The Old N1 and EPR again!
Join Date: Feb 2005
Location: flyover country USA
Age: 82
Posts: 4,579
Likes: 0
Received 0 Likes
on
0 Posts
AirRabbit, I don't think anyone's disputing your analysis. (But, if what you say were generally true, why aren't there many more similar accidents/incidents, not only on the 737, but other swept-wing aircraft?)
The official (NTSB Identification: DCA82AA011) report however cites the EPR PT2 probe icing as contributing to the false thrust setting.
The official (NTSB Identification: DCA82AA011) report however cites the EPR PT2 probe icing as contributing to the false thrust setting.
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
Hey Barit1 – Thanks for the comment.
Believe me, I KNOW what the NTSB report says. But what it doesn’t say is that on the same day, (actually it was about 8 hours earlier) half-way around the world in Scandinavia, another B737 was subjected to a rather fierce snow/freezing rain scenario while taxiing out and awaiting takeoff clearance. As the crew initiated the takeoff and brought the control column “to neutral or slightly aft of neutral to prepare for the rotation” (quote from the Boeing manual), that B737 auto-rotated just like the B737 in Washington. However, because this airplane was subjected to a crosswind, the deformation on the wings was asymmetrical. The airplane pitched up to about the same pitch attitude as in Washington (in the neighborhood of 22-23 degrees) but because of the asymmetrical deformation, the pitch was asymmetrical as well, and resulted in a pitch up and roll off. The crew had the control column against the forward stops. They slammed the throttles to the firewall, went to full opposite aileron, full opposite rudder – to no avail. They were along for the ride. However, as the roll continued over toward 90-degrees, the nose began to fall. As the nose began to come down toward the horizon, the airplane began to accelerate. As it accelerated, the outboard portion of the affected wing began to produce lift and the aileron became effective. The crew rolled level less than 100 feet above the ground. When advised of the existence of this circumstance, the NTSB chose not to look into it.
As for why doesn’t this happen more often … Well, the Washington accident airplane was deiced – but it was deiced with hot water – sprayed very evenly over the entire airplane, which promptly froze in the 22-degree weather. There was a malfunction in the Trump De-icing Truck and the ground crew made improper repairs that resulted in drawing only from the water tank when spraying anything above the “ON” position dribble out of the nozzle. So, this circumstance is not seen very often because not very many airplanes are deiced with water – carefully, all over the airplane, with particular attention given to the lifting surfaces, when its way below freezing outside!
As you would probably guess there was a flurry (no pun intended) of lawsuits filed and everyone was pointing fingers at what “came out” during the NTSB public hearings. It was interesting when some found out that during the flight-testing of the B737, Boeing had recorded several instances of a “pitch up” or a “pitch-up/roll-off” occurring during periods of freezing precipitation (snow/freezing rain primarily). In these situations the test pilot indicated that the airplane was not controllable, and it was all written off as “autopilot anomalies.” This, in combination with the revelation that Boeing had been working on the development of an anti-icing/deicing system to be used during takeoff for the B737, led to the fact that the settlements reached were split between the operator’s insurance company and Boeing. The NTSB report acknowledges “the known, inherent pitch-up and roll-off tendency of the B737” but attached no particular importance to it.
I know that the report cites the PT2 probes as being iced over. I also believe that to be the case. But the power setting is not what gets an airplane into the air. Flight crews rotate when they get to the previously computed rotate airspeed. It is airspeed that allows the airplane to fly. It is airspeed, or lack thereof, that causes an airplane to stall. In fact, given that the takeoff would have been somewhat longer with only 75% power from both engines from initiation of the takeoff roll, it still should have achieved flight and certainly should have been able to sustain flight – it had 50% more power than with what the airplane had been certificated to operate. The problem is that if only the inboard portion of the wing will generate lift or it generates substantially more lift than the outboard portion of the wing, when the forward control column pressure is relaxed, the inboard (forward) portion of the wing acts just like it should – and the outboard (aft) portion does as well. The result is a pitch up – and if it is asymmetrical, you get a roll as well.
I’ve had the opportunity to make several “simulated single engine” takeoffs in DC-9 aircraft (not the simulator), from initiation of takeoff roll, using one engine at idle power and the other at takeoff power. The throttle has to be advanced slowly and steadily while maintaining a good grip on the NWS until the rudder becomes effective. Yes, the takeoff roll is longer, but it got into the air every time.
There are other “interesting” factors about the NTSB report that I could go into, but I’ve probably already said more than enough to satisfy anyone who was confused about what really happened that day in January 1982.
Let me know if you “want more” -- because there certainly is more…
__________
AirRabbit
Believe me, I KNOW what the NTSB report says. But what it doesn’t say is that on the same day, (actually it was about 8 hours earlier) half-way around the world in Scandinavia, another B737 was subjected to a rather fierce snow/freezing rain scenario while taxiing out and awaiting takeoff clearance. As the crew initiated the takeoff and brought the control column “to neutral or slightly aft of neutral to prepare for the rotation” (quote from the Boeing manual), that B737 auto-rotated just like the B737 in Washington. However, because this airplane was subjected to a crosswind, the deformation on the wings was asymmetrical. The airplane pitched up to about the same pitch attitude as in Washington (in the neighborhood of 22-23 degrees) but because of the asymmetrical deformation, the pitch was asymmetrical as well, and resulted in a pitch up and roll off. The crew had the control column against the forward stops. They slammed the throttles to the firewall, went to full opposite aileron, full opposite rudder – to no avail. They were along for the ride. However, as the roll continued over toward 90-degrees, the nose began to fall. As the nose began to come down toward the horizon, the airplane began to accelerate. As it accelerated, the outboard portion of the affected wing began to produce lift and the aileron became effective. The crew rolled level less than 100 feet above the ground. When advised of the existence of this circumstance, the NTSB chose not to look into it.
As for why doesn’t this happen more often … Well, the Washington accident airplane was deiced – but it was deiced with hot water – sprayed very evenly over the entire airplane, which promptly froze in the 22-degree weather. There was a malfunction in the Trump De-icing Truck and the ground crew made improper repairs that resulted in drawing only from the water tank when spraying anything above the “ON” position dribble out of the nozzle. So, this circumstance is not seen very often because not very many airplanes are deiced with water – carefully, all over the airplane, with particular attention given to the lifting surfaces, when its way below freezing outside!
As you would probably guess there was a flurry (no pun intended) of lawsuits filed and everyone was pointing fingers at what “came out” during the NTSB public hearings. It was interesting when some found out that during the flight-testing of the B737, Boeing had recorded several instances of a “pitch up” or a “pitch-up/roll-off” occurring during periods of freezing precipitation (snow/freezing rain primarily). In these situations the test pilot indicated that the airplane was not controllable, and it was all written off as “autopilot anomalies.” This, in combination with the revelation that Boeing had been working on the development of an anti-icing/deicing system to be used during takeoff for the B737, led to the fact that the settlements reached were split between the operator’s insurance company and Boeing. The NTSB report acknowledges “the known, inherent pitch-up and roll-off tendency of the B737” but attached no particular importance to it.
I know that the report cites the PT2 probes as being iced over. I also believe that to be the case. But the power setting is not what gets an airplane into the air. Flight crews rotate when they get to the previously computed rotate airspeed. It is airspeed that allows the airplane to fly. It is airspeed, or lack thereof, that causes an airplane to stall. In fact, given that the takeoff would have been somewhat longer with only 75% power from both engines from initiation of the takeoff roll, it still should have achieved flight and certainly should have been able to sustain flight – it had 50% more power than with what the airplane had been certificated to operate. The problem is that if only the inboard portion of the wing will generate lift or it generates substantially more lift than the outboard portion of the wing, when the forward control column pressure is relaxed, the inboard (forward) portion of the wing acts just like it should – and the outboard (aft) portion does as well. The result is a pitch up – and if it is asymmetrical, you get a roll as well.
I’ve had the opportunity to make several “simulated single engine” takeoffs in DC-9 aircraft (not the simulator), from initiation of takeoff roll, using one engine at idle power and the other at takeoff power. The throttle has to be advanced slowly and steadily while maintaining a good grip on the NWS until the rudder becomes effective. Yes, the takeoff roll is longer, but it got into the air every time.
There are other “interesting” factors about the NTSB report that I could go into, but I’ve probably already said more than enough to satisfy anyone who was confused about what really happened that day in January 1982.
Let me know if you “want more” -- because there certainly is more…
__________
AirRabbit
Last edited by AirRabbit; 27th Sep 2005 at 02:36.
Join Date: Feb 2005
Location: flyover country USA
Age: 82
Posts: 4,579
Likes: 0
Received 0 Likes
on
0 Posts
I’ve had the opportunity to make several “simulated single engine” takeoffs in DC-9 aircraft (not the simulator), from initiation of takeoff roll, using one engine at idle power and the other at takeoff power. The throttle has to be advanced slowly and steadily while maintaining a good grip on the NWS until the rudder becomes effective. Yes, the takeoff roll is longer, but it got into the air every time.
Last edited by barit1; 28th Sep 2005 at 00:42.
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
barit1 :
Its not a matter of takeoff power or TOGW or runway length. Its a matter of a deformed leading edge that causes a pitch-up in the B737. You can choose to believe that or not. The AF accident was more than 20 years ago -- there has been a lot learned about hold-over times and deicing procedures in general. For that I am thankful. However, once that crew advanced the power with the intention of taking off, they were doomed. Full power from the initiation of the takeoff roll would have served no purpose unless they held the airplane on the ground until it accelerated to an airspeed where the all of the wing surfaces would support the weight of the airplane. As long as they indended to rotate at the computed rotate speed they were destined to crash -- full power or no power. Full power would likely have carried them beyond the 14th street bridge and those killed or injured ON the bridge would not have been -- but the rest of the scenerio would have been the same.
_______
AirRabbit
Its not a matter of takeoff power or TOGW or runway length. Its a matter of a deformed leading edge that causes a pitch-up in the B737. You can choose to believe that or not. The AF accident was more than 20 years ago -- there has been a lot learned about hold-over times and deicing procedures in general. For that I am thankful. However, once that crew advanced the power with the intention of taking off, they were doomed. Full power from the initiation of the takeoff roll would have served no purpose unless they held the airplane on the ground until it accelerated to an airspeed where the all of the wing surfaces would support the weight of the airplane. As long as they indended to rotate at the computed rotate speed they were destined to crash -- full power or no power. Full power would likely have carried them beyond the 14th street bridge and those killed or injured ON the bridge would not have been -- but the rest of the scenerio would have been the same.
_______
AirRabbit
Join Date: Feb 2005
Location: flyover country USA
Age: 82
Posts: 4,579
Likes: 0
Received 0 Likes
on
0 Posts
Please don't mistake what I'm saying. You may be perfectly right re the AF90 accident cause. I'm only saying that:
1. They had subpar acceleration, because of bad EPR readings.
2. You tend to discount the above point simply because you flew a simulation of the above. But you had a long enough runway, and/or low enough TOGW, to tolerate a subpar acceleration with a good wing airfoil.
Enlightening, but not yet conclusive.
1. They had subpar acceleration, because of bad EPR readings.
2. You tend to discount the above point simply because you flew a simulation of the above. But you had a long enough runway, and/or low enough TOGW, to tolerate a subpar acceleration with a good wing airfoil.
Enlightening, but not yet conclusive.
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
Hey barit1:
Yes, I completely agree with you that the AF B737 had sub-par acceleration. And, I also agree that the sub-par acceleration was due to having bad EPR readings. There is very little doubt about that. While there was no way to confirm that ice was the culprit (any ice would have melted in the water and the probes were clear when the engines were recovered), I do believe that ice was present and was the source of the erroneous EPR readings. The primary source of confirmation for me is the sound spectrum analysis done on the sound of the engines as recorded on the CVR. The frequencies registered approximately 75% of the expected RPM.
I think you and I exchanged opinions on ground acceleration checks and the methods by which that may be done in an earlier thread – and I still support having some kind of acceleration measurement, particularly if it is independent of airplane systems. [I flew KC-135s for quite a while during my “military life.” We used a minimum acceleration check time between 80 and 120 knots – as called out by the pilot monitoring and timed by the navigator. It was simple and effective.]
Having said all that, however, the AF B737 did accelerate (slowly, yes) but enough to get to V1. They also got to V2, but that call was made just as the stick shaker started and the stall buffet was entered. They did get airborne. What I’m trying to say is that while they got airborne, they did so thorough no effort on their part. The F/O (who was flying) specifically said he was going to “takeoff the nose gear and let her fly off.” Going from 3-on-the-ground to being in the stall buffet in approximately 2 seconds (requiring a rotation rate of something like 4 – 6 times normal) is a whole lot more than merely “taking off the nose gear and letting her fly off.” What they were concerned about at that very instant – and they were very concerned – was the pitch attitude. “Come on, forward. Forward. Just barely climb. We only want 500 hundred. Forward.”
I’m looking at straightforward aerodynamics, and what role aerodynamics played in what happened. To do so, I’ll ask that you allow me the luxury of postulating for a moment. We know that during the initial portion of any acceleration for takeoff, the wing is beginning to develop lift. But it is not until the pilot rotates the airplane, getting the wing to an AoA that generates enough lift, that the airplane gets into the air. I’m sure that it isn’t any super revelation to state that the wing does not generate lift uniformly and the entire wing doesn’t generate lift simultaneously. You and I know that you can get an airplane into the air, in ground effect, before it is really ready to fly outside of ground effect. We know that Vmu tests are done where the controls are essentially held back from early in the acceleration run, forcing the tail onto the ground (or very close to it) at a speed well below what is necessary to fly – to see what the minimum lift off speed will be. The reason pilots don’t rotate the airplane prior to reaching “rotate speed” is that they don’t want the airplane getting into the air until the wings support the airplane properly and controllably. This is what they learn. This is what they expect. This is what they do.
But, what would happen if we changed the equation a bit – right at that critical moment – when the pilot moves the control column to a “neutral position, or slightly aft of the neutral position, in preparation to rotate.” (A quote from the Boeing manual.) What if, at that moment, the pilot realized that pulling further back on the controls would not get the airplane rotated? The pilot, pulling like crazy on the controls, gets no rotation. Well, in my KC-135 days, instructors and evaluators were trained on the unique use of the spoiler panels. Under the glare shield were 2 guarded switches that controlled 2 valves to open or shut off hydraulic pressure to either the inboard spoilers (L) or the outboard spoilers (R) on each wing. In this case (no rotation – and you had to know that the KC-135 had p*ss poor brakes and no reverse thrust – quick stops were, well, rare.) and you wanted to get into the air, you would turn off the inboard spoilers, grab the speed brake lever, and gradually, very gradually, raise the speed brakes. With the inboard panels shut off, the only spoiler panels being raised were the outboard panels, creating differential lift – lift on the inboard portion (forward) and no lift on the outboard portion (aft). The airplane would rotate just like “normal” and when you got the pitch attitude you needed, you lowered the speed brakes, just as gradually. If you yanked the spoiler handle up quickly, you would very likely smack the tail on the ground. Differential lift at takeoff can be very interesting, to say the least.
In the AF B737 situation, a very similar circumstance was handed to an unsuspecting crew. When the F/O moved the control column to the neutral position, the inboard portion of the wing was generating lift – and outboard portion was not. Presto. Rotation. Quick rotation. Very quick rotation. Rotation all the way into the stall buffet in approximately 2 seconds. But here, the crew couldn’t get rid of the differential lift. It was “glued” onto the wing leading edges. Straightforward aerodynamics.
If the airplane had been kept on the ground long enough to accelerate to a speed that would have allowed the outboard portion of the wing to generate enough lift to counter act the rotational moment they probably would have recognized a “sluggish” airplane – with only 75% power. But they didn’t know to do that. They wouldn’t have been able to do that – even if they had shoved the throttles to the firewall at brake release. They planned to rotate at the computed speed. They got ready to rotate at the computed speed. They expected the airplane to fly at the computed speed. What they didn’t expect was a wing deiced with water; a wing that now had a thin coating of ice that deformed the leading edges; deformed just enough to cause this very unique aerodynamic problem – from which they were incapable of recovering.
Again, thanks for your comments, and for allowing me to praddle on.
_______
AirRabbit
Yes, I completely agree with you that the AF B737 had sub-par acceleration. And, I also agree that the sub-par acceleration was due to having bad EPR readings. There is very little doubt about that. While there was no way to confirm that ice was the culprit (any ice would have melted in the water and the probes were clear when the engines were recovered), I do believe that ice was present and was the source of the erroneous EPR readings. The primary source of confirmation for me is the sound spectrum analysis done on the sound of the engines as recorded on the CVR. The frequencies registered approximately 75% of the expected RPM.
I think you and I exchanged opinions on ground acceleration checks and the methods by which that may be done in an earlier thread – and I still support having some kind of acceleration measurement, particularly if it is independent of airplane systems. [I flew KC-135s for quite a while during my “military life.” We used a minimum acceleration check time between 80 and 120 knots – as called out by the pilot monitoring and timed by the navigator. It was simple and effective.]
Having said all that, however, the AF B737 did accelerate (slowly, yes) but enough to get to V1. They also got to V2, but that call was made just as the stick shaker started and the stall buffet was entered. They did get airborne. What I’m trying to say is that while they got airborne, they did so thorough no effort on their part. The F/O (who was flying) specifically said he was going to “takeoff the nose gear and let her fly off.” Going from 3-on-the-ground to being in the stall buffet in approximately 2 seconds (requiring a rotation rate of something like 4 – 6 times normal) is a whole lot more than merely “taking off the nose gear and letting her fly off.” What they were concerned about at that very instant – and they were very concerned – was the pitch attitude. “Come on, forward. Forward. Just barely climb. We only want 500 hundred. Forward.”
I’m looking at straightforward aerodynamics, and what role aerodynamics played in what happened. To do so, I’ll ask that you allow me the luxury of postulating for a moment. We know that during the initial portion of any acceleration for takeoff, the wing is beginning to develop lift. But it is not until the pilot rotates the airplane, getting the wing to an AoA that generates enough lift, that the airplane gets into the air. I’m sure that it isn’t any super revelation to state that the wing does not generate lift uniformly and the entire wing doesn’t generate lift simultaneously. You and I know that you can get an airplane into the air, in ground effect, before it is really ready to fly outside of ground effect. We know that Vmu tests are done where the controls are essentially held back from early in the acceleration run, forcing the tail onto the ground (or very close to it) at a speed well below what is necessary to fly – to see what the minimum lift off speed will be. The reason pilots don’t rotate the airplane prior to reaching “rotate speed” is that they don’t want the airplane getting into the air until the wings support the airplane properly and controllably. This is what they learn. This is what they expect. This is what they do.
But, what would happen if we changed the equation a bit – right at that critical moment – when the pilot moves the control column to a “neutral position, or slightly aft of the neutral position, in preparation to rotate.” (A quote from the Boeing manual.) What if, at that moment, the pilot realized that pulling further back on the controls would not get the airplane rotated? The pilot, pulling like crazy on the controls, gets no rotation. Well, in my KC-135 days, instructors and evaluators were trained on the unique use of the spoiler panels. Under the glare shield were 2 guarded switches that controlled 2 valves to open or shut off hydraulic pressure to either the inboard spoilers (L) or the outboard spoilers (R) on each wing. In this case (no rotation – and you had to know that the KC-135 had p*ss poor brakes and no reverse thrust – quick stops were, well, rare.) and you wanted to get into the air, you would turn off the inboard spoilers, grab the speed brake lever, and gradually, very gradually, raise the speed brakes. With the inboard panels shut off, the only spoiler panels being raised were the outboard panels, creating differential lift – lift on the inboard portion (forward) and no lift on the outboard portion (aft). The airplane would rotate just like “normal” and when you got the pitch attitude you needed, you lowered the speed brakes, just as gradually. If you yanked the spoiler handle up quickly, you would very likely smack the tail on the ground. Differential lift at takeoff can be very interesting, to say the least.
In the AF B737 situation, a very similar circumstance was handed to an unsuspecting crew. When the F/O moved the control column to the neutral position, the inboard portion of the wing was generating lift – and outboard portion was not. Presto. Rotation. Quick rotation. Very quick rotation. Rotation all the way into the stall buffet in approximately 2 seconds. But here, the crew couldn’t get rid of the differential lift. It was “glued” onto the wing leading edges. Straightforward aerodynamics.
If the airplane had been kept on the ground long enough to accelerate to a speed that would have allowed the outboard portion of the wing to generate enough lift to counter act the rotational moment they probably would have recognized a “sluggish” airplane – with only 75% power. But they didn’t know to do that. They wouldn’t have been able to do that – even if they had shoved the throttles to the firewall at brake release. They planned to rotate at the computed speed. They got ready to rotate at the computed speed. They expected the airplane to fly at the computed speed. What they didn’t expect was a wing deiced with water; a wing that now had a thin coating of ice that deformed the leading edges; deformed just enough to cause this very unique aerodynamic problem – from which they were incapable of recovering.
Again, thanks for your comments, and for allowing me to praddle on.
_______
AirRabbit
Join Date: Feb 2005
Location: flyover country USA
Age: 82
Posts: 4,579
Likes: 0
Received 0 Likes
on
0 Posts
REALLY interesting background and analysis, AirRabbit, and I can see why you're so disappointed with the AF90 report.
Have you ever flown the KC-135R? I'm curious what they had to do to handle Vmc with the bigger engines. I know they put a bigger horizontal tail on the beast - but I didn't see the vertical tail changing.
Have you ever flown the KC-135R? I'm curious what they had to do to handle Vmc with the bigger engines. I know they put a bigger horizontal tail on the beast - but I didn't see the vertical tail changing.
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
Hey barit1:
Unfortunately, I had the pleasure of flying only the A's and Q's - both had water injection (a whole story by itself ) and they both had the hydraulically powered rudder. I don't know if the existing powered rudder was sufficient to handle the bigger engines or not. But, I would imagine that the R's would have been a little less "attention grabbing" during takeoff roll.
_______
AirRabbit
PS - Thanks for the comment. I appreciate your understanding.
Unfortunately, I had the pleasure of flying only the A's and Q's - both had water injection (a whole story by itself ) and they both had the hydraulically powered rudder. I don't know if the existing powered rudder was sufficient to handle the bigger engines or not. But, I would imagine that the R's would have been a little less "attention grabbing" during takeoff roll.
_______
AirRabbit
PS - Thanks for the comment. I appreciate your understanding.
