KSFO ATC - PAN PAN PAN - China Airlines Boeing 747 Bird Strike

[okgood]

See the ATC exchange here

A China Airlines Boeing 747-400 freighter (flight CI-5107) from San Francisco to Taipei with 7 crew, was in the initial climb out of San Francisco's runway 28R, when just upon contacting departure and receiving instructions from ATC for further climb the crew declared PAN PAN PAN reporting a bird strike, that had taken out some of their flight instruments. The aircraft stopped the climb at 6000 feet, the crew indicated they were still considering how to proceed, there were no procedures for the faults. The crew decided to dump fuel and return to San Francisco (KSFO). The aircraft landed safely back on San Francisco's runway 28R about 75 minutes after departure. Both angle of attack probes were found to be damaged.

[Evan]

See the ATC exchange here

A China Airlines Boeing 747-400 freighter (flight CI-5107) from San Francisco to Taipei with 7 crew, was in the initial climb out of San Francisco's runway 28R, when just upon contacting departure and receiving instructions from ATC for further climb the crew declared PAN PAN PAN reporting a bird strike, that had taken out some of their flight instruments. The aircraft stopped the climb at 6000 feet, the crew indicated they were still considering how to proceed, there were no procedures for the faults. The crew decided to dump fuel and return to San Francisco (KSFO). The aircraft landed safely back on San Francisco's runway 28R about 75 minutes after departure. Both angle of attack probes were found to be damaged.

Just trying to figure out how damaged AoA vanes could 'take out' any instruments (aside from AoA indicators if present).

[okgood]

Just trying to figure out how damaged AoA vanes could 'take out' any instruments (aside from AoA indicators if present).

AOA is used to drive stall warning (stick shaker), stall margin information on airspeed indicators, and the pitch limit indicator (PLI) on the primary attitude displays. AOA information is combined with other data and displayed as an integral part of flight deck displays.

[Gabriel]

AOA is used to drive stall warning (stick shaker), stall margin information on airspeed indicators, and the pitch limit indicator (PLI) on the primary attitude displays. AOA information is combined with other data and displayed as an integral part of flight deck displays.

I don't know in the 744-400, but most modern airplanes (and "modernized' like the 737) also use it to "fine tune" the airspeed and altitude indication since same airspeed and altitude but at different angles of attack generate slightly different pressures in the static and dynamic ports. That's how you get speed disagree and altitude disagree when you have 1 AoA vane sending wrong data.

[Evan]

AOA is used to drive stall warning (stick shaker), stall margin information on airspeed indicators, and the pitch limit indicator (PLI) on the primary attitude displays. AOA information is combined with other data and displayed as an integral part of flight deck displays.

As I said, aside from AoA indicators. The most critical function of the AoA vanes is in providing data for autoflight and manual flight envelope protections (unless you have some half-baked stabilizer-hijacking software installed. The 74 thankfully does not). There would be some instrument disagree but the description makes it sound like they lost entire instruments. Still, especially the loss of the 'barberpole' on the speed tape is most likely a mandatory return. But I expect the larger issue would be the inability to engage the autopilot.

[Gabriel]

As I said, aside from AoA indicators. The most critical function of the AoA vanes is in providing data for autoflight and manual flight envelope protections (unless you have some half-baked stabilizer-hijacking software installed. The 74 thankfully does not). There would be some instrument disagree but the description makes it sound like they lost entire instruments. Still, especially the loss of the 'barberpole' on the speed tape is most likely a mandatory return. But I expect the larger issue would be the inability to engage the autopilot.

I am not sure I like how the technology is evolving, in a way that different sensors are not dedicated to individual indications but logically interconnected to produce a combined picture that, while superior to the old-school one, falls apart when one (or a couple) of sensors fail generating a cascade of side effects.

There is no reason why the autopilot and autohtrottle cannot hold the altitude and the speed without an AoA indication.

[BoeingBobby]

It's almost like watching CNN here now. " Breaking news" SMH

[Evan]

I am not sure I like how the technology is evolving, in a way that different sensors are not dedicated to individual indications but logically interconnected to produce a combined picture that, while superior to the old-school one, falls apart when one (or a couple) of sensors fail generating a cascade of side effects.

I don't know enough about the 747, but I did read a pprune post claiming that pulling the CB for one of the AoA sensors resulted in altitude and airspeed "flags" on the onside PFD. I assume that 'flag' here means loss of instrumentation. If so, losing both could result in a loss of instruments on both sides.

There is no reason why the autopilot and autohtrottle cannot hold the altitude and the speed without an AoA indication.

I think the mad scientists are thinking that, as AoA is not simply constant to a speed and attitude, the FMC should have actual sensor data on that parameter to assure the aircraft remains safely in the envelope. But I honestly don't know whether the autoflight on the 747 can remain operational without it.

[kent olsen]

I was giving IOE to a friend, who eventually became my Asst Chief Pilot, on a return from Germany in a DC-8 freighter. Somewhere over Ireland the autopilot failed. I asked him if he wanted to trade off hand flying. "Nope" he said and hand flew all the way back to Dover AFB. He had the touch.

When I worked for Evergreen, 1987 - 2001, we hand flew all of our airplanes (DC-9, B-727, DC-8, B-747) up to FL180 and FL 180 back down including the approach, (if you were comfortable). Then at NetJets I flew the Hawker 1000 and Citation X until 2015. I had the reputation of hand flying and using the rudder, more than anyone else, at least in the X. Am I bragging, YES!!!

[BoeingBobby]

I was giving IOE to a friend, who eventually became my Asst Chief Pilot, on a return from Germany in a DC-8 freighter. Somewhere over Ireland the autopilot failed. I asked him if he wanted to trade off hand flying. "Nope" he said and hand flew all the way back to Dover AFB. He had the touch.

When I worked for Evergreen, 1987 - 2001, we hand flew all of our airplanes (DC-9, B-727, DC-8, B-747) up to FL180 and FL 180 back down including the approach, (if you were comfortable). Then at NetJets I flew the Hawker 1000 and Citation X until 2015. I had the reputation of hand flying and using the rudder, more than anyone else, at least in the X. Am I bragging, YES!!!

I had quite a few F/O's not only object, but I had one call the chief pilot and complain about me hand flying the -8 to FL 240. Needless to say, he and I had a good laugh about it when he called me to let me know.

[Evan]

After a bit more digging, I found a report for a similar issue, but involving only one AoA sensor failure (improperly refurbished with a gear train slippage). The AP was available despite an ALT DISAGREE and a IAS DISAGREE message appearing on EICAS. There were slight disagreements between the L and R instruments due to the different error correction calculations carried out by the ADC's (Air Data Computers). There was no loss of onside instruments.

The failure of both AoA sensors (both being damaged and resolving erronoeous AoA values) would likely result in a similar disagree if the ADC's still considered the sensor values to be valid. The crew can selectively switch the ADC providing data to onside instruments but with both in error there would be no way to determine which was valid. Running through the QRH procedures for the two faults reported on EICAS would be fruitless. I expect the best course of action would be to disconnect the autoflight and hand-fly a return with caution given to modest instrumentation errors (and AoA).

I still don't see how this would cause the loss of instrumentation, just some inaccuracy, but there nothing I could find of this happening in-situ.

There is no reason why the autopilot and autohtrottle cannot hold the altitude and the speed without an AoA indication.

The reason is ADC data supplied to the FMC's. That data would have erroneous altitude and airspeed corrections applied (or no corrections applied). So, for instance, the AoA corrected altitude might be 300-500ft off, mach would be off, speeds would be off. If you are going to entrust the flight to automation, it should not be available if air data is unreliable.

You might be referring to old-school, pre-digital, pre-ADC autopilot, but then remember that in those systems, at high AoA, the static and pitot sensor readings will be off. Not a good place for automation to be making errors.

The attachment is unfortunately low-res but you can squint out to basics, which show the circuitry paths of how AoA sensor signals affect things like static port error corrections that effect altitude, airspeed and mach calculation (the red box is the AoA sensor unit) .

[Gabriel]

You might be referring to old-school, pre-digital, pre-ADC autopilot

Precisely. There was a time where the airspeed indicator was just a differential manometer (calibrated in fancy speed units) measuring the difference between total pressure in the hose connected to the pitot minus the pressure in the hose connected to the static port, and the altimeter was just an absolute manometer measuring the pressure in the static port (aka barometer, again just calibrated in fancy length units), and the Mach number was a bit more complicated since there was in fact a computation involving the total pressure, static pressure and outside air temperature. In those conditions an AoA vane failure would not render the AP and AT inop. In fact, I suspect that the AP and AT would not even know that there was an AoA vane.

While I see the advantages of today technology, we need to also recognize the drawbacks: A failure in an AoA vane generates speed disagree, alt disagree, AP disconnect, AT disconnect, I don't know what would happen to the the PLI (it would be either unavailable or show erroneous info) and the FD, and to the barber pole stripes in the airspeed indicator, perhaps a stall warning, and sometimes a spooky uncommanded nose-down trim. It would be a really bad day for a stupid malfunction.

[Evan]

In fact, I suspect that the AP and AT would not even know that there was an AoA vane.

Was there an AoA vane prior to digital autoflight?

I'm not sure what you're getting at. RVSM was made possible by these air data computer error corrections. You're not suggesting that RVSM was a bad idea...

[Gabriel]

Was there an AoA vane prior to digital autoflight?

Yes. It was there for the different types of stall protection systems, some of which are quite complex. The MD-80 (and I think the DC-9 too) had first an auto-slats, the stickshaker, then the aural stall warning (at the actual critical AoA) and then the stick pusher.

I'm not sure what you're getting at. RVSM was made possible by these air data computer error corrections. You're not suggesting that RVSM was a bad idea...

As I said: "While I see the advantages of today's technology, we need to also recognize the drawbacks"

I believe that there is still a lot of room for making the current systems more robust. What is the pilot suppose to do when the automation gives up and what clues is the pilot supposed to use to control the plane in that situation, that the automation cannot do it by itself? We talked about that around AF-447.

Oh, and by the way, I am not sure that RVSM requires these corrections. There are still 737-200, DC-9s flying. I don't think that they have AoA-corrected airspeed and altitude, and I think they can fly in RVSM airspace.