"Guarding" the controls while on autopilot
 

Things never cease to amaze one. When the automatics are engaged with the aircraft below a radio altitude of 2500 ft, an operator in this part of the world requires that the PF actually rest one hand on the thrust levers and the other on the control wheel "following through" on the automatic pilot and autothrottle movements. This is termed "guarding" the controls.

Presumably this is in case the automatics suddenly give up the ghost.

One would think if this is a valid reason for "guarding" things, then the same policy should apply if the PF is hand flying below 2500RA and the PNF should have his hands on the controls as well as the PF and also over the PF's hand on the thrust levers just in case the PF should suffer sudden incapacitation. What a strange way to fly an aeroplane.

Most pilots dislike having the other pilot "hovering" over the controls at any time.

The 737FCTM under Cat 3 Operations recommends guarding the controls on approach and through landing.

How do you define the term "guarding". Does that mean hands actually grasping wheel and thrust levers in a claw like grasp - or does it mean hands resting relaxed on knee ready for a quick take over?

Hi Centaurus,

Yes we do that on the 767 too. Yes its to guard against the aircraft suddenly doing a dirty dart due to some type of aircraft or ground aid failure..more likely when doing a practice autoland when Low Vis Procedures are not in force.

There have been cases of aircraft suddenly doing a dirty dart stage left/right and I guess this gives you the extra nanosecond required to correct the problem...we actually hold the relevant controls as if flying the aircraft...only lightly of course.

Mind you in every failure scenario I can remember in the sim the aircraft continues on it's stately way via the IRSs....can remember in initial AWOPs training not noticing a complete ILS transmitter failure below 1000 radio...the aircraft continued on as if nothing had failed and landed itself normally:}

Did you do Cat3b stuff in the 737s you flew in Europe?

We do 3b no decision...ie 75m vis 0 cloud base...taxiing is the harder bit;)

I must say it does make sense to me to have the PF's hands on the controls during an autoland....for starters you need to have your thumb poised over the GA button if someone strays into the protected area and bends the localiser signal. That happened to me in the sim...scares the crap outa you when you see the HIALS out the left window at 50R.

Chuck.

The issue of guarding the controls is also related to the knowledge about automatic systems in general – how automatic is automatic, and what are the design and certification assumptions?

Thus for older aircraft with ‘fail hard’ autopilots it was definitely hands on.

For more modern ‘fail passive’ (fail soft) systems then hands off is quite acceptable in the cruise. In the event of a failure, the certification requirements should limit the pitch excursion to approx 1 deg / sec (300 ft alt in 4 sec) and less than 15 deg / sec roll (60 deg AOB in 4 sec); most autopilots will disconnect automatically. The cert assumption is that the crew will be in the loop within 4 sec. With some aircraft there has to be a balance between the crew deciding to be hands on and them not fighting the autopilot or attempting to overpower the system (with potentially hazardous results).
During the approach and landing, whilst the cert requirements are tighter (crew involved within 2 sec) it is more sensible to guard the controls. Some aircraft specifically require the crew to be hands on i.e. for the Avro RJ to accommodate extreme conditions, hands on (follow through) is required to damp/prevent a low amplitude oscillation, which the cert standard would not allow, but I doubt if it would ever be seen in service.

For landing with fail operational systems then hands off would be more appropriate as the auto-land (even after the first failure) will perform far better than any crew i.e. the crew are more likely to fail / screw up than the system.

Personally I choose to follow through on the controls whenever possible; for those systems that back drive the controls the pilot gets a feel for how hard the auto pilot is working and whether it is in touch with the aircraft motion; a bit like supervising a student pilot.

Now apply that principle to other automatic systems; how do you maintain hands on for an automatic fuel transfer system? Is the system fully automatic or only just automatic?

With apologies to co-pilots and autopilots:
“An autopilot is like a co-pilot, but never learns”
“A co-pilot is like an autopilot, but can forget”
--------------------
Airspeed and Upwardness

Our SOP's state that the PNF must guard the controls below 1000ft RADALT, regardless of whether autopilot is engaged or the PF is flying manually. This means that the PNF places one hand behind the throttles (but doesn't physically touch them) and the other hand on his knee with hand upturned ready to grasp the control column if required.

Seen the autopilot flick off numerous times due to turbulence/shear below 1000' and thats in an aircraft designed only 10 years ago!!

With autothrottle connected, the PF will generally have his hands actually on the throttles as our autothrottle lags behind by up to 10 knots (both fast and slow) and when configuring with autothrottle, can quite easily overspeed the gear.

Air Berlin SOP is that the PF keeps it's hands on controls AND throttle until 10.000', auto-pilot engaged or not.

Despegue. Are the automatics really that unreliable in the 737 that you really have to have all hands on wheel and thrust levers to 10,000ft? Surely the automatics would not have passed certification flight tests if they were that bad. Seems like a gross overkill to me - especially as the Boeing advice is to only guard the controls during a Category 3 landing.

Menen,
My opinion too, but try and tell that to the guys who make the SOP's...
Furthermore, I like to have both hands on the control wheel whenever autothrottle is engaged.

despegue
 

In that case wouldn't one hand on the yolk and the other on the throttles, (ready for a quick disengage if required), be a better option?

With our outfit PF covers controls whenever flaps are extended (i.e. closeish to the ground)

Blimy, hope you don't fly longhaul. You would struggle a bit after a 13 hour flight.

:P

L337

If your reactions are such that you cant move your hands from your knee to the column in a split second, should you really be flying an a/c?

Always makes me smile how many pilots are not relaxed enough to be on a flight deck, you know the type that feel they have to hold their mike every time they transmit!

whatunion
 

Just wondering if you have ever used the old style Airmed headset or similar? On those the boom mike often dropped down below an acceptable level under the chin and had to be held up for satisfactory transmission. Later, more expensive, model headsets seem to have cured this problem.

Yes i have and know what you mean but in our company everyone uses the same headsets and the mikes stay were they are set.

Seriously, i have studied those who always hold their mikes when transmitting and these are the ones that get up tight about everthing from turbulence to slots!

Funny old thing, I was was just thinking about this very subject when I happened upon this thread...

I have observed a large range of behaviour from many people over the years, ranging from only touching the controls if absolutely neccessary to hanging on for grim death at all times!

These days I sit in a large FBW beast with (a) very reliable autopilot(s), so I tend towards the former and let go as soon as the A/P is engaged.

I do profess to a (irrational?) slight degree of annoyance when the PF insists on staying physically connected to the aeroplane even when the electronics have taken over. It makes the PNF's job harder as you often find a marked reluctance from the other side to 'let go' of the comforter and make FMC & MCP changes and the like. It also runs somewhat contrary to our SOP's.

My real fascination is with what these guys think they are achieving with their white knuckles (showing through the gloves). I have seen blokes with their hands on the yoke for half an hour in the cruise in light turbulence.

There are two main states of man and machine:

1) A/P out, pilot flies aeroplane.
2) A/P in, A/P flies aeroplane, pilot tells A/P how to do it.

Granted, in the latter stages of a CATIII approach, it would seem to be wise to 'guard' the relevant controls but at FL370?

My point seems to be shaping up thus: If you are happy with the A/P, leave it alone; if you are unhappy, take it out. There is no 'middle ground'.

You cannot instill a idea into the A/P or A/T by pressing on the controls or hanging on the thrust levers. At best, nothing happens; at worst the A/P goes into some other mode (CWS, V/S or whatever) and you don't notice because it hasn't disconnected...

In short, to get the A/P (or any automated system for that matter) to do what you want, use the INTERFACE PROVIDED BY THE MANUFACTURER and follow the protocol demanded.

As I said earlier, it's more a point of interest and I don't really care what other people do. I'm sure the earlier generations of aircraft had to be watched continually and A/P failures were frequent. After nearly 5000 hours on my present model and no A/P dropouts maybe I'm getting complacent...

Keep pressing the buttons.

What does your S.O.P say about guaring rudders and brakes?

Most common cause of sounding like your head is in a bucket is holding the boom mike - don't do it! If it keeps drooping I hear modern medicine has a cure ;)

Having had a 737 decide to dump it's autopilots just before the flare on a CAT 3A landing and a 757 trying to barrel roll on a triple channel approach when following a 747, you won't see me not following through on the controls when established on the approach - I value my life too much!

I find that the command RETARD wakes me up sufficiently to remember to put my coffee down and touch something..................

After that it's anyones guess what the landing will be like :ok:

Let's see now. A/P engaged and pilot has hands on knee monitoring the automatics. Something dreadfully critical happens and after the normal one second human response delay, one of the two monitoring pilots grabs the controls. That would take maybe one second.

Scene 2. As above except automatics engaged but PF (or PM, if you so desire) has mitts on controls as well. A dreadful event occurs and after the one second evaluation period the PF tightens his grip on the controls. That takes maybe half a second.

So now we are looking at an emergency that is so deadly that one half of one second is all there is between life or death.

Seems there is no shortage of nervous nellies masquerading as airline pilots.

During a bog standard CatIII approach, (a regular feature in Europe for several months of the year), the PF will have one hand on the control yoke, guarding the auto-pilot dis-connect button, and one hand on the throttles guarding the auto-throttle disconnect/ TOGA switch, in the event that at any time any indications shouldl move outside the accepted parameters for an autoland under CatIII conditions, then the PF will take immediate control of the aircraft and carry out a go-around/over shoot. On the B747-400 this situation could occur at a height of 20' agl and no visual reference outside the aircraft.

Under such circumstances having ones hands on ones knees is not recommended, but if in doubt, try it out!