Windmilling or non-feathered propellers in flight.

[GyvAir]

There has been quite a bit of discussion recently about the effects of having an inoperative engine and a prop that fails to feather. Recent threads touching (speculatively at least) on the subject include Bearskin in Red Lake, Buffalo’s DC3 in Yellowknife, the PA 31 in Luton, UK and NT KA 100 in Vancouver (engine idled), to name a few.
I see a lot of questions, misinformation and skewed perceptions on the effects of a windmilling prop and also on what takes place in different powerplant/prop control systems to ensure that the propeller indeed feathers when it should.
Other questions I see and hear regularly on the topic:
Is it better to let a piston engine prop windmill, or to try to slow down enough to allow it to stop?
How much drag does a windmilling prop induce? Is it really more than the drag that would be induced by an equivalent diameter solid disc bolted to the leading edge of the wing?

My question right now is, does someone have a reference or link to a good layman’s terms explanation of the aerodynamics involved with a prop windmilling, versus stopped in fine, versus stopped in feather?

[GyvAir]

Interesting article on the the risks involved in attempting to restart an engine after feathering it:

Windmilling Warning

[photofly]

Rate of work done on the blades by the air is the blade drag x blade velocity through the air - that's the rate at which energy is extracted from the descending aircraft.

If the blades are stationary the work done is the airspeed of the aircraft x the extra drag of the blades.

If the blades are turning the work done is the airspeed over the blades, now much higher, times some different drag coefficient.

It's hard to imagine the drag coefficient is hugely different (although it will be different, because the angle of attack over the blades is different) but the blade airspeed is much higher when the prop is turning. So more energy is extracted and the glide is less efficient for it.

[PositiveRate27]

I don't have any info to give on the exact amount of drag a feathered prop makes vs a windmilling prop. All I can comment on is personal experience when bringing a functioning turbine engine to idle. The amount of drag (disking as some guys like to call it) when you bring a turbine to flight idle and the props go fine is quite noticeable. In the case of NT, I would be willing to bet the engine at idle with the prop full fine would produce more drag than had the engine been shut down and feathered. This would have skewed their single engine Vref speed, hence the unexpected trip below Vmc that was unrecoverable. From what I gather, and my own experience with my company's SOP's, there were no published numbers for Vref with an engine simply at fight idle, only failed and feathered. Pretty poor time to be unsuspecting test pilots unfortunately.

[Flyboy757]

I had a prop fail to feather due to a TOTAL and immediate electrical failure on an RCAF Grumman Tracker. We ended up ditching in Whitewater Lk near Sudbury, Ont.
After the fact and when the findings were published they determined that the Vref, Vmca, etc numbers were for a FEATHERED prop and in our case non applicable.
During the attempt to fly the "beast" to YSB at the "published " numbers we had full power on the good engine and full rudder to try and maintain directional control and altitude. It didn`t work!!!!

[PilotDAR]

For the common GA piston twins, if you do not get the prop feathered before its rotational speed slows below about 800 RPM, it will not be possible to feather it. This warning appears in some flight manuals, but not all. I have insisted that be added in some flight manuals I have influenced, and FMS's I have written. Therefore, yes, if you have an engine seizure on these aircraft, you will be stuck with a flat pitch, stopped prop. I'm told this is more drag on most twins than the gear and flaps extended combined (though I have no quantitative proof of this).

When I have had to glide C/S singles, I have selected full coarse, which seemed to slightly reduce the rate of descent.

Most PT-6's left to themselves, will feather after shutdown, if you do nothing, but it takes minutes, not seconds. I did once on a Twin Otter, just out of curiosity.

If you are gliding a Caravan, all nicely trimmed, and at the correct speed, and you then feather the prop, it's like a giant push from behind, and your rate of descent reduces by nearly half.

[iflyforpie]

I'm kind of curious about those who were lucky enough to fly planes like the Airspeed Oxford, Avro Anson, and De Havilland Drover. Did those planes have propeller brakes or were you stuck with a windmilling prop?

I remember reading Barry Schiff's article on the terrible Champion Lancer. According to the POH, upon engine failure you were to shut down the remaining engine and look for a place to set down. He said that you could get it to fly on one if you brought the remaining engine to idle, slowed the aircraft below Vmc until the dead engine stopped windmilling, then accelerate above Vmc and giver hell on the good engine.

[oldtimer]

The G159 Gulfstream 1 had published Vmc with an engine failure during a dry take-off with prop feathered and rudder servo and prop feathered but without rudder servo and the same for a wet take-off with the same situations. It was so long ago that I cannot remember the actual numbers but the difference between a windmilling prop and a feathered prop were quite significant. Something in the neighborhood of 100 knots with prop feathered and rudder servo to 148 knots without feather and rudder servo. The Dowty prop driven by a RR Dart engine was just an older design oil prop optimized for the Dart engine so it required engine oil pressure for both fine pitch and feather. If the prop did not feather, there was an electric feathering pump to supply oil pressure to completely feather the prop. Nowadays, modern props are designed to be more fail safe in that the lack of oil pressure will result in the prop achieving the feathered position. What the pilot does when he/she moves the prop control to the feather position is to simply dump all oil pressure out of the prop and then a combimation of blade twist, counterweights, Nitrogen pressure and or feathering springs push the prop to the feathered position. The problem that can sometime come up is where locks designed to prevent feathering during normal shutdown will sometimes get in the way of an emergency shutdown. That, plus the pilots reluctance to commit to an engine shutdown can in some situations, lead to problems.
This is at least my opinion derived from reading accident reports.
Now with the Metro, a complete in flight intentional engine shutdown is required after an engine or propeller change to verify 3 things, will the Negative Torque Sensing system work as required, once feathered will the prop stop rotating and finally, will the prop unfeathering pump bring the prop out of feather enought to begin windmilling to accomplish an airstart.
It has always been my opinion that a propeller has to feather on a twin not only to reduce drag sufficient to allow continued flight but also to stop rotation because one golden rule with any turbine engine is that if it starts to vibrate, shut it down and stop rotation before the engine shuts its self down in a less than pretty fashion.
I hope this all makes sense.
It is interesting to see the picture of the Champion Lancer. I never had the courage to fly one but I had a quick peek at the AFM and in the perfomance section, there were speeds for best rate of climb at various weights and emergency procedures for an engine failure but a disclaimer at the end saying all take-off speeds were below Vmc and at the best rate of climb speed, all performance numbers were negative values. The guy who owned the Lancer wanted to run a school out of Fernie but TC inspectors would not fly in the airplane so he crashed it into a mountain, fortunatly with only minor injuries. A real piece of shit airplane.

[iflyforpie]

My CP was DFTE for that guy's multi ride after TC refused to get into the aircraft.

[Doc]

After the fact and when the findings were published they determined that the Vref, Vmca, etc numbers were for a FEATHERED prop and in our case non applicable. !

And this came as "news" to them??? Christ mate, who did your training?? You're lucky to be alive, going anywhere near VMC with one NOT feathered! Seriously, you should have been on your back WAY above VMC, if you had the running engine anywhere near max power!

BTW, I have some Tracker time....really like the airplane!

[Flyboy757]

After the fact and when the findings were published they determined that the Vref, Vmca, etc numbers were for a FEATHERED prop and in our case non applicable. !

And this came as "news" to them??? Christ mate, who did your training?? You're lucky to be alive, going anywhere near VMC with one NOT feathered! Seriously, you should have been on your back WAY above VMC, if you had the running engine anywhere near max power!

BTW, I have some Tracker time....really like the airplane!

Yea Doc it was a handful for sure! Full power, rudder and aileron and we were flying sideways!!! The OTU (406 sqn) tried to duplicate our problem by flying at 10,000 ft then introducing our problem. The plane flicked over on her back. We, for sure, were lucky. During the investigation nobody had ever considered a total, battery included, electrical failure....RCAF, USN, Grumman..... Everything was "electrically actuated". I flew with and was trained by some very long in the tooth Tracker guys, they had never even heard of let alone considered such a scenario.
Live and learn.

[iflyforpie]

Weren't the RCAF Trackers shorter than the standard ones so they could fit on the Bonnie? I would imagine that would make things even hairier...

[AOW]

WRT the King Air A100 (I have a few hours in that one), especially with the Raisbeck props, idle power is producing considerably more drag than a feathered prop. Someone suggested it was somewhere in the neighbourhood of negative 300 ft.lbs of torque (-17% Tq). For training purposes, we set 400 ft.lbs (23% Tq) to simulate zero thrust on the simulated feathered engine. Even in that configuration, with a full load and flaps and gear extended, you cannot climb with full power on the good engine. You need to clean up some drag, or else hope to do no better than maintain altitude. With one at idle, below Vref, landing configuration, they had no hope of recovery with just the good engine. What still baffles me is the fact that they didn't just add power on the idle engine - otherwise what was the point of leaving it running?

Anyways, this isn't the place for that discussion...

[Flyboy757]

Weren't the RCAF Trackers shorter than the standard ones so they could fit on the Bonnie? I would imagine that would make things even hairier...

I don`t know if they were shorter, don`t recall that, but they were about 4000lbs lighter after the" Bonnie" was turned into "razor blades"! They pulled all the ASW and MAD gear out and wiring plus the boom and tail hook.
I think they were around 26,500lbs after the stuff pulled out. Don`t hold me to that # as it was Dec 1978 when I last flew them. They certainly had lots of power for the lighter weight. I recall sim sessions where on T/O as the gear was retracting that you had to pull power back on the good engine to stop from rolling over.
During my emergency, without the electrics we had no rudder boost which was also part of the Vmca, Vref criteria.
Yup, we were lucky. Wasn`t our time.

[iflyforpie]

I found it.

In 1954, de Havilland Canada entered into a contract to build Trackers under license to replace the outmoded TBM-3E Avengers being used by the Royal Canadian Navy. A total of 99 Canadian-built Trackers entered service, starting in 1956. From 1957 onwards, these aircraft operated from the newly deployed aircraft carrier HMCS Bonaventure and various shore bases. All the Canadian Trackers were built to the earlier "A" model airframe design with a length of 42 feet (12.80 m) [11] (c.f. 43' 6" for later model Trackers) in order to fit in the Bonnie's hangar.

http://en.wikipedia.org/wiki/Grumman_S- ... al_history

Don`t hold me to that # as it was Dec 1978 when I last flew them.

Don't worry, I won't. I was born that month so I don't remember much either.

[Old Dog Flying]

I worked in Downsview Tower when the first CS2F-1 flew in Oct 1956; s/n 1502

Barney'

[Siddley Hawker]

Great model Barney.
(Thanks for helping that guy out on Swanny's. It was me that sicced him onto you.)

Oldtimer did you ever do that drift down procedure with the G-string with one prop unable to feather? I seem to recall 136kt was Vmca with the rudder servo operable and the airplane would drift down at a couple hundred feet per minute. You got one shot at landing.

[Ki-ll]

How much drag does a windmilling prop induce? Is it really more than the drag that would be induced by an equivalent diameter solid disc bolted to the leading edge of the wing?

My question right now is, does someone have a reference or link to a good layman’s terms explanation of the aerodynamics involved with a prop windmilling, versus stopped in fine, versus stopped in feather?

I don't know a source of good information in English, however I do have some materials in foreign language that might help. The following two diagrams should hopefully clear up some aerodynamics of the windmilling vs. stopped propellers.
Here is a chart which compares the windmilling and stopped propeller drag against the bade angle:

The vertical axis is Coefficient of Drag, the horizontal axis is Blade Angle. The solid line represents a windmilling propeller, the dashed one is a stopped propeller. The diagram represents a three-bladed propeller at an altitude of 3000 feet and speed of 120 knots.
The second graph represents the relationship between the Coefficient of Drag and Airspeed for different states of a propeller. The vertical axis is the ratio of propeller drag for given propeller vs. the same propeller stopped with blade angle of 17 degrees, the horizontal axis is airspeed. Dashed lines are stopped propellers and solid lines are windmilling propellers. φ is the blade angle.


The answer to the second question

How much drag does a windmilling prop induce?

will depend on a lot of things. Before getting to that here is a diagram of the different modes of operation that a propeller can have:

I know that looks a bit overwhelming, but it is in fact fairly simple. The propeller is basically a wing which spins, so the ame aerodynamics apply. The “Lift” which propeller produces is basically the thrust, and the “Drag” it produces is what is being overcome by the torque applied by the engine on the propeller. Just like with a wing both of these forces (propeller’s Lift and Drag) depend on the direction and magnitude of relative airflow (essentially angle of attack).
The vertical axis on the graph represents Power Required to spin a propeller, created Propellers Thrust and Propeller Efficiency. The horizontal axis is airspeed. The blade angle of this propeller is fixed at a certain value, so the only thing which will be varied is airspeed (effectively changing the angle of attack of the blade).
n is rpm
φ is the blade angle
N is Power Required
P is Thrust Created
η is Efficiency
V is airspeed
The top five drawings represent the angle of attack diagrams for different parts of the graph.
On these diagrams:
P is Thrust
Q is Drag
W is the total airflow angle, or essentially angle of attack
α is angle of attack
φ is blade angle
u is rotational velocity of the propeller blade

The first diagram corresponds to graphs between points 1 and 3, this is the normal mode of operation, or propeller mode. The Thrust is positive in this range of angles of attack and the Power Required to spin the propeller is also positive, meaning it is required to come from the engine. This is the area where the angle of attack of the blade is positive.
The second diagram represents point 3 on the graph, where the Thrust is zero but the propeller still requires some Power to be spun at desired rpm. This is a zero-lift angle of attack of the blade.
The third diagram represents a range between points 3 and 4 where thrust is negative, but the Power Required to spin the propeller is still positive and comes from the engine. In this range the propeller starts working as an air break while still using engine a a source of power. The angles of attack of the blade in this case are slightly negative
The fourth diagram represents point 4 on the graphs, at this point the Thrust is negative but the Power Required to spin this propeller is zero, means it spins on it own regardless of what the engine is doing. This is the autorotation mode. Angle of attack of the blade has reached such a negative value that no “Drag” is being produced by the propeller.
The fifth diagram is all the rest of the graphs beyond point 4, here the propeller thrust is VERY negative and the propeller is actually spinning the engine by extracting the required power from the airflow. The angles of attack of the blade are very negative.
What is important to remember is that all these modes pertain not to particular speeds, but to particular angles of attack, just like on a wing. If an angle of attack of a wing is very negative, it will still create lift, just in different direction.

The most important condition which determined how much drag is produced by a windmilling propeller is what this propeller is attached to? Is it a piston engine, free turbine engine or a geared one?
The most drag can be created by the propeller in combination with a geared turboprop engine and here is why.
In any turboprop engine a compressor and accessories use about 70% of the total energy produced, in a nutshell it means that if an engine is producing 1000hp, 700 of it is used to drive the compressor and accessories. Only 300hp is transferred into useful work by the propeller.
So if on take off a geared engine is producing 1000hp, 700hp is used to drive the engine at required rpm and 300hp is used to overcome propeller’s “Drag” to produce useful thrust. If for any reason an engien flames out, then compressor and accessories still want to recieve their 700hp from somewhere and keep spinning at whatever rpm they were being spun and the only place they can get that from is the propeller. And since the propeller is also mechanically connected to the compressor the governor will reduce the blade angle to the minimum possible value to deliver the 700hp which is being asked from it. If there are no stops incorporated into the design the propeller will go right to a low pitch stop while being spun at 100%rpm and it will do it really-really fast. So if you can imagine a spinning propeller at low pitch (if you can’t look at the very first diagram) will create A LOT of drag if it is not stopped. I would imagine that this A LOT will be bigger than a flat disk of the same diameter, since a windmilling propeller is a much more effective producer of drag than a flat disk. This A LOT can be visualized on the following diagram.

Point 1 on this graph represents maximum thrust an engine will deliver at 300km/h, point 5 represents the maximum drag this engine can create if it is allowed to windmill at minimum angle. So as you can see both of those values are equal to 3 tons, so if an engine quits on take off the propeller will create as much drag as it will create thrust at maximum fuel flow. That’s very significant!
This large amount of drag is tansient in nature, in the literature they call it dynamic negative thrust. Its value will decrease fairly quickly after it peaks at a large negative value, the reason for that is that even at its minimum angle the propeller cannot deliver 700hp which are being asked from it, so the rpm of the engine will decrease and so will the drag. To illustrate:

The top diagram is RPM vs. time, the middle one is blade angle vs. time and the last one is Ratio of Thust created to maximum continuous hp that can be produced by the engine vs. time.
The solid line is the propeller which is able to be stopped at an intermediate angle (about 12-20 degrees) and the dashed line is the propeller which goes right to a low pitch stop.
As you can see the maximum drag is created 3-4 seconds after flameout and if a propeller can be stopped at an intermediate angle it creates A LOT less drag (almost 50% less). Overtime however it is more benefitial to get the propeller to the low pitch stop (that is if you cannot feather it which of course is the preferred course of action), since it will produce less drag windmilling at a lower angle. This seems counterintuitive, since the first graph kind of does not support it, but if you remember that the total Drag produced depends no only on a coefficient of drag but also on airspeed, then it all makes sense since a propeller at a low pitch stop windmills at a much lower rpm than the one at an intermediate angle.
And this is all the reason why geared engines have an intermediate pitch stop, even the Garret 331 engines do.
All of these graphs for geared engines represent a propeller which is different than the design currently in use. On the contemporary propellers loss of oil pressure drives them toawrds feathering (counterweights, feathering springs, nitrogen pressure etc.), this has not always been the case as one of the posters mentioned here. On old engine/propeller combinations the propeller’s pitch angle would be changed by oil pressure in both directions – low and high, which makes this negative thrust issue even more acute.
How does this relate to piston and free turbine engines?
The power required to drive a stopped piston engine at maximum rpm is only 10-15% of the maximum power developed by the engine. This is because exhaust strokes in some cylinders are happening together with compression in others, so the engine helps itself. This is different from a turbine engine where the compression is happening constantly. So the drag created by a windmilling piston engine will not be as large as the one on the geared turboprop.
The free turbine turboprop engine would probably be the best of all of the three (of comparable horse power). This is because after a flameout there is virtually nothing holding that propeller, since it does not drive anything but the gearbox and some accessories. So while it will still windmill it cannot create the same drag as a geared engine.
I hope I did not go into too much detail here... There is more to it, but these are the basics.
And of course the handling of the airplane will be determined by some other factors such as size and power of the controls, C of G etc...

[GyvAir]

Thanks for the posts everyone. Some eye opening stories! Kk-ll: Wow.. that's a lot of translation and explanation effort! Thanks! Haven't absorbed it all yet. Will have another go after a night's sleep and with a larger screen.