Sammy Mason on Stalling & Spinning

Many descriptions have been posted here in several threads on spinning
over the last couple of weeks and I have found all of them to be
fascinating. I've been prompted to think through again my own
understanding of the stall and the spin in gliders.

One description that has been strongly asserted in several threads is
that there is a difference in angle of attack on the two wings on an
aircraft in a descending turn AND that the difference in AOA is due to
the slower horizontal speed of the inside wing as it traverses a
shorter circle than does the outside wing in the descending turn.
The slower speed of the inside wing thus coinsides with a higher AOA
on the inside wing than is experienced by the outside wing. This of
course has implications for which wing may reach stall AOA first.

I am only a pilot and have no training in aerodynamics whatsoever.
For what it's worth (not much probably), I find the above explanation
of the cause of different AOA on the left and right wings of an
aircraft in a descending turn to be easily understandable and very
likely accurate, but incomplete.

Many years ago I read in Sammy Mason's book Stalls, Spins and Safety,
an additional description of the AOA on the wings of an aircraft in
ascending and descending turns. I find that his description is
persuasive and illunimating and contributes to a more complete
understanding of the stall and spin.

Here is a short quote from Sammy Mason's book:

"During a level, coordinated turn, once the bank is established,
the airplane will continue to turn about the yaw axis and pitch upward
about the pitch axis. It will not be rolling about the roll axis.
When a stall is encountered in a level turn, the reaction will
normally be very little different than during a wings-level stall.."
"During a descending turn, or spiral, in addition to pitch and
yaw, the airplane will be rolling about the roll axis in the direction
of the turn. As the airplane rolls, it induces an upflow of air into
the descending wing. This results in the descending wing having the
greatest angle of attack. If a stall is encountered, the airplane
will likely roll into the turn." pp.40-41.

Sammy Mason goes on to describe the opposite differences in AOA on the
two wings of an aircraft in an ascending turn.

In a descending turn are both wings going down? Of course they are,
relative to an outside frame of reference and assuming the rate of
aircraft descent is greater than the rate of roll (or however one
describes the rate at which wings are going around in a roll). From
the frame of reference of the aircraft are the two wings proceeding in
opposite rotational directions? Sammy Mason's description of the
aircraft rolling about its roll axis in a descending turn describes
just such a difference -- and its contribution to the differing AOA on
the two wings. This suggests that the difference in AOA on the two
wings is not due only to their differences in horizontal speed in
their differing size circles.

Your milage may vary, of course, and each pilot likely benefits from
some image of what is happening to an aircraft in a stall and spin.
In my view, Sammy Mason's descriptions add additional insights for me
to the nature of stalls and spins.

Sammy Mason's flying career is described in the book as having begun
in the 1930's, and included WWII, and then working with C.L."Kelly"
Johnson of the Lockheed "Skunkworks". He was a jet aircraft test
pilot for Lockheed and became Lockheed's authority on stall/spin
testing.

"During a descending turn, or spiral, in addition to pitch and
yaw, the airplane will be rolling about the roll axis in the direction
of the turn. As the airplane rolls, it induces an upflow of air into
the descending wing. This results in the descending wing having the
greatest angle of attack. If a stall is encountered, the airplane
will likely roll into the turn." pp.40-41.


I'm having some trouble visualizing this.

Is it possible that Sammy has posited a reference frame that looks
only at AOA, ignoring bank and relative speed across the span? If I
wanted to keep my model and my math simple, rather than describing the
turn as a hollowed cylinder with inner- and outer-walls transcribed by
the wings during descent, I could look solely at AOA, in which case
the model of a turn would look similar to, if not exactly like, a slow
rolling motion. Our reference frame has no horizon. In fact, it is
purely scalar. AOA simply has a range of values across the wing. If
this is the case, I can see how it would be useful for a snapshot --
such as just prior to the stall, but confusing when describing the
dynamics of a turn in its fuller context. This is a kind of partial
differential: an alternative way of describing a turn, but only
predicts outcomes based on AOA. Good for analysis in a narrow band...
Certainly counterproductive if integrated haphazardly into a more
intuitive three axis model.

So it goes like this maybe. The observed effect of constant sink rate
and differential airspeed across the span of a turning airfoil when
described in terms of differential AOA can be likened to the rolling
motion produced by the ailerons in level flight. The downward moving
wing, during the rolling motion, exhibits an increasingly higher AOA
as you go out the span (ignoring that part of the wing with deflected
aileron) than the rising wing, which shows a descending value of AOA
with span. Thus, during a descending spiral, if the airfoil were to
stall, this "psuedo-rolling moment" could be said to contribute to the
wing drop typically experiended during a turning stall.

I'm not sure I see how this changes with level flight or a climb.
After all, if we establish the longitudinal axis as the basis for
measurement, up or down with respect to the ground shouldn't matter.

This seems to me a more useful short cut for the engineer than the
aviator. Just remember, similitude is not exact. But it is an
interesting concept nonetheless.

Maybe someone could do the math for change in AOA for a 15M glider
traveling at 100kph and rolling at a rate of 30 degrees per second,
ignoring the ailerons, of course.

Wow, that was fun. Thanks.

Chris O'C

On 10 Feb 2004 14:38:27 -0800, (Chris
OCallaghan) wrote:

"During a descending turn, or spiral, in addition to pitch and
yaw, the airplane will be rolling about the roll axis in the direction
of the turn. As the airplane rolls, it induces an upflow of air into
the descending wing. This results in the descending wing having the
greatest angle of attack. If a stall is encountered, the airplane
will likely roll into the turn." pp.40-41.


I'm having some trouble visualizing this.


Try it as a wide barrel roll heading straight down.

Rolling into the turn no?

Mike Borgelt

On 10 Feb 2004 14:38:27 -0800, (Chris
OCallaghan) wrote:

"During a descending turn, or spiral, in addition to pitch and
yaw, the airplane will be rolling about the roll axis in the direction
of the turn. As the airplane rolls, it induces an upflow of air into
the descending wing. This results in the descending wing having the
greatest angle of attack. If a stall is encountered, the airplane
will likely roll into the turn." pp.40-41.


I'm having some trouble visualizing this.

Is it possible that Sammy has posited a reference frame that looks
only at AOA, ignoring bank and relative speed across the span? If I
wanted to keep my model and my math simple, rather than describing the
turn as a hollowed cylinder with inner- and outer-walls transcribed by
the wings during descent, I could look solely at AOA, in which case
the model of a turn would look similar to, if not exactly like, a slow
rolling motion. Our reference frame has no horizon. In fact, it is
purely scalar. AOA simply has a range of values across the wing. If
this is the case, I can see how it would be useful for a snapshot --
such as just prior to the stall, but confusing when describing the
dynamics of a turn in its fuller context. This is a kind of partial
differential: an alternative way of describing a turn, but only
predicts outcomes based on AOA. Good for analysis in a narrow band...
Certainly counterproductive if integrated haphazardly into a more
intuitive three axis model.

So it goes like this maybe. The observed effect of constant sink rate
and differential airspeed across the span of a turning airfoil when
described in terms of differential AOA can be likened to the rolling
motion produced by the ailerons in level flight. The downward moving
wing, during the rolling motion, exhibits an increasingly higher AOA
as you go out the span (ignoring that part of the wing with deflected
aileron) than the rising wing, which shows a descending value of AOA
with span. Thus, during a descending spiral, if the airfoil were to
stall, this "psuedo-rolling moment" could be said to contribute to the
wing drop typically experiended during a turning stall.

I'm not sure I see how this changes with level flight or a climb.
After all, if we establish the longitudinal axis as the basis for
measurement, up or down with respect to the ground shouldn't matter.

This seems to me a more useful short cut for the engineer than the
aviator. Just remember, similitude is not exact. But it is an
interesting concept nonetheless.

Maybe someone could do the math for change in AOA for a 15M glider
traveling at 100kph and rolling at a rate of 30 degrees per second,
ignoring the ailerons, of course.

Wow, that was fun. Thanks.

Chris O'C

It is wonderful isn't it! I'm a complete dolt about aerodynamics but
I love flying so much, especially flying gliders, that I just can't
get enough of the whole world of it.

When I try modeling the ascending turn without continually rolling
against the turn as the aircraft ascends the aircraft ends up on its
back! Only by a continual rolling input opposite to the ascending
turn does the aircraft to remain in a constant turn! Just wonderful
stuff and fascinating! Mason's book is a treat, bless his heart.

Todd Pattist wrote:

Jim wrote:

(Chris OCallaghan) wrote:
I'm having some trouble visualizing this.

Is it possible that Sammy has posited a reference frame that looks
only at AOA,

I think he's looking at the ground reference. If an
aircraft is making a pure level skidding turn, it's turning
in yaw only. If it's rolled knife edge in the level turn,
it's making a pure pitching turn. If it's pointed straight
down (Mike's "vertical barrel roll") it's making a pure
rolling turn. Different banks and descent/climb rates from
the level turn produce the combinations of the above that
Sammy describes.
...

But nevertheless it seems there is a confusion about what
are the pitch, yaw and roll axis. The usual convention is
that these axis are tied to the airframe, independantly of
the attitude and the airflow. In this case, most gliders
when they are near stall have a nose up attitude despite
their strongly descending path. In this case Sammy's description
would be an ascending turn, with a difference of AOA between
wings opposite to what is observed (remember the extreme case
of the half wingspan equal to turn radius yielding an AOA
over 90 degrees (90 degrees + nose up attitude + incidence)
for the inner wingtip). Sammy's description would probably
be correct if we redefine the roll axis as parallel to the
airflow, pitch axis unchanged from wintip to wingtip and yaw
axis perpendicular to both others. This is in some way implied
by this description as being the case of a descending turn, as the
fact that the aircraft is descending is tied to the direction
of the airflow, not the pitch attitude. In this case this description
and the previous one involving the speed difference of both
wing tips are equivalent.

On Wed, 11 Feb 2004 10:23:18 -0500, Todd Pattist
wrote:

Jim wrote:

(Chris OCallaghan) wrote:
I'm having some trouble visualizing this.

Is it possible that Sammy has posited a reference frame that looks
only at AOA,

I think he's looking at the ground reference. If an
aircraft is making a pure level skidding turn, it's turning
in yaw only.

I had never thought of this. Will an aircraft actually turn, that is,
change its direction of flight, if it is not allowed to bank at all?
I know that a clumsy kind of turn might be accomplished by use of
rudder only but I thought that was because the yaw would eventually
lead to a bank - due to the increased lift of the 'outer' wing caused
by the yaw.


If it's rolled knife edge in the level turn,
it's making a pure pitching turn. If it's pointed straight
down (Mike's "vertical barrel roll") it's making a pure
rolling turn. Different banks and descent/climb rates from
the level turn produce the combinations of the above that
Sammy describes.

Todd Pattist - "WH" Ventus C
(Remove DONTSPAMME from address to email reply.)

On Tue, 17 Feb 2004 16:00:26 -0500, Todd Pattist
wrote:

Jim wrote:

Will an aircraft actually turn, that is,
change its direction of flight, if it is not allowed to bank at all?

Yes. The turn is produced by "lift" in the horizontal plane
by the AOA of the fuselage (and some wing dihedral). The
"AOA of the fuselage" is just the yaw angle produced by the
rudder. The "lift" is caused by the sides of the fuselage.
It's certainly not a turn that anyone makes intentionally,
unless they are playing around to see one. A modern glider
will produce very little sideways lift from the fuselage, so
the turn will take a while. :-)

I know that a clumsy kind of turn might be accomplished by use of
rudder only but I thought that was because the yaw would eventually
lead to a bank - due to the increased lift of the 'outer' wing caused
by the yaw.

You have to prevent the roll by using opposite aileron to
reduce the lift on the faster outer wing to equal the lift
on the slower inner wing. Otherwise, you are right, it will
roll, but then it's not a "wings level" turn.


Todd Pattist - "WH" Ventus C
(Remove DONTSPAMME from address to email reply.)


Wonderful stuff. Thank you.

Jim wrote in message . ..

I had never thought of this. Will an aircraft actually turn, that is,
change its direction of flight, if it is not allowed to bank at all?
I know that a clumsy kind of turn might be accomplished by use of
rudder only but I thought that was because the yaw would eventually
lead to a bank - due to the increased lift of the 'outer' wing caused
by the yaw.

Absolutely. In fact, in the early days of aviation, when airplanes
were so underpowered that any bank resulted in a descent, flat
"skidding" turns were considered by many to be the only "relatively"
safe way to turn.

I flew last friday, and having some altitude to spare, tried some flat
turns. Level wings, feed in rudder, opposite aileron to maintain
wings level, nose down to maintain airspeed. After a while, let off
the rudder, and sure enough, you will have turned a little. Very
inefficiently, by the way!

It helps to have a fuselage with a lot of vertical surface, which my
LS6 obviously does not. And while turning, it is very hard to tell
what your heading will be when you let off the rudder. Kinda fun to
do, and not all that easy to do well.

Kirk

Just think about the relative air flow in relation
to all the aero dynamic surfaces. It is then quite
clear. Remember that in a spin there is pitch, roll
and yaw so the raf changes. Think about what happens
when it reaches the aerodynamic limit of the relevant
surface. If it makes it easier just think about the
raf in a straight stall first where no attempt at recovery
is made. What happens? Then add rolling and yawing.

At 22:42 10 February 2004, Chris Ocallaghan wrote:
'During a descending turn, or spiral, in addition to
pitch and
yaw, the airplane will be rolling about the roll axis
in the direction
of the turn. As the airplane rolls, it induces an
upflow of air into
the descending wing. This results in the descending
wing having the
greatest angle of attack. If a stall is encountered,
the airplane
will likely roll into the turn.' pp.40-41.


I'm having some trouble visualizing this.

Is it possible that Sammy has posited a reference frame
that looks
only at AOA, ignoring bank and relative speed across
the span? If I
wanted to keep my model and my math simple, rather
than describing the
turn as a hollowed cylinder with inner- and outer-walls
transcribed by
the wings during descent, I could look solely at AOA,
in which case
the model of a turn would look similar to, if not exactly
like, a slow
rolling motion. Our reference frame has no horizon.
In fact, it is
purely scalar. AOA simply has a range of values across
the wing. If
this is the case, I can see how it would be useful
for a snapshot --
such as just prior to the stall, but confusing when
describing the
dynamics of a turn in its fuller context. This is a
kind of partial
differential: an alternative way of describing a turn,
but only
predicts outcomes based on AOA. Good for analysis in
a narrow band...
Certainly counterproductive if integrated haphazardly
into a more
intuitive three axis model.

So it goes like this maybe. The observed effect of
constant sink rate
and differential airspeed across the span of a turning
airfoil when
described in terms of differential AOA can be likened
to the rolling
motion produced by the ailerons in level flight. The
downward moving
wing, during the rolling motion, exhibits an increasingly
higher AOA
as you go out the span (ignoring that part of the wing
with deflected
aileron) than the rising wing, which shows a descending
value of AOA
with span. Thus, during a descending spiral, if the
airfoil were to
stall, this 'psuedo-rolling moment' could be said to
contribute to the
wing drop typically experiended during a turning stall.

I'm not sure I see how this changes with level flight
or a climb.
After all, if we establish the longitudinal axis as
the basis for
measurement, up or down with respect to the ground
shouldn't matter.

This seems to me a more useful short cut for the engineer
than the
aviator. Just remember, similitude is not exact. But
it is an
interesting concept nonetheless.

Maybe someone could do the math for change in AOA for
a 15M glider
traveling at 100kph and rolling at a rate of 30 degrees
per second,
ignoring the ailerons, of course.

Wow, that was fun. Thanks.

Chris O'C

Just think about the relative air flow in relation
to all the aero dynamic surfaces. It is then quite
clear. Remember that in a spin there is pitch, roll
and yaw so the raf changes. Think about what happens
when it reaches the aerodynamic limit of the relevant
surface. If it makes it easier just think about the
raf in a straight stall first where no attempt at recovery
is made. What happens? Then add rolling and yawing.

At 22:42 10 February 2004, Chris Ocallaghan wrote:
'During a descending turn, or spiral, in addition to
pitch and
yaw, the airplane will be rolling about the roll axis
in the direction
of the turn. As the airplane rolls, it induces an
upflow of air into
the descending wing. This results in the descending
wing having the
greatest angle of attack. If a stall is encountered,
the airplane
will likely roll into the turn.' pp.40-41.


I'm having some trouble visualizing this.

Is it possible that Sammy has posited a reference frame
that looks
only at AOA, ignoring bank and relative speed across
the span? If I
wanted to keep my model and my math simple, rather
than describing the
turn as a hollowed cylinder with inner- and outer-walls
transcribed by
the wings during descent, I could look solely at AOA,
in which case
the model of a turn would look similar to, if not exactly
like, a slow
rolling motion. Our reference frame has no horizon.
In fact, it is
purely scalar. AOA simply has a range of values across
the wing. If
this is the case, I can see how it would be useful
for a snapshot --
such as just prior to the stall, but confusing when
describing the
dynamics of a turn in its fuller context. This is a
kind of partial
differential: an alternative way of describing a turn,
but only
predicts outcomes based on AOA. Good for analysis in
a narrow band...
Certainly counterproductive if integrated haphazardly
into a more
intuitive three axis model.

So it goes like this maybe. The observed effect of
constant sink rate
and differential airspeed across the span of a turning
airfoil when
described in terms of differential AOA can be likened
to the rolling
motion produced by the ailerons in level flight. The
downward moving
wing, during the rolling motion, exhibits an increasingly
higher AOA
as you go out the span (ignoring that part of the wing
with deflected
aileron) than the rising wing, which shows a descending
value of AOA
with span. Thus, during a descending spiral, if the
airfoil were to
stall, this 'psuedo-rolling moment' could be said to
contribute to the
wing drop typically experiended during a turning stall.

I'm not sure I see how this changes with level flight
or a climb.
After all, if we establish the longitudinal axis as
the basis for
measurement, up or down with respect to the ground
shouldn't matter.

This seems to me a more useful short cut for the engineer
than the
aviator. Just remember, similitude is not exact. But
it is an
interesting concept nonetheless.

Maybe someone could do the math for change in AOA for
a 15M glider
traveling at 100kph and rolling at a rate of 30 degrees
per second,
ignoring the ailerons, of course.

Wow, that was fun. Thanks.

Chris O'C