explain the "auto throttle" to me/us
 

From the Asiana threads, I am not sure how the "auto throttle" really works, or even the intent.

If the thing is set to hold an airspeed and change the engine thrust to hold that airspeed, then sounds like what we USAF pilots usually flew. You know, point the nose and move the throttles to hold speed. The nasal radiators did it different. Hold an AoA that gave you best approach speed for weight and such, then move throttles to control rate of descent. We SLUF ( A-7D) pilots found that the Navy procedure was great, but we still did a slight flare just before touchdown as we were not trying to catch a wire on the deck of a pitchin/rolling boat.

The Asiana crew appears to have used the procedure/technique we USAF dweebs used - point the jet, and move throttles or let HAL move them.

I would appreciate some explanations and rationale for different implementations.

The Auto throttle does different things dependent on the mode it is in and the phase of flight.

As a general rule, though, I would suggest that in Airline flying, your "Point the jet use thrust to control speed" is what is normally used on approach.

As an example, the most usual approach in an Airliner would be an ILS- if done with the automatics, the AP would use pitch to follow the Glide Path, Roll to follow the Localiser, and the A/T would be in SPD mode, adjusting thrust to maintain the selected speed.

There are modes when this is not true, however- for instance if wishing to make a flight-idle descent, Flight Level Change mode simply reduces the thrust to idle, then uses pitch to maintain selected speed.

A pertinent factor about THAT mode is that after a few seconds at idle, the A/T mode changes to "Hold", which means it is basically dis-engaged and in standby. This allows manual changes to thrust to be made if you wish to vary your ROD.

In the Asiana case, it SEEMS that what the crew did was manually use pitch and roll to follow the approach path, EXPECTING the A/T to manage the speed, but the A/T was in fact in "Hold", and thus not doing so.

Hope that helps somewhat!!

Never understood why Nasals used the secondary effect of controls to fly approaches. :E

When engaged, the AT moves the throttles as required to try to achieve the speed.

Only when the flight guidance system mode is in "Open Descent" (Airbii)/"Flight Level CHange"/Real Boeing/"Level Change"(others) does the AT generally go to Idle/Standby: in these modes the pitch position is used to control the speed (ala Nasals). The trap in this mode is, if hand-flying and ignoring the Flight Director (which probably is saying fly down) by pulling up the nose, the speed will reduce but the AT won't react. If the speed gets too slow because Bloggs continues to pull/hold the nose up, any decent AT system will spring to life regardless of previous pilot selections and save the day.
Apparently the 777 doesn't do this.

Also VNAV SPD in the Boeing, and it then reverts to "Hold"

Good chance the PFs F/D was off.

It will in any mode EXCEPT "Hold"- INCLUDING disengaged!

Yes, I think it's a shortcoming.

Interesting differences to the tiny boeing (aka 737). It reverts to ARM instead of HOLD and low speed will revert it to active, except below 27ft RA where it will revert to RETARD. THR HLD is only used during takeoff on that plane. But then, according to Capn Bloggs above it is not a "real" boeing as it uses Level Change and not Flight Level Change.

My profuse apologies, Denti! You do fly a "real" Boeing. :ok:

When going slower than the normal Ref+5 that an airliner would do, ROD is more effectively controlled with power, which is what the Navy does. Even I transition to that technique in my 185 when the necessity for going slow into a short strip is required. However, when increasing speed from the backside is necessary, then power comes to the rescue. Pitch only isn't enough.

In normal operations with an autopilot, the A/P controls pitch/glideslope and the autothrottles (if equipped) control power/airspeed, and of course, a combination of pitch and power is a blend for maintaining speed and profile.

Yeah, Desert, guess we were used to a higher AoA than the big jets. Our Viper used about 13 degrees AoA, so we were close to the area of reverse command WRT power.

Of course, we trimmed like crazy to get to and then hold the AoA, then began using the throttle for precise touchdown point and rate of descent.

I still have trouble with all the throttle modes and such due to "connections" with the other HAL - the autopilot. And that's where my questions come from, as I only flew one jet with an Otto-coupled ILS mode and we had to use the throttle manually. And you better have had the sucker trimmed for best approach speed/AoA before "coupling". Figure 180 knots IAS basic, then more for gas/loadout.

It also looks like there's a plethora of implementations depending on what heavy you're flying.

Gums,

You must be referring to the Mighty Voodoo, which I had the great pleasure of flying for 12 years/2,000+ hrs.

As a bit of a sidebar, I have considerable difficulty with the current culture/attitude regarding automation in general and auto-throttle use in particular. Acknowledging that I have never flown with an A/T, I find it very scary how it would appear that some folks on the flight deck these days seem to have an overwhelming reliance on this piece of automation.

For example, it would appear to me that in both this Asiana accident and the Turkish Airlines Flight 1951 in Amsterdam, the flight deck crew (three pilots in both cases) were under the impression that the A/T was handling the throttles, when in fact it was not, for different reasons. Regardless, in both cases, these six pilots did not recognize that the IAS dropped 40 KIAS below Vref, resulting in … guess what … a stall !

The problem that I have with the culture/attitude I referred to above is because, in my opinion at least, many pilots in this culture will say: “well, the faulty RADALT disconnected the A/T … or whatever … so it was not my fault”.

Wrong !

It is, it has been, and it always will be the pilots responsibility to ensure that the aircraft flies within its aerodynamic parameters. If the automatics aren’t doing the job, then the guy who signed for the aircraft has to take over. It's his responsibility, his job; it's why he is the aircraft captain !

I acknowledge that I am clearly a dinosaur as far as today’s automatics are concerned. But I am a pilot and I firmly believe that all pilots must be capable of keeping their aircraft flying.

Yep, actual Dash-1 was 175 plus some knots for fuel over 3,000 pounds, but what the hell. Only the 104 had similar speeds on approach, best I recall.

Only had about 500 hours in the jet before the "war" called me, and I remained an attack puke next 16 years.

We sure appreciated the Cannucks flying 200 miles north of us toward the North Pole, but from our Grand Forks' orbit north of Gimli we could still see Baffin Bay, heh heh.


I am not a dinsosaur steeped in "hand flying" all the time, and I flew two of the latest and greatest jets with advanced avionics and nav features that the airlines did not get for years.

But I still have a problem understanding exactly how that auto-throttle fits in when merged with the auto pilot modes. I am confused. And I can't get a sim ride to see how it all works.

The folks on the AF447 thread know me, and they know my opinions regarding FBW, so we can skip all that stuff. They also know my opinion about a HUD, and I think a good HUD would have helped a lot for the Asiana pilot.

NY Times
 

Recent article quoted FAA as saying auto throttle will adjust power to maintain speed. Boing aircraft, the thrust lever actually moves when AT steps in to add power. Airbus, the throttle remains fixed.
The pilot said he noted the throttle position was not moving, but he thought the AT was going to step in and fix the problem.
No link found. Read it.

G'day Gums,

The venerable Douglas/Boeing 717, a DC9 on steroids, works thus:

Selecting Autoflight on the takeoff roll once symmetrical power increase is observed engages the Auto-throttles to the target takeoff power, be it REF power or flex/derate.

Passing the acceleration altitude entered into the FMS as part of the pre-departure procedures will see the take-off power reduce/increase to climb power, with the aircraft accelerating towards (usually) 250 kias. During this phase the AFS uses pitch to accelerate to or maintain 250 kias, and once through transition altitude to further increase to enroute climb speed. The Flight Mode Annunciator panel (FMA) will indicate SPEED on PITCH & T/O or CLIMB THRUST.

Once cruise level is reached the FMA will change to indicate SPEED (FMS or manually selected speed) on THRUST and LEVEL HOLD, and the auto throttle system will adjust power within the bounds of its capabilities to maintain the set speed.

When commencing a normal descent, the auto throttle system will for the most part continue to use thrust to manage speed, whereas pitch will manage the flight path. Should the aeroplane get higher than the computed flight path, the auto throttle will reduce power to idle, and may increase the IAS/MN to the maximum allowable minus a margin that is too small for comfort. On regaining profile the auto throttle will revert to SPEED on THRUST

A normally flown instrument approach or visual approach will see the SPEED on THRUST relationship remain, and in the flare the thrust will be reduced to idle, with reverse thrust being manually selected.

In any of the above modes, if the FMS detects the speed at Vmin - 5 kts, the auto throttle will advance the thrust, or if it had been turned off (but not broken) it will engage the auto throttle, and use a combination of thrust and pitch to maintain Vmin-5 and sacrifice altitude until the pilot intervenes and selects a faster speed.

In essence the 717 philosophy is to engage auto throttles during takeoff and leave them engaged until landing.

I think in my 8 years on type I've only had two occasions where the auto throttle system didn't work and both were a result of FADEC failures.

speed with thrust is a much better, although less Boing/manly, way of describing the speed control mode, Clarrie... :)

gums, tlb and all Fj- guys,
we had some kind of auto throttle as well, computed in our brain by reference to speed and attitude (present and desired) and implemented in our left elbow and hand.
Or did either of us need to do any kind of lengthy computing what kind of thrust setting would be suitable, aside from the first hours in the respective aircraft? How did we close formation takeoffs/ landings and WX penetration, how air refueling and AG bomb runs partly in single seat aircraft while navigating, comunicating and in bad days watching not to get shot down?

Now in the days of automation and magenta line training this ability to correlate speed and attitude to the thrust necessary has been lost, and a few (as i hope only a few) occupants of the front office ( i dare to name those pilots) even don't know anymore, what a AS indication is good for and what keeps an aircraft flying. Too many accidents happened that way.

Adding eyes's scan it is what I call "Gesture algorithm".:D
NTSB hearings show they have no intention and are no more able to modify that :mad: misbehaviour, individually or in crew. They call it their culture....Designing his HUD, GILBERT KLOPFSTEIN wrote and showed that it was easier to be aware of the actual State of the Bird considered as a dynamic system and to control the path in one sight with correct speed, thrust, attitude.

Thanks, Clarie.

Now for the bigger jets.

And is there an overall philosophy according to the manufacturer?

Yes there is a tendency regarding the use of automation in "airline flying".
Save money in selection and training of pilots.
It is easier( = cheaper) to tell somebody to let the autopilot and autothrust do all the work, than to train him and to keep him proficient in manual flying.

I didn't want to hear that, LW.

I used otto a lot in my single-seaters, and even the family model I flew as a nugget. But it was simply a task management thing so I could pull out an approach plate for a strange field or figure out a new course or.....

My back seater in the VooDoo was never hesitant to ask me what the hell are you doing if I did something strange, heh heh. Same for my wingies in the SLUF and Viper.

Basic Boeing philosophy on the old types (737) - use AT with AP engaged. With AP disengaged, disconnect AT.

The automatics on the 737 are not top drawer though - the AT needs "help" in gusty conditions. The thrust levers move according to the AT demands.

The modes available for the 737 NG AT:

N1 - AT holds commanded N1 limit. Used in takeoff and climb, with aircraft pitching for speed

MCP SPD - AT holds target speed set on the MCP using thrust up to current N1 limit. Used in level flight, approach mode and in climb/descent when V/S is the active pitch mode

FMC SPD - AT holds target speed from the FMC - normally in cruise or in a path descent on a geometric descent path with VNAV PATH as the pitch mode

ARM - thrust levers static and able to be positioned manually. Normally idle thrust in descent with MCP SPD as the pitch mode (Level Change) Also provides very limited low speed protection.

GA - provides reduced thrust if the TOGA button is pressed once on a go around equivalent to around 2000fpm climb rate

It's a fairly simple system, with the main point being that if you are flying manually, you should be using manual thrust. I have never tried manual flight on approach with the AT in, and I imagine with the pitch couple it is fairly interesting.

In the olden days (pre-FADEC), the autothrottles had a pretty important task in protecting engine ratings and keeping things like the 'climb' rating during the changing altitude/temperature during a climb.


With the advent of FADEC, that's no longer a significant issue - climb rating occurs at a constant throttle position, and it's pretty hard to have a significant overboost in Normal FADEC mode without really trying.


BTW, the autothrottle on all Boeing airplanes used to be more 'aggressive' in turbulence - it was intentionally slowed down for a number of reasons.

Thanks Scud.

Not what I wanted to hear, but if the heavy folks can live with such an implentation ( or die with 200 SLF in the back), then seems their choice( the SLF don't vote).

I can understand some basic profile settings, like climb for optimum rate of climb or range when cruising, and descent to a specified atltitude, tho I don't like that one, and the Asiana crew witnessed that sub-mode of Otto and throttle otto).

My only experience with the climb and descent otto, held the speed or mach, and I set the throttle to whatever I wanted or needed.

I can also understand the jet adjusting the power once level to get the best cruise mach and such. Ditto for a climb profile, but don't know why otto would adjust the throttle(s) as the crew could just set the power to the basic setting from the manaul.

But seems to me that there are still a plethora of modes and sub-modes that are connected with the autopilot modes, and it gets confusing.

If the policy is to fly the USAF technique of point the jet and use throttle for speed, regardless of the optimum AoA for an approach, then I understand that. But the last thing I wanted was a system that was hard to understand depending upon a host of other systems that were supposed to "help" you.

Even if the whole profile was flown by HAL, then the faux pilots would have to monitor system operation and detect a problem or whatever. Why not just fly the jet yourself and take pride in nailing the mach for climb and cruise and setting power for descent to reach the initial approach fix at the right altitude and.....

Gums, if you don't mind my asking, what's an implentation?

Sorry Flyboy, I mean "implementation", how a concept is implemented.

Start with the operational requirement, get some ideas from the players, refine the requirements, get some proposals, then decide to go with brand "x" or brand "y". Basic systens engineering and procurement stuff. Gone are the days that a company can roll out a plane and get some customers.

Bus and Boing seem to figure out what the "need" is, and propose something. If the major carriers don't seem interested, then they refine and improve what is already flying.

So...you're saying this is a bad thing?

Different philosophy in airliners. The aim of climbing is to do it as quickly as possibly to get up into the thin, fuel-efficient air. So the throttles (ATS or not) go to full/climb power and the speed is controlled by the nose position.

For jets, the best situation is flying at a specified speed eg Long Range Cruise. So the power (ATS or otherwise) is used to control/maintain the speed. Apparently, turboprops use the opposite technique: level off then set cruise power and accept whatever speed you get. Claret will confirm: he has plenty of boat time. ;)

Approach speeds are determined to the knot; that gives the optimum AoA. Pitch for attitude and thrust for speed is just another way (IMO more logical) of flying at the correct AoA. After all, that's what the autopilot does on a coupled approach. Maybe when sitting on the back of the drag curve using power to control the descent worked OK; we didn't use that technique in our 190kt-approach land-based delta...

Auto throttles are very simple things and they all work pretty much the same way. Speed on thrust or speed on elevator depending on the selected vertical mode. The will also have a hold or clamp mode if they are used for take off, this is to prevent the AT system buggering up the take off and a retard mode that is triggered at some radio height over the threshold.

Nothing to it really.

C'mon, Bloggs

The optimum AoA for approach, climb, range etc. is what it is. The speed is a result of your configuration at the AoA ( slotted flaps, leading edge devices, etc factored in).

The jet stalls at "x" AoA, and speed could be above or below the manual's number - think accelerated stall and such.

I did not appreciate the concept of the AoA "indexer" lights on the gunsight/HUD until flying the SLUF. Then, we had the actual AoA bracket next to our flight path marker in the HUD. Sheesh, how could it have been easier, with no flight directors or auto-anything? Our AoA was the real deal, and compensated for gross weight without us doing anything. Some configuration factors came into play, but the basic AoA didn't change much - figure a degree or so AoA.

We usually set a power of 95% or so for our wingies when climbing. Then we climbed at a mach or a generic speed while trimming for such. Sheesh. Easy. Upon level, got to the best mach and used whatever power we had to, duhhhh? Why is that so hard?

It would seem to me that basic throttle settings for climb, cruise and descent would be easy to know. For approach, if you need otto to help, then fine. but quit connecting the throttle to the AP modes. If you want to "point" and use throttle for speed, your choice.

The advantage to modern auto flight systems is just like the cruise control in your car is that can operate very efficiently making very small corrections. The basic operating is very simple, but really understanding them is much more difficult. As shown by the Koreans --they could use but didn't really understand about the hold mode. The other problem is that you need to actually know how to fly when things don't go perfectly. Apparently the Koreans couldn't do that either. I flew the A 320 for a long time and always told new guys that the magistrate stuff works great so use it, but don't ever completely trust it. It is kind of like flying with a student--never let it put in a position that you can't recover from. Another problem the Koreans had and something they don't teach you in training.

Gums, you've lost me. For any configuration, there is only one speed that will give you a particular AoA. That is why Vref to the knot is used by airliners. Somebody decided many moons ago that speed would be a better "target" to use.

Airliners generally don't do accelerated stalls :eek: so the stalling speed is always the same for a particular configuration and in any case is calculated by the FMS (taking into account all sorts of things like C of G) and is displayed on the ASI by way of Vref, being the stall plus 30%.

Who said it was hard? What you've described is exactly what happens in a jet airliner; climb at constant power using pitch to control the speed, then in the cruise, use power to stay on speed.

?? As has been pointed out already, having the autothrottle control the speed reduces workload. The autothrottle can be used with or without the AP being engaged. It is not "connected" to the AP modes. It does different things depending on what the AP is doing with the aircraft but that is generally reactionary eg AP pulls nose up to get back on GS, autothrottle pushes up power to keep speed on target.

Some types, like mine, recommend using the autothrottle all the time, AP in or out. And it does a far better job than I ever could. "pointing and using the throttle for speed" is simply flying without the AP or the autothrottle. Problem?

From his posts gums is addressing two very different things in my opinion. One is the use of automatics including autothrust and the other is basic airmanship.

Using automatics has improved safety considerably, however as we now see for a couple years, despite a record breaking low accident count, is the fact that pilots still need to be pilots and need basic airmanship. That said automatic flight systems are tools. Nothing more, nothing less. A professional pilot needs to know his automatics inside out as well as he needs to be able to fly manually if the need arises and of course has the ability to monitor what the aircraft and its system are doing. The latter one is actually the hardest task and human beings are not very good at monitoring,

Autothrust is very good at what it does, and over the long and incredibly boring hours of commercial flight, it is much better than any human being could ever be. However as any system it has its limitations, design criteria which it follows and different modes that work according to what state of flight they were designed to be used in. Using Level Change, or Flight Level Change, for a visual approach seems to be a rather odd use of autothrust mode in my opinion, certainly no mode i would ever use in the tiny boeing, however as i'm not rated on the huge twin i can't possibly comment if it is a normal use of AT mode there. In my outfit we would usually fly a visual in manual flight with the AT in arm (which provides low speed protection) and knowing that we have to set thrust manually.

Answering Wizofoz saying it will in any mode EXCEPT hold , including disengaged

It´s not true ...you seem to mix ARMED and DIDENGAGED/DISCONNECT functions

Even in HOLD mode , A/T will react , I agree lately , but it will react provided A/T arm switch on the MCP is armed which is always the case .

Precisely, when speed is approaching middle part of speed amber scale , your autothrottle becomes "awake" , Speed mode engages et your auto throttle is keeping speed on top of speed amber scale which is MMS minimum manoeuvring speed.

More info on PLI activation, caution air speed low and trim inihibits are available by your nearest retailer ..

NO NO NO! That is only true in steady flight. Stall is a transient situation.
See the PRANDTL basic curve : all the points are mesured at the same speed, but different AoA.
Speed is a bad indicator of stall. Only AoA is true
gums is true.

I can't understand the aversion to using AoA to control flight in the take-off, approach and landing phases in heavies, airlines, FAA-world. I used it in the T-38, the Citation (same indexer display) and the C-5 (no indexer, but 7 units for approach, from configuring to landing). Not very useful in cruise, I'll admit. An AoA indication was installed on our Bombardier planes, but NO flight-verified recommended indications for any operations. We were told to "let us know how it works and what approach indications are, the FAA won't certify it".

The wing flies on AoA, not airspeed.

When 4D gets here, I'm going to shooting range. :E. I didn't say airspeed Could be replaced, just that AoA can be better in some regimes. And, not subject to incorrect weight calculations, for example. A thousand hours in the A-10 and hardly ever used that Tweet airspeed indicator on approach.

That's 'cause you couldn't go as fast as a Tweet. :p

I agree with you entirely about the use of AOA....and you don't have that much time before heading to the shooting range. :}

edit: BTW 1000 hrs in the Hog doesn't cover a lot of mileage.

Well, not much mileage, but it was several years in C-5 before I flew a leg exceeding the 10.2 hour, Yenisheir to Lajes leg in Hawg.

Back to A/T, the bizjet systems work like the Boeings--they move. The modes are Thrust (idle, if descending, CLB, if climbing) or Speed, if the vertical FD mode selected is Vert Spd or on a VNAV PATH. We have the same "FLC Trap", that is, in THRUST mode in a descent, the throttle setting commanded will be IDLE until reaching an altitude or intercepting a VNAV PATH, when they switch to SPEED and fly the commanded speed by using the throttles. IF altitude is the runway and no path is selected, no power until they see a min speed about 5 KIAS above low speed cue.

I'm not a commercial pilot but having read this post with interest and being amazed at the plethora of modes that there are which all need learning, wouldn't it be a lot easier just to use manual control of the throttles? I mean, there are two of you...

In ten letters, no.

Thanks GF and Okie. I am getting a clearer picture from the ex-lite folks than the steeped heavy folks.

Despite my basic views of auto-throttle, I can tellya that if trying to get absolute best range in the SLUF, we would gradually pull the throttle back using the TOT ( turbine outlet temp, not EGT). The deal was as TAS wnet up, we would pull back 3 degrees of temp at a time. I forget the increase in TAS we used. But once flew from McChord to The Beach unrefueled using four bags and step climbing like the heavies seem to do for long hauls. An auto throttle rule based on mach and altitude hold would have been nice. Seems Concorde was only heavy that let altitude increase to get the best performance. Ditto for SR-71.

My main concern is if the throttle moving or not, and then the modes where the thing isn't really doing anything and the pilot thinks the sucker is working, but ain't. So think Asiana. My buddy that flew B757 and then B747 said it was easy to tell if the thing was working, as the throttle moved. 'course, he was one that kept his right hand on the throttles.

My focus upon AoA is that for a given configuration, then it is a great cross check for the calculated speed. We lites didn't have a hundred thousand pounds or more of basic weight, so a 1,000 pounds of gas was significant. After a quick WAG and rules of thuimb from the flight manual, it was comforting to see the optimum AoA in the HUD agree with the calculation. From then, you only needed to glance at speed every few seconds, and the flare was also easier looking at the AoA next to the flight path marker.

Thanks to all so far, but I still have questions about implementations and company philosophy.

Short story, gums,

Landing on a beautiful day in Dallas area in a Global. Former schoolhouse IP is playing with FMS entering a " hand drawn" visual by copying the runway threshold and adding an distance/altitude final fix and cranking down the altitude selector to the field. Hand flying with A/T engaged. Because I was flying a visual pattern, I was a steeper than the defaulted 2.5 degree path and would never intercept that 2.5 degree path. As we rolled out on 2 mile final, hands on throttles, I noted speed slowing to approach expecting the throttles to start "waking" up to about 1200-1300 pph of fuel flow which they weren't doing. I clicked them off and noted THRUST in the HUD. It dawned on me what was going on. The A/T were in idle because we lacked a vertical constraint--altitude or path. Good debrief followed.

Now, I never fail to announce modes shown in the HUD or the mode annuciator.

The problem you, I and others who came out of the old generation is that we have thousands of hours of hand flying, due to the nature of the planes we flew. That experience isn't easily or, more to the point, economically, replicated today. The conversion from those planes to "magenta line" isn't always easy for types like us, but we can fallback on those previous hours of flying. Imagine never flying those hours and they take away the automatics, then add in the fact that you rarely saw throttles move.