Stall speed

[SCaro]

I would like to collect some ideas on why two airplanes of the same model, which look the same, at the same weight & CG can have almost 10% difference in stall speed (power off, full flaps)? I have tested a number of things, but have not found an explanation. Any inputs are appreciated, has anyone come across this before?

[effortless]

Perhaps you could tell us which ones?

[SCaro]

The airplane with the higher stall speed has been used as a test airplane. The stall is aft stick limited. I have eliminated the horizontal tail and elevator authority, it is sufficient to get it stalled properly. I tried different angles of stall strips, that did not change stall speed. It is a single engine, piston. Adding a small amount of manifold pressure as opposed to throttle fully closed decreases stall speed by several knots. The wing leading edge contour was slightly different from the other airplane, we tried to rework it to match, that reduced stall speed slightly, but not enough.

[Mad (Flt) Scientist]

1. How sure are you of the aircraft weights?

2. What is the history of the annual stall speed checks (if these are CAA aircraft)

3. Have you been applying mods to just the one aircraft, or to both. Are you modding the faster or the slower staller, if just the one.

Im a bit suspicious of your statement that moving the stall strips is doing nothing. Does it USUALLY on this type. If so, then it sounds very much like something else is trigerring a stall earlier than the range of stall strip positions, and that LE contour issue you mention is a likely culprit Id have thought.

have you tried stall checks in other configs, such as clean. perhaps its a flap rigging issue?

[John Farley]

Are both ASIs reading correctly?

[B Fraser]

The wing leading edge contour was slightly different from the other airplane, we tried to rework it to match, that reduced stall speed slightly, but not enough.


Is there still a difference in profile ?

Do they weigh the same ?

[SCaro]

I should have posted more details, sorry. I checked the airspeed indications on both airplanes with GPS, they are right on. Both are new airplanes (this is in the US). The one with the higher stall speed was weighed several times, the weight seems to be accurate. We have only modified the higher stall speed airplane. The stall speed flaps up and take off flaps is also higher than normal, but not as much as with full flaps (slotted flaps). I measured the angle of attack at which they stall, the one with the higher stall speed also stalls at a lower angle of attack. We checked the flap rigging, and set it to the largest flap deflection within the tolerance, no change in stall speed. I tufted the flap upper surface (low wing), the flow on the flaps was attached at full flaps before the wing leading edge modification, afterwards it was not, even though the stall speed was lower (another mystery, normally this would be a loss of lift). The wing contour is now as close to type design as we could get it, although on one side it semms to be slightly thinner than it should be. Stalls are straight. I'm out of ideas.

[barit1]

Are engine idle speeds the same? If a constant-speed prop, are the low-pitch stops the same?

[B Fraser]

You did not check the airspeed with a GPS, you checked the groundspeed.

[Matthew Parsons]

What do you mean by "The stall is aft stick limited"?

If you're hitting control stops then control rigging could be a factor.

[barit1]

I think it's time for an alpha (AOA) strap-on meter. Something very simple like a telltale streamer well in front of the wing LE may work fine.

[Genghis the Engineer]

I should have posted more details, sorry. I checked the airspeed indications on both airplanes with GPS, they are right on.

How? This could be relevant, there are many ways of checking IAS against GPS, some can be very accurate, some (most?!) can be very misleading.

<snip> I measured the angle of attack at which they stall, the one with the higher stall speed also stalls at a lower angle of attack.

How measured?

The stall is aft stick limited. I have eliminated the horizontal tail and elevator authority, it is sufficient to get it stalled properly

Surely you contradict yourself here. If the stall is aft-stick (rather than aircraft response) limited then elevator authority is extremely relevant to stall speed. What are the actual stall warning and stall characteristics?

We checked the flap rigging, and set it to the largest flap deflection within the tolerance, no change in stall speed.

Have you also checked aileron rigging? Drooped ailerons can decrease stall speed, or conversely reflexed ailerons can increase it. (Particularly if the aircraft has a swept wing, this can also be evidenced by a change in apparent LSS).

I tufted the flap upper surface (low wing), the flow on the flaps was attached at full flaps before the wing leading edge modification,

Leading edge shape is very very relevant to stall alpha, and this might be the issue. If you've got a tame aerodynamicist about the place, it might be worth their taking the two wing forms and running them through something like X-foil for a quick and dirty look at the differences.

afterwards it was not, even though the stall speed was lower (another mystery, normally this would be a loss of lift). The wing contour is now as close to type design as we could get it, although on one side it semms to be slightly thinner than it should be. Stalls are straight. I'm out of ideas.

Hmm, "seems to be slightly thinner" is interesting - anything that is (presumably) apparent to the naked eye in this regard could be very significant. I'd spend some time carefully measuring it up.


In the meantime, could you give us a couple of clues...

- Composite / metal / other?
- High/low/mid wing?
- High/low/mid tailplane?
- What sort of values are we looking at in the weights and stall speeds?

G

[SCaro]

It looks like I will not get anywhere with this without telling you everything. I'm an aeronautical engineer and DER test pilot.

The airspeed readings were verified with the GPS triangle method and were within +-1 kt. I measured angle of attack with a string attached to a boom on the wingtip and a scale next to it. The accuracy was verified by measuring the angle of the fuselage at different airspeeds in level flight. It correlates well up to about 8 deg alpha, at higher angles the string on the wingtip indicates too high, but this was mainly to get a comparison between the two airplanes. The difference in stall alpha was at least 3 deg on the wing tip gage.

The stall can be either defined by the airplane pitching down or by the stick reaching the aft stop. In this case the stick reaches the aft stop and can be held there, but the lowest speed is reached right at aft stop and then increases together with the sink rate. The nose stays up. I had increased the elevator up deflection until I got the airplane to pitch gently, but this had no effect on stall speed.

The wings are not swept and the aileron rigging was checked (within type design tolerance).

Engine speed at idle is about the same, I have to add about 3-4 inHg of manifold pressure before the stall speed decreases, but it does that for all of our airplanes.

I have run some airfoil variations through XFOIL, but the analysis showed no difference. At stall angles of attack any analysis is more of a guess anyway. We took "splashes" of the wing leading edges at various stations (wood templates bonded to the wing with filler, then removed, to get the exact airfoil shape) and compared them to plots of what the shape should be, that's how I know about the "thinner" issue. It is not apparent to the "naked eye".

[barit1]

Do the two ships retain aileron effectiveness similarly through the stall? --

Which is maybe another way of asking if the stall originates at the same spanwise position. I like to see the stall begin at the root & progress outboard.

[Genghis the Engineer]

A belated welcome aboard SCaro, and an interesting problem you have there. Sorry if I sounded a little patronising, but it's not always possible when somebody appears in the FT forum to gauge how much of a "grown up" they necessarily are.

A couple of thoughts...

- I used to own a little composite canard single seater, and have flown a couple of others of the same type. A common problem on that was differing washout between wings/aircraft which could alter both stall characteristics and trim. Any risk this might be a player?

- Presumably the CG values are similar? I've previously seen two otherwise identical aeroplanes (single engine, high wing) show quite different stalling speeds as a result of differing CG values. (Fwd CG leading to higher Vs).

I'll admit those are longshots. I shall go away and think.... :sad:

G

[camlobe]

" I measured the angle of attack at which they stall, the one with the higher stall speed also stalls at a lower angle of attack."

Commonly, aft C of G as mentioned by Gengis.

Long shot. Wing/mainplane angle if incidence?

Hope this may be of some help.

[SCaro]

The stall originates at the stall strips, which are about 30 inches outboard from the fuselage. I tested stall speed at forward and aft CG, in this case no difference, although usually there is about a one knot difference (slower at aft CG). This airplane has very benign stall characteristics, less than 5 deg of roll, and very little aileron input is needed to keep the wings level.

We measured incidence of wing and tail and everything is within type design tolerance.

[Mad (Flt) Scientist]

The stall originates at the stall strips, which are about 30 inches outboard from the fuselage.
.....
I tried different angles of stall strips, that did not change stall speed.

together, those two statements are VERY suspicious.

If I understand you, the stall on this model is supposed to originate from the stall strips, YET on this specific tail number the stall speed did not react to the position of the stall strips. This implies that the stall is actually being triggered by a mechanism independent of the stall strips for this aircraft.

How far did you move the stall strips, and in which sense? If you moved them further down (under) the leading edge, which would correspond to a higher stall AoA than normal, hoping to delay the stall and get a lower speed, and nothing actually happened, this would be consistent with the theory that something else is tripping the flow.

That something is, I'm afraid, quite possibly your deformed LE profile.

If you wanted to try a test to confirm the theory (albeit not really solve anything...) you could calculate the delta stall strip position consistent with a 3 degree ADVANCED AoA, and place the strip in front of this position. This should then cause a change in stall AoA as the strip finally gets a chance to work before the other unidentified trip mechanism takes effect.

That would confirm the behaviour, it won't FIX anything. I fear you may be looking at further work on the LE - perhaps even reskinning, although I know thats an ugly word.

[barit1]

Did you try it with stall strips removed?

[SCaro]

We are planning to do something along what Mad Flt S. suggests, starting with more extensive tufting of various parts of the airplane. We know that without the stall strips the stall speed would be several kts lower, but that is not an option. This airplane is certified as spin resistant, and needs them for that purpose.

[barit1]

... We know that without the stall strips the stall speed would be several kts lower, ...

But--if what Mad (Flt) Scientist proposes is true, removing stall strips MAY NOT decrease the stall speed. :8

[MilFlyBoy]

Scaro,

You still haven't replied to a couple of the Questions:

1. What type of ac is it?
2. What are your actual differences in stall speed? (10% at around 30kts would seem virtually insignificant)

An observation:

1. A stall is never defined (to my knowledge) as being defined by the location of the elevator or stick. It is the point where significant boundary layer flow separates from the wing and does not reattach, hence causing a steep decline in the L/D relationship.

Some thoughts:

1. Aerofoil Section: As has been pointed out in several replies the aerofoil section is absolutely critical to stall characteristics. The LE being the most critical element. Any slight deformation causing a rapid loss in chordwise positive pressure gradient could trigger a separation and stall. The surface of your LE is also critical as any difference in smoothness will obviously alter the transition to turbulence and therefore the ability of the flow to remain attached. Is the surface absolutely identical? Are you working at low Reynolds Number where laminar flow is even more critical?

2. Engine Thrust: Can you be certain that the engines produce identical power? Are the props identical to transfer all the energy? A small difference in flow over the inboard wing at 30ins fuselage, where you suggest stall begins could have a huge effect on stall speed.

3. Test Aircraft Plastic Wing Deformation: A long shot but has the ac been subjected to loads that may have plastically deformed the wings during frequent high loading? (The thinner section you mention will effect the pressure profile certainly as well) can you be sure that the aerofoil section does not deform in flight different to the other ac? More info required on the ac structures and flight profiles I think!

4. CofG Vs Mass distribution: An equal CofG does not necessarily mean an equal mass distribution. It just means that the average of the masses is in the same location. You may need to investigate the mass distribution of your ac about the wings and the fuselage. Can you take the wings off to test? For example if the Mass is distributed more outboard on the wings on one ac then you can expect a different wing pressure planform and thus stalling characteristics.

5. Some possible tests and some slightly more crazy ones: I am not sure how much money/time you have or the legalities of these suggestions, but I am sure you can find out if you are that desperate! Take the wings off and find the CofG of the fuselage and Wings separately. Swop the wings over on the 2 ac!!?? Even just to reweigh! Can you swop the engines? Get an accurate readout of thrust at least for the 2 engines at idle.

:ok:

Regards

[SCaro]

We finally got the stall speed down to where it should be. It turned out to be a combination of two separate issues that influenced each other. The wing leading edge contour was not correct and the flap leading edge contour was not correct either. The latter deviation was not visible even with templates, but showed itself by tuft tests. The flap upper surfaces had completely turbulent flow at full flap deflection. We sanded the flap leading edges until they were nice and round, and it got the flow reattached.

[GOLF_BRAVO_ZULU]

You techy boys are great. Knowing the conditions of stall is fascinating. What I wan't to know is what IAS am I going to depart from controlled flight

[John Farley]

What I wan't to know is what IAS am I going to depart from controlled flight

Of course you do.

Sadly you will first have to let the 'techies' know the present weight, bank angle, g, IMN and rate of change of IAS (to name just the main variables)

Which is why you should demand to be given an AoA gauge. No automatic control system will agree to work without such information so why men (and women) do is beyond me.

[barit1]

But--if what Mad (Flt) Scientist proposes is true, removing stall strips MAY NOT decrease the stall speed. :8

Responding to my own answer, I still believe that removing the stall strips would have given you a strong clue the airfoil was not right.

[old-timer]

An interesting thread & a good point John - data is required to be completely sure.
Going back to the original question about differences between aircraft of the same model having different stall speeds maybe it's due to manufacturing tolerances ? a few thou' here & there on the overall
aerofoil shape etc plus different paint finishes / polish etc must a difference -Gliders I've flown were more efficient when polished & the
same must be true for power too - an old instructer at southend I knew used to swear by mr sheen on his 150's & they did indeed fly very well - far better than the 'matt & dirty' finish competitors !

I wonder what effect these things have on fast jets too - it must be even more important to have a good finish - any thoughts folks ?
I remember mr Boeing advising that even small imperfections make a huge difference. (although that was with regard to icing - similar except potentially worse as it increases in bad conditions ! )
All Interesting stuff anyway...... :-)

[slowrotor]

Old timer is right about wing surface smoothness.
My motorglider would roll to the left when flying in rain and not when dry.
I determined the reason was the difference in the way rain beaded up on the left wing which had been painted and the right wing wing that had old paint and the rain laid smooth. Very minor difference, but it scared me the first time it happened. Both wings painted now and it flies straight and level.

Also, if the root gap tape is missing the aircraft will shake at low speed. The root gap is only about .060" without tape, but it does make a difference!

[barit1]

Boeing has determined that the 787 nacelle (http://www.boeing.com/commercial/news/2006/q3/060710d_nr.html)should be one color only (no trim features) to prevent disruption of laminar flow at the step of a new paint layer. :=

[old-timer]

Interesting subject - thanks for feedback folks,
ref the root gap this reminds of when I used to fly Condors, (FUN !)
there is (was ? - post CAA-now EASA/PFA) an AD to ensure that the gap between the elvator & tailpane was taped over - if not, the gap adversely affected the elevator effectiveness due to bleed through airflow - not quite sure of the exact terminology but I know of another Condor operator who chose to ignore this AD which very nearly resulted in a tragedy.
I believe the 'bleed through' airflow caused loss of elevator authority ?