One approach to cockpit warnings is that it is better to have a warning, and be able to override it, than not have the warning. Dozy, I understand what you are saying based on the spec, and as we noted previously, you can use a weight on wheel switch to inhibit the stall warning when on the ground.

We have discussed this over and over, and the hamster is getting awfully tired. :cool:

Dozy,

I'm not a systems expert but how about, in the case of loss of all AoA data, the stall warning system retaining the last valid value of AoA until that parameter is valid again? There also should be an ECAM message suggesting selection of FPV to allow comparison of FPA and pitch attitude.

@LW_50:

I get what you're saying, but this isn't a case of Stall Warning being inhibited, it's a case of the data being unusable. You can override the SW, but if the failure is of a nature whereby it comes straight back on, it could cause problems.

@HN39:
Retaining the last valid value can cause the same issue - sure, in a scenario like AF447 the values can come back, but in other failure scenarios (damaged vane or cable) SW could be stuck on and cause problems, because the value will never be valid again.

And not having the warning can cause problems. See AF447. (Though that wasn't their only problem, not by a damned sight.)

It's a lose lose situation, other than the fact that there are multiple AoA probes on the A330. We went over that as well. We could even run the hamster for a while on the AoA indicator in the C/P ... but that got us nowhere other than "who uses it?" "if they aren't scanning primary, what else won't they scan?" and "it could cause problems."

Nothing new here, and that poor hamster is exhausted. :p

This all may shed some light on why the general training standard is prevent stall in large transport aircraft. Doing so renders a few of these problems moot.

10000 feet a minute is a good breeze enough for all AoA probes to indicate a pretty similar value. If the probes mechanical limit is at 85 deg I believe 40 to be in their range of operation.

The stall warning has to keep going.

Agreed - however...

The AF447 scenario was not a case of the Stall Warning suddenly disappearing after sounding briefly, the warning sounded for almost a *minute* with no apparent acknowledgement by the crew. The AoA readings became invalid because the aircraft had been pulled into an attitude so extreme that the sensors could no longer keep up.

This is what would be considered in engineering terms an extreme edge case, and thus far it has happened precisely once. If a technical solution was simple to implement without serious knock-on effects, it would be done - but the fact is that it is an extremely difficult problem to solve.

[EDIT : Additionally, while it may seem logical to assume that a warning in the face of uncertain data is better than none, I can't help but think of Birgenair 301 - where the captain fixated on a false Overspeed warning to the exclusion of all else and ended up stalling the aircraft. ]

Well, arguably that training standard was at the root of the problems because it was implemented at the expense of stall recognition and recovery, and as such the standards have been changed in the wake of the accident.

At the end of the day, we're not talking about a stall warning that stopped sounding in an unreasonably short period of time after stall was initiated, we're talking about an aircraft that was pulled even deeper into the stall while the warning was sounding - until it was so far outside the flight envelope that it exceeded even the theoretical limits of instrument behaviour.

Gee, a weight-on-wheels switch is too damned hard to wire up. Insanely difficult technical obstacle. You just went from "you only have to follow the numbers on the spec" to 'it's too hard." Further comments :mad: as I am done with this hamster abuse.

How would a WoW switch affect the fact that the sensors don't work in the air if certain parameters are exceeded?

DozyWannabe,

The stall warning system looks at three values of AoA from three independent sensors. It sounds a warning if the highest of the valid AoA values exceeds the threshold defined by the highest of the valid airspeeds. If I understand correctly, the system can loose any two of those values and it will still operate on the single remaining one. In other words, the system is strongly biased towards giving a warning when required, at the detriment of unwanted operation. Yet, when all three AoA values suddenly disappear, the last valid value being 41.5 degrees, the system just goes silent. No warning, not even a message on the ECAM.

Against my better judgment:

The decision to accept or reject the signal and be silent at <60 kts is what is being pointed to as a design oversight. This isn't new. To avoid spurious warnings at low speed on T/O or Ldg with a WoW avoids spurious warning. Any other time, weight off wheels, that you approach those values, the warning will have already been going off long before the "threshold" value of 60 kts.

HazelNuts covered the rest.

See also Confiture's point about air flow when falling at 10,000 fpm.

@HN39:

I'm sure you're correct. However we're talking about a situation whereby all three airspeed sensors (pitot tubes) are being fouled by disturbed air as a result of extreme AoA.

You're also absolutely right that implementing a warning - either visual through ECAM and/or aural - that AoA is no longer valid would be do-able, but it's then incumbent on the crew to understand what that means.

@LW_50:

I'm not disputing what you're saying - but the <60kts limit was not implemented purely for the purpose of avoiding spurious warnings on the ground, it was also because the data from the vanes is unreliable in the air in that scenario. CONF iture is working on the assumption that the airflow from the descent would be enough to generate a valid reading, but I'd need to know what the specification said to know whether that was true or not.

How about Vmo? :}

It makes sense that the vane is unreliable below 60 KTS, but I'm not seeing how you could have a false stall warning at that speed (absent weight on the wheels). Since the aircraft would be stalled, any warning would be accurate.

Unreliable data might mean that the alarm would not always sound below 60KTS even if the aircraft was stalled. Apparently the designers decided a stall warning that would never work below 60KTS was better than one that would sometimes fail below that speed.

No guarantee that the aircraft would be stalled if the <60kts reading was false.

I don't think the designers made a conscious decision in that regard - it's so far outside the flight envelope that it probably wasn't considered at the time.

Since no aircraft certified for airliners can fly below 60 knots I can see why stall warning is inhibited. If the crew can't figure that out it is their problem. We need smarter pilots. I don't want to point any fingers but sometimes pilots do not act as pilots, they just react to stall warnings and don't think.

Pretty sure the AoA sensors can very well keep up at 100 kt.

To clarify that statement I just wanted to say AF447 pulled up into a full stall for no reason. They were probably taught the Airbus wouldn't stall but it did. It had different laws and what they did was very stupid but they did what they were told to do by some instructor probably. I have been told very stupid procedures by instructors that never fly an airplane so ignore them. Be careful about what they tell you. They just teach, they don't fly.

In terms of the horizontal component of travel yes - I was more concerned with the vertical. The aircraft was oriented such that the vertical airflow was fouling the pitot tubes - I suspect it would also have been sufficient to give false AoA readings.

Of course, this is all speculative as we don't have the spec to hand. I also think we're splitting hairs at this point, because we can't prove one way or another - and the last thing I want to do is give the hamster wheel another push.

Quote:
Apparently the designers decided a stall warning that would never work below 60KTS was better than one that would sometimes fail below that speed.
Dozy, I'll agree with that statement, however the whole loss of stall warning below 60 knots was a conceptual design error.

Airspeed and AOA are each independent measures of aircraft performance. To force AOA to become dependent on airspeed is an error.

If you will recall, I ran some experiments on a surplus airline type AOA vane sensor, and the thing came alive below 20 mph.

There is no reason to suspect that there was inadequate dynamic pressure to activate the vanes on AF447, nor is there reason to suspect that the location of the sensors while at that high AOA would subject them to flow reversal or turbulence. The worst that could happen would be that the AOA would exceed the range of the sensor, but at least they would be at the limit in the proper direction.

This shut down of the stall warning system was not without consequence. It deprived the captain of an essential piece of information when he attempted to size up the situation upon arrival back on the flight deck (Remember? Stall warning shut down almost co-incident with his arrival on the flight deck).

The question I have now is, has the conceptual design error been quietly corrected?

Dozy,

I agree with you about the designers --there was no reason for them to foresee a need for a stall warning below 60KTS. And if they had, presumably they would have chosen a technology that was reliable below that speed.

But if you're saying the stall warning should be suppressed below 60KTS because the airspeed data might be wrong, I don't agree. No matter what the pitot tubes are reading, the airspeed is either over 60KTS, in which case the AOA sensor is reading correctly, or it's below 60KTS, in which case the aircraft is stalled regardless of what the AOA sensor says.