Continental 737 Off Runway at DEN
Do we have any theories as to why we see left rudder when the a/c is apparently heading rapidly left towards the edge of the runway? I reckon my right foot would have been sticking out of the radome at that point.
Given how much practice we get at steering aircraft on the ground using the rudder, it's almost reflex action to apply it in the correct manner. I think it would be very difficult to steer in the opposite direction... But that's what the FDR says.
It also says that pretty much full right roll control was applied about 3-4 seconds before leaving the runway. Not to mention that during the reject, the right engine stayed at TO power for several seconds longer than the left. + + =
Re 'theories'
There appears to be a sudden increase in lateral acceleration without significant heading change, i.e. sideways drift but little yaw. This could be an aircraft type related effect – the nosewheel sticks to the runway (good wheel related directional stability) or due to the nature of the side force on the fuselage/winglets etc (little yaw tendency). FDR 18.20.7 onwards, 18.23.2 in particular.
‘Standard’ simulator training (often IFR) may only consider yaw – expecting a change in heading to be controlled by rudder. In general, simulators do not have side force – no lateral acceleration. Thus, crews may be unfamiliar with the reality of a strong side force body cue, and hence be confused about the motion of the aircraft. Which way is it heading (probably assessed visually – slightly into the runway?), which way is the aircraft accelerating – forward and sideways (towards the edge of the runway).
There appears to be a sudden increase in lateral acceleration without significant heading change, i.e. sideways drift but little yaw. This could be an aircraft type related effect – the nosewheel sticks to the runway (good wheel related directional stability) or due to the nature of the side force on the fuselage/winglets etc (little yaw tendency). FDR 18.20.7 onwards, 18.23.2 in particular.
‘Standard’ simulator training (often IFR) may only consider yaw – expecting a change in heading to be controlled by rudder. In general, simulators do not have side force – no lateral acceleration. Thus, crews may be unfamiliar with the reality of a strong side force body cue, and hence be confused about the motion of the aircraft. Which way is it heading (probably assessed visually – slightly into the runway?), which way is the aircraft accelerating – forward and sideways (towards the edge of the runway).
Join Date: Feb 2007
Location: Not here
Posts: 222
Likes: 0
Received 0 Likes
on
0 Posts
Fullwings, re.: " Not to mention that during the reject, the right engine stayed at TO power for several seconds longer than the left. "
The left engine lost half it's rpm in 1 second after hitting the snow bank on the edge of the runway. See 1 of the reports; wreckage report.
Also, at the point where the most right rudder was applied (but for a short time), the heading properly increased; albeit not long enough.
I remember flying in Utah and Arizona with very strong crosswinds !! It's hell to know that the landing gear "it's as if it's gonna be torn off" with the side forces. (C172).
.
The left engine lost half it's rpm in 1 second after hitting the snow bank on the edge of the runway. See 1 of the reports; wreckage report.
Also, at the point where the most right rudder was applied (but for a short time), the heading properly increased; albeit not long enough.
I remember flying in Utah and Arizona with very strong crosswinds !! It's hell to know that the landing gear "it's as if it's gonna be torn off" with the side forces. (C172).
.
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
At the beginning of the takeoff roll (approximately 18:17:40) the control column and the control wheel were positioned at the neutral (0 degrees) position. At 18:18:01, the rudder pedal began a movement to the right, gradually increasing to approximately 4½ degrees (18:18:04½), where it was held essentially constant for approximately 2 seconds (18:18:06½) when it was moved further to the right to approximately 11 degrees (18:18:07½), the maximum rudder pedal deflection achieved. Over the next 11 seconds, the rudder pedals are moved to an essentially neutral (0 degrees) position and returned to slightly less than 10 degrees right. There wasn’t any control column movement (up or down) until 18:18:08 (airspeed slightly above 70 kts) when it appears to have been moved aft (nose up) approximately 1 degree, at the same time the left aileron was moved UP and the right aileron DOWN (a left turn position) and held in that position until 18:18:14 (for 6 seconds) – corresponding to approximately 105 kts), at which time the control column is again placed in the neutral (0 degrees deflection) position and the aileron controls are reversed (right turn position). The control column doesn’t appear to have been moved to the forward (nose down) position until 18:18:18 (at approximately 120 kts – the highest airspeed attained), which apparently coincides with the time the airplane departed the runway.
Am I the only one who would have expected to see the control column moved to the full forward position at the beginning of the takeoff roll, and held at least until aerodynamic control was sufficient to maintain solid nosegear contact with the ground? Without the nosegear having positive contact, and with a crosswind of the amount cited, the aircraft would have most assuredly been more susceptible to the effects of a crosswind. Additionally, with such a strong crosswind blowing across a rather large, snow-covered field directly adjacent to the runway, it would be more than likely that some snow (or at least ice crystals) was blown onto the runway surface. Did that contribute anything? We may never know. Unfortunately, the FDR does not provide a readout of tiller position. Experience is that use of the tiller early on in crosswind conditions exacerbates any crosswind effect on directional control down the runway for any airplane – simply because of the magnitude of nosegear tire movement with tiller application, preventing solid contact with the surface.I'm just readin' what I'm readin'
Join Date: Jun 2009
Location: US
Posts: 497
Likes: 0
Received 0 Likes
on
0 Posts
When the aircraft was certified the test pilots proved the crosswind capabilities of the aircraft I am sure using standard crosswind techniques like using down elevator to put weight on the nose wheels to help prevent weather vaning problems until the rudder became effective and using rudder to maintain centerline. I don't think they test it with neutral or slightly up elevator like this flight did.
p51, you should not assume that all flight testing uses different procedures when determining certification procedures. Whatever manufacturers’ claim (written in the manuals) has to be demonstrated to an airworthiness authority and it has to be repeatable by an ‘average’ pilot.
Similarly, you assume ‘a standard technique’. Aircraft may have different recommendations, even amongst type variants, but generally, pushing forward on the stick to load the nose wheel may not be the best technique.
If this idea is due to a mistaken belief that nose wheel steering is improved, this may not be true as airspeed increases, where the aerodynamic control becomes increasingly more powerful – use the rudder (same on landing – use the rudder at high speed).
Also there is a need to keep weight on the main wheels to provide a countering side force to the wind generated fuselage side force, preventing the aircraft being blow off the runway sideways.
Some aspects are shown in 8.7 — Crosswind Landings and in Safety – Landing Techniques – crosswind.
Similarly, you assume ‘a standard technique’. Aircraft may have different recommendations, even amongst type variants, but generally, pushing forward on the stick to load the nose wheel may not be the best technique.
If this idea is due to a mistaken belief that nose wheel steering is improved, this may not be true as airspeed increases, where the aerodynamic control becomes increasingly more powerful – use the rudder (same on landing – use the rudder at high speed).
Also there is a need to keep weight on the main wheels to provide a countering side force to the wind generated fuselage side force, preventing the aircraft being blow off the runway sideways.
Some aspects are shown in 8.7 — Crosswind Landings and in Safety – Landing Techniques – crosswind.
Join Date: Jun 2009
Location: US
Posts: 497
Likes: 0
Received 0 Likes
on
0 Posts
Normal technique as I have always understood is adequate forward yoke to keep positive weight on nose gear tires until rudder becomes effective. Obviously you don't want to push so hard the main gear gets light. Below 80 knots it would be difficult to push that hard. Basic airmanship early in flight training both in the military and civilian world has everybody I have ever flown with using basically the same technique. Since this flight used no down elevator at all and at one time 1 degree up and lost directional control in the strong crosswind and possibly a very strong gust at a fairly low speed could mean the rudder alone couldn't handle it if the nose gear started to skid with side loads.
Join Date: Feb 2007
Location: Not here
Posts: 222
Likes: 0
Received 0 Likes
on
0 Posts
Join Date: Mar 2009
Location: Feriton
Posts: 73
Likes: 0
Received 0 Likes
on
0 Posts
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
Yeah, BOAC 
... here is one of many links:
Accident: Continental B735 at Denver on Dec 20th 2008, veered off departure runway and burst into flames
After the pretty decent photos and the transcript of the CVR, the FDR material is at the bottom.
... here is one of many links:Accident: Continental B735 at Denver on Dec 20th 2008, veered off departure runway and burst into flames
After the pretty decent photos and the transcript of the CVR, the FDR material is at the bottom.
Looking at the FDR data, the decrease in heading follows the decrease in rudder pedal and rudder deflection pretty well, from about 18:12 onwards. Add the drift angle to the heading and it produces a slowly decreasing ground track.
Just going on this, it *seems* like the airframe did exactly what you'd expect with hardly any rudder in a strong crosswind and departed stage left. It's still odd, though, as the earlier part of the recording shows some pretty positive corrections to maintain the centreline...
It reads as though the P1 forgot how to steer an aeroplane about 20s into the takeoff roll... Which is surprising, to say the least.
I can think of a couple of things but they're rather 'out there' in theory terms: what would happen mechanically with a steering failure? Could it possibly leave you unable to get much in the way of rudder deflection or are the systems not linked that closely? Given the reports of very strong and gusting winds and the mountains nearby, it's possible that there was some sort of (detached) rotor effect reaching ground level which would put the wind all over the place and maybe even reverse it for a short time...?
Just going on this, it *seems* like the airframe did exactly what you'd expect with hardly any rudder in a strong crosswind and departed stage left. It's still odd, though, as the earlier part of the recording shows some pretty positive corrections to maintain the centreline...
It reads as though the P1 forgot how to steer an aeroplane about 20s into the takeoff roll... Which is surprising, to say the least.
I can think of a couple of things but they're rather 'out there' in theory terms: what would happen mechanically with a steering failure? Could it possibly leave you unable to get much in the way of rudder deflection or are the systems not linked that closely? Given the reports of very strong and gusting winds and the mountains nearby, it's possible that there was some sort of (detached) rotor effect reaching ground level which would put the wind all over the place and maybe even reverse it for a short time...?
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Thanks all - the clearest FDR trace is from alph2Z http://www.ntsb.gov/Dockets/Aviation...021/418501.pdf The Av Herald is a little blurred.
Given possible sampling rate anomalies, I am somewhat puzzled by the divergence of rudder pos v pedal around 18:18:17 onwards In fact there seems to be a 'spike' around 18:18:05 as well) and likewise aileron v wheel from 18:18:19 on? Any experts to comment?
Given possible sampling rate anomalies, I am somewhat puzzled by the divergence of rudder pos v pedal around 18:18:17 onwards In fact there seems to be a 'spike' around 18:18:05 as well) and likewise aileron v wheel from 18:18:19 on? Any experts to comment?
Join Date: May 2000
Location: Glorious West Sussex
Age: 76
Posts: 1,020
Likes: 0
Received 0 Likes
on
0 Posts
BOAC, Greetings from a non-expert...
The rudder spikes may be caused by the yaw damper if the heading is changing rapidly.
Tha left and right aileron positions are sampled once per second, the left on the second, the right on the half-second (approx). If the control wheel is moving rapidly they will show different graph shapes. If you imagine them superimposed it sort of fits!
At 18:18:20 the left engine N1 decreases before the thrust lever is retarded, presumably because of the earth it is ingesting, and who knows what the ailerons will do after that...
Thinking of the broken nosewheel steering cable what we really need is the nosewheel angle - but good old Sod decided that the FDR would not record it!
Not much mention of the surface wind in the docket - surely the microburst capital of the world has very sophisticated wind recorders all over the airfield?
The rudder spikes may be caused by the yaw damper if the heading is changing rapidly.
Tha left and right aileron positions are sampled once per second, the left on the second, the right on the half-second (approx). If the control wheel is moving rapidly they will show different graph shapes. If you imagine them superimposed it sort of fits!
At 18:18:20 the left engine N1 decreases before the thrust lever is retarded, presumably because of the earth it is ingesting, and who knows what the ailerons will do after that...
Thinking of the broken nosewheel steering cable what we really need is the nosewheel angle - but good old Sod decided that the FDR would not record it!
Not much mention of the surface wind in the docket - surely the microburst capital of the world has very sophisticated wind recorders all over the airfield?
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Good 'non-expert' stuff, TP. The yaw damper authority is, I believe, 3 deg - which is exceeded in a few places. I guess the aileron plots are 'normal' (within sampling criteria) until later when I guess they were possibly being 'influenced'. by ground contact.
Join Date: Jul 2002
Location: Confusio Helvetica
Posts: 311
Likes: 0
Received 0 Likes
on
0 Posts
BOAC -- the chart you've listed is generated from the data also published on the site. That's where I got the numbers I put up earlier.
The CVR transcript also records two snapping sounds at 18:18:14 and 15 respectively. Could be anything, I suppose, but the time correlation is interesting. This correlation probably explains its inclusion on the transcript.
I didn't bother with the data after 18:18:18, since that's when the FDR seems to show the aircraft leaving the runway, and this is backed up by the statements of the powerplant group.
The CVR transcript also records two snapping sounds at 18:18:14 and 15 respectively. Could be anything, I suppose, but the time correlation is interesting. This correlation probably explains its inclusion on the transcript.
I didn't bother with the data after 18:18:18, since that's when the FDR seems to show the aircraft leaving the runway, and this is backed up by the statements of the powerplant group.
Join Date: Apr 2005
Location: Southeast USA
Posts: 801
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by BOAC
Given possible sampling rate anomalies, I am somewhat puzzled by the divergence of rudder pos v pedal around 18:18:17 onwards In fact there seems to be a 'spike' around 18:18:05 as well) and likewise aileron v wheel from 18:18:19 on? Any experts to comment?
I’d ask the same question I asked before, am I the only one who would have expected to see the control column moved to the full forward position at the beginning of the takeoff roll, and held at least until aerodynamic control was sufficient to maintain solid nosegear contact with the ground?