Is Raw Data Outmoded?

[Centaurus]

Coffee conversation with colleague recently. Flies 737 NG. Very experienced on type. He was concerned that his raw data skills are steadily eroding after three years on automation. The way he describes it, as a first officer, he is completely dependant on the captain's "generosity" in permitting him to turn off the flight director and/or autothrottle during climb to altitude or during descent in appropriate weather conditions or even during a visual circuit.

He reveals interesting reactions from captains. A few say no problem have a go, and monitor the first officers raw data tracking etc. Other captains look aghast and say no way- don't touch a thing and leave the automatics in.

Still others look uncomfortable and mutter things about only turning off the flight director for just a few minutes and only in bright sunny weather. In his experience the majority of his captains baulk at any request to switch of the FD even for short time.

As boxes have to be ticked in cyclic sim sessions and the usual emergencies covered, it gives no time to practice raw data in the sim. In any case if a FD off approach is made in the simulator it only lasts 10 minutes at the most - scarcely time to catch up lost raw data skills once every three months or more.

I would have thought that an airline pilot would be required by the Regs to be competent at hand flying his aircraft both with FD on or off and autothrottle on and off. Not so, perhaps? Automatics skills yes - hand flying skills not vital to safe operation?

Are captains of this era so uncertain of their own raw data skills because of years of automatic dependancy? Research reports would suggest so. Or is it all too much trouble - or worse still is it the dreaded label of "Non Standard" that makes their knees tremble if a keen young first officer asks if he can have a go without the crutch of an FD?

Serious comments only would be appreciated - no ungentlemenly rubbishing, please.

[ManaAdaSystem]

There is a time and place for everything. They can do raw data approaches all they want for me, but not around busy airports, and not in IMC.
FO on raw data greatly increases MY workload, and the risk of errors, track deviations and altitude busts increases.

[Seat1APlease]

I am not having a go but It's brought a smile to my face.

If I could have a pound for every time a FO did a manual approach because of an imminent sim check I would be very rich.

On my first conversion course 30 odd years ago I remember being told that we could try anything we liked on the sim, in complete safety, because this is where we build up skills and knowledge to take forwards onto the line.

Now in 2005 we are blessed with school league tables, hospital star rating, continuous monitoring, and pilots using the aircraft to practice things, so they can then show how well they can fly the simulator.

Its a funny old world.

[safetypee]

A good instrument scan is a fast disappearing skill, predominantly due to the use of automation (FD /Autopilot). The sole use of automation is not a hazard provided the SOPs, (normal and abnormal), and the aircraft systems (normal and failure cases) support this form of operation. In many cases the systems failure case is not supported and training is required to compensate; this can be done during routine operations and thus raw data flying should be encouraged. However some operators (and regulators) who overlook this point and are either oblivious to the increased risk or accept it for the rare occasions.

A greater concern is the loss of scanning skills. This is reflected by loss of situation awareness for both automatic systems and for other aspects of flight operations. “Dependency on automatics leads crews to accept what the aircraft is doing without proper monitoring” – chapter 4 of CAA PAPER 2004/10 Flight Crew Reliance on Automation (http://www.caa.co.uk/application.aspx?categoryid=33&pagetype=65&applicationid=11&mode=detail&id=1598).

I believe that there are many problems which stem from poor instrument scanning. In the first instance how does an operator ensure that the habit (good or bad) from basic training on a ‘steam’ aircraft are transferred / negated during transition to a glass cockpit. Similarly when changing types.

Then what training is provided for use of alternative scan patterns for different flight situations; climb, cruise, and approach, these would not necessarily require the same scan pattern, and there are significant differences between precision approaches and NPAs.

Finally I have concern that the loss of scanning skills is reflected in the current weakness of gaining and maintaining situation awareness; do pilots know what is important, what to look for, where it can be seen, and when they should be looking for it; all of these skills can be developed and improved during raw data flying.

Operators / captains should remember that raw data flight need not be as accurate as automatic flight; it only has to be sufficient for safety. This understanding should remove unnecessary pressure both from the PF and the captain, but in turn this requires knowledge of what is sufficient for safety – see rules and procedures.

Cautionary note; some aircraft designs do not easily enable ‘split’ auto/ manual operations, thus it would be advisable that both pilots use the same display format during raw data operations. Not only would it keep them out of trouble (at least one serious loss of control incident in Europe due to split FD/raw data flight), but it would also enable the monitoring pilot to improve his skill of using raw data as the monitoring source – something that should be considered as normal.

[AtoBsafely]

I fly the Dash 8 (captain) and do a fair bit of hands-on (departure, visual approach, FD ILS). I surprised myself by trying a no flight director ILS and struggled to fly a half decent approach. These tools help us fly accurately, but if you you lose them make sure you don't get lost! What's the chances of really needing them? What does it cost to practice raw data (not just for the sim)? I say practise when you can!

[SR71]

ManaAdaSystem

FO on raw data greatly increase MY workload, and the risk of errors, track deviations and altitude busts increases.


And the corollary is, perhaps, even more pertinent:

CAPT on raw data greatly increase MY workload, and the risk of errors, track deviations and altitude busts increases.


Embracing this philosophy, are we going to be feeling better or worse when, in IMC with a FD failure, possibly OEI, you have do this for real?

There is a time and place for everything?

But the original poster recounts the experience of a colleague who says that, whilst this is an admirable sentiment, time isn't afforded him/her, even in the sim, to practice raw data flying. So just when does he/her get this experience?

I didn't learn to get my knee down at 120mph on my bike without having a go. Thankfully, I only fell off once during the learning process!

Its another example of paying lip service to safety.

When was the last time you had a FD failure during an EFATO or at any subsequent point during a check?

Centaurus,

Your colleagues experience agrees with my own. Personally I'll take a raw-data/visual approach whenever I can but the range of reactions from the guy/gal in the other seat vary tremendously.

Switch everything off, go direct to 10 mile final from VILNA for RW10 at ALC in day VMC?

Oooooo...don't know about that.

You what????

Am I driving this thing, or is it driving me?

Operators / captains should remember that raw data flight need not be as accurate as automatic flight; it only has to be sufficient for safety. This understanding should remove unnecessary pressure both from the PF and the captain, but in turn this requires knowledge of what is sufficient for safety – see rules and procedures.

Brilliant point.

I made some comments in another thread about how the real test of your safe operation is how your perform when things start going wrong.

16hrs/year isn't enough to ensure that, bearing in mind the complexity of a modern commercial airliner flightdeck, you can deal with the effects of certain failures, that whilst being statistically rare, are operationally, enormously significant.

Seat1APlease,

Its a topsy-turvy world!

:ok:

[Jetstream Rider]

I quite like it when the Capt swithces the auitomatics off. It increases my workload SLIGHTLY not too much and I pay more attention, making an alt bust less likely. That is unless he is a fool or well out of practice.

It is possible in the aircraft I fly to have a SINGLE failure and lose the three autopilots and the flight directors (if the MCP panel fails). I want to be able to practice manual flying, even at busy places and IMC, as it is DIFFERENT to flying it on a nice day.

If it is too dangerous, then by definition we should declare an emergency when we take the automatics out. If you are at the stage where it is that dangerous then you really should build up your skills again. The sim is not the place to do it, as we are hardly ever in it. Perhaps we get 10 mins at the end to practice something after the boxes have been ticked. That is not nearly enough.

A good scan and the ability to fly the aircraft as well as chat, listen to the radio, keep situational awareness, ident navaids etc will increase capacity (size of bucket) greatly.

Having said that, the autopilot is a great tool and improves capacity (by taking balls out) and should be used where it helps, especially in non normal situations.

If I can't fly an approach to CAT 1 minimums reliably, every time, I want to take LOADS more fuel so I never get in the situation where I will need to.

[ManaAdaSystem]

OK, I think you have misunderstood what I said, SR 71.

Effoh can (as far as I'm concerned) practise raw data as long as we are in an easy environment where errors will not cause anything more than a slight embarrasment. Raw data (I'm not talking FD) approaches can be flown heads down simulating IMC, there is no need to be in actual IMC to practise this.

I will not allow handflying, raw data into LHR in IMC. Period.

Last I had a FD fail was four days ago. The Effoh side was ok, so I handed controls to him. Should I fly raw data just to prove I can?

We practise raw data every sim session. And I practise in the aircraft. Just not when the workload is at peak WITH all systems working.

[Intruder]

I just transitioned from 744 FO to 742 Capt. I had a significant amount of raw data ILS practice, mainly due to 1 Capt I often flew with who insisted I fly raw data when the weather was good (and, later on, even at night or reasonable IFR).

As a result, my transition to steam gauges was a snap, even after 7 years of glass.

If the Captain retains his FD and ND while the FO turns his off and/or goes to ILS mode, there should not be a significant increase in workload for the Capt.

[FullWings]

Ever since I started commercial flying there have been worries about loss of some skills due to automation. I think it's just a fact of life now and the people who write the SOPs are well aware of it.

In the last ten years I have gone from a jet which was still airworthy with no working automation, regularly flying fully manual limiting approaches, to a FBW beast in which we're not even allowed to disconnect the autothrottle...

If I was asked which aeroplane/crew combination was safer for the general public to travel in, I would unerringly point to the more modern airliner and the fact that although it is highly automated, it is incredibly reliable in that respect. I'm not as good as I used to be at "shooting a quick visual" but then I don't think that's a big a part of the job as it used to be. :(

We have GPS, TCAS, EGPWS, Envelope Protection, etc. which have done far more for flight safety (IMHO) than having "Top Gun" level manual flying skills. OK, so a combination of all would be great but my airline can't afford to give me 10hrs a month in an F-16. :{

There are more and more flight regimes where manual flight is not encouraged, i.e. most sub-CAT I conditions, RVSM, V-NAV approaches, to name a few. However there is still a need, i.e. circling, ILS PRM breakouts, windshear, etc. for some nifty handling when the chips are down. Not to mention little things like crosswinds, rain and snow.

My personal opinion is that we are relying on a diminishing pool of experience. "Handling events" occur quite regularly in my airline and the management response seems to be to reduce the amount of manual flying allowed in order to limit the risk exposure. Of course, in the long term this will cause more problems but I would never accuse airline managers of thinking long term!

Does the glass cockpit layout encourage much of an instrument scan? I wouldn't say so but we must be careful not to lose the plot here. The 'classic' instrument layout and scan was forced on aviation by lack of technology at the time. Now we can present most of the useful information in one place in any format we like. Do you really 'scan' a HUD? I would guess not as the data is just 'there' for you to assimilate. The better designed the human interface to the aeroplane, the less you need to learn a method of gathering the required information or spend conscious brain power interpreting it.

As an aside, I have to say I've never seen much point in manual flight using a FD. You either follow it, which doesn't need much skill (or practice), or not, in which case why have it on at all?

[barit1]

An old hardware designer - 35 years ago - cautioned me that the more he "Murphy-proofed" his hardware, the more he contributed to raising a new generation of "bigger and better Murphies". Someone who has not experienced soft failures will have no idea how to prevent catastrophic ones.

The present thread is a corollary of his admonition, it seems. If I couldn't hand-fly the automated bird often enough to remain competent, I would be unprepared to handle the real thing.

[BEagle]

Hmmmm....

All my large a/c flying was in 4-jets with 1960's-era autopilot systems, latterly augmented with a basic LNAV steering option. No autoland, no VNAV, no autothrottle above 180 KIAS. Using the primitive 'automatics' was often less simple than manual flying - and we regularly practised NDB let downs with just a basic RMI. Most crews flew manually once they began to decelerate in the instrument pattern towards flap extension speed; the 'Auto-ILS' was crude and prone to spurious disconnects, even handled carefully.

But such museum pieces are now becoming extinct. More modern a/c with vastly superior AFS systems, full flight regime ATS, FMS-profile VNAV and LNAV are now the norm. And excellent though they are, a different skill set is needed to operate them to their full advantage.

Which is fine under normal circumstances. I recently observed a flight in a relatively modern 'demi-glass' 2-jet a/c where, after a low-level 200 ft flypast, a TOGA was flown to level at 3000ft. ATC then changed the clearance to 5000ft. The PF (Captain) rapidly got himself totally behind the a/c; despite frantic twiddling of knobs various and pushing various FCU pushbuttons, he busted the level off by a good 400ft. This in a busy ATC environment in IMC at an international airport. Having mishandled the AFS, it simply never occured to him to go back to basics - disconnect the AFS, reduce thrust and level off manually.....

Had they not flown their fly-by but just landed normally, all would have been fine. But by doing something unusual, they soon got out of their depth. This is the hidden danger with such high performance twinjets at light weight; the AFS normally flies so smoothly and accurately that the crew can be lulled into a false sense of security - and then quickly get behind things when attempting something out of the ordinary.... Even something which should be very straightforward such as a missed approach and go-around under ATC radar control. Something which we used to practise all the time in our old 1960's-era 4-jet, but which is an exceptionally rare event in a modern twin.

So some old-fashioned manual flying skills are probably still needed on modern a/c - but they're best practised in the simulator, not in busy terminal airspace, I would suggest.

[Cerberus]

Do we actually do 'Raw Data' anymore on modern jets?

When we turn off the automatics you actually sit there with a flightpath director and speed trend arrow. These would be considered wonder electric-jet modes 15 years ago. So the 4 skills of keeping level, on-speed, adjusting for wind and setting the right thrust are taken care of. Then we have the Nav Display so actually getting onto a radial is a none event. To aircraft designers 'Real' raw data is so unimportant that the 330s RMI has been out of use for about 4 years and the F/O doesn't even have a standby ADI. The skills have changed! If you put GPS into the mix as well, I wonder how many guys have flown an NDB approach without identing the beacon or being aware of their rate of decent on final?

My instrument scan on EFIS is less than it was on a HUD and less than on AHARs and less than on the JP and heaven forbid less than on the mighty Chipmunk. What I would agree with Beagle about though; is that the best form of reversion is total. If you are gonna overspeed or bust a level, disconnect the automatics and fly the aircraft! Without the designers 'automatic control damping system' between you and the Flt Control Computers, it is amazing how quickly the beast can be brought back under control before re-engaging it all again.

'Needle left, ball right, numbers decreasing and both horizons steady!!'

Cerberus:ok:

[SR71]

ManaAdaSystem,

Should I fly raw data just to prove I can?

We practise raw data every sim session

I think your expressed statement describes the uneviable position we all find ourselves in.

Presumptuously, perhaps I may crudely paraphrase the above as:

I can, but I won't/don't feel the need to.

But wouldn't it be slightly more accurate to say,

I'm not quite so sure I can, so I won't/don't feel the need to.?

Considering your rhetorical(?) question, one ought not to be considered either a hero or a daredevil for completing/flying a manoeuvre that it is quite legitimate to expect a pilot to have in his repetoire of skills i.e., a raw-data approach in a TMA, IMC.

After all, in your particular scenario, arguably (s)he is trained to complete the manoeuvre in far more taxing circumstances, whereupon, why exactly the prohibition?

Surely, the answer is that, actually, we're not really that confident in either our own or the other guys abilities to do the job bearing in mind the 30 mins/year we spend doing it and a healthy dislike for paperwork.

Bearing in mind the significance of the task at hand when the chips are down, most seem to adovocate prohibition rather than more training.

Is it that we are all flying around thinking that, if its happened for real, and I've declared a MAYDAY, I'm OEI in IMC with a 35kt crosswind on the deck in snow flurries, inop TCAS, A/P and FD and min fuel, none of the normal rules apply regards the accuracy of our flying?

Various other posters have already suggested that, even with two donks turning, something slightly unusual can easily leave us way behind the drag curve. Sailing past the GS in ALT ACQ/ALT HOLD without having armed APP is another. By far the easiest solution to the problem is...

The safety of the operation is a function of the ability of the machine to do what it is supposed to and our decision making skills and training.

These days the machines are pretty good, but when they aren't, its all down to us, and its then that our claims to be safe really are tested.

I want more training because I feel like Murphy.

:ok:

[411A]

Looking at this from another angle, I have recently completed line qualification for a Captain who, having been in modern glass twin-jets, now finds himself in command of the Lockheed TriStar once again.
Now, while the TriStar was the best there was in automation years ago, it surely is not as automated as many today (such as the B737NG this guy was recently flying), he did not present any problem whatsoever in going back to steam-driven gauges...with raw data NDB approaches the norm on most flights at the moment.

OTOH, this Captain had lots of prior experience in older transport jets, so the transistion was a non-event.
Could younger guys cut the mustard, with minimal exposure to round dials/raw data?
Most likely not...therefore, perhaps they (and their 'overworked' Captains) had best stick to the automation plan, for better success.

Hey, the automatics never fail...right?:} :E

[McD]

Hey, the automatics never fail...right?Not too long ago, on a "modern" aircraft, we experienced dual failure -- right in the midst of a rather complicated STAR into a very busy airport.

(No, the failure wasn't self-induced or operator error ;) )

Thankfully, because we were both still proficient at raw-data flying, had a decent scan, etc., the lack of our normal features didn't cause us any real problems at all, although it was a bit "busier" of course. But it just goes to show you -- you must find ways to stay proficient with raw data, because you never know when your automatics might choose to fail.

[Alpine Flyer]

IMHO raw data practice is required as long as raw data flying might be required in an emergency situation.

On most aircraft I have flown so far the flight directors are gone in a battery-power-only situation, so it seems prudent to mee to practice flying without them.

In an ideal world this practice would be done in a simulator, but as many companies won't schedule more sim time than the regulations require, and that time is completely filled with other "required training", this probably won't happen.

I don't think raw data flying compromises safety when limited to VMC and "other than high density" traffic situations.

I wouldn't practice it at LHR or FRA.

I am not a grizzled veteran (flying the line for 15 years) and rather feel that copilots should switch off FDs (even APs/ATS) more often. I very often see FOs switching the AP off at 700 AGL. I probably couldn't hand-fly any ILS if I always let the AP do it's job....)

I also agree that "real" raw data is hard to train, as most glass cockpits offer a lot of helpful stuff even when the FD is off. (speed trend, flight path vector, etc.)

[Centaurus]

I must say thank you for so many thoughtfully written replies.
When I originated this subject, it presumed that commonsense and good airmanship would prevail when it came to when to disconnect the automatics (FD, AT etc) in order to keep up with raw data skills. Clearly in bad weather, low cloud or other marginal weather, circumstances would dictate automatics as the way to go. But to say that raw data in IMC is just not on under any conditions (subject to the caveat above), indicates perhaps a certain apprehension that one's skills are a tiny bit rusty.

Which is more taxing? A one engine inoperative autopilot engaged, ILS or NPA approach, using manual throttle and manual rudder trim? (typical 737 set up) - or an all engines non-automatics raw data ILS in IMC (actual or in the simulator).

Each of us who have flown the 737 will have done many of the former quite happily - after all it happens all the time during type rating or recurrent training in the simulator. Given the choice, I would take the "normal" raw data approach every time because it is easier and less workload.

I fail to understand why so many pilots sometimes show ill disguised apprehension (fear?) of simple straight forward raw data flying especially when research has revealed that automatic complacency (read laziness) is a known hazard.

It should not be forgotten that cross-reference skills and situational awareness will not be significantly improved by staring fixedly at the cross hairs of a flight director while at the same time unconsciously relying on support calls by an alert crew member to remind you what is going on outside the magic FD picture.

Perhaps one of the most alarming statements I have ever seen was in the FCOM of a European B737 Classic operator. The policy section on use of automatics said "Manual flying should not be attempted unless absolutely necessary."

And that's even with the flight director turned on!

[Jetstream Rider]

Do you need to scan with a HUD - YES!

Ok the information is "there" but part of scanning is a controlled focussing of attention. Just staring at the wiggly lines means nothing, however close together the information is.

Why can't I practice IMC or night raw data? If it is dangerous to practice, then it is dangerous to do it for real should the need arise. The need won't arise on a nice sunny day. Flying in cloud is different, even if you look down, the aircraft response to the bumps for a start.

On the Airbus, things are a little different, than the 757 or 767. We do not have a trend vector or a speed vector. In an Airbus, the required level of manual reversion is different to that of a different type. I have no need to practice DECCA or LORAN as I have no possible way of ever needing them in the aircraft I fly. There is a need to fly without the FD or the AP though, so let me practice.

The human is extremely good at doing complex tasks, but requires practice. The circus unicycle rider can do some pretty fantastic stuff, safely and talk at the same time. If he did it for 30 mins a year, he would soon lose the ability to do it as well.

I have seen many Captains try to fly a visual approach with the heading bug and go through the centreline a number of times on the way down. Much better to unplug the AP and point the aircraft at the right place.

We fly to Grand Cayman and at the moment there are no electrical aids at all, except lights. (might have changed by now). Some Captians are scared of this!!!!!! OK, you need to break cloud above MSA and if you can't get in divert - that is sensible. But to be frightened of it on a gin clear day (like that part of the world generally is) I find interesting.

The only reason some of these things above are considered dangerous is because we never practice. In the olden days (!) BEA said the Trident should be Autolanded wherever possible, as the failure rate for sims went up, the rule was changed. The failure rates then went down.

In BA the Airbus guys cannot disconnect the autothrottle. One of my mates has just passed the Airbus course with 1 days worth of training with manual throttle. Is he safe to fly on the line when the AT fails - or indeed to use the MEL 3 day autothrottle inop section?

If you don't practice these things, you may be OK to fly without them, but add something else like a garbled radio call and all of a sudden something like an alt bust becomes iunevitable as your capacity bucket is full. If you practice enough, added extra things do not approach a dangerous situation at all.

[barit1]

IIRC that AC crew dead-sticked a 767 into Gimli, a few years back, without benefit of navaids. Not so sure about Air Transat's A330, though.

[ManaAdaSystem]

Presumptuously, perhaps I may crudely paraphrase the above as: I can, but I won't/don't feel the need to. But wouldn't it be slightly more accurate to say, I'm not quite so sure I can, so I won't/don't feel the need to.?

Not in this case. We had traffic and terrain considerations and a minor tech problem. Just my way of practising CRM.

[BOAC]

F/D on or off, manual or F/D - it is ok with me subject of course to workload (and perceived ability), and visual approaches are encouraged where appropriate where they will 'facilitate' or improve skills.

It is, however, a good thread, and we have to look at where 'modern' a/c design is taking us. How long before the average C152 has GPS/RNAV etc? The path to the RHS of a jet then follows computer generated inputs from day 1. It is not unusual to see an F/O who cannot grasp the concept of finding one's way around on DR/VOR/DME, or even who is 'lost' without a map display with waypoints and track line - in other words, basic position skills are going. Time and time again in my airline career I have seen the seductive 'head down, sort out the FMC' style when what was needed was 'disconnect and point' - and been 'seduced' myself. Frightening to me, but with the advance of system reliability and automation, should we be looking at only flying automatics in simulator tests? I know not of airbus, but believe that the sim cycle is flown using the automatics? As said above, I have heard that BA are 'discouraging' the use of manual throttle on the AB, and I believe, 777.

Is system reliabilty at that point yet or do some airlines need to review policy?

[flying scotsman]

I would have thought in this era where it is simple to build solid state standby instrumentation it's not so essential to be as good as we all used to be BUT..........


why isn't there more of an effort to make the primary instruments far easier and more basic. I accept that the ng instrumentation we use now are far better than on an old caravelle etc but the philosophy is the same. maybe it's time for a radical rethink eg the way airbus changed things in the eighties with FMS and computetr systems.

horizons and loads of gauges all over the place make things harder during an abnormal. wouldn't it be nice in a critical situation to have information presented more accurately.

why not a voice activated QRH from the egpws. for example............

just a thought.

[haughtney1]

IMHO, I think it really comes down to how you want to fly, personally I have no problems with raw data..its actually what my company SOP's call for to be used on departures and arrivals.
Some of the places we fly to, require all the automatics to be switched off, and you are left flying a 120 tonne jet purely in manual, a little intimidating at first...but when you get used to it, its really no big deal.
So far as keeping current, I think the biggest problem here is scan-rate...it might sound a bit odd to some, but I use flight-sim every now and then..it helps me develop the scan rate effectively.
I think the other key area is knowing your a/c configuration, and how this effects AOA, power settings, and ROD or ROC. If you put all this into place then you cant go far wrong..and if you do..it just needs small adjustments to regain the correct flight profile.

I dont think raw data is outdated, I think it will continue to be the primary means of accurately flying/navigating an aircraft, after all the FMC's/Air data computers of this world merely filter the info in such a way as to make it more representable to the human brain..nothing more. Basic flying skills..and basic scan rates can do exactly the same thing.


Just my thoughts:ok:

[ikan_terbang]

Raw data flying is in Steven Covey metalanguage ... "sharpenning the sword".
When you do get the chance, or if the opportunity arises, take it and enjoy it.

It is pretty safe ....no doubts

[Centaurus]

This makes the subject of raw data quite topical. Take a quick gander at the Australian Forum regarding an incident today to a MAS B777 that struck trouble some time after departing from Perth. Passengers reported heavy buffeting and aircraft gyrations. Aircraft returned to Perth. Reading between the lines as an armchair expert, it could be that the crew were un-current on basic raw data after an ADIRU problem?