777 PIO?

[Check Airman]

Hi all,

I came across this video of at Ethiopian 777 experiencing some roll oscillations on short final. I seem to recall this happening before. Is it my imagination? Why does the 777 seem prone to this kind of overcontrolling?

You Tube

[Mr Good Cat]

Originally Posted by Check Airman

Hi all,

I came across this video of at Ethiopian 777 experiencing some roll oscillations on short final. I seem to recall this happening before. Is it my imagination? Why does the 777 seem prone to this kind of overcontrolling?

https://youtu.be/zIaqt-8G9Rs?si=3DmcYvIIIcTI2_Lp

It is only prone to over controlling if you come from non fly-by-wire types (Ethiopian will have a lot of 777 pilots who have transferred from the 737). If you apply the same frequency and amount of yoke input on a 777 that you would on a non-FBW type you'll be in for a surprise. Airbus is no different, and there are plenty of videos of crosswind PIOs on Airbus types.

[Check Airman]

Any idea why the 777 would be different? The ailerons move in proportion to control wheel movement, so wouldn't that be identical to a 737 setup?

[wiggy]

Check Airman:

I seem to recall this happening before. Is it my imagination?

Not your imagination..give me five minutes to compose a PM..

Edit to add - Which I've done but your inbox is full...

[FullWings]

I think the video above is more indicative of less than perfect technique than any susceptibility of the 777 to PIOs. I’ve flown it for some time and IMO it feels and responds like an aircraft with conventional controls, apart from not needing back elevator in turns and removing most of the effects of thrust and configuration changes on the trim.

It has very good aileron authority at approach speeds so if you give a disproportionate response to a roll input, you could end up over controlling. That applies to most aircraft, though.

[Mr Good Cat]

Originally Posted by Check Airman

Any idea why the 777 would be different? The ailerons move in proportion to control wheel movement, so wouldn't that be identical to a 737 setup?

A similar movement of the ailerons (or more specifically a similar roll demand) on the 737 requires a much bigger input of roll on the yoke. There is also a larger dead band on the 737 which has to be overcome. The 777 also has a much larger momentum, which means that it isn’t disturbed by gusts as much (even though you can sense them). If you handle the yoke in a similar way to a 737 on the approach you will just end up waggling the wings a lot, like in the video. As you then counter the ‘waggle’ with an opposite large roll input you’ll just end up in PIO.

[MrBernoulli]

Originally Posted by FullWings

...responds like an aircraft with conventional controls, apart from not needing back elevator in turns ...

Only in turns up to 30 degrees angle of bank; for turns using over 30 degrees AoB, some elevator back pressure is required. 😁

Prior to flying the 777, all my previous aircraft types were non-FBW. I found the 777 to be twitchy in roll, particularly at the higher weights of the usual operational take-offs. When conducting simulator exercises, I found the 'slimmed down' secondary control modes (as opposed to normal control modes) to be more to my liking in terms of feel.

In particular, the secondary roll mode functioned in a manner that was much more to my liking, possibly largely due to spoilers 5 and 10, out of a total of 14 spoilers, being locked out.

Originally Posted by wiggy

Not your imagination..give me five minutes to compose a PM..

Edit to add - Which I've done but your inbox is full...

Any chance you could share that with the rest of us, wiggy? 😉

[Kraftstoffvondesibel]

Asking as a non-pilot, but as one involved in human/machine interfacing in other very specific technical disciplines: Is there a consensus and/or established theory as to why this happens with FBW systems? Is there a delayed onset of feeback due to inertia of the larger aircraft(the 737 was compared above) or some sort of computer induced latency? Is there a known, fixed latency between control input and resulting deflection of a control surface, or does this vary depending on «how involved» the computer gets?

[STBYRUD]

PIO isn't limited to FBW aircraft, you can get a 737 into beautiful oscillations as well - all you need is a system with some sort of inherent delay to an input, combined with a degree of dynamic stability - and a pilot doing his utmost to make a benign landing seem dramatic by not anticipating how the airplane will react to his inputs

When NASA was experimenting with the digital fly by wire F-8, they turned up the input delay purposely to study when this would result in unacceptable PIO, made for some slightly scary footage:

You Tube

[FullWings]

Originally Posted by MrBernoulli

Only in turns up to 30 degrees angle of bank; for turns using over 30 degrees AoB, some elevator back pressure is required. 😁

Yes, that’s true, but for normal passenger flying you aren’t doing that unless you really have to. If you want to get really geeky it also has (soft) envelope protection, but again you shouldn’t really be encountering that in normal ops.

Prior to flying the 777, all my previous aircraft types were non-FBW. I found the 777 to be twitchy in roll, particularly at the higher weights of the usual operational take-offs. When conducting simulator exercises, I found the 'slimmed down' secondary control modes (as opposed to normal control modes) to be more to my liking in terms of feel.

And that is what makes aviation interesting, as a lot of it is still subjective. With >10,000hrs on the 777 “twitchy” is the last word I would use to describe the handling, but each to their own.

In particular, the secondary roll mode functioned in a manner that was much more to my liking, possibly largely due to spoilers 5 and 10, out of a total of 14 spoilers, being locked out.

PFCs off is even better, if you want to go completely hardcore.

I suppose it depends whether you are comparing it to similar non-FBW airliners, or aircraft in general. An Extra 300 owner might call the 777 slow and ponderous but compared to the Spruce Goose it’s more like an F-16...

[Check Airman]

Originally Posted by Mr Good Cat

A similar movement of the ailerons (or more specifically a similar roll demand) on the 737 requires a much bigger input of roll on the yoke. There is also a larger dead band on the 737 which has to be overcome. The 777 also has a much larger momentum, which means that it isn’t disturbed by gusts as much (even though you can sense them). If you handle the yoke in a similar way to a 737 on the approach you will just end up waggling the wings a lot, like in the video. As you then counter the ‘waggle’ with an opposite large roll input you’ll just end up in PIO.

I think this may be it. If I'm not mistaken, they also operate 737's. I've seen videos of some pretty significant control displacements in 737's on approach, so perhaps that could explain it.

Thanks to all who shared their experience.

[CVividasku]

There is something that will cause PIO easily. On any FBW type, if the airplane tries to follow the stick command, and the stick goes too quick, it can cause some sort of PIO.
Try this on the A320 during flight control check. Full right, then full left and back to full right as quickly as possible.
The stick is quicker than the ailerons. So if the ailerons try to follow the same path (ie, up to full left then back), they will be lagging.
If the ailerons decide not to follow the same path but to follow the stick position at any given moment, they won't reach full left but will have almost no lag.

[safetypee]

Much of this discussion is considering over controlling - technique opposed to control system response.
Aircraft PIO are specifically addressed by certification.
"The stick is quicker than the ailerons" not as important as the stick-to-aircraft response characteristics.

[Sim25]

Originally Posted by CVividasku

The stick is quicker than the ailerons. So if the ailerons try to follow the same path (ie, up to full left then back), they will be lagging.

That's not how rate limiting is applied/ works...
The second example is the exact one.
Apart from that: On ground (direct law active) spoiler deployment is limited to 40°/s, elevator deflection 50°/s with a command demanding more than 30°/s activating the double pressurization function... Getting to those speeds you will hit the stick hard

[Sim25]

Originally Posted by Kraftstoffvondesibel

Is there a known, fixed latency between control input and resulting deflection of a control surface, or does this vary depending on «how involved» the computer gets?

Yep, for Airbus systems the latency is fixed and pretty small. Load of a computer doesn't change that much.

[B2N2]

​​​​I found the 777 to be twitchy in roll, particularly at the higher weights of the usual operational take-offs.​​​

Exactly this ^^^
The force feedback is obviously fake in a fly-by-wire aircraft and on the 777 I find the aileron ‘feel’ too light compared to the pitch ‘feel’.
With roll lighter then pitch it’s very common to see a little wing rocking when pilots are hand flying.

[CVividasku]

Originally Posted by Sim25

That's not how rate limiting is applied/ works...
The second example is the exact one.

Well, I found an incident on A350 where they did exactly that

[tdracer]

Originally Posted by Sim25

Yep, for Airbus systems the latency is fixed and pretty small. Load of a computer doesn't change that much.

Exactly - the latency in the computer portion of FBW is measured in milliseconds. The ability of the hydraulics to respond is much greater contribution to the system latency.

[Sim25]

Originally Posted by CVividasku

Well, I found an incident on A350 where they did exactly that

You are talking about this incident: https://assets.publishing.service.go...XWBC_09-22.pdf ?

Ok, then let's first define what I was talking about: I was talking about direct law => direct stick to surface relationship. Well it isn't really direct (proportional), most of the times a LUT is between the stick and the surface position demand, plus rate limiting.
When we're comparing fly by wire with conventional controls in terms of deflection speed and latency of control surfaces (as a result of stick input), frankly we can only compare direct laws/modes... (As conventional/ mechanical is direct, which seems also beeing true for the 777 "normal law" in roll)
When flight control loops are present in the path behavior of the surfaces can't be directly derived from stick input. You were talking about the surface action on the ground i.e. direct law...

Now let's have a look on the incident (the one you were most likely talking about):


In red I marked the zone where the aircraft was most likely in direct law (ground mode). It's a guess, nowhere is mentioned if, or when the aircraft was in ground mode. I added black markings as an aid to compare both graphs better with each other with respect to time.
I don't see any of the behavior you mentioned, it's perfectly direct. No surface running after the stick and trying to reach the position even though the stick is travelling in the opposite direction. Everything before and after the red rectangle is with flight control loops beeing engaged...
(It's also pretty likely that the aircraft is already fading into normal law at the end of the rectangle as RA height is increasing, it rotated)

[FullWings]

Originally Posted by B2N2

Exactly this ^^^
The force feedback is obviously fake in a fly-by-wire aircraft and on the 777 I find the aileron ‘feel’ too light compared to the pitch ‘feel’.
With roll lighter then pitch it’s very common to see a little wing rocking when pilots are hand flying.

I think some of that comes from the style of input. Flying the aeroplane through the use of short, sharp jabs simply sends ripples up and down the wing and fuselage and does very little to change the attitude; measured and prolonged control inputs are what you need and the airframe behaves exactly as it should, which for something over 200ft long and wide that can weigh nearly 350T is actually appropriate. If you find yourself rapidly going from one side to the other on the controls you need to damp it out. Low gain pilot and all that.

I have witnessed numerous times people fighting “turbulence” on takeoff and soon as the AP goes in, the “turbulence” disappears, as if by magic! I think some pilots have unfortunately trained themselves to be continually moving the controls, even in smooth air. It just isn’t necessary, especially as the 777 will pretty much stay where you put it as that’s what it’s designed to do.

I don’t know of any commercial aircraft with fully powered flying controls that doesn’t “fake” the control forces, FBW or not. Personally, I find all the controls a bit heavy, but that’s me.