Fly By Wire Design Problems

[SASless]

I participated in an interesting discussion at Kevin Barry's, a really interesting Irish Pub in Savannah, Georgia, re" Fly By Wire" design issues. The chat was about the now dead Commanche Program and a problem that was discovered when the US Army requirement specifiying two identical cockpits for the aircraft two crew members. The original thought was to simply crew training requirements and provide for redundancy.

What seemed to be a major concern by the Army was the possibility the FBW flight control system could be confused by two simultaneous but different control inputs.....say.....rear seater applies left cyclic....and front seater applies right cyclic such as in an emergency manueuver to avoid a tree while flying at low level.

As I heard it....Sikorsky say it was not a problem....as during testing no such problems were encountered but the Army was not satisfied with the test criteria and the resulting statement by Sikorsky.

Anyone know anything about this kind of thing?

[Rigga]

As this sort of thing is already installed in many fixed-wing aircraft I dont see how this could be an issue?
Both Pilots have inputs to multiplex control system, but the PICs (whichever) inputs are favoured.
There must be more to this issue than portrayed from the US Army.

Incidentally:
Kevin Barry...hanged for membership of the IRA circa 1920 - my dad was named after him!!!

[Brian Abraham]

The Airbus system.

Each sidestick has a takeover button. In normal flight, simultaneous inputs from both pilots will be summed algebraically, but priority will be given to the pilot who depresses, and holds depressed, the takeover button on his stick. A green light on the glareshield tells the pilot he has priority, while a red light tells the nonpriority pilot that his stick is dead. He can then retake control by depressing his takeover button. If he elects instead to release his stick, the green light in front of the priority pilot will go out. The priority pilot can then release the takeover button, extinguishing the red light and returning the aircraft to normal flight control.

If the non-priority pilot does not release his stick, however, because he is incapacitated, the priority pilot keeps his takeover button depressed for 30sec and then the dead stick is electronically latched. As soon as the obstruction is removed and the dead stick centred, it is automatically delatched.

[Hullaballoo]

So when it comes to training, "follow lightly with me on the controls" is no longer plausible?

[SASless]

The discussion did not mention any sort of "Take Over" system.....but that would have worked it would seem. Any one from Sikorsky know how this was addressed in the software or later modifications before the project got killed politically?

[The Sultan]

SAS

The Comanche's death was not political. No one wanted to pay for a $100M helicopter which one bullet could take out.

The Sultan

[heli-cal]

No one wanted to pay for a $100M helicopter which one bullet could take out.

So, its like a rotary winged version of the Death Star, capable of being destroyed by a single round! :confused:

[Matthew Parsons]

I think it is best to call this a consideration vice a concern. As long as you build a logic that makes sense, and is understood by all the operators, then you shouldn't have any problems. There is difficulty in building something that makes sense, and definitely a problem in determining what you do if you leave one dead, then switch to the other (trivial if they're both centered, but if they are opposing at the time of the switch...).

[birrddog]

What are the votes on how important it is to be able to "follow through on the controls" in a FBW configuration?

The caveat being this is for highly qualified operational pilots, and not ab-initio training...

[Brian Abraham]

What are the votes on how important it is to be able to "follow through on the controls" in a FBW configuration?
That has proved to be one of the problems with the Airbus if you read the incident reports. There is an indicator in the panel showing control input but who's going to be looking at that when taken by surprise or in a moment of stress. Another recurring problem has been the failure to punch the priority button when the PNF jumps on the controls in an attempt to put things right and salvage a situation.

[Thridle Op Des]

Brian has the Airbus FBW logic explained very well, however it is prone to serious misunderstanding by Airbus Line Pilots. The PNF can 'Dual Input' against a PF control anytime, but you get scolded very severely by the gentleman who shouts 'Dual Input'.

As someone who trains Ab-Initio pilots with either no previous FBW experience or very low total hours (our National Cadets for example) the issue of no feedback can be a little challenging. It was always reassuring as a trainer on a Bell 212 for example to feel or see the control inputs from the trainee, however the secret to the Airbus is to watch the attitude of the aircraft (and sometimes the trainee) and be poised on the sidestick (cyclic) with the thumb on the take-over push button (where the winch cable cut/ext load disconnect would be on a helicopter cyclic).

A classic issue with us is the flare for landing, either done too early or late - both equally "bad", the passengers usually notice the latter (as they limp off the aircraft with a slipped disk) and the "spy in the cab - FOQA" usually notices the former (long landing or tail strike as the float is progressively held off and the pitch attitude is increased until there is a strange scraping sound overlaid with the screams from the cabin crew in the back....)
In a sense FW are easier to take over since you are mainly guarding against over or under pitch with roll being secondary but still very relevant beyond 7 degrees. (see the korean-landing-in-hong-kong-horror-video).
With rotary issues there are more axes to take over in and would require considerable vigilance IMHO.

The beauty of FBW I have to say is the total lack of requirement to trim. If anyone has flown a well set up SFENA equipped Bell 212 (I know SAS, that probably happened three blue moons ago), it's very similar and was called 'Transparent'. With the Airbus FBW commanding a "G" load in pitch and a roll rate in roll the aircraft trajectory remains where you stop inputting onto the controls, it is literally point-and-shoot (within reason of course - top of a wing over excepted). The other great benefit are the provisions of protections which you cannot do with an analogue classic helicopter control system. Until you actually witness the effect of hurtling towards a Swiss cliff at 250KIAS, waiting for the EGPWS to give you the Whoop, Whoop and then applying full back sidestick (cyclic) and fully forward on the thrust levers to TOGA thrust and literally hang on the edge of the stall (with a safe margin) climbing at 6-7000 fpm out of danger it is hard sometimes to appreciate what 'protections' really mean.

In regard to the comments about 'feeling the vibration' back fed through the controls, I would venture to offer that in a modern helicopter like the 225 or 92 with 3000 ish PSI of hydraulic pressure doing the work for you, then the chance of 'feeling the vibration' would be near to zero.;)

TOD

[Runway101]

I am not an aircraft designer, but in a helicopter FBW design it would make sense to actually start the FBW infrastructure _behind_ a normal mechanical dual control system.

Such a system would allow normal handling by pilots during operation and during instruction, and still provide the benefits of FBW (thanks for the swiss cliff example). In other words, the stronger one wins without confusing the electronics....

I fail to understand why this is not done on planks as well, but maybe the designers thought all this complicated logic makes definitely more sense than taking the risk of the co-pilot bumping into the controls while he gets up to get a coffee from the back.

[maxtork]

Is it just me or does this seem to take an already complex system and make it even more so? I would think a much easier solution to all this would be to simply link the pilot and copilot controls together mechanically. You could still be FBW from the cockpit to the aircraft servos but you would eliminate the conflicting control inputs from two pilots, and also give the feedback for training. Granted this feedback would only allow one pilot to feel the movements of the other pilot on the controls but as already mentioned, with 3000 psi hydraulic pressure how much feedback could one expect to get from the aircraft? Maybe I am missing something and this is not possible.

I guess I am old school and I worry about electrifying and digitizing things that worked ok without. I work with electronic control systems everyday (FADEC) and I can say they do provide great benefits when working correctly but when you have an issue it is much harder to troubleshoot. I can see when a push pull tube is bent or broken but I can't see when electrons are leaking out of a wire at a bad connection. I don't know how many times I have gotten a call about an aircraft that "did something funny" in flight but the fault detection system reported nothing and the pilots have now lost confidence in the machine and dont want to fly it again until fixed. This is understandable in many cases but cleaning all the cannon plugs and a maintenance test flight that does not duplicate the problem doesn't always restore that confidence.

My point is that all of the computer technology we install seems great as much of it brings a significant added value. As long as the added value is there I'm all for it but, there has to be a point of deminishing return. I admit I know virtually nothing about the Comanche or an Airbus (other than how uncomfortable the seats are on a long flight), but I would think from my outside view that a simple mechanical linkage would do this particular job better or at least easier.

Max