I dunno, i turn the A/T off at 2000ft usually, but i never fly with weather all my approaches are hand flown from high altitude.

 

 

I don't know of the accident with a stuck throttle on TO

 

http://en.wikipedia.org/wiki/TAROM_Flight_371

Fascinating! Amazing how the fundamentals if multi engine flying can be forgotten so easily.

 

 

Sent from my iPhone using Tapatalk

Look at that crash with jammed throttle leaver during take off. Pilot could prevent it if he monitored performance of AT, but he fell into false sense of security thinking that AT is decreasing his workload. In fact there were one more thing to monitor than with manual take off.

 

IMO, it's completely fine to fly an approach manually, as long as it's performed completely manually. AP off - AT off.

You should have your hand on the levers, A/T on or off. This isn't an Airbus, the thrust levers move and you can feel what the A/T is doing. It's also our company policy that the PM guards the thrust levers on takeoff and approach. If the A/T is doing something wrong you will know. It's easily overridden with little effort.

You should have your hand on the levers, A/T on or off. This isn't an Airbus, the thrust levers move and you can feel what the A/T is doing. It's also our company policy that the PM guards the thrust levers on takeoff and approach. If the A/T is doing something wrong you will know. It's easily overridden with little effort.

Agree, I mean that by "monitoring AT performance"

If you extend this logic, why not fly the approach completely manually?

Why not indeed. The pilot should be able to do a better job than the A/T and A/P in gusty, turbulent conditions or other situations such as when the landing aircraft disturbs the localizer. Flying the approach manually should be pretty standard stuff. Although, most times, most pilots will disconnect at around 1,000'.

 

 

 

but my argument is A/T do not decreases workload in every situation

Kind of agree. Specifically monitoring speed on landing shouldn't be too taxing, but generally, monitoring automation is probably the most difficult jobs the pilot has to do. We are fundamentally bad at it. Take the invisible gorilla phenomenon, where people are completely blind to completely obvious and yet unexpected events when being in receipt of lots of other expected signals. Automation problems can be like that, one little thing silently going wrong when everything else is looking good.