Narrative:

Below is an article I wrote a number of years ago while working in the training department of [an airline company]. It was distributed to all of the pilots there; but to my knowledge has been passed along to the greater aviation community. I regret not have been more aggressive in trying to get this information out to the larger population of pilots. Sadly; since that time; there have been several of these unexpected upsets and one fatal accident which I am convinced were a result of this threat. I hope this finds its way to someone with enough interest in flight safety to verify the phenomenon of mach induced stall; and to take actions that will help protect pilots; crews and the traveling public from this threat. Please contact me with any questions or concerns.a few years ago a friend; who was a B737 captain at the time; told me of a very unsettling few seconds he and his crew of a B737 had experienced inbound to the new york area. Cruising at FL410 the flight was instructed to enter holding over pxt due to next sector saturation. The hold was programmed into the FMC and as the designated point was approached; the auto flight system began to slow the aircraft to the computed hold speed of 214 kcas. Upon crossing the holding fix; the aircraft rolled into the 'standard' 25-30 degree bank turn. This is when the unexpected happened. In a split second; with no warning; the aircraft went from smooth flight into heavy buffet (no stick shaker); the ap (autopilot) disconnected and the plane dropped 800 ft before control could be regained by the startled crew. Two weeks later; after relating his experience; he still had no idea what had really happened. As it turns out; this is a phenomenon with which I had some experience; and had followed up with some enlightening research. My initial encounter occurred while flying an F-4 simulating a high altitude subsonic intruder for intercept training. Our altitude was in the low 40s and required that we fly at a relatively slow (for the F-4) indicated airspeed and high angle of attack (aoa/alpha) in order to remain below mach 1. As we climbed; we reached a condition in which the aircraft suddenly became very unstable in pitch axis. Though we were well below the aoa at which buffeting normally began; I found that any slight increase in aft stick pressure would result in an instant onset of buffeting accompanied by the nose continuing to pitch up and the aircraft sinking. Because of the excellent control response of the phantom; I was able to fly it for some time right on the edge; where we were in and out of this buffeting; and develop some feel for what the plane was doing and how to counter it. I just didn't know 'why' at the time. After landing; I described the experience to many of the 'old heads;' but none had any similar experiences or explanations. So; I tucked it away with the rationalization that it must have been a peculiarity of that particular aircraft - maybe it was a little bent. Some years later; flying an F-105 in the low 40s (exceptionally high for a thud) I had a similar experience. As this was too much like the F-4 event; I launched into some research; the crux of which revealed that this can occur to any aircraft flying at high subsonic mach and high aoa. A few years later I verified that even an F-16 exhibits the same characteristics under similar conditions.for any given airfoil as the aoa is increased; the speed of the airflow over the upper surface increases because the relative air must travel an increasingly longer distance. Furthermore; as the aoa increases the attachment of the boundary layer to the wing becomes more and more tenuous as it has to bend more radically around the wing. At low mach numbers with this increasing aoa we eventually reach a point at which the boundary layer can no longer hold its conformation to the surface of the wing and begins to separate. This results in buffeting which is recognized by pilots as the prelude to a stall. For a given airfoil this always occurs at the same aoa. Stall warning systems are built into most aircraft which warn as this critical aoa is approached. These stall warning systems vary from simple reed type buzzers; all the way to digital fly-by-wire systems that limit aoa. Pedal shakers; stick shakers; stick/yoke pushers; horns; voice commands and others are used for this purpose in other aircraft. What these all have in common; is they are bases on normal aoa induced boundary layer separation of the typical stall. However; if the boundary layer flow is disturbed by an external source such as airfoil icing; use of flight spoilers or a sonic shock wave; separation can occur at a significantly lower aoa. Most jet transports are certified with a mach as well as calibrated or indicated speed limit. The indicated mach number is the speed read in the cockpit and typically represents the speed of the air entering the pitot tube(s). However the relative speed of air curving around the various surfaces of the airframe is significantly higher than that entering the pitot system. Normally the mach limit is set just below the speed at which the first bit of relative air moving around the airframe goes supersonic. In most transport aircraft that is the air moving over the inboard (thickest) part of the wing. Should the limit be exceeded and the flow reach mach 1 then a shock wave is generated which disrupts the flow of the air passing through it. The resulting turbulence can be felt as a buzz or high pitched rumble in the airframe. At low aoa this turbulence is not sufficient to cause separation of the boundary layer. Usually all that is needed is to reduce speed somewhat to return to normal flight. These mach limits are established based on high calibrated/indicated airspeed and low aoa found in a typical descent where most overspeeds occur. The problem arises in that the difference between the speed of air at the pitot tube and the speed of the air over the upper wing increases as aoa is increased. As we climb the plane higher and higher we have to fly a progressively slower calibrated/indicated airspeed to remain below our mach limit. As we slow our airspeed we must increase our aoa to sustain level flight. In swept wing aircraft; this increase is not linier but progressive. Eventually in this climb we reach a point at which the relative speed over the airfoil exceeds mach 1 even though we may be indicating a mach number well below the published limit. More commonly; while operating in level flight above FL400 we allow the aoa to inadvertently increase to sonic speed by turning too steeply; encountering turbulence; or simply allowing the aircraft to slow for some reason (such as slowing for holding as my friend did). What happens in this case is that the aoa regime for subsonic flight above FL400 is such that for swept wing aircraft a reduction in calibrated/indicated airspeed results in a proportionately higher increase in aoa to the point that the relative speed/mach of air across the airfoil is increasing at a higher rate than the indicated mach number is decreasing. Now the airfoil is at an aoa still well above the calculated stall but beyond the point where the boundary layer can avoid separation should it be disrupted. In this environment; unplanned turbulence from a shock wave causes the boundary layer to separate and the airfoil behind the shockwave to stall. Since the wing in front of the shock wave is still producing lift; the center of lift/pressure instantly shifts forward; causing the plane to pitch up. This results in a further increase in aoa causing more of the wing to become supersonic with increasingly negative results that end in a full aerodynamic stall. The real trap occurs when speed is reduced for 'spacing' or worse; for 'holding speeds.' if the buffet is not encountered in level flight it may be imminent and can occur abruptly with an increase in alpha such as occurs in a turn or in turbulence. The resulting 'disruption' can become a full blown stall almost instantly with no stall warning! When this happens at high altitude level flight cannot be maintained. Autopilots and fmcs are not programmed to adequately compensate for this threat! Crews are not trained for this type event. Airline certified full flight simulators are not programmed to replicate this threat! This is a significant threat to the safety of the increasing number of private and commercial jets operation at altitudes above FL390. Air crews need to be made aware of this threat and trained in its avoidance. Flight control and FMS (flight management system) software should be programmed to compensate for this threat.

Related reports:

Google
 

Original NASA ASRS Text

Title: Ground instructor submitted an article on Mach Induced Stall.

Narrative: Below is an article I wrote a number of years ago while working in the Training Department of [an airline company]. It was distributed to all of the pilots there; but to my knowledge has been passed along to the greater aviation community. I regret not have been more aggressive in trying to get this information out to the larger population of pilots. Sadly; since that time; there have been several of these unexpected upsets and one fatal accident which I am convinced were a result of this threat. I hope this finds its way to someone with enough interest in flight safety to verify the phenomenon of Mach Induced Stall; and to take actions that will help protect pilots; crews and the traveling public from this threat. Please contact me with any questions or concerns.A few years ago a friend; who was a B737 Captain at the time; told me of a very unsettling few seconds he and his crew of a B737 had experienced inbound to the New York area. Cruising at FL410 the flight was instructed to enter holding over PXT due to next sector saturation. The hold was programmed into the FMC and as the designated point was approached; the auto flight system began to slow the aircraft to the computed hold speed of 214 KCAS. Upon crossing the holding fix; the aircraft rolled into the 'standard' 25-30 degree bank turn. This is when the unexpected happened. In a split second; with no warning; the aircraft went from smooth flight into heavy buffet (no stick shaker); the AP (Autopilot) disconnected and the plane dropped 800 ft before control could be regained by the startled crew. Two weeks later; after relating his experience; he still had no idea what had really happened. As it turns out; this is a phenomenon with which I had some experience; and had followed up with some enlightening research. My initial encounter occurred while flying an F-4 simulating a high altitude subsonic intruder for intercept training. Our altitude was in the low 40s and required that we fly at a relatively slow (for the F-4) indicated airspeed and high angle of attack (AOA/Alpha) in order to remain below Mach 1. As we climbed; we reached a condition in which the aircraft suddenly became very unstable in pitch axis. Though we were well below the AOA at which buffeting normally began; I found that any slight increase in aft stick pressure would result in an instant onset of buffeting accompanied by the nose continuing to pitch up and the aircraft sinking. Because of the excellent control response of the Phantom; I was able to fly it for some time right on the edge; where we were in and out of this buffeting; and develop some feel for what the plane was doing and how to counter it. I just didn't know 'why' at the time. After landing; I described the experience to many of the 'old heads;' but none had any similar experiences or explanations. So; I tucked it away with the rationalization that it must have been a peculiarity of that particular aircraft - maybe it was a little bent. Some years later; flying an F-105 in the low 40s (exceptionally high for a Thud) I had a similar experience. As this was too much like the F-4 event; I launched into some research; the crux of which revealed that this can occur to any aircraft flying at high subsonic Mach and high AOA. A few years later I verified that even an F-16 exhibits the same characteristics under similar conditions.For any given airfoil as the AOA is increased; the speed of the airflow over the upper surface increases because the relative air must travel an increasingly longer distance. Furthermore; as the AOA increases the attachment of the boundary layer to the wing becomes more and more tenuous as it has to bend more radically around the wing. At low Mach numbers with this increasing AOA we eventually reach a point at which the boundary layer can no longer hold its conformation to the surface of the wing and begins to separate. This results in buffeting which is recognized by pilots as the prelude to a stall. For a given airfoil this always occurs at the same AOA. Stall warning systems are built into most aircraft which warn as this critical AOA is approached. These stall warning systems vary from simple reed type buzzers; all the way to digital fly-by-wire systems that limit AOA. Pedal shakers; stick shakers; stick/yoke pushers; horns; voice commands and others are used for this purpose in other aircraft. What these all have in common; is they are bases on normal AOA induced boundary layer separation of the typical stall. However; if the boundary layer flow is disturbed by an external source such as airfoil icing; use of flight spoilers or a sonic shock wave; separation can occur at a significantly lower AOA. Most jet transports are certified with a Mach as well as Calibrated or Indicated speed limit. The indicated Mach number is the speed read in the cockpit and typically represents the speed of the air entering the pitot tube(s). However the relative speed of air curving around the various surfaces of the airframe is significantly higher than that entering the pitot system. Normally the Mach limit is set just below the speed at which the first bit of relative air moving around the airframe goes supersonic. In most transport aircraft that is the air moving over the inboard (thickest) part of the wing. Should the limit be exceeded and the flow reach Mach 1 then a shock wave is generated which disrupts the flow of the air passing through it. The resulting turbulence can be felt as a buzz or high pitched rumble in the airframe. At low AOA this turbulence is not sufficient to cause separation of the boundary layer. Usually all that is needed is to reduce speed somewhat to return to normal flight. These Mach limits are established based on high calibrated/indicated airspeed and low AOA found in a typical descent where most overspeeds occur. The problem arises in that the difference between the speed of air at the pitot tube and the speed of the air over the upper wing increases as AOA is increased. As we climb the plane higher and higher we have to fly a progressively slower calibrated/indicated airspeed to remain below our mach limit. As we slow our airspeed we must increase our AOA to sustain level flight. In swept wing aircraft; this increase is not linier but progressive. Eventually in this climb we reach a point at which the relative speed over the airfoil exceeds Mach 1 even though we may be indicating a Mach number well below the published limit. More commonly; while operating in level flight above FL400 we allow the AOA to inadvertently increase to sonic speed by turning too steeply; encountering turbulence; or simply allowing the aircraft to slow for some reason (such as slowing for holding as my friend did). What happens in this case is that the AOA regime for subsonic flight above FL400 is such that for swept wing aircraft a reduction in calibrated/indicated airspeed results in a proportionately higher increase in AOA to the point that the relative speed/Mach of air across the airfoil is increasing at a higher rate than the indicated Mach number is decreasing. Now the airfoil is at an AOA still well above the calculated stall but beyond the point where the boundary layer can avoid separation should it be disrupted. In this environment; unplanned turbulence from a shock wave causes the boundary layer to separate and the airfoil behind the shockwave to stall. Since the wing in front of the shock wave is still producing lift; the center of lift/pressure instantly shifts forward; causing the plane to pitch up. This results in a further increase in AOA causing more of the wing to become supersonic with increasingly negative results that end in a full aerodynamic stall. The real trap occurs when speed is reduced for 'spacing' or worse; for 'holding speeds.' If the buffet is not encountered in level flight it may be imminent and can occur abruptly with an increase in Alpha such as occurs in a turn or in turbulence. The resulting 'disruption' can become a full blown stall almost instantly with no stall warning! When this happens at high altitude level flight cannot be maintained. Autopilots and FMCs are not programmed to adequately compensate for this threat! Crews are not trained for this type event. Airline certified Full Flight Simulators are not programmed to replicate this threat! This is a significant threat to the safety of the increasing number of private and commercial jets operation at altitudes above FL390. Air crews need to be made aware of this threat and trained in its avoidance. Flight Control and FMS (Flight Management System) Software should be programmed to compensate for this threat.

Data retrieved from NASA's ASRS site and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.