Hi all.  Would anyone know how to determine stall speed of an airplane using APM2.5 w/ airspeed sensor?  I have a Skywalker 1680, but would like to apply this to other airframes as well.  One reason would be to set a safe minimum airspeed and another would be for use in calculating optimum cruise speed and prop selection.

When I cut throttle, it just glides.  Would I need to cut throttle in a climb?  Is there a safe process to induce the stall?  Assuming one would need to cut throttle in a climb, would the altimeter be accurate enough to record the moment of stall in order to pick off the air speed?  I can sense the stall visually and "feel" the loss of aileron effectiveness, but I'm looking for something that will indicate a stall condition in the tlogs.



Views: 1964

Replies to This Discussion

Hi Trung,

I wouldn't have the throttle on during the climb . . .

When I was learning to fly gliders (the big ones that you sit in) we would slowly raise the nose to make the aircraft reduce speed. Just before it stalls, the controls would become unresponsive due to the lack of good airflow over the control surfaces. Then it would stall and the nose would drop on its own accord and start to speed back up again.

So I would suggest throttle off and slowly pull back on the stick and watch the aircraft slow down. At the point of stalling either the nose will drop, or if it isn't quite stable then one of the wings will drop first.

cheers, Graham

If you have telemetry to APM, you could just fly lots of stalls, then look at the data afterwards (from the t.log file.)

Plot the x+y axis gyro readings and airspeed. When the aircraft stalls there will be a sudden step in the gyro reading, and just look at what the airspeed was at this point.


Thanks that's a good suggestion for the method to induce a stall.  I was looking for a data logged indicator though, so that I can identify a stall condition in the tlog files during analysis.


Thanks for the suggestion.  That makes sense to use gyro readings.  I will do a flight to capture a bunch of stalls.  Any thoughts on whether this should be done in Manual or Stabilized mode?

Other questions:

Is it correct that the accelerometers measure linear acceleration and the gyros measure angular acceleration?

For fixed wing aircraft, what do the x, y, z axis correspond to?  (assuming properly installed APM2.5)

x=longitudinal axis?  (nose to tail)

y=lateral axis? (wing tip to wing tip)

z=vertical axis? 

Would that then mean that:

X gyro = roll?

Y gyro = pitch?

Z gyro = yaw?

As has been said just keep pulling with the power off until it stops flying, then repeat again you will soon see the speed. Its how I do it.

As others have already said, you increase up pitch without power to slow the plane until it stalls.  

In straight and level flight, maintaining altitude by increased pitch without throttle will eventually do it, and give you what is likely the lowest possible speed.  

For minimum safe, if that's what you're trying to discover, be advised, a stall from straight and level flight is at a different (lower) speed than one under heavier load, as in pulling G's or just maintaining altitude while in a banked turn.  Under those conditions, your effective weight is higher, and therefore the wing must produce more lift, requiring more airspeed.  (Altitude is another factor, but unless you're flying in the mountains, somewhat irrelevant for our typical flight profile).  

Analysis of stall speed at these higher loads is possible, and since you already have the accelerometer (Z), shouldn't be hard to get some decent data points, i.e. 2G's 4G's etc.  A 45Deg bank is about 2G's, if you maintain altitude (you will have to again increase pitch as the plane slows, then falls out of the turn in a stall).  Recovery is then wings level, slightly nose down, full power, recover to cruise.  

The difference isn't likely to be a lot, probably a few knots, but it would be interesting to see what the data shows at this scale.  

Thanks, Jay, for the insight and the point about the effect of loading.  I will add that to the test plan run list.  (This is starting to sound too much like a work discussion - somehow still fun though.)

Also thanks for the recovery tip.  I was apprehensive about stall recovery for the straight and level case, and even more so now for a banked turn.  I think it will be wise to enlist the help of some experienced local pilots with the stall/spin recovery.

One thing I noticed in a plot of a previous flight were repeating peaks in acc and gyr levels during an unpowered spiral.  (Just RTL - not a death spiral)  I believe this was due to a slight wind causing a difference in sink rate as the plane circled while descending.

So my question is:  Could the potential for changing relative wind direction in a turn skew the results when looking for banked stall speed?  I know this is getting rather academic.  Just a thought that came up while I was familiarizing myself with the acc and gyro data.  I would be satisfied with determining the straight and level stall speed and adding some safety factor.

I am eager to see how precisely the on board sensors can indicate the onset of stall in the data logs.  If the results look useful, I'll post them here.

If you're doing this manually, you can fly the plane in a constant bank (unlike when doing ground reference maneuvers, which is what the above basically is due to the use of GPS).  

Your circular flight will drift with the wind, which should reduce or eliminate the effects of said wind on the results, unless conditions are gusty (in which case you just have to wait for a time they're not, mornings or late evenings usually).  This would be a lot easier to do FPV I suspect, if you have a HUD with altimeter and can watch your bank angle.  I'd also try to be away from large hills/trees, or thermal sources/sinks such as lakes or parking lots.  

Look forward to seeing what you find!  

yes yo can find the stall speed of skywalker. take the aircraft to about 100m altitude. cut throttle. let the glide begin. slightly pull up the elevator stick so the aircraft is about horizontal. do this couple of times. after slight intervals. now chk the telemetry log. you can easily see stall speed by replaying telemrty log ans observing the AHRS. you would notice at which speed the nose drops by looking at AHRS. that would be the stall speed.

Hi Trung

here is one good practical way to determine stall speed

1- fly to a safe altitude

2-switch to FBW-A or Stabilize mode

3- fly level

4-decreasing throttle try to reduce speed

5-almost reaching the stall speed airplane will oscillate around it's roll axis about 2 degrees (sensing this oscillation is not easy so continue to 6)

6-when oscillations reaches the amplitude of 10 degrees consider it stall speed


The weather did not cooperate this weekend.  There was plenty of wind and rain.  I was able to get out for a quick flight Sunday for some stall tests.  The wind was gusting near 20mph, so I consider this just a test of a test.


  • It is difficult to slowly ease into a stall when flying LOS at a safe height.  (100-200m)  My elevator inputs were more abrupt than I would have liked because it was hard to tell if I was raising the nose without big inputs.
  • Stalls ranged from benign - dropping the nose - to violent/scary - dropping into an inverted dive/corkscrew.  (After the first spiral dive, I went to 200m)
  • I think the windy conditions contributed to the variation and severity of the stalls.
  • I am not very good at stall recovery.  Ardupilot Stabilize may have saved my butt.  Everything was flown in Stabilized mode.  I was not willing to try Manual.
  • Airspeed continues to drop momentarily after/while the nose pitches down.  I believe the wing has already stalled when the nose starts to pitch down, but the plane's inertia must be overcome before the plane starts to dive.  (my theory)  Therefore the resultant speed at the point of pitch inflection (stall speed) is slightly higher than the minimum airspeed.
  • Airspeed read 2.3 to 2.9 m/s on the ground.  In doors as well.  I need to rezero before flight with the pitot covered.

Here's some data plotting Airspeed vs. Pitch.  The green lines indicate where I think the stall occurs and the magenta lines indicate the corresponding airspeed.

I found the Pitch parameter to be the best indication of the onset of stall.  This is debatable.  The other parameters (accel & gyros) sampled at a much lower rate than airspeed and I felt were therefore unreliable for picking off a stall speed.

I am also a bit puzzled by the stall speed.  If I were to offset the data by the ~2.5m/s that it seemed to be off, the stall speed would be around 5.5m/s.  That seems very slow to me.  Does anyone have a point of reference?  I would like a sanity check, perhaps from a similar airframe.  I have read one person post a stall speed of 9m/s for the Skywalker.

Due to the windy conditions and uncalibrated airspeed, I'm throwing out this data and will try again.  Maybe next weekend.

Hi Trung, 

I suspect your actual stall speed is a bit before where you have the line at the obvious break, not by much, but given how steeply your airspeed is dropping, it will have an effect on your numbers.  I bet that first little ripple before the big drop is when the stall is beginning.

As for your comparative stall speed against others, it's highly dependent on wing loading, so depending on the weight of your total aircraft, you may be stalling slower than somebody with heavier gear, and also where their CG is.  And of course that's multiplied by G loading, per my earlier comment.

5.5Mps translates to 12.3MPH or 10.6 Knots, but 8Mps (which is where the ripple starts if I'm looking at this right) is about 17.9MPH or 15.5KTS, which is probably closer to reality.

As for calculating optimum cruise, here's a good reference:  http://www.eaa1000.av.org/technicl/perfspds/perfspds.htm

Look forward to seeing your data next weekend!  Good luck, may the WX be with you!  




Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service