Flight directors may cause more problems than they are designed to solve
 

Flight International 5-11 March 2013 has an editorial and and also an article by David Learmount on automation dependency. The editorial heading says "A decision by an airline several years ago to test empirically how pilots use flight instruments to monitor aircraft performance provided information that was unwelcome - and unexpected".

The editorial is then taken up at page 30 by David Learmount under the main heading "Dodging Disaster. Seems that after the Bournemouth close shave by the Thomson Airways 737 where the aircraft nearly went in during a go-around, the company arranged for eye tracking to be carried out in the 737 simulator to see how pilots scanned their flight instruments during a go-around.

According to the FI article, the investigation discovered that many pilots had a disorganised scan which frequently left out vital displays such as the airspeed indicator for critically long periods. Such a fundamental failure in the exercise of a skill which, it has always assumed, was basic to all licenced professional pilots was a shock, raising the question as to how widespread this failure is in the industry as a whole.

It would be interesting to know if these simulator go-around exercises were conducted using the flight director or were they done with the crews flying raw data?
Believe me there is a big difference in scanning between the two situations.

The very nature of flight directors require the pilot to concentrate squarely on the FD needles and normal flight instrument scan becomes practically non-existent. It is no wonder pilots instrument scan on the airspeed indicator and other flight instruments are "disorganised" when 99 percent of an airline pilot's flying in his career will be with reference to the flight director. That is dictated by company policy and recommended by the manufacturer.

The solution to the perceived problem of disorganised instrument scan is to encourage the switching off of the flight directors when not needed for critical parts of a flight. But manual flying would also be needed to keep current on basic instrument flying skills. Pilots scan would soon revert to that taught when they first learned to fly on instruments. The perceived danger involved with IMC go arounds would be a lot less. Flight directors are fine in most circumstances but have proved to be distracting during critical events such as low altitude go-around where trim changes and altitude captures require fast scanning of all the flight instruments - not just the flight director needles. Many pilots find it difficult to `look through` flight director needles in an attempt to see the little aeroplane symbol behind the FD needles.

FPA
Speed error tape
Acceleration pointer.


All the data you need is right in the middle of the ADI.

Not long ago, in an airbus, the pilot did a go around but they did not actually engaged the mode (which you do by setting the levers in TOGA). He only advanced them to climb power.

Thereafter he followed the bars, which were still in LAND mode, or LOC GS, and the pilot pitched down, going to terrain. They finally averted disaster, but the lesson I get from that is that when workload and stress are at a peak, FD bars are what a pilot will follow, disregarding other sources of information.

I advocate more time with no FDs when circumstances permit. Just a few ILSs and perhaps some no FD take offs, too would be very helpful, along with the sim sessions.

I think it depends on how you were trained. If you've had it firmly nailed into you from early training to fly the airplane, and that the secret to doing so is always pitch + power = performance, you will be looking "through the bars" for the magic combination that works for your aircraft. On mine during GA I expect to see a little less than 15˚ ANU and thrust 1.45/1.55 depending on engine type (or even simpler, T/Ls wherever they end up with my right arm straight). If the FD bars are wildly off, you bloody well ignore them...

There are only about six or seven attitude and thrust combinations you need to memorise in order to save your ∫utt in pretty much any situation, certainly less than ten, and any pilot taking coin and calling himself a professional owes it to himself and others to learn them.

Grumpy rant mode OFF :}

I agree, but after 10 years on the ULR, I bet even the best trained pilots will tend to follow the FDs blindly

Indeed. But I also suggest four 4 hour sim sessions a year with nothing but raw data manual approaches, two engine go arounds with early level offs, unusual attitude, stall recovery and unreliable airspeed. Pure training, no checking, no marking, no records, no systems, no stress, just feedback and performed by TRIs and not TREs so as not to influence the actual checks.

I'm also thinking at least a two hour observation sim session prior to a four hour training slot at least once maybe twice a year.

Poor instrument scan is a problem in initial instrument training and in simulator training too, but I cannot understand how that translates to a need to reduce used of FD's. Rather, training that reemphasizes airspeed control, attitude flying in absence of reliable airspeed indication and, indeed, in normal flying is a critical need as well as engine monitoring and so on.

I have no idea how many pilots I have trained for jet transitions and ATP's.
Basic scan has rarely been a major problem in training.

However, doing type ratings for very experienced pilots has sometimes presented scan problems, but seemingly only for pilots who've never faced an actual emergency. Thus, I think good simulator training in emergency procedures with instrumentation failures is highly effective in maintaining proper scans.

Nothing induces good scan techniques so well as partial panel IMC approaches IME.

You have hit the nail squarely on the head. In fact, it is probable that most pilots would be delighted at the opportunity to have that sort of practice. If pushed by ICAO - and by association by regulators - it would surely increase pilot handling proficiency - particularly of newly graduated low hour first officers now being steadily recruited into the airlines. More simulator training as suggested above would also reduce the insidious slide into automation dependency. It would be money well spent.

It is not in the sim that one learns how to fly, it is on the line where one must keep proficient in basic stick and rudder. Standard operation should be without FD. only when traffic is dense or weather should automatics be used, and in RVSM.
The sim is for training emergencies, not stick and rudder. In an emergency, you Should use all help available to you, including automation.

If that is true, then why is it that modern Level D simulators are used for zero/zero flying training and at completion of the type rating course the pilot receives a type rating on his licence. They are very much used for "stick and rudder" training including max crosswind landings, and automatic landings.

You have hit the nail squarely on the head. In fact, it is probable that most pilots would be delighted at the opportunity to have that sort of practice. If pushed by ICAO - and by association by regulators - it would surely increase pilot handling proficiency - particularly of newly graduated low hour first officers now being steadily recruited into the airlines. More simulator training as suggested above would also reduce the insidious slide into automation dependency. It would be money well spent.

I am with you on this, but there in lies the rub. The airlines DO NOT WANT to spend the money on unnecessary, i.e. not required training. If I was a chief pilot I would demand my pilots were proficient at these skills and allow them to perform them on a daily basis. That's how I learnt and still keep those skills even with FD on. You have to scan the basics to ensure the FD is not telling porkies. I'm sure the FD's were beautifully centred when Turkish Airlines stalled at AMS. In todays airlines, where visual approaches are discouraged for 'safety issues' and to reduce the number of screw ups and G/A's = time + money, it will take a massive change of culture to shift back to basic piloting skills. There is no will, so there will be no way. And there will certainly be no extra sim training allocated to such frivolities. Meanwhile survivable crashes will continue and Air Crash Investigation on Nat Geo will thrive much to our frustration. One of the latest scenarios was survival if only basics had been employed. If they'd never been taught nor practised then the smoking hole is inevitable.

True, however once on the line you only get back into the sim two or three times a year and spend most of that time doing regulatory items. So in order to keep your scan from getting rusty, F/D off approaches should be regularly practiced on the line (shock, gasp, horror :} )...

I only flew with one pilot that blindly flew the flight director. I told her to level out, you are too low. She advised me look at my FD it says fly down. I said look at the runway. She leveled off and intercepted the glide slope but stormed off the airplane because it wasn't her fault the FD stuck, unbelievable.

What a fatuous and utterly unsubstantiable statement.

Take them away and find out how many score of problems occur per day as a result.

And that is exactly the problem. if you can not fly without the FD, stay out of the flightdeck and get behind the Macdonalds counter.

Said it before and so I'll say it once more;

FDs are a great tool. Like any tool they need to be checked and used correctly and, sometimes, not to be trusted. Treat them as an aid and remember what Uncle Ronnie Reagan used to say, "Trust, but Verify".

Reverting to an earlier post on a different thread and at the risk of being (am) redundant;

When being told to "follow the command bars", the reply was "those are suggestion bars, I'll do the commanding thank you".

This is an example of what I mean:

Despegue, on that logic we'd better bin autopilots too, on the basis that "if you can not fly without ... stay out of the flightdeck and get behind the Macdonalds counter."?

I'm sure you don't actually mean that...

Hell, lets dump instruments, get rid of the windows and put the pilots' heads out in the breeze again. Make the lazy buggers learn to fly by the seat of their pants again.

It aint progress buddy...

See this is what I don't understand - when I go-around I want to go UP and I ensure the machine does
so, whatever the wx and whatever the circumstance. FLY THE DAMN AEROPLANE! has - and always will
be - Emphatic Survival Rule #1. If the FDs say something else and are an annoyance then bugger 'em
- I turn them off and fly the basic combination of attitude power speed and navigation.

In short...Firewall throttles - pull back - get a climb rate - get that gear up - clean up when clear of the
hills and safe to do so (crew workload, airspace, etc). Just because you've plugged 1500ft AGL into the
box as your end of 2nd segment doesn't mean you roboticly shove the nose down and slam into some
thing or someone. Those who fly around the back boondocks of China will know what I'm talking about.

FDs are TOOLS - and very useful ones at that - not gods. In fact I see through them a lot to make sure
they're not bull****ting me - I was trained to do this well on the DC9 back in the days of the old laggy
FD108s. It still holds true for every aeroplane including Airbi.

320 FCOM advises attitudes to fly without FD (EFATO, WS etc). Its primarily what I initially target, fine
tune then confirm with FD.

I can't help but think over the years that good pilotage has to be reinvented every 10 years or so. No
doubt other readers are also aware the primary causes of it.

I'm pushing our Training Dept to train our kids using sole raw data at the Basic level and introduce the
AP/FD/AT later in their sim training. The experimental first batch went well and the children took to it
like a duck to water! Our current sim renewals are raw data approaches and go-arounds for ILS & NPAs
(with AT), for both AEO and EO, Capts and Senior FOs. Juniors get the FD for NPAs only. AT off is used
and practised in the LOFT component.

Bubbers44 - I can certainly believe it. One kid decided to follow his FD into a hill because he screwed
up the crossing restrictions in the box. I let him go so he could realise and correct his mistake. When
I finally spoke up he said everything's fine because the GPWS hasn't yelled anything! :ugh:


PS apologies if I've echoed and essentially repeated other posters' statements.

Years ago, when asked to write a TQ course for B737, I accomplished as much system demonstration and sign off as possible in the FBS phase. Of course some manoeuvres have to be signed off in FFS, but later on. FFS should be focused on handling and operating. Then, in FFS 1 I designed an aerial ballet of turns, climbs, descents, accelerations all mixed up, without FD's to educate the student how to control the a/c manually and learn the scan. ILS's & G/A's followed. This was very useful when converting from needles & dials to EFIS and Maps, and again when converting from up & over screens with ASI & VSI to the side by side screens with speed tape and VSI tape. This allowed 2 hours of acclimatisation and gaining fundamental knowledge of how this a/c wanted to fly. FFS 2 introduced some of the same with FD's. but only 1 hour. Then it was ILS's and G/A's, stalls etc. Now the student had full grasp of the basic a/c and the AFDS. FFS 3 moved onto SE work, a dedicated session. FFS 4 introduced the systems non-normals etc.
Since then I've had to teach courses where FFS 1 had a few moments of poling the a/c and then straight into complicated demos of systems non-normals and all with FD's. FFS 2 was a mixture of various QRH items and then 2 hours for SE intro. The students heads were exploding form doing 6 different QRH's in 1 session and then having to deal with SE intro when their basic a/c handling was still so-so. The foundations had not yet been built. I suggested a couple of extra FFS sessions at the beginning for more GH. Not allowed: "the students are paying and it'll make it too expensive; the course includes all the mandatory items and minimum hours and therefore extra is not necessary. If students are not to LST standard they can pay for extra training later on."
So there is the current philosophy. Sadly, they will not gain much manual flying on the line, as it is discouraged, and command times are 1/2 of what they used to be. Thus many commanders can not perform many basic piloting manoeuvres and there is no will to change. Without regulation and investment it ail get worse not better. Auto-systems will improve and pilots will indeed become button pushers and it will all go well and we shall all survive. There will be no incentive to have good GH pilots. Until............
Things will only change when a top politician or high profile person is the victim of a survivable scenario, but dies and then the questions will be asked: why did the pilot not save the day? Then something may happen.

In 1990, several British glass-cockpit 737 captains were contracted to fly for Hapag Lloyd and based in Hamburg. The reason was a rapid expansion by H/L with new 737-400/500's and their first officers didn't have the minimum hours for command training.

As part of the interview process we were assessed by the chief pilot in a 737 simulator. The test included timed climbing and descending turns at Vref flap 40 at set rates of climb and descent - as well as clean steep turns. This was all done raw data and manual throttle and was certainly a very good exercise in basic instrument flying ability. It was similar to the Pattern D exercise in the old D4 Link Trainer.

It was immediately apparent that those current on raw data flying (whether through choice or previous company SOP) showed up well. Others steeped in automation had significant problems once the flight directors were off and were lucky to just scrape through the simulator assessment. The chief pilot's reason for those raw data exercises was that in his opinion, it gave him an idea who could really fly the aeroplane as against those pilots who were more or less tied to automatics flying.