View Full Version : Double engine failure turnback

Wee Weasley Welshman
29th Apr 2002, 11:08
Whilst brushing my teeth this morning I started to ponder the feasibility of a 'turnback' on a B733.

What height would be required for such a manouevre to be successful? Indeed, is it even possible assuming typical weights and weather?


29th Apr 2002, 11:28
For most situations couldn't you just turn and land back on the runway you've just taken off from, only in the reverse direction? The worse the tail wind the more you get blown back to the airfield in the turn which is a help. Offhand I would say the optimum glide sink rate is lower than the normal climb rate so you just need enough height to give you the turn. Otherwise you always make a profit.

Wee Weasley Welshman
29th Apr 2002, 11:35
I was thinking more of finding the MINIMUM height required for a turnback to land on the reciprocal to the take-off runway.


Max Angle
29th Apr 2002, 11:41
Try using Microsoft flightsim, the 737 model seems quite good and it would give you a pretty good idea I suspect, your problem will be the radius of the turn required change direction, you are going to end up a long way off the centre line. Good luck, I'm going to give it a go now.

29th Apr 2002, 11:58
Only try this at home folks... :eek:

Wee Weasley Welshman
29th Apr 2002, 12:00
I don't have a flightsim. Cheers,


29th Apr 2002, 14:24
It's not simply a question of height. It is also a question of how quickly that height was obtained.

Far better to land in a farmers field with the wings level than cartwheel over the airport fence.

At a guess, I would say that 1000ft ft per nm from the end of the runway with a minimum of 2000ft and 2nm in no wind could be done.

The minimum would see the aircraft align with the centerline as it hit the runway 300m in.

Must try it. :)


29th Apr 2002, 19:58
Interesting question WWW. Next sim-check Ill play around and see what happens. Your assuming a 2-eng exercise holding max power at V2 flap 5? Or a normal everyday T.O.?

If its the latter:

Any particular TO flap setting? 5 or 15? We arent alowed flap 1 for T.O. and the sim wont accept it either.

3rd segment? 800 ft ok?

Weight? MTOW or MLW or a mid-range weight of say 56 tonnes?

Wind? Say a typical 10kt head?

Power? Max or reduced?

Runway length? A 4000 foot cricket-pitch or a 13,123 foot freeway? I assume a full-length T.O. run?

29th Apr 2002, 20:51
WWW - is there something we need to know about Go-Fly......

Don't know much about the little 73, but in man-sized aeroplanes you also lose a considerable number of hydraulic and other systems with total engine failure - it's outside the realms of consideration unless on an ETOPS twin you have a massive fuel leak and an associated single engine failure on that side and then elect to run the 'fuel imbalance' checklist and lose all your remaining fuel as a result....

Once took off from Cyprus and lost 2 engines; whilst completing the drills another failed at around 2500ft . Set full power on the last and flew a 250 kt descent until we were on the reciprocal glidepath, then dropped the gear and reduced to around 210 until sure of getting in, selected flaps and slats t/o and out and kept full power until we landed in that configuration. In the simulator, of course!

GearUp CheerUp
29th Apr 2002, 20:59
WWW - why not try it in the sim, maybe from the power cutback (say) 1000' AAL and see how you get on.

BEagle - you say its outside the realms of consideration but how about the SAS MD80 (or whatever it was) that suffered a double engine failiure due to ice which had formed over the cold soaked fuel tanks and then broke off and entered the rear mounted engines. Happened about 10 years ago and they got id down reasonably OK in afield (straight ahead but the problem occured at low height straight after takeoff)


29th Apr 2002, 22:00
> how about the SAS MD80 (or whatever it was) that suffered a double engine failiure due to ice which had formed over the cold soaked fuel tanks and then broke off and entered the rear mounted engines. Happened about 10 years ago and they got id down reasonably OK in afield (straight ahead but the problem occured at low height straight after takeoff)


It wasn't ice formed over cold soaked fuel. It was the first flight of the day after snowing overnight. All aircraft would have had the same ice that day.

Wee Weasley Welshman
1st May 2002, 08:40
Lets say Full power take off 49,000kgs 10kt headwind, flap 5 T/O Boeing speeds, ISA VMC conditions.

I don't have a sim for a while. Its purely a theoretical exercise - obviously a controlled straight ahead 'landing' is much more likely to be the better option.

Back of a fag packet calcs suggest to me 2,200ft...


1st May 2002, 10:37
... don't know about height, but as far as getting back on the centerline I would suggest an 80/260 PT.;)

1st May 2002, 13:45
Are there any aeronautical experts out there that can tell us what the ideal bank angle would be??? Something that gives the best rate of turn / rate of descent ratio I suppose.

My gut feeling says 45 degrees bank angle. Don't ask me why.

If you were doing say 210kts and you just tried to do a 180 degree turn back, the radius of turn would be about 1.4 nm. So if say you were 3 nm up wind, you'd end up about 45 degrees off the runway centre line.

So to track back the the upwind end of the runway you would have to turn 225 degrees and then try and judge a 45 degree turn right at the last moment to track down the centreline. A total heading change of 270 degrees.

If you did a 80/260 turn you'd have to turn a total of 340 degrees. You'd be on the centreline for sure but I just wonder if the extra heading change might be the difference between making it back or not. Depends on how much altitude you've got to play with of course.

Great topic.

Oops. :rolleyes: That radius of turn assumes 25 degrees bank angle. At 45 degrees bank angle the radius is 0.63nm. Therefore at say 3nm upwind that would put you 23 degrees off the centreline. The total change in heading would be 226 degrees (180+23+23).

I think too that it would have to be a situation where you had a sudden double engine failure after climbing away on two engines. If it was the case where you had an engine failure early on in the piece say, soon after v1 and then you had a second failure after climbing at around 2.4 to 3 percent climb gradient, you wouldn't have a hope of turning back would you? Climb angle of 3 percent plus a glide angle of at least 5 percent, means that unless you find that nice open field in front of you, there's an exellent chance you'll scratch the paint.:D

1st May 2002, 13:52
Did the same in sim of a 737ng, but it was engine sepration, at 1700ft, turned back and did a belly landing, almost on the runway.

Wee Weasley Welshman
1st May 2002, 16:57
Double engine seperations - wow - that is a bad day. I thought maybe a flock of geese...

Good point on turn radius and optimum angle of bank.

At some airports that have a 180 degree Emergency Turn procedure I guess a landing back on would be a viable option if you encountered the mythical formation of kamikazee Geese.

Needs someone with more perf knowledge than myself...


2nd May 2002, 05:38
WWW you want 49T? Thats 700 kg under 733 MZFW. How about 55T as a more realistic fugure? Sure your on 300s and not 500s mate?

PS Ill try mucking around in the sim and come out with a figure that gives you a heading rollout on final at 50ft alt at Vref flap 5 for 55T (168 kt).

2nd May 2002, 09:06
You could probably work it out from the Performance Charts. Go on, WWW, when was the last time you worked it all out from scratch? :D

I have the following in my collection. Admittedly it is for a light single, but many of the principles remain the same.

In May 1992 at Tumbler Ridge, B.C., a Piper Cherokee aircraft with five passengers crashed following an attempt by the pilot to return to the runway after experiencing an engine failure. The aircraft was only 100 feet above the ground when the engine problem occurred. The aircraft stalled during the attempted turn, causing loss of control at an altitude from which recovery was impossible.

This was not an isolated incident. During a 10-year period from 1982 to 1991, there were 176 accidents resulting from engine failure after take-off in single-engine aircraft. In about half of these, the pilot tried to turn back to the departure runway instead of landing straight ahead. In most of these accidents, the pilot lost aircraft control while attempting the impossible return to the runway.

An analysis of these accidents showed that an aircraft crash caused by loss of control as a result of excessive manoeuvring is 10 times more likely to cause fatalities, and five times more likely to cause serious injuries than if the pilot had elected to land straight ahead. Lower groundspeed associated with a straight-ahead into-wind forced landing, as well as being under control prior to impact with the terrain, reduces the risk. Surprisingly, the data also revealed that experienced pilots are just as likely as novices to attempt the impossible.

Using data from the Cessna 172 Aircraft Flight Manual, our test pilots crunched some numbers to help convince you that straight ahead and under control is your only real option.

Using the following conditions, the analysis was done for an engine failure at 500 ft and 1000 ft.

Conditions and Assumptions

Altitude - Sea level
Temperature - ISA
Wind - Calm
Climb speed - 75 kts IAS
Rate of climb to 500' - 688 ft/min
Rate of climb to 1000' - 675 ft/min
Glide speed after engine failure - 65 kts IAS
Glide performance - 1.5 nm/1000'
Bank angle in turn back - 30 degrees

The analysis assumes a straight climb-out followed by a 270 degree turn, a reversed 90 degree turn and a straight return to the runway. It also assumed that the climb starts at the end of the runway at 50 ft and at the specified climb speed. Flap extension for landing was not considered.

Failure at 500'
Failure at 1000'

Time to climb
39 secs
84 secs
Distance covered
4,937 ft
10,634 ft
Radius of turn
648 ft
648 ft
Return distance covered during turns
1,296 ft
1,296 ft
Distance remaining to runway
3,641 ft
9,338 ft
Total distance from failure back to runway
7,711 ft
13,408 ft
Glide capability after engine failure
4,560 ft
9,120 ft

The analysis shows that from 500 ft a turn back would result in landing 3150 ft short of the runway, and from 1000 ft the landing would be 4300 ft short.

You can argue that a tighter turn reduces the distance back, but it also increases the load factor and therefore degrades glide performance, gaining you no advantage.

If a 10-knot headwind is considered and the numbers recrunched, the results show that the landing would still be 1840 ft short of the runway and 1460 ft short for the 1000 ft case.

The calculations indicate that given sufficient wind a return to the field may be theoretically possible, but the hazards of a downwind landing in such strong wind would not make this advisable, specially if one considers the very low groundspeed expected during a forced landing directly into a strong wind.

In summary, for a single-engine aircraft, given reasonable wind conditions, it is not possible to return to the take-off runway following an engine failure. Straight ahead and into wind is the only option.

Young Paul
2nd May 2002, 23:50
Pretty dependent on lots of things. Are we assuming both engines conk out simultaneously?

I was given a leaflet suggesting procedure for double engine failure on a B737 a while ago. It gave a circling visual approach, similar to the old "high key" / "low key" thing, and a glide approach clean one dot above an ILS glideslope.

The danger with turnbacks - as it is with light aircraft - is that you run the risk of stalling and spinning, or not getting where you want to. I suppose if the alternative is a mountain in front of you, you might give it ago. I'd like about 4000' below me to turn the aircraft through 180' - but that wouldn't get me back to the runway.

You ought to be able to work out what gradient you would need to climb out at to allow you to have enough height to turn around and land back on the reciprocal. My suspicion is that it would be something considerably more than you are likely to have until about 7 or 8 thousand feet.

Code Blue
3rd May 2002, 00:51
With regard to the light single I would recommend reading:

The Possible 'Impossible" Turn By David F Rogers, AIAA Journal of Aircraft, 1995. Vol 32: pp 392-397.

Basically 45º bank turn at 1.05 V stall(clean).

There is a lot of math in the article but you might be able to use it to give data for heavy metal.


3rd May 2002, 13:29
Had a bit of a play with the B737 in MS Flight Simulator 2002 earlier.

Who knows how acurate it simulates the performance of the real thing but this is what I found.

Also I only brought the thrust back to idle. I didn't fail them.

About the best performance I could get out of it was Flap Up, 210kts, 45 bank angle. From 10,000ft I did a 360 degree turn and lost 4000ft.

I tried various combinations of 25, 45, 60 degrees bank angle, 210 kts, 250 kts, 300 kts with Flaps UP, 180 kts with Flap 5, 150 kts with Flap 15 (yes the gear horn was blowing the whole time),

Any slower than about 190 kts with Flap up and 45 degrees I got the stick shaker. (Perhaps I should have tried Flap 1.)

So I figured for a cleaned up aircraft at 210 kts, you need at least 2500 feet for the turn plus at least 1000 ft for every 3 nm you are upwind of the runway.

I popped the gear down and selected Flap 1 at about 200 ft and landed with that.

The raduis of turn and angle from the runway centreline that I mentioned earlier seems to work out in the flight simulator.

Thankfully the Boeing family of planes (and Airbus too no doubt) glide much better that the lighties we all know. What a shame they must glide sooooo much faster.:(

3rd May 2002, 20:24
The "IMPOSSIBLE TURN" document that HugMonster quotes is hopelessly pessimistic in its assumptions. In particular the claim that a tighter turn offers no advantage is something that the original author clearly didn't bother to support with the nitty-gritty aerodynamics.

The optimum angle-turned/height-loss comes with 45 degrees of bank at maximum lift coefficient. That suggests that a practical application of the manoeuvre would have one operating as close to stall as possible, perhaps on the onset of the stick shaker. By holding the AOA precisely, the height loss for a 180 degree turn should theoretically be just hundreds of feet even for WWW's 737, though the height lost depends on v^2 for the speed corresponding to that max Cl.

I have no reason to doubt the accident stats quoted though -- clearly it usually is substantially safer to pick a reasonable landing spot straight ahead. I can think of a just few airports though (Antwerp, Tempelhof spring to mind) where trying a turnback from a few hundred feet in a light aircraft might just be preferable to a forced landing in the environs.

Dave Rogers article (cited by Code Blue) can be found here (http://web.usna.navy.mil/~dfr/aiaa1col.pdf).

3rd May 2002, 21:09
The Rogers article is impressively analytical - but it assumes a constant rate of bank - a simplification which costs you in this special situation.

I experimented with this at some length in a single many years back and determined that (under the prevailing circumstances) one could gain considerably in performance by using an intentionally non-uniform bank/speed profile. To do so requires a prepared mind.

Pilots' information and analytical skills may be poor amid the confusing moments just after everything has clearly gone to heck. This is the strongest argument for accepting a forward + / - 90 degree target for an-off airport landing.

But if you're gonna turn around, this is what I figured out:

When the reverse course decision is made, job 1 is to reduce the vector velocity away from your intended landing spot. Turning is the only way to do this. The earlier and faster you turn, the greater the likelihood of reaching your goal.

By turning hard and early, you a) shorten the net distance, b) allow time to stabilize and assess the view, c) improve the odds for at least a successful off-airport landing nearer to those valuable emergency services on the field.

Slowing to anywhere near stall speed in a steep emergency turn is a bad practice in light aircraft and really bad in heavy ones.

The safer alternative is to crank in the 'maximum allowable' bank as soon as the decision is made, simultaneously pushing the nose down for acceleration. It is counterintuitive to dive when you only have x much precious altitude left, but doing so allows steepening the turn so that travel away from the airport can be stopped at the earliest possible time. If this is done in a smooth and 'aerodynamic' manner - flying through the turn rather than skidding - , then the total energy lost is not going to be vastly higher than optimum. (a la Bob Hoover) You will have traded some of the altitude for increased velocity, and the rest for the vector velocity cost of the turn.

Assuming your departure was straight out, the first 90 degrees of turn (into the crosswind) gets you to zero increase of distance from the airport. Ideally, this must be completed as soon as possible. The next 45 gets you onto a very nice converging downwind angle with your target, with time for shallow bank adjustments on the way in. When you are on the 45 and set up for max glide distance speed, you have a familiar, stable context and can then quickly intuit if it's going to work or if you need to find an alternative 'plan C'.

3rd May 2002, 21:42
Did it once in the A-320 Sim successfully. Can't remember the details but do remember clean wing, speed just above stall and a teardrop type approach. Gear down at 500 ft agl. & 0 flap landing.

4th May 2002, 00:35
Just to be pragmatic, which I'm not normally known for; I’d use a speed slower than Vsuby in the turn and a Vsubx speed in the return. Questions, see the 757 post a couple units down, covers the same problem. Basically VsubY is based on the minimum power required to sustain flight speed, it is a speed higher then lowest point on the speed vs. power curve. Whereas the greatest distance per altitude lose is at around Vsubx, which is the nadir point on the drag curve. It's confusing to me, but so is most everything else.

I realise (sic) this is a predominantly British forum, please excuse my colonial spelling, courtesy of Noah Webster, if I find myself in possession of 150 quid, I will invest in the computer version of the OED.

[edited: italicized items are new because Tinstaafl pointed out my erroneous ways, the max endurance speed will be slower then the max range speed (I suppose that wasn’t clear in the above) keep reading it gets more clear as the posts progress – no really!]

4th May 2002, 01:20
%MAC, unless I've misread your post I think you have it backwards.

Vg (range) usually corresponds with ~ Vy
Vg (min sink) with ~Vx.

That's my understanding. Of course I could always have misremembered something! :o

4th May 2002, 03:43
Tinny, Shame on you, You’ve dislodged me from my drunken stupor, American whiskey (Knob Creek) this time, Single malt, (Lagavulin) in Autumn and Winter. :)

Not as I understand, here is my reasoning, correct me if I’m wrong.
The minimum sink rate in glide occurs at the minimum power speed, i.e. that speed in which the power which must be supplied by loss of potential energy is at a minimum, which is conveniently the lowest sink rate = a speed a bit lower than max climb rate for a jet Vy. (This offset occurs because the available engine power increases more quickly with speed at lower kinetic energy then the required power). Best endurance is the lowest point on the Power Vs. Speed Graph.

The minimum glide angle is that which minimum weight occurs in direction of motion, id est the minimum drag speed on the drag Vs. speed graph. This gives the greatest horizontal distance per vertical as the former gives the greatest loiter time. The drag graph corresponds to Vx. Id est weight just balancing drag.

One thing I can count on, you guys won’t let me get away with bull crap, so edify!!

Captain Stable
4th May 2002, 09:01
Bookworm, the max coeficient of lift / rate of turn ratio is not a constant. It depends upon the design of the wing, the distance between the centre of lift and the centre of gravity, the airspeed and a great many other factors. As has also been stated, it is not appropriate to assume a constant angle of bank. The maximum roll rate of the aircraft is a major factor here as well.

What, as I read it, that article is highlighting is the inherent dangers of attempting to carry out such a maneouvre without having practiced it in your current type when other, far less dangerous options are usually open to you.

I have only once seen a pilot attempt to turn back to the runway after a power loss (it was not total) on the climb-out. He made it back to the runway (just), but the aircraft was a total loss. And this was an experienced pilot who (in my opinion) should have known better.

I know that the RAF certainly used to teach turn-back techniques (I don't know if they still do) but it is not something that should be attempted for real unless it has been practised.

4th May 2002, 11:49
Capt. Stable

You raise several interesting points.

I wasn't suggesting that max coefficient of lift / rate of turn ratio is a constant, merely that for a given aircraft in a given situation, minimum height loss per unit angle turned is achieved at max coefficient of lift.

I can well imagine that the optimum manoeuvre involves a changing angle of bank, but the point is that by rolling on 45 degrees of bank, a la Rogers, you achieve a result considerably better than the one given in the article that HugMonster cites. I can see that, particularly in larger aircraft, roll rate can be an issue. What's the maximum rate of roll to 45 degrees in WWW's 737?

I have no issue with the views that the turn-back manoeuvre is not usually worth the risk of getting it wrong and is also difficult to practise safely. But I do feel that the risk management should be based on proper physics rather than needlessly pessimistic data.


The Rogers article indeed does assume a constant angle of bank, and it may be possible to do better. But I'm not convinced that your argument for increasing speed holds water. The rate of turn is very sensitive to speed, and I think Rogers shows in his paper that the trade-off you suggest is not a good one.

One further point: the state is not unaccelerated, and the assertion that maximum lift coefficient is optimal means flying an AOA rather than a speed. Thus you're absolutely right that the optimum is not going to be at constant airspeed.

Finally I don't doubt that given a few hundred feet more than the absolute minimum, your method may be a much safer and more effective option.

4th May 2002, 16:05
Some decades ago, I witnessed a Grunau Baby glider (12:1 glide ratio?) release from an aerotow about 150 yards past the airfield boundary.

The pilot made it back to the field without difficulty, but the downwind approach resulted in a fast, heavy landing and minor damage. He would likely have done better landing straight ahead in the maize field, especially with the high wing. In an airliner, the extra 20 kt. is much less a big deal.

The Blanik with its 18:1 glide ratio similar to airliners would be a good machine to familiarise oneself with power-off approaches. And 45 degree bank angles are optimal.

It was suggested earlier to turn upwind immediately on an engine failure. Derek Pigott's advice for a midfield rope-break is to first turn downwind so that you can approach the runway with some headwind component. With an 18:1 glide ratio, you need some room as you are not coming straight down.

And yes, when making a landing approach in a glider, the practice is to accelerate to best glide speed from the minimum sink we were using until our lift/engine ran out and we had to head home.

Now for the big question: will there be enough hydraulics, electricity, air to operate the spoilers:confused:

stator vane
4th May 2002, 21:28
i have thought of this many times, but they would rather us practice circle approaches rather than this in the sim.

then i always think of the aircraft that just took off after we did.

if we went straight ahead at least we could minimize the possibility of taking others with us into the hole in the ground.

i would think it very possible only with a lot of practice.

bob hoover could have done it, and maybe even roll it over on the way back.

international hog driver
5th May 2002, 01:27
Had a long think about this over a couple bottles with a few mates last night and our conclusion is this.

IN reality not less than about 2200 AGL

Drop the RAT (apu?)

Drop your Guts

80/260 @ 45 degree AOB

If you were part of a stream departure off the same runway, tell everyone to clear off.

If @ F1 or F5 leave it if @ F15 consider how much headwind you have/had.

Gear as late as as you can

You should make it back to the hard surface (well the flyover at least).

Call for the tug...........

In reference to the NG and the guy that said he had a double separation then he would have had less drag & weight.

Best WX would be a LH X wind.

In the places we discussed straight ahead is usually not a great option, hills, Built up populus or water. The euros and seppos might have a better time where the terrain may be more forgiving.

5th May 2002, 06:09
I hear what's been said about the unattractivess of a straight ahead, or approximately so, landing when the whole area is filled with offices space, hotels, suburbs or evil terrain. But I also hear what's being said about the precision with which the aircraft has to be flown, to have ANY chance of success.

I doubt that the average line crew trains for this, let alone the average single-pilot operation. And I feel sure the very rigorous and frequent training would be necessary. And, even then, how likely is it that the training would cover all possible scenarios that might lead to consideration of a turn around.

Many of the airstrips in my part of the world are equally unattractive for straight ahead landing. But, then, there is often not enough free airspace to permit any consideration of radius of turn, to return to the strip.

Over the years, I've happily settled for the rationalisation that the chances of loosing both engines is too remote to contemplate in my operational environment. The reason I've lurked on this thread, however, is that the unthinkable could happen. While it would certainly mean that I'd be having a VERY BAD day, its worthwhile to read all the comments here.

For the most part, we have SOPs that basically say its better to make a controlled crash at stall than to spear in, out of control in an unrecoverable stall/spin scenario. I'm not sure that I've yet heard anything here to change this philosophy but remain hopeful. I have to go back to the point that, unless the whole crew is trained and regularly refreshed on an emergency turn-back, the chances of success are very slight indeed.

I know that our SOPs won't work too well in many other parts of the world, so "horses for courses" is the phrase that comes to mind. I shall continue to lurk on this thread with interest.

5th May 2002, 14:22
%MAC, to be fair I was using piston engine curves for the Vx/Vy comparison with sink rate/glide range. I shouldn't have done that.

However, now that I have :p , the engine's power & thrust curves are irrelevent because we're also referring to a no-engine scenario.

This means that we have to consider airframe drag/thrust/power requirement curves only and not the more usual 'greatest difference between available & required' ie maximum excess which is governed by the powerplant type.

We're talking about two different measurements - one involving angle of descent as a function, the other involving rate of descent. Rate includes a time factor.

The appropriate curves for each of these is the drag curve (for range) and the power required curve (for duration). Power is is appropriate for duration (min. sink rate) because power is a rate ie includes time as an element.

Best range (as far as the airframe is concerned) is achieved by flying at min drag and provides the best L/D ratio. I think of this as the least cost in D to achieve the lift, hence the range effect ie if drag was zero we'd be fly S&L indefinitely and if D=L then we'd be descending vertically.

This is also the same as flying at a tangent to the power required curve but by using this method we've effectively introduced a distance parameter into the situation (the tangent commences from the graph's origin). Handy if you know the headwind or tailwind component because you can adjust the tangent's commencing point to find the Best range speed for that wind component.

Best endurance (ie min sink) will be found by flying at the min. power required. This is slightly slower than L/D. This I think of as the least cost in energy and hence its effect on duration.

Bear in mind that this spiel is done entirely from memory and I could be wrong.

Wee Weasley Welshman
5th May 2002, 15:00
Thank you gentlemen. Would Boeing themselves have figures on this at all?


5th May 2002, 19:21
I'm really not convinced by this don't turn back thing we're all taught. The Cessna twin (402 or 404) that crashed at Glasgow in Sep 1999 (G-ILGW) was less than 5 miles from the airport but it took 10 mins for the fire truck to arrive. Meanwhile flames had roasted most of the passengers and crew.

I can't see why if you have a bit of height, you shouldn't attempt to plonk the thing where everybody is ready for you and where kind people have created the biggest piece of open space for miles around. This is all the more so with jets where the direction you hit the ground relative to the direction of the wind is of less significance.

6th May 2002, 00:47
Guten Tag Tinstaafl

Your post is right on, well written and informative. Whereas my posts were neither as eloquent nor as erudite, I am trying to say the same thing. Namely, these two different speeds are associated with different constraints. As you pointed out Best Range is a function of drag, whereas endurance is a function of power. Vx is a function of drag and Vy is a function of power. Though I realise that x and y speeds don’t appear on part 25 aircraft, it only confounded the issue. When I wrote that “lowest sink rate = a speed a bit lower than max climb rate for a jet Vy” I was being much too qualitative. And when I said I’d fly Vy in the turn I was just damn wrong. (It has since been edited for those that were going to try it). No, I would fly at max endurance in the turn and max range to get back to the aerodrome. To try to clarify for those I have unwittingly confused, mostly myself:

Max Range

Max angle of climb for a jet is derived where thrust subtract drag is the greatest, since thrust for a jet is relatively constant with speed, max T-D is at the lowest point on the drag curve vs. speed curve. The lowest drag gives the best range in glide: Vx = Best Range (jet)

Max angle of climb for a prop is derived in the same way except thrust falls off rapidly with speed so Vx in a prop is at a speed less than minimum drag speed: Vx < Best Range (prop)

Max Endurance

Vy speed in a jet is determined where the max excess power exists. In a jet the available engine power increases very quickly with speed (more so than the required power), so the greatest distance between the two curves is at a speed greater than minimum power speed: Vy >> Best Endurance (jet)

Vy speed in a prop is determined the same way, but in a prop the available engine power is relatively constant with speed so max excess power does occur at the lowest point on the power required curve: Vy = Best Endurance (prop)

Aha, and my math teacher said I couldn’t compare apples and oranges

Okay back to the original thread, let’s forget about Vx and Vy because they are only semi-relevant to the discussion of glide, we need to have a best glide endurance speed, and a best glide range speed available.

6th May 2002, 06:05
WWW Ive just been rostered for sim-duty tomorrow so Ill have the numbers for you late tomorrow or next day.

Ill also try to post pics of the sim read-outs in both dimensions of the successful profile.

Just to recap:
* flap 5
* APU off/no APU generater avbl
* engine and wing TAI valves closed
* max power TO in CAVOK conditions by day
* 9000ft EOL
* flat terrain
* wc +10
* GW 55T (if I have time Ill do one at 46T as you requested since a lighter weight would be more critical)
* MACTOW 22%, trim 4.2 (typical CG)
* ISA +10
* gear up at positive ROC
* Gear-lever left in UP position because the OFF selection is normaly done after flap retraction
* V2 to V2+5 (146 to 151 kts)
* no climb thrust reduction at normal 3rd seg altitude.
* sudden complete failure of both engines (the engine failures will be simulated by simpley pulling the start levers rather than lengthy engine-failure pre-programing).
* imediate initiatiation of loss of thrust on both engines recall items by PNF, followed by SYS A +B flight control sws to STBY RUD (which youd do in real life)
* PNF should then imediatley drops gear lever to OFF, then DN (in case pulling the handles is necessary), but the aim is to make the runway so gear will be select to OFF only, to protect what little hyd press you have left for primary flt controls. Belly land.
* All through this the PF turns to the left for return, and lands at Vref5 for 55T (168kt) and Id expect possible manual reversion.

This is gonna be bloodey interesting!

6th May 2002, 10:50
Hello all,

Actually tested in a 737-300 sim.

Wind 2 o'lock at 20 knots.
Min turn altitude: 1500 feet (2000 is much better)
flaps 1
apu operating

As soon as you notice turn into the wind bank 45 pitch +5 to the runway you just took off from.

With the loss of gen 1 and 2 only bat power is available.
Take apu on bus one to activate B hyd to extend the flaps to 5 (or 10 in case flaps 5 t/o).
As soon as this is done (you're already very low) disconnect bus one and connect to bus two to active A hyd and lower the gear.

In my opninion you have no chance taking off without APU since it takes to long to start it.

It's still a difficult manoever. The most difficult is that since you're coming at the runway from a 30 degree angle in heading you need to start the turn early or you'll overshoot.
But the most important is to level te wings otherwise you don't have the slightest chance to survive. Rather land within the airport perimeter but not on the runway, than crash into it.

Hope I could be helpfull,
regards. Tijmen

7th May 2002, 19:38
Alright WWW with the parameters I stated above, the absolute minimum height is 1,350 ft AGL if your actions are immediate.

Im sorry Im not able to post the sim profile read-outs. Having trouble copying it clearley to the HD and photo-storage site.

PS and WWW fire away your questions and Ill do my best to answer them. Oh and just for info (if its useful to you) WWW, the exercise was conducted in a 1992 Hughes Rediffusion 6 DOF hydrostatic with Motorola computer under Full Flight Regime Simulation with Collins EFIS.

8th May 2002, 07:58
Slash ... sometimes the FotoTime site goes down for short periods. It could be that the photo is already on there, just not yet showing up. Go into the relevant album as "yourself" and do a "refresh". That usually works for me.

If all else fails, if the file ain't tooooo big, send it to me and I'll try putting it up on my FotoTime site.

8th May 2002, 09:05
Thanks Ozex but it now appears the problem is in the read-outs themselves. I obtained both the lateral and vertical profile diagrams as a print-out and had to scan each and save in jpg format. Fototime has them ok (file-size is correct) but its just too damn bloodey faint to be anything ledgable.

Sim-consoles bamboozle me even during the best of times and there was no tech guy around to help me directley save them to a floppey or anything.

9th May 2002, 09:24
That's a real shame Slash, coz now ya'll just havta go and do it all again! :D But, hey, ya came up with a very good number that surprised me beyond belief. Of course, the operative expression is - as you said - "if your actions are immediate". Don't think I'd like to be the one putting it to the test tho!:eek:

As it's in the same ballpark as the results notified by tijm, it would seem that you both went about the test in pretty much the same way. Is your sim configured as a -300? I can't look that far back on the thread while writing this reply.

9th May 2002, 17:22
Yeh but Ozex we simulated an actual everyday TO with only minor variations as per that "recap" list above. Tij said he left the APU running. Most Companys require the APU off after engine start (save fuel etc) unless you require it on for certain MEL reasons (busted engine generater, engine bleeds-off TO, etc). So we turned it off after start. When engine-driven AC electrics died all we had left was the battery and it takes too bloodey long to start the APU and get its genny on one of the main busses before scraping the treetops.

Also Tij dropped the gear, where our profile left it up and we belly-landed. This obviousley led to our lower figure of 1350 ft as against his 1500-2000 ft. And WWW did ask what was the minimum height to make a 180 and get in.

PS Yeh actions have to be done asap alright, and yeh its a 300. And we tried startin the APU but it didnt fire up and have its genny available until we were slidin around on the deck showerin sparks everywhere.

9th May 2002, 21:52
I tried the 500' & 1000' agl engine loss after takeoff with
Cessna 172 & 182 models on MS 2002. The 172 crashed
short both times :( , while the 182 got back to the runway
just on the 500' and easily on the 1000':) . I think the
key is the difference in best rate of climb, the 182 was
showing 1500fpm while the 172 was more like 500fpm.

So given a 90-knot best climb for the 182 a 1000' fail was
taking place within a mile of the runway, while the 172
at 75-knots was failing at 2.5 miles.

All in all a good argument for climbing to en-route altitude
as rapidly as you can...and having a powerful engine to start

Of course this is MS software so your mileage may vary.

-- Andrew

10th May 2002, 18:56
Interesting thread.

I have often wondered about this one when flying jets/tugs/gliders. I see the major problem when you are flying 'faster' aircraft as being restricted as to the rate of turn you can achieve in order to get pointing back towards the airfield. This automatically positions you some distance from the runway, and leaves you flying for some time at high rates of descent.

Problem: You want to reverse the direction of the a/c using the minimum distance, and in the shortest possible time.

Possible solution: If you were an aerobatically inclined pilot with vitals of steel, you might consider a cross between a stall turn and a chandelle, depending on height/speed available. If you 'unload' the wings during the manoevre, you can fly as slowly as you like. You could 'flop over the top' as long as you pointed towards the ground for a bit afterwards to get speed back for conventional flight. The turn rates for this trick would be much higher than using a normal 1.x Vs turn or whatever, and because you would not be asking the wings to generate vast quantities of lift for extended periods, probably quite efficient in overall energy terms. It would also leave you pointing almost directly down the runway.

Maybe one of our aerobatic aces could try this at altitude and report on the height losses etc.;)

Captain Stable
10th May 2002, 22:59
Hmmmm - I can just see the Aresti diagram... :D

Extra points for not falling off the top of the loop, keeping within the box and not wiping any of the pax across the top of the cabin!

I know that both Barry Tempest and Brian Lecomber subscribe to PPRuNe... maybe one of them might break cover!? Or Louisa? You out there?

14th May 2002, 02:54
WWW are you interested in this thread you started or not pal?

I get a bit annoyed when someone asks a question, someone takes the time to duly investigate it, only to find the original enquirer has disappeared like a fart in the wind and couldnt give a stuff.

I have techniques, speeds, bank angles etc used in that sim exercise if your still interested WWW. If your NOT interested then I will ignore ANY further threads you start from hereon.

14th May 2002, 07:50

I’m sure many would like to see the data, not only 3W. I was very excited about your research and ran home early to catch your post about the results. So, while the originator of the thread may have lost enthusiasm, more fill the void.

Still I hear what you’re saying: It is a pain to compose responses and it’s time intensive in general. Of course, you went beyond the normal post and did independent research. You have a right to feel unappreciated, the feedback you have received has been paltry (Expat Excluded). However, your work may save my scrawny neck sometime, and possibly others. Many will read your words of wisdom, and some will put into practice your hard borne experiences. In aviation we learn vicariously, there are too many mistakes out there to make them all individually. But, data is proprietary, it is your choice to disseminate.

Anyway I wish to extend my thanks for the work which you have already submitted, it has enlightened my sense of possibilities, and awakened within me a realization that this is not as easy an exercise as one might be initially led to believe.

Safe flying

14th May 2002, 14:39
G'day Slash,

Now for a variance to your configuration:

What's the minimum height needed to land back on with the wheels down?

As an aside, I haven't a clue how long it takes to get the gear down on a 737. Is it longer than the time it took to land back on from the double failure in your previous experiment or is it shorter?

All other conditions the same as your first effort.

Wee Weasley Welshman
14th May 2002, 17:29
Yeah I'm still here Slasher. Thank you for using a sim to investigate my question. You answered my question - 1,350ft.

If you could clarify just how you flew the manouvre that would be great. I am guessing the gear down would have cost a couple of hundred feet...



15th May 2002, 08:07
Ok then.

Unfortunatley we didnt have enough sim time to go through various configuration and weight scenarios such as gear down and lighter (46T) gross weights. We therefore stuck to a gear-up landing as this would dictate the minimum height for a return, which was your original question WWW. Various heights were experimented. 1000 ft was not enough and 1500 ft was an overkill.

Please refer to my previous "Recap" for aircraft configuration and status before TO.

* Upon the double engine-failure the stick must be IMMEDIATLEY pushed forward from 21 deg NU to at least an initial 5deg down, while at the exact same time violentley initiating (yes even though your picking up your into-turn flight spoilers at that point), then smoothley executing, a turn at 50 deg AOB while increasing speed to vref 5 (168kt). With 22% MAC the nose was quite agreeable to pitch down and increase speed at that bank angle which required only a .1 unit stab-trim input ND once at wings-level. In our exercise I overshot slightly to 172 kts IAS on the standby ASI but letting go the stick at that point was beneficial and 168 kts was quickly regained.

This manouver is the most costly of the exercise in terms of height. Rollout in intercept heading (about 30 deg) was at a height of 320 ft.

*APU start-attempt was tried, but as predicted we were alreadey on deck by the time the genny was avbl. So no AC electrics around throughout.

* Auto-activation of the standby hyd system valves had occured but the pump itself of course was dead.

* Around 2 deg ND (as read from the Standby AI) was necessary to maintain Vref 5. I was quite surprised at this as I thought 3-4 would be required.

* About 60 deg before intercept-heading rollout we were on manual reversion. This is why the initial manouver must be flown quickly and relativley violentley in order to utilise the flight controls before you lose all hydraulic sys A+B pressure. Of course from then on you fly rudder for bank and stab-trim for pitch where possible as you well know.

* Rollout on final was at 50ft, just in from the piano-keys. Flare and landing as per gear-up procedure in the Boeing FCTM.

Summary: If the crunch comes in real life Id recommend absolutley no lower than 1500 ft in height if your hard decision is to turn back. In our exercise no recognition time was allowed which of course is about 2 secs, and another 1 second to initiate the required action even assuming your wide-awake. Anything lower than 1500 ft Id aim straight ahead and pick a road or aim in between the houses or cows.

Notes: In our aircraft, standby pump only powered by xfer bus 2 (or 1 through the xfer sys) and FMC is not avbl when battery and standby inverter are only sources of power.

15th May 2002, 09:14
Slash... I guess it's also true to say that you were ready for this simulated emergency. How many pilots would ever be quite THAT ready for such a situation? I guess that, if you have the opportunity again in the near future, can you try a few seconds delay ... to simulate the standard first call in the emergency checklist - Sh!t... WTF was THAT??!:eek:

Perhaps end up needing something more like 1800 FT eh? But then I'm guessing you may run into the "tyranny of distance" problem in getting back to the runway... :(

Wee Weasley Welshman
15th May 2002, 09:21
Slasher - v interesting. Many pilots I talked to about this thought it impossible below something like 3 - 4,000ft.

I might have a go at your profile myself if I can get 5 spare minutes at the end of a sim detail.



15th May 2002, 16:15
Slash... As you know, I'm not rated on the 73, but I'm interested in that bit about using stab trim for manual reversion. Are you referring to electric or manual trim? I'd have thought that electric stab trim wouldn't be available with a total loss of AC.

15th May 2002, 23:59
Thats a very perceptive question Ozex considering your not even rated on type! Arent you the smart cookie!

Yes mate you are correct. Electric stab trim isnt available with a loss of AC, so physical manual movement of the stab trim wheel is necessary. As any 737 driver will tell you the elevator is very heavey in manual reversion and stab trim helps augment control in pitch.

Hmmm while on the subject, WWW would you happen to know which bus powers electric stab trim? We know rudder and aileron trims are off AC Xfer Bus 1, but Im not sure of stab trim.

PS Oh and Ozex your "what the faaaaark!" factor was mentioned by me earlier referance the 2 + 1 secs recognition in my "Summary" paragraph.

17th May 2002, 04:30
Electrical power for the stabilizer trim actuators is provided from the No. 2 115 v AC transfer bus and the No. 2 28 v DC bus.

17th May 2002, 08:59
Okay Slash, I've started me stopwatch! :D

The Scarlet Pimpernel
19th May 2002, 17:08
Very interesting thread. I can only comment on turnbacks I have practised in the military (Tucano/Hawk) many moons ago. I believe the Air Force has now stopped practice turnbacks after a couple of nasty accidents.

Notwithstanding the urgent need to get the aircraft back from whence it came, there are inherent dangers with turning directly towards the runway, funnily enough. Firstly, many people try to turn way too hard to get back to the centreline as quickly as possible, with the risk of stall/wingdrop - remember you only have 2 things in your favour, speed and height...turning too quickly washes off the former thereby losing the latter more quickly. Secondly, the final alignment turn is done perilously close to the ground at low speed which is another cause for concern. Slasher's roll out at 50ft is a case in point.

Now (I hope I can remember this correctly!) on the turnbacks I was taught, the first thing to do was to actually turn away from the runway! This allows you to do 2 things - see the runway before you turn back sooner and set the aircraft up for the turn -Vimd usually worked. I'm not saying a prolonged outbound leg, but enough to get you a little displacement to execute a decent turn.

I only proffer this up as food for thought. There will be those that say that you are wasting energy going in the wrong direction - I don't know the answer. Perhaps you may need more height for the manouevre - maybe someone can try it out in the sim (My next one's in Aug....so you'll have to wait for the Airbus answer!!!)


20th May 2002, 09:29
Thanks Checks.

Ozex my stopwatch battery just went flat! Yours? ;)

20th May 2002, 11:14
Lots of interesting stuff here, but there is a bottom line. That is :

Unless you have 1500ft in anything swept wing forget it.

The wash off is just too great to get a stable configuration before 300-500 ft and you have got to get very lucky indeed to make a "non-fatal" touchdown after that.

Better to go with the "Old Rules". Land ahead in the best available area. Gear up/down is a matter of choice. Gear down would be better as it takes some of the crash load (surface permitting).

A controlled arrival onto bad ground is better than an uncontrolled one anywhere.

That said; we can screw around with the Sims all day long - and I have for years. However the numbers programmed therein are not reliable outside the normal operating params and it can be misleading to rely on them.


Wee Weasley Welshman
20th May 2002, 11:18
What would be a typical climb gradient and what would be a typical glide gradient of our average 737?

I was out of Alicante yesterday doing 3,500 feet per minute to 10,000ft and to be honest a glide back to the runway looked feasible from 2500ft looking back over my shoulder.

21st May 2002, 07:08
MG no-ones going to seriousley consider turning back unless its certain youll die if you proceed straight ahead. The original threads question was rather academic dont you think? Sims are academic so the answer was also.

WWW Im a bit confused here. Are you a 737 pilot under training?

Wee Weasley Welshman
21st May 2002, 09:33
Nah just being lazy and not looking it up.


21st May 2002, 15:11
Slash ...

Ozex my stopwatch battery just went flat! Yours? ;)

Yer never gunna believe this mate, but MY watch battery died on me yesterday. Took me all of 30 minutes and a couple of bucks to get a replacement battery, but its all too late now of course, ain't it... ;)