Propellor feathering on light twin-engine aircraft

[Centaurus]

There wouldn't be a pilot who, during asymmetric training on his initial light twin-engine aircraft, that did not have the mantra drummed into him of: Pitch up, mixture up, power up, flap up, gear up, identify, dead side dead leg, throttle close to confirm identification then feather. Maybe not in exactly the same order depending on aircraft type and instructor preference.

Chances are for the rest of your career on light twins that order of actions will be stuck in your mind. It is almost certain your instructor in those early years would not have told you that the above sequence of actions take time and that you should never hurry them. Take your time Bloggs, lest you inadvertently identify the wrong engine and close down the good one.

That may be good general advice especially for a cruise engine failure but does the same order of things apply to an engine failure after take off where the cause of the engine failure could be a major mechanical fault; for example seizing bearings due to loss of oil?

Back in 1980, Aviation Safety Digest published the following article entitled Propeller feathering on light twin-engine aircraft. It introduced the article thus: The following article was produced as Aeronautical Information Circular 9/1979 by the Civil Aviation Authority, United Kingdom. It concerns the possibility of feathering difficulties with propellers fitted to light twin-engine aircraft. The message it contains is applicable anywhere in the aviation world

Most feathering propellers (hydraulically actuated, constant speed, such as the Hartzell and McCauley types) fitted to twin-piston engine light aircraft are designed in such a way that it is not possible to feather the blades below a certain low rpm (typically 700-1000 rpm). This is because at these low rpm centrifugal latches operate to hold the blades in fine pitch to ensure that when the engine is shut down on the ground, the subsequent restart is not made with the propellers feathered.

In cases where the normal windmilling rpm at low airspeed may fall low enough to prevent feathering, the Flight Manual, Owner's handbook or Pilot's Operating Handbook warns the pilot that feathering cannot be accomplished below a certain rpm. However, the full implications of the situation may not always be clear, and other factors of which the pilot should be aware of are:

(a) In the event of an engine failure caused by a major mechanical fault (e.g. seizing bearings due to loss of oil), the rate of deceleration of the engine can be rapid and it is thus imperative that the pilot take immediate action to feather the propeller before the rpm falls to the 1000 rpm region.

(b) On most twins the usual procedure when shutting down an engine which has failed is initially close the throttle of the inoperative engine. This serves to confirm which engine has failed before commencing the feathering actions. However, if the windmilling rpm has reduced towards the critical region where feathering may not be successful, then re-opening the throttle will usually increase the rpm slightly and improve the probability of being able to feather.

(c) In the event of an engine failure, it is important not to let the airspeed reduce below the scheduled engine-out climb speed. This will help ensure that the propeller continues to windmill at sufficiently high rpm for feathering to be successful. if optimum performance is required it is vital to achieve and maintain this best engine-out climb speed.

(d) The loss of performance associated with a stopped propeller in fine pitch or more importantly with a windmilling propeller is potentially serious. The additional drag will considerably reduce the single-engine climb performance from that available with a fully feathered propeller. The directional control will also be reduced, though adequate control should still be available down to the minimum control speed (Vmca), as Vmca is determined with the propeller in the condition existing prior to feathering action by the pilot (i.e. normally with a windmilling propeller). It will probably not be possible to trim the aircraft on the rudder trim at the best rate-of-climb speed and considerable foot force may have to be held to maintain heading. However, it cannot be over-emphasised that, if it is necessary to gain or conserve altitude, the best available performance is essential and for this the best engine-out rate of climb must be maintained.
............................................................ ......................................
Paragraph (a) bears repeating: (a) In the event of an engine failure caused by a major mechanical fault (e.g. seizing bearings due to loss of oil), the rate of deceleration of the engine can be rapid and it is thus imperative that the pilot take immediate action to feather the propeller before the rpm falls to the 1000 rpm region.



Centaurus comment: Flying school instructors qualified to instruct on light multi-engine piston aircraft should bring the attention of their students to the above UK CAA AIC 9/179 and explain that with an engine failure shortly after take-off caused by a major mechanical fault, the subsequent actions prior to feathering may have to be modified in order expedite feathering before the rpm drops below a critical figure.

The current updated CAA advice can be located in the following link:

http://www.ead.eurocontrol.int/eadbasic/pamslight-0350E825EFDCDCE3552027740D306FAC/7FE5QZZF3FXUS/EN/AIC/P/100-2005/EG_Circ_2005_P_100_en_2005-12-08.pdf

[Checkboard]

There is no performance guarantee in light piston twins for Engine Failure After Take-off, and so looking for the best crash site is also up there on the "important" list.

Spend important time to correctly feather the engine, and then crash with a cry of surprised dismay because the aircraft is STILL descending, won't help you much, either...

[Oktas8]

I see a dilemma.

The article referenced by Centaurus makes a valid point.

However, the advice not to rush a feathering drill is also correct, and written in blood.

So, will you teach inexperienced pilots to do drills quickly, or methodically? In my experience, you can't teach both, even if you stipulate the different scenarios applicable. The brain doesn't work that way.

Given the accident rate and the type of accidents commonly experienced in GA, it seems to me that doing emergency drills methodically and diligently is the way to save lives.

That article can then be a discussion point (as Centaurus intends, I'm sure!) for commercial pilots who would like to expand their existing toolbox of life saving tricks.

[VH-Cheer Up]

Here begins the argument for four engines.

[Homesick-Angel]

Every time Ive tried to rush in an aircraft it hasn't made things happen any faster (or easier). I also feel less effective and more stressed.

I get in the Sim at work as often as possible and spend the entire time in worst case scenarios. The more I practice the quicker and easier it becomes, and the more time I seem to find simply by being more proficient and confident with the procedure.

I also like what Bob Hoover said about airspeed and twin performance. Breaking it down he said If you've got and keep the former, then you'll get the latter. Im sure many of you will have seen this, but watch this vid (https://www.youtube.com/watch?v=g7R7jZmliGc) for the way he's handling the feather and shutdown.

I find it helps to have selected any possible landing areas prior to the roll and Il mention them in the TOSB. There's usually a preferable way to turn even if its only a slight change to heading (terrain, buildings, wind etc) . Another time saver.

[DeltaT]

If I have an engine failure after take off and haven't reached blue line speed, what am I achieving by worrying about feathering?

[MakeItHappenCaptain]

Going to start this reply with the caveat, "Not every situation will be the same."

Very good point by Delta T (and in no way detracting from the OP's intent) in that you won't keep it under control below Vmca. You need to increase speed (most often by lowering the nose) if below Vyse, so if there is runway remaining, it would be wise to get it back on the ground. In between, you are in dangerous territory.
I would suggest that even going through the fence at 20-30 kts will have a much better result (ie. Less likely to kill yourself, disregard aircraft damage, that's what insurance is for) that trying to continue flying low and slow.

There is conjecture about utilising balanced field lengths when flying FAR23 certified aircraft with regard to a desicion point. Decision speed is therefore an alternative. FAR25 (heavier transport category) are guaranteed performance. FAR23 only guarantees control in the event of an engine failure.

Suggestion however wrt throttle confirmation of failed engine. If the engine is still producing any power (evidenced by a last minute change of noise during the final portion of the throttle closing, it is still in a better state than feathered. This will be more easily detected by slowly confirming with throttle rather than belting it into the closed position. A turbo failure for example will still produce ~75% power normally aspirated at sea level in a PA-31, but still give a noticeable yaw when the power reduces, possibly inducing the pilot to initiate a complete shutdown.

Again, not a one size fits all discussion.

[Checkboard]

If I have an engine failure after take off and haven't reached blue line speed, and I am not already above the trees/buildings
Fixed it for you. :ok:

Forget about single engine climbing in a light twin - the height you have at engine failure is the most height you'll get, in 90% of cases.

[Capt Casper]

Pontificating about responses to an engine failure after take off in an aircraft not certified under CAR 125 or equivalent is futile.
One response only - Are you above Vmca by a good margin? & Are you above V2 by a good margin?
If both apply,[and the numbers should be assessed pre take off ] - commit to continue - [my idea of a good margin is 20%].
IF NOT-, commit to an emergency landing in the best area ahead.
Statistics pretty much verify that loss of control is fatal.
Only test piots have the skill and "corporate knowledge" to operate at the boundaries!!

[Nomde plume]

90% is a pessimistic figure. Single engine performance is largely a function of weight and density, of course some aircraft are more critical than others.
You should know the single engine limits of your aeroplane for weight/density before you take off (allowing for engine/airframe degradation of course). If you are outside the limits, take a single! Half the chance of engine failure and certain doom! :ok:

[Bankstown Boy]

Very interesting article and does provide good food for thought.

I fall into the camp that aircraft like PA31s, in the Australian summer particularly, should more be considered like a powerful single, rather than a twin, below 'comfort' height (this 'height' depends on ALL of the circumstances).

Hey, forget about about the reduced ROC or power output with a failed turbo, or the like - I've been in PA31s (very well maintained ones too) that I've popped off the end of some runway in the apparent dead centre of the GAFA and have been convinced that neither turbo is working, despite seeing a healthy 40+" of MP! Anybody who's flown these things much, knows that feeling.

I always based my plan, in those conditions, that I would expect to get a couple of hundred fpm ROD rather than alt hold - on the basis of, plan for the worst - hope for the best.

These things, 40 years ago, when all nice shiny and new, were pretty much promised to do a few fpm ROC at 15c, with Bob Hoover behind the wheel.

Consider, the original P charts only show a SE ROC for a PA31 up to a max of 35C at sea level - hotter and/or higher than that - you are the test pilot. 35c in the GAFA in summer, would have you reaching for a jumper.

[JustJoinedToSearch]

The 17 most popular ways to fall out of the sky video had a good line about light twins.

It was something along the lines of: "Unless you restrict the weight, you're essentially flying a single engine airplane with power divided into two packages. While you're chance of a failure in the first place is doubled."

[DeltaT]

Vmca -red line
Vyse - blue line

It might be possible to lower the nose build up speed to then be able to climb away...so therein a reason to feather, though if you are at that kinda height to achieve that the aircraft should be above blue line.
If you are so low as to be below below line, I put forward that spending time worrying about feathering the engine is a moot point and you are better closing the throttles and preparing for the crash landing ahead.

From memory in my training it was something like 300ft close the throttles landing ahead, above that and blue line then feather, and when above something like 1500ft continue the drills for basic troubleshooting.
As for knowing the speed specific for the weight, again if memory serves, in a light piston it is only a variation of some 5knts.

[Centaurus]

From memory in my training it was something like 300ft close the throttles landing ahead, above that and blue line then feather, and when above something like 1500ft continue the drills for basic troubleshooting

And therein lies the problem for a candidate undergoing dual instruction on his initial twin engine piston type. Every instructor has his own opinion or parrots what he was taught by his first twin instructor who in turn parrots what he too was taught and so on. Lots of opinions but few facts. We see that for example in the differing instructor opinions on minimum height for an asymmetric go-around; yet rarely is that mentioned in the manufacturer's handbook, AFM or POH. Most POH cover the basics of engine close down but nothing specific about the need to feather before the prop reaches a critically low rpm

The purpose of the opening post was to remind instructors that swifter than normal action is required to feather the propeller of a failed engine that has packed up due to a major mechanical fault. That applies during initial climb after lift off as well as at any other phase of flight.

[Nomde plume]

Centaurus, you are right in principle, a seized prop isn't as good as a feathered one. but how exactly is one supposed to diagnose a 'major mechanical fault' before commencing any immediate action drills, particularly after takeoff?

Why risk botching it up and feathering the wrong engine?

The drills all the way through to feather shouldn't take more than 5 seconds for a well-practiced pilot. And getting the gear and flaps up is just as important as feathering the prop.

Why delay all these items to take the time to consider whether the prop will seize first? You simply can't train someone to knee jerk react with proper checks and drills as well as have them consider the possibility of having to feather immediately if the revs are dropping below min feather speed (somewhere around 700rpm or so I seem to recall from most I have seen). They are conflicting commands. You can either stop and think about it. Or do the checks. I'd rather fly the plane and do the checks immediately. If the prop stops that quick, there probably wasn't much I could do about it anyway.

As for in the cruise, well is a seized prop in fine pitch really that critical? You're going to have to come down anyway!

[megan]

I have an extremely vague memory that back in the old days a check flight involved loading the aircraft to something like 75 or 80 % of gross. I know we used to load sand bags to do just that, but the practice was not continued for what ever reason.

Have once again extremely vague recollection that CASA required loading to some % of gross. Getting somewhat aged so failures of memory are excusable.

Talking to recent twin pilots, all they know is how the aircraft performs when empty on a check flight.

False training?

[jas24zzk]

Nomde,
chances are a major mechanical fault is going to be one that has a hell of a bang, so you probably won't need to confirm much.

The gear up flaps up thing is a bit of a misnomer. Subsequent training/experience (tho not as extensive as many here) has taught me that If you are in the take-off phase and the airframe isn't cleaned up, or your speeds haven't been reached, then you are a single engine...Close both throttles and land ahead....at least you have control and can concentrate on your crash.

If on approach, you are going to continue the approach anyway regardless of alt. The only difference is going to be how much time you commit to getting that prop feathered.

[Nomde plume]

a major fault may or may not manifest itself as spectularly as that. The whole mixture pitch power gear flap thing is basically just confirming everything is where it should be. I've seen many pilots unwittingly leave the gear down til 1000' or takeoff with the mixtures chopped for the taxi out. The key is to properly identify the correct engine then feather in a swift manner. Why you are feathering it has no significance except for when you are safely back on the ground!

If you are on approach feathering is the least of your concerns, just add a bit of power to pick up the slack of the dead one, bit of rudder and the aircraft will not know the difference. Go arounds are another can of worms as Centaurus mentioned.

[43Inches]

Paragraph (a) bears repeating: (a) In the event of an engine failure caused by a major mechanical fault (e.g. seizing bearings due to loss of oil), the rate of deceleration of the engine can be rapid and it is thus imperative that the pilot take immediate action to feather the propeller before the rpm falls to the 1000 rpm region.

In the case of Piper aircraft they usually include the minimum feather RPM in the checklist for feathering, as well as a description of it in systems. It would be very worrying if instructors today are not teaching this, as its vital to understand what the feathering locks are and how they operate, this in additional to how the prop control mechanism works, oil flows etc. I remember it used to be emphasised during the full feather demonstration where we would shut down the engine in flight, fly around for a while single engine to get a feel for it, do a restart, and in ages long gone, shut it down again and land feathered. Even if you are adverse to feathering engines you should do a sequence of simulated single engine flight at zero thrust including level, climbing, turning, climbing turns, to see the performance and how it can be improved.

From memory in my training it was something like 300ft close the throttles landing ahead, above that and blue line then feather, and when above something like 1500ft continue the drills for basic troubleshooting.
As for knowing the speed specific for the weight, again if memory serves, in a light piston it is only a variation of some 5knts.

Years ago we used to practice engine failures at 200ft (some lower than that) after take-off by cutting the mixture, ie you had a real failure, you would then climb away on zero thrust at blue line to circuit height to prove the aircraft can do it. All my training was done this way in everything up to a Navajo, mixture cuts included (not advocating mixture cuts but just that its is an actual failure, if you don't trust the throttle method actually replicates a failure). You should be close to blue line if not above by 50ft in most light twins, if not why not, by 100ft you should be above blue line and climbing away nicely. A PA-44 training is usually close to max weight most of the time anyway. Even most 6 seaters these days are only 4 seaters with max fuel because of all the toys they carry, so 2 up they're only 150kg off max, chuck in a little ballast and you're almost there.

The key here is you have to make a decision once the runway runs out, can you make it, depending on a lot of factors you should consider before take-off. Above blue line, without sufficient runway I would be thinking "go", control, clean it up(drills), assess whether my performance meets my expectations. If performance satisfactory, continue the "go" plan, if things look ugly, use what performance I have to follow the "stop" plan (includes knowledge of likely landing areas, similar to what you would think SE.

This process must be 100% methodical and not rushed, but that does not mean its is a slow process and it should not be delayed in any respect at any point of flight if there is oil issues or anything affecting the prop mechanism or engine that could possibly cause RPM below minimum feathering. On take-off the initial actions should already be complete, everything is just a confirmation 10-20 seconds and you are ready to feather, rushing will only save a few seconds that may lead to big mistakes and in the end will not change the outcome.

If you are on approach feathering is the least of your concerns, just add a bit of power to pick up the slack of the dead one, bit of rudder and the aircraft will not know the difference. Go arounds are another can of worms as Centaurus mentioned.

Where I've seen most cause for concern is engine failures late final, usually requires a good deal of power to overcome the loss of power and windmilling prop combo. If you have time I would always recommend identify and feather the dead, there's always a chance you might not have the performance to make the field from above 200ft on final. There are also a lot of accidents caused by pilots choosing not to feather and having to suddenly add power late final and losing control. It's also important to still follow the drill, without getting too distracted from flying, control, mixture, pitch and power up (to maintain glide path), then you are almost there to identify & feather anyway. In some aircraft in a descending turn at low power you can get the aircraft to yaw towards the live engine, if you follow the drill and power up it will remove any false yaw.

I also remember one pilot who had an engine failure turning final, he was high and fast, elected to do the drills after the turn to final complete and found he'd pulled a mixture back with the throttle (Beechcraft), power restored and now a normal two engine landing resumed.

[MakeItHappenCaptain]

I definitely discuss what the cenrifugal latches (I even have a set to show) do and how to deal with an engine that can't be feathered.
Consider Avro Ansons. There's one in the display case at Tamworth. They had fixed pitch props.:ok:

[Progressive]

This debate a bout how to feather an engine after a major mechanical fault on takeoff is entertaining but misses a key point in feathering propeller design.

All current single action feathering propellers (light a/c) use an "oil pressure to fine pitch" design.

Loss of oil pressure (IE: the major mechanical failure) above minimum feathering RPM will result in the propeller Automatically feathering as oil pressure drops off (the reason low pitch latches are required).

The only conditions that will result in the propeller completely locking are a loss of oil pressure AFTER the engine has decreased below min feathering speed AND the prop has reached fine pitch (An instantaneous ceasing of rotation with the prop already at fine pitch - very rare for an inflight condition).

Or a mechanical failure of the propeller or governor resulting in the inability to drain oil from the prop (in which case no pilot action will make it feather)

So: Do the feather drill carefully and make sure you identify the correct engine. If you have had a mechanical fault chances are the engine will feather itself before you touch the lever anyway.

[Nomde plume]

Do the feather drill carefully and make sure you identify the correct engine. If you have had a mechanical fault chances are the engine will feather itself before you touch the lever anyway
thankyou! exactly what I was getting at!

[Centaurus]

The original post discussed the need for instructors to be aware that a major (severe damage) engine failure in a light piston engine twin requires prompt action to feather the propeller before the rpm falls below a critical figure – nominally between 800 and 1000 rpm. The current mantra taught by the majority of flying schools correctly emphasis the need to correctly identify the defective engine before feathering its propeller.

One of the steps recommended is to slowly close the throttle of the suspected engine to confirm it is indeed the correct one to feather. A slow closure is normally effected since a fast inadvertent closure of the wrong (live) engine by a pilot could leave the aircraft with no power even for a few seconds. The resultant airspeed loss would be rapid.

Realistically, the throttle precautionary closure of the suspected failed engine serves no other purpose at that instant except to confirm that the original assessment of “dead leg – dead side” was correct. Some light twin POH recommend the confirmation by use of throttle while others do not mention the technique. While this technique at a safe altitude and plenty of airspeed is acceptable, it is another thing altogether at a low airspeed and low altitude such as the initial climb after lift off.

It was in the days of four engine aircraft powered with piston engines (e.g DC4 Skymasters , Super Constellations or the military Lancaster bomber), that the confirmation of which engine had failed, came into being. For example, if an outboard engine failed, the strong swing towards the dead engine was usually quite obvious through the universally accepted dead leg – dead side confirmation technique. If an inboard engine failed, the swing was less marked although it was still obvious that an engine had failed on one side.

A problem arose if an outboard engine suffered partial failure. In that case from the swing alone it could be easily mis-identified as a failure of an inboard engine although a glance at the engine power instruments would normally give an indication which engine was the problem. So the technique came into being with four-engine aircraft that once the failed engine side had been identified by the dead leg, dead side technique, a slow throttle closure was effected on the suspect engine to confirm which of the two engines on one side had failed. If there was no further yaw when the assumed dead engine was throttled back, then it was generally safe to assume the dead engine had been further identified.

Conversely if closing of the suspect dead engine produced a further really severe swing then both engines on the same side were now out and the pilot had misidentified which was the dead engine.
Even if the two engines on the same side (one of which was truly dead while the other was merely momentarily throttled back and therefore windmilling) were out of action but for different causes, a four engine aircraft usually had sufficient performance to cope for a short time until power was quickly restored to the live engine. But pull back the wrong engine throttle on a light twin with already marginal climb performance on initial climb after lift off, then things soon get out of hand.

The original post resulted in contributors expressing their opinions on light twin take off performance and that is a good thing. This is where Pprune comes into its own with the number of readers having their say while the majority observe and ponder who to believe. But it still boils down to the need to get a defective engine propeller feathered before airspeed loss becomes potentially fatal due to the insurmountable windmilling drag. Reducing the number of actions before feathering is often the key to a successful single engine climb out performance.

In the type of piston engine light twins under discussion and with an engine failure shortly after lift- off, there will invariably be a "dead man's gap" of about 5-10 seconds where the speed and configuration means the aircraft is in a no-man's land for a few seconds. This is not new. Early military aircraft such as the twin-jet Canberra bomber and Meteor fighter sometimes had a 30 knots or more between lift off speed and minimum control speed. In that situation there are so many variables that it is impossible to consider each one before making a decision to forced land straight ahead or attempt a climb out. Of the variables, pilot skill is one vital factor. That means intimate knowledge of the asymmetric performance of his aircraft.

Lack of published single engine performance information in some POH, means rate of climb on single engine may be difficult to quantify; especially if the pilot has never experienced the doubtful pleasure of a real engine failure at that point in the take off flight path. That is the beauty of simulators.
The gear up, flaps up mantra, can be a bit of a trap. If the aircraft type requires a set take off flap (rather than a flaps up take off), and an engine fails shortly after getting airborne and the pilot whips the flap lever to up as part of the gear up, flap up drills, the loss of lift could be serious and the aircraft could sink back into the ground (problem solved re decision to stop or continue:E

Most pilots would agree that unwanted drag after lift off should be reduced as soon as it is safe to do so. That includes retracting the landing gear once the aircraft has attained a positive rate of climb. That way acceleration towards blue line is quicker. On the other hand, some pilots on initial twin endorsements are advised by their instructors to consider deliberately delaying retraction of landing gear until they guesstimate it is no longer viable to safely land ahead on the remaining runway length (taking into account forward vision over the nose, day or night-time, wet or dry runway, head or tailwind component, pilot reaction time, and maybe no time for the flaps to reach full down if airborne abort).

Rarely are figures published to calculate with any accuracy how much runway is needed for such an airborne abort. Hence the reference to “guesstimate”. Assuming the pilot has been certified competent to operate in command, there is clearly a necessity to making a correct and prompt decision if an engine fails suddenly shortly after lift off in a light piston engine twin. Should he close the throttles and deliberately elect to crash land because he thinks the aircraft doesn't have the performance to climb on one engine? Or does he think he can get away with it as long as he cleans up the drag and that includes prompt feathering of the failed engine?

To summarise: If, immediately after lift off, there is the proverbial loud bang and severe vibration of a badly damaged engine, there should be no need to go through the complete flying school engine failure mantra of the example detailed in the opening post. After all, the mixture levers should already be full rich (or as needed depending on density altitude), the pitch levers should already be at full forward, the failed engine already identified by the subsequent yaw, and no pressing need to confirm which engine has gone by slowly closing its throttle. These are all time wasters at low altitude since the immediate problem is windmilling propeller drag which in some aircraft is greater than the drag caused by an extended landing gear.

The prime identification has already been confirmed by the pilot’s immediate action of preventing further yaw towards the dead engine. Hence dead leg – dead side. Where an engine failure has occurred within seconds after lift off, the pilot may not have the time or airspeed available to afford the luxury of slowly closing a throttle to confirm which engine has failed. The pilot must get it right first time.
Caution: The comments above are personal opinion only. .

[JustJoinedToSearch]

Just for general info: I did my initial multi at a Moorabbin based school in '09ish and at least then, this is what was being taught:

Blue line speed = decision speed so there is no 'in between zone.' Failure below: close both throttles, land. Failure above with sufficient runway: close both throttles, land. Failure above without sufficient runway: go through the checks.

I was also taught to keep the gear down until there was insufficient runway remaining to land.

Seemed sensible enough to me at the time.

[Judd]

Blue line speed = decision speed so there is no 'in between zone.' Failure below: close both throttles, land
All too simplistic and typical of some flying schools. In any case, Blue line speed is nothing more than best rate of climb on single engine with clean aircraft and prop feathered. Another speed quoted in some POH is best angle of climb speed to clear obstacles which still gives you a positive rate of climb although not necessarily as good as Blue Line. There is no published "Decision" speed on light twins.


The so-called "Decision Speed" quoted by some instructors is a personal speed and nothing else. Same as an earlier post regarding "Decision height" for asymmetric go-around. None is published in the manufacturer's AFM/POH so someone in a flying school invents one and lo and behold it catches on and quoted widely as a measured fact. "When ignorance is bliss, 'tis folly to be wise" comes to mind...

[JustJoinedToSearch]

Wasn't actually just made up by one of the instructors, it was in the manual of one of the major Moorabbin schools.

I am fully aware there is no published 'decision speed', but I imagine the reason for using that phrase is to suggest that if you haven't even gotten to Vyse when an engine fails, it's best just to treat it as a single failure and just get it back on the ground.

I would think that fewer would come to grief trying to get a twin landed and stopped before the fence (or go through it if necessary) than trying to get away on one engine before even achieving the best ROC speed. Especially if they are somewhat out of practice on asymmetric flying.

[Tee Emm]

Wasn't actually just made up by one of the instructors, it was in the manual of one of the major Moorabbin schools.


Someone wrote the flying school manual and probably copied some other flying school manual with a few minor variations. Either way, a flying instructor- possibly the CFI or the CFI before him used his position to push his personal preferences (flying techniques of course) which then are published in the company manuals. As a student at a flying school you are more or less are bound to follow the company published procedures even though they are always a matter of personal opinions.
However, that should not prevent you from questioning those procedures if you have a contrary point of view based upon your own experience and study of the subject.

[JustJoinedToSearch]

Yes, I meant it wasn't just my one instructor making it up on his own. It was presumably being taught that way to all the students at the time.

At the place I currently hire from I have questioned a few of their 'local requirements' to the boss, provided evidence as to my position (typically to do with blanket things that don't fit with a specific type/models AFM) and they've had no problem changing things. One of the reasons I go there.

I prefer to get my 'how to' from the manufacturers manuals in general. I presume (hope) they spend plenty of time and money developing the procedures to put in the things.

I mostly made that post to illustrate how things were being taught at a major MB school somewhat recently for the info of those reading to comment on.

I (at my comparably limited experience to many on here) still believe that you're going to have a much better chance looking for somewhere to put it down if you have a failure before getting to Vyse than trying to get everything sorted at a speed which is probably going to give you at best very marginal performance until you can accelerate. Even more appropriate if your asymmetric flying is not quite as current as it should be.

However I'm certainly open to hearing other opinions, your life is going to be far too short if you are going to fly and be obtuse.

[Lookleft]

Great discussion although I always pointed out to students converting onto twins, the only certification requirement was 1%climb gradient at 5000' ( along time ago so the numbers might not be entirely correct). EFATO was best handled by looking for a clear space ahead and controlling it to the point of impact.

A further discussion might be what is actually causing light twins to crash?
CFIT, fuel exhaustion, EF in cruise. Possibly a bit more focus on those areas of piston multi operation might be worthwhile.

[c100driver]

Some flying schools and charter operations still persist with the illusion that light twins can have an engine failure and continue. Maybe yes; maybe no: either way you and your passengers have just become test pilots and observers.

The current 14 CFR Part 23 single-engine climb performance requirements for reciprocating-engine twins are as follows:
• More than 6,000 pounds maximum certificated takeoff weight and/or Vso of more than 61 knots. The single-engine rate of climb in feet per minute at 5,000 mean sea level (MSL) must be equal to at least .027 Vso squared. For twins type-certificated on February 4, 1991, or thereafter, the single-engine climb requirement is expressed in terms of a climb gradient, 1.5 percent.
• 6,000 pounds or less maximum certificated takeoff weight and Vso of 61 knots or less. The single-en- gine rate of climb or climb gradient at 5,000 MSL must simply be determined. The rate of climb could be a negative number. There is no requirement for a positive single-engine rate of climb at 5,000 feet or any other altitude.

Regarding climb performance, the light twin with OEI will perform marginally at best and may not be capable of climbing at all under existing conditions. There is no requirement that a light twin in the takeoff or landing configuration must be able to maintain altitude, even at sea level, with OEI.

And my personal favourite from the FAA!

Landing Gear Selected Up, OEI Climb Performance Inadequate
When operating near or above the single-engine ceiling and an engine failure is experienced shortly after lift-off, a landing must be accomplished on essentially whatever lies ahead. The greatest hazard in an OEI takeoff is attempting to fly when it is not within the performance capability of the airplane to do so.

A recent study revealed a very high success rate for off-airport engine-inoperative landings when the airplane was landed under control. The same study also revealed a very high fatality rate in stall-spin accidents when the pilot attempted flight beyond the performance capability of the airplane.

[DeltaT]

I think it is worth pointing out that many an airline doesn't follow what the manufacturer has in the manual for various things. The airlines Operations Manual gets approved and then that is followed.
Aeroclubs are no different.

[john_tullamarine]

Some purely personal observations ..

All too simplistic and typical of some flying schools.

That may be true however simple rules form the basis of SOP - when the action happens and the adrenaline starts pumping is not the time to be going back to developing strategy. A rule-based protocol may well not be the best strategy but it might offer a reasonably considered option in most situations of interest

In any case, Blue line speed is nothing more than best rate of climb on single engine ...

OEI, the speed range for climb capability is narrow and narrows rapidly as weight and DH increase .. rapidly becoming non-existent. Best climb speed sounds like a pretty important number to me.

Another speed quoted in some POH is best angle of climb speed to clear obstacles which still gives you a positive rate of climb although not necessarily as good as Blue Line.

I probably wouldn't like to put my trust in that philosophy ...

There is no published "Decision" speed on light twins.

Depends on what the OEM chooses to include in the POH.

Same as an earlier post regarding "Decision height" for asymmetric go-around. None is published in the manufacturer's AFM/POH ..

But may be. This aspect of OEI operation is looked at with some attention for military acceptance - unfortunately not routinely in civil practice.

Blue line speed = decision speed so there is no 'in between zone.'

Unless the overrun is a cliff drop off or some other dreadful thing ... sounds like a useful rule-based strategy for light twins. Given the rapidity with which one can lose those few knots .. I preferred to target a speed a little in excess of blue line ..

[Slippery_Pete]

Centaurus, I usually really dig your posts. This is all well and good theory but practically, it's a storm in a tea cup, relating to an impossibly rare occurrence.

Firstly - As someone else noted, a catastrophic mechanical failure is reasonably likely to be caused or accompanied imediately by a complete loss of oil pressure - and the prop will feather before the propeller slows and the pitch locks engage.

Secondly - can someone please tell me the last time an Australian registered twin had an engine failure shortly after takeoff where failure to feather quickly was the cause of loss of control and crash? I can think of NONE in twenty years.

Now compare that stat to light twin accidents from
- Inadvertent IMC
- disorientation at night
- fuel starvation (exhaustion or selection)
- weather related phenomena

This is an interesting system technicality and nothing more.

If you have a light twin and want to polish your skills with an instructor, go and do some night circuits at a black hole aerodrome, practice some UAs in a synthetic trainer, or get an instructor to induce the leans for you (easy to do) and learn to fight it and trust the clocks no matter how bad and disoriented you feel.

Statistically, this training is probably 100 times more likely to keep you alive than learning how to quickly feather an engine without running a 5 second drill first, in the unbelievably unlikely scenario that in the first few hundred feet after takeoff, you suffer a catastrophic engine failure and fast loss of propeller rotation while your engine still develops oil pressure.

You're probably statistically more likely to win lotto. Twice.

[Tee Emm]

Secondly - can someone please tell me the last time an Australian registered twin had an engine failure shortly after takeoff where failure to feather quickly was the cause of loss of control and crash? I can think of NONE in twenty years. The Duchess fatal at Camden during a practice engine failure after take off. the instructor foolishly cut the mixture as the gear was retracting but failed to re-start the engine and set zero thrust. The drag from the windmilling propeller caused drastic airspeed loss and the aircraft descended into the ground after clipping trees. . Failure to feather quickly (aka set zero thrust which for simulation is the same thing) was the cause of loss of control.
Whether it is a VH rego or not has nothing to do with the subject.

The Dove crash at Essendon. Engine failure shortly after lift off. The pilot was so engrossed in attempting to first retract the gear and flaps that he never got around to feathering the failed engine.

The Cessna 404 (?) crash at Essendon after taking off from runway 35. Total or partial engine failure shortly after becoming airborne. The pilot was so engrossed in talking to and replying to ATC about the engine failure that he never got around to feathering the propeller and due to windmilling drag got below Vmca and flicked inverted and went in.

Cessna 404 crash in Scotland. Engine failure after take off beyond Blue Line speed at the time. Pilot did not feather the propeller and got below Vmca and went in killing all 8 aboard.

I think these accidents illustrate the importance of prompt feathering following engine failure.

relating to an impossibly rare occurrence. Read above fatal accidents for "impossibly rare occurrence".

[Oktas8]

Tee Emm, I think those accidents illustrate failure to prioritise rather than mechanical failure leading to inability to feather. All except for the Camden accident, which was instructor stupidity.

Whilst you have answered slippery_pete's question, I think he was actually referring to Centaurus' scenario of catastrophic bearing failure leading to a seized propeller. Like you, I'm not aware of any scenarios like that in recent years.

I did once suffer a mechanical malfunction in a light twin, causing cycling of propeller RPM without any change in engine power. I feathered the propeller at the high point of RPM without going through the full drill, for the reasons Centaurus explained. But that was in the cruise, at low workload. I devoted my full & undivided attention to which propeller was malfunctioning and which propeller control I was touching...

[john_tullamarine]

Perhaps Centaurus' concern also is with those folk who take far too long generally in doing the identification exercise regardless of the nature of the failure ?

I had a simple procedure which worked fine on light twins and the identification and shutdown could be actioned in a matter of a few seconds. I used to go fly some practice time prior to renewals with whomever might have been available to be safety pilot .. on a number of occasions the other chap commented on the short time taken to identify and go through the simulated actions.

On the other hand, I have observed a number of GA pilots (including instructors) taking far too long .. fine at cruise but a recipe for mishap whilse manoeuvring at lower levels ?

[Centaurus]

Perhaps Centaurus' concern also is with those folk who take far too long generally in doing the identification exercise regardless of the nature of the failure ?

Thanks JT. You got it in one :ok:

[Oktas8]

Paragraph (a) bears repeating: (a) In the event of an engine failure caused by a major mechanical fault (e.g. seizing bearings due to loss of oil), the rate of deceleration of the engine can be rapid and it is thus imperative that the pilot take immediate action to feather the propeller before the rpm falls to the 1000 rpm region.

To be quite blunt, I don't see the evidence of this failure mode being a major problem.

Human factors issues though... I can readily accept a need for improvement.

Training for good, quick, emergency response to a critical EF is difficult. I had a good ME instructor who once (near end of training) gave me an unbriefed EF during a radio call. I fixated very nicely on the radio call and learned a good lesson from the result.

Not sure how to fix that problem though, except the same way it was fixed with me. Better instruction. :)

[john_tullamarine]

I had a good ME instructor who once (near end of training) gave me an unbriefed EF during a radio call. I fixated very nicely on the radio call and learned a good lesson from the result.

Likewise, shortly after check to the line on my first airline Type, the line captain - a positioning flight with only the two of us on board - pulled an engine on me during the rotation - night but visual.

Bit different when one isn't expecting it, is it not ?

Not terribly good form on his part in the general way of things but the exercise certainly made a good point regarding prioritisation and command thinking - stayed with me thereafter. (Fortunately, I did the appropriate things and it went well.)

Not sure how to fix that problem though, except the same way it was fixed with me. Better instruction

I suggest training and repetitious exposure in a safe environment .. and then a bit more of the same.

How many times, in our early Link training, did the instructor come to the hatch with a coffee (or other innovative distraction) .. just in time to see us go through the level, radial, or whatever ? Eventually, we all got the message - some quicker than others.

The FFS offers a wonderful tool for this sort of conditioning work.

[A Squared]

Read above fatal accidents for "impossibly rare occurrence".

You're taking that comment out of context. That was referring specifically to an engine failure of such a nature that oil pressure would be lost, but the propeller would *not* feather. Given the mechanics of a light twin feathering system, *that* specific event has a very low probability.

[A Squared]

I think he was actually referring to Centaurus' scenario of catastrophic bearing failure leading to a seized propeller. Like you, I'm not aware of any scenarios like that in recent years.

Additionaly, my impression is that a stopped, but not feathered propeller is not a huge drag source, obviously feathered is best, but the *big* thing is to get the windmilling stopped, as a windmilling prop creates a huge amount of drag, far out of proportion to the form drag of the prop blades themselves

Anyone have actual test data comparing the drag of a feathered prop to a stopped but not feathered prop?

[MakeItHappenCaptain]

Quote:
Secondly - can someone please tell me the last time an Australian registered twin had an engine failure shortly after takeoff where failure to feather quickly was the cause of loss of control and crash? I can think of NONE in twenty years.


PA-31-350 went in off East Point in Darwin on 6 Feb 2009. Seem to remember some conjecture on this site as to the contributing circumstances.:confused:

Another PA-31 IHR was buried just short of Mt Isa on 17 Jul '08 with no fuel in auxies, but full mains. Didn't follow all failure drills, but granted, not an EFATO event.:rolleyes:

[Slippery_Pete]

I just wrote a big long winded reply, but *($#& iOS8 on my iPad dumped the lot as usual.

Tee Emm... Oktas8 and others are completely correct in that you have missed the point of my post.

I wasn't referring to EFATO, I was referring to EFATO where shortly after becoming airborne the engine seizes quickly, but still generates sufficient oil pressure, resulting in pitch lock and inability to feather. NONE of your scenarios demonstrated this.

Whether it is a VH rego or not has nothing to do with the subject.

It does. I was making a point of comparing pitch lock = failure to feather scenarios with other accidents in Australia to make a point about the statistical probability. If you want to consider international feathering accidents that's fine, but you'd need to compare this to international inadvertent IMC/fuel exhaustion accidents etc.

Incident #1 - Just piss poor instructing (and perhaps poor CFI oversight and mentoring to instructors). Failure to feather due to delay causing pitch lock, nope. Just failure to realise that mixture cuts at low altitude are idiotic.

Incident #2 - The manufacturer data indicated it may have been possible to keep the aircraft flying in this scenario had everything worked as it should have. And I quote from the ATSB " However, when the landing gear failed to retract on the first attempt, any possibility of the pilot being able to attain the required aircraft performance was lost."
Had the pilot ignored the gear issue and left it down and then promptly feathered the engine, it still would have crashed. BOTH of those things - gear up and feathering - needed to occur.
No mention of rapid prop speed loss and pitch lock engagement causing inability to feather.

Incident #3 - Just a complete loss of prioritisation as alluded to by others. Didn't matter if he feathered it fast, or after a 5-10 second drill. He just never feathered it at all. Game over.

Incident #4 - I'm not terribly familiar with this, but then it seems pretty straight forward. A failure above blue line should have provided ample time to complete an identification drill and then feather it. The pilot didn't feather it AT ANY STAGE.

illustrate the importance of prompt feathering following engine failure.

I never argued against feathering an engine efficiently and without unnecessary delay.

I simply stated that an EFATO, with a rapid engine seize, where the engine is developing oil pressure, with a short delay running an identification drill, leading to pitch lock, and subsequent inability to feather, and resulting loss of control - as an impossibly rare occurrence.

None of the incidents you linked to show evidence of this.

And in fact, the C404 incident in Scotland is in fact a very good example against your case. The INABILITY TO CLIMB was caused by the failure to feather, but the aircraft was controllable. What happened then, was the incorrect engine was feathered leading to loss of control and fatal crash. Perhaps the incorrect engine was feathered because the pilot rushed the shutdown through fear of the dreaded "pitch-lock" :}

History is scattered with rushed shutdowns of the operating engine. It is not scattered with pitch lock induced inability to feather.

Anyone have actual test data comparing the drag of a feathered prop to a stopped but not feathered prop?

The turboprops I used to fly had this data. From memory, the resultant drag from a windmilling prop was relative to an increase in aircraft weight of about 30%. So if you departed at a 20,000kg MTOW for example, and suffered an engine failure and the propeller was allowed to windmill, with MCP set on the live engine, the aircraft performed as if it was 26,000kg... ie 6000kg above MTOW, and well outside any guaranteed single engine performance.

EDIT: Sorry, missed your point on this. No, haven't seen this data before. I remember reading an aerodynamics text book at uni saying windmilling drag was equivalent to a solid disc the same diameter as the prop, but can't remember which text it was. You are right however, a completely stopped prop in fine pitch would have a mere fraction of the drag of a windmilling one.

Perhaps Centaurus' concern also is with those folk who take far too long generally in doing the identification exercise regardless of the nature of the failure ?

Thanks JT. You got it in one :ok:

Well that's fair enough Centaurus, but don't blame "pitch lock". And regardless, I still think there'd be at least as much (if not more) evidence of incorrect shutdowns causing fatals than "feathering too slow" causing fatals - but I admit I don't really know and that's just an educated guess.

If you have a light twin and want to polish your skills with an instructor, go and do some night circuits at a black hole aerodrome, practice some UAs in a synthetic trainer, or get an instructor to induce the leans for you.

I stand by this comment, and the data supports it.