge line not sanded in at time of picture). I hope I'll have time some time to run XFOIL on something like that. Possibly the RN effects change with thickness at lower values of each...?

Cool post.

SK

[Howard Rush]

Is your angle of attack taken from a chord from the leading edge of the stabilizer to the trailing edge of the elevator or just for the stabilizer?

I intended it to be taken from the stabilizer CL, but sometimes Profili redefines it.  That's a handy feature, but a nuisance when comparing wings with different flap deflections or stabilizers with different elevator deflections.  These plots look like they came out as intended. 

Also what causes the stabilizer lift to jump suddenly or vary some at 20- and 25 degree deflections - would that be felt at the handle or seen?

The boundary layer plots, which I don't understand, appear to show flow separation on the upper surface just behind the LE.  I would reckon that when the elevator is down, the stab angle of attack is usually negative, so it might not be a problem.  It might be why the transition from outside to inside loops in my horizontal eights is flaky, but that's probably because I only practice that one half as much as the the transition from inside to outside loops.   

I think I remember Igor mentioning changes in optimal stabilizer/elevator thicknesses for different chords at really low RN. TR 586 doesn't really show what he saw. I do remember his discussing sharper leading edges for the thinnest sections, but I don't know whether he settled on that. I think he did some computer modeling, but don't have time to search. Even though you were concerned with Impact qualities, If you venture further afield, it would be interesting to see if results are any different for a section like the one shown below, which operates at RN's closer to 210K. It's tapered with flat top and bottom except for the leading edge and slightly convex thinning just before the hinge line not sanded in at time of picture). I hope I'll have time some time to run XFOIL on something like that. Possibly the RN effects change with thickness at lower values of each...?

Other folks have suggested venturing afield, too.  I tried the airfoil in the picture below, a quick approximation to some West Coast tails, but I couldn't get the program to accept it.  Yours looks less radical, so it might work. 

Igor sent me some stuff which I need to study.

[Howard Rush]

The visual I get is that the elevator may be sort of swimming around a few degrees in the wash of the stab and needs to deflect to a place where it is presenting a surface more cleanly to the airflow.  What do you think Profili would do with a stab and elevator of equal thickness and little elevator taper until at least the last bit.  It would appear to provide pressure with nearly no added deflection. Do you think the required droop is something of a function of the difference in stab and elevator thickness?  

There's a popular idea that the elevator is hiding behind the stab and doesn't have much effect until it peeks out.  I don't see that happening from these data, or from a later run of elevator deflections every degree from 1 to 5 degrees, but the airfoil has been smoothed a lot around the hinge line compared to the actual airfoil, so if there is a deadzone effect it might be getting covered up.  I think the droop is there because the airplane needs a little down when the flaps are zero.  Some folks use positive stab incidence.  This is another way to do the same thing.

I am a little surprised that landing gear drag doesn't offer enough forward pitching moment that induced moment is needed.

I don't know what induced moment is, but I'm surprised, too, that landing gear drag doesn't do much.  I calculated it awhile back, and I think I posted it here somewhere.  I guess now we can figure out how much elevator it takes to cancel gear drag as well as prop gyroscopic effect and other stuff.  

[Dave_Trible]

Ah the answer!  We simply carry eight different size wheels in our box and trim with wheel size.  Of course.  Have to honestly say I'm not seeing in most cases what you are trying to trim out. Maybe I'm used to it or doing it some other way and not realizing it.  Then again I' m one of those who don't fit the controls very tight and always has a bit of play in the controls which may be doing the same thing. I think you understood what I was meaning to say that I supposed the landing gear drag would tend to pitch the plane forward to some degree so I found it interesting that it wasn't enough and so you were adding or 'inducing' more forward pitch yet with down trim. Thanks again.

Dave

Maybe the big grass wheels I have to use for my flying site aren't so bad after all! 

[Brett Buck]

There's a popular idea that the elevator is hiding behind the stab and doesn't have much effect until it peeks out.  I don't see that happening from these data, or from a later run of elevator deflections every degree from 1 to 5 degrees, but the airfoil has been smoothed a lot around the hinge line compared to the actual airfoil, so if there is a deadzone effect it might be getting covered up.  I think the droop is there because the airplane needs a little down when the flaps are zero.  Some folks use positive stab incidence.  This is another way to do the same thing.

I don't know what induced moment is, but I'm surprised, too, that landing gear drag doesn't do much.  I calculated it awhile back, and I think I posted it here somewhere.  I guess now we can figure out how much elevator it takes to cancel gear drag as well as prop gyroscopic effect and other stuff.  

    Don't overlook the effect of thrust line missing the CG. That one is seemingly non-trivial, particularly given that changing it by 1/4" seems to matter.


    Brett

[Kim Mortimore]

There's a popular idea that the elevator is hiding behind the stab and doesn't have much effect until it peeks out.  I don't see that happening from these data, or from a later run of elevator deflections every degree from 1 to 5 degrees, but the airfoil has been smoothed a lot around the hinge line compared to the actual airfoil, so if there is a deadzone effect it might be getting covered up.  I think the droop is there because the airplane needs a little down when the flaps are zero.  Some folks use positive stab incidence.  This is another way to do the same thing....  Howard

Wouldn't the droop eliminate any deadzone effect compared with stab incidence, assuming such effect exists?  I'm curious if anyone has run any field tests to determine whether or not the deadzone exists.  I find hunting a persistent and evasive problem.  It would be good to know what it is about the Impact that produces the stability.