B737 Simulator compared with the real aircraft. Question

[Tee Emm]

In another life we trained on the real B737 because a suitable simulator was unavailable. Training was on a long runway and included simulated engine failure at V1 conducted by pulling back one throttle and continuing the take-off. The aircraft was easy to control with no undue yawing or rolling.

In contrast, in the several B737 simulators I have "flown", all the simulators seem significantly more tricky when conducting an engine failure after V1 in that the aircraft rolls and yaws more rapidly that one would expect after having done the same exercise on a real aircraft.

There is anecdotal evidence and also from personal observation, that many students undergoing a type rating or recurrent training on a 737 simulator,often experience significant difficulty in handling the engine failure at the moment of cut and initial climb out.

If a flight director is used this often seems to exacerbate the problem for students because they chase the FD needles instead of looking "behind" the FD needles. Next second they are all over the sky with loss of directional control. In any case pilots should not have to look "behind" the FD needles to fly an accurate profile during engine failure climb. It means the FD needles are too distracting and it may be easier to switch off the FD and use basic instrument scan instead of close concentration on a single instrument.

Is the problem that FD needles are too sensitive in pitch and roll in today's simulators? Pilots used to flying a V-Bar presentation seem have less general difficulties in handling directional control on climb than those using the two needle presentation. .

While simulators are subject to scheduled fidelity checks where one pilot must be current on type, it is almost certain the pilot conducting the fidelity check would have never experienced a real engine failure immediately after V1 in a B737 variant. After all, not all approved fidelity check airman are qualified test pilots who may have ben involved in actual testing of this sequence in the real thing. That being so, I wonder how many simulators can be relied upon to demonstrate true fidelity on engine failure immediately after V1 when the check airman has never done one in measured conditions in the real thing? The simulator could go for years being unlike the real thing at engine failure at V1 simply because the so called expert check airman responsible for simulator fidelity checks has never known the difference between the actual aircraft characteristics and the simulator characteristics.

Students have been scrubbed for failing to perform up to expectations while undergoing training on engine failures in a simulator immediately after V1. After all a simulator is nothing more than a sophisticated computer albeit remarkably close to the real thing within certain tolerances. Does this all boil down to simulator handling skills rather than aircraft handling skills? With constant practice on the same simulator you get to know its characteristics. Students may not get that opportunity however. Some pick up the knack easily while others not so easily.
Your comments would be appreciated. Have you ever experienced the problems discussed above? Apart from the usual canned advice of keep stesdy rudder pressure applied and avoid excessive ailerons, how many of you have had trouble with following the FD needles as yaw and roll occurs at instant of engine failure. Is the blind use of the FD the best way to keep it straight? How do you teach some to "look behind" the FD needles when the FD needles hide the view of the "little aeroplane" ?

[john_tullamarine]

For a new chum it might be necessary to start off with a very long runway and the usual post failure steering exercise until runway head and freeze.

Once the rudder work is up to speed, a useful technique is to use repetitive takeoffs (via freeze at, say, 300ft and reposition to BRP)

(a) run the training exercise raw data

(b) start with a reasonably high weight, forward CG which gives reasonable control and a high V1. (If the student needs an easier introduction, try using an overspeed schedule).

(c) step sequence to lower weight forward CG which retains reasonable control with a reducing V1

(d) step sequence either to minimum weight at forward CG for minimum (stall limited) V1 or the weight at which V1 becomes constant (Vmc limited)

(e) step sequence slowly to aft CG. Generally the main problem seen is a rapidly increasing roll response. The trick for this is aggressive rudder and aileron

(f) once the student gets on top of this introduce reducing minima. This ought not to be a real problem as the last bunch of failures have been on the clocks anyway but it does boost student confidence.

(g) then return to min weight forward CG and run the failure so that the yaw response is delayed until well into the rotation.

(h) step sequence to aft CG

That's as hard as we can make it.

The only other useful workload increase is if the particular sim has a realistic catastrophic failure. On a contract around 10 years or so ago, the 732 sim concerned had an FDR record crane impact modelled and that was interesting .. especially until we got used to the button's location and avoided selecting it in the dark during cruise ... As an aside, this particular operator's procedures generally emphasised high overspeed schedules so the general training work for their pilots had been skewed to the easier end of things ...

I've used this approach for a number of students who were a tad frightened of engine failures (and, as a consequence, were poor performing) and the results were uniformly productive.

Subsequent "normal" sim failures were a doddle and my students had no trouble with such exercises during their endorsement checks.

Of course, the instructor penalty usually is a few slabs for the sim techs who turned a blind eye to extra time and sessions where required ..

The degree of sim fidelity is pertinent. However, providing that the apparent response is reasonable, the generic training value will be useful. The handling delta between min V1 and aft CG and "normal" will be significant.

[BOAC]

I concur with your experience of the handling differences - my 73-2 base training was 'for real' at CWL and far easier to control than the sim. Mind you, the 'visual' was a darn site more realistic at CWL:)

[aozc]

How I do it and have been taught is to instantly level the wings using aileron and then "step on the ball" almost simultaneously while keeping the pitch on 12-13 degrees. And then smoothly manouver the airplane onto the runway track on the ND.

I don't follow the FD commands initially since it's pitch+roll only where as I have to fly the airplane with pitch+roll+yaw. Once the ball is in the center and the airplane is on the rwy extended track I go back using the F/Ds with the turn coordinator centered.

When only using thrust to idle with the lever. It reduces thrust pretty smoothly and also even at idle thrust, the engine produces a considerable amount of thrust. With an instant engine surge it is possible that the aircraft would handle more abrupt and violently.

[aerobat77]

you point the problem out with your own explanation :

you will struggle to find somebody who has real life experience how a 737 really feels when exactly at V1 an engine suddenly quits.

the simulators are to train procedures for it.

your story about engine failure training on a real 737 by pulling back the lever and continuing the take off sounds cute but must really be from another life :)

[763 jock]

A similar experience, albeit on a different type. The V1 cut on the real aircraft was extremely easy to deal with. As was a S/E landing for real years later.

I have yet to "fly" a sim that truly represents the benign handling of the aircraft when flying engine out. However, if you can handle the more tricky sim, the aircraft will be a pleasant surprise on the day.

[zlin77]

None ever "fly" like the real airplane, we must remember they are simulators not duplicators!

[cosmo kramer]

My take is that the problem arises from the sensations of the motion. That is really unrealistic.

Personally I "switch off" my sensing of the motion and just fly the instruments, be it FD or raw data, using pure visual input. If you combine FD/raw data with what your inner ear is sensing, for your corrections, you are bound to be all over the place.

[john_tullamarine]

instantly level the wings using aileron and then "step on the ball" almost simultaneously while keeping the pitch on 12-13 degrees. And then smoothly manouver the airplane onto the runway track on the ND

One of the main training problems with engine failures (sim or aircraft) is that the exercise is done at a reasonably high V1 and mid to forward CG. The negative training value significantly increases if routine training is at overspeed schedules.

Points to note -

(a) the normal pussycat becomes a roaring tiger at min V1 and aft CG

(b) at min V1 (especially if Vmc limited) V2 often will be Vmca-limited and wings level might just push the actual Vmc up enough to catch you out .. inverted and into the ground. One needs to keep in mind that Vmca is strongly bank dependent and, near invariably (I have to say that as there may be the odd aircraft for which geometry may limit the bank angle), is scheduled for 5 deg wing down into the operating engine(s)

I might note that my students, after a bit of practice, were able to do the exercise under nil vis conditions including back tracking the opposite end localiser. The confidence building effects were quite stunning.

[cosmo kramer]

V1 will never be VMCa limited. It will be VMCg limited, it's damn hard to maintain 5 degs bank on ground :8

[john_tullamarine]

Mea culpa .. a bit late at night for me. Quite correct. However, the associated V2 will be Vmca limited and that's what I was getting at.

Previous post clarified with some words in red.

[tom775257]

In 2006 I had the CP pull an engine to idle on me at V1 during base training. Very easy to control for real (A320).

[BUGS/BEARINGS/BOXES]

I too have found the aircraft easier to handle than the sim during a Daily Mail 'death, shock, near disaster' scenario. The aircraft was a pleasure to fly and strangely enough, I enjoyed the real life handling experience. However... I am one of the 'train hard, fight easy' brigade. If a student can't do it in a more complex (:oh:) simulator, I would rather they didn't get to try it on the real thing.
Was the training representative? Well yes it really was, the SOPS came naturally and without any massive yaw or control inputs required, as suggested by the sim, we both had plenty of excess capacity to get all sorts of things sorted that we never have time for in the sim.
TIME A TRE'S NO1 ENEMY.

[Johnny Tightlips]

We have both level D and B 737 sims. I never had any problems with V1 cuts in the level D sim, but i find them quite difficult in the level B sim. It just does not feel right and I am never happy with my performance, even though no TRE has ever complained about them in the de-brief....

[DC-ATE]

Well, it's been awhile [nearly 30 years !!] since I last flew a 737 [200] sim, but I don't recall too much difference between the sim and the a/c. I can't recall now the make/model of sim, but it was a full-motion sime and seemed to be very realistic. Maybe the newer ones are different?

[Fullblast]

On a different plane years ago i flew the OPC in the real plane because there was no sim available in the world, V1 cut, SE approach and go-around lot easier than sim.

FB

[aerobat77]

[QUOTE][A similar experience, albeit on a different type. The V1 cut on the real aircraft was extremely easy to deal with. As was a S/E landing for real years later.[QUOTE]

its so cute and the most fun of pprune in reading computer games stories and dream wishes in hoeroing situations which would in real life occour in news papers and goverment investigations on airline transport airplanes. please keep on goaing gents :ok:

[nitpicker330]

I've had the pleasure of Sims in the 737 747 and 777 and then base training in the Aircraft where we did multiple circuits with failures. I can tell you without a shadow of a doubt that in all cases the Aircraft was easier to fly. It is a lot more stable and as someone mentioned above the visuals are a lot better not to mention the motion you feel matches the picture your eyes see.

Remember in the real World the Earth ( horizon ) is fixed and we ( aircraft ) move around it. In a Simulator the opposite can occur, we ( Simulator ) are fixed and the Earth ( Horizon ) moves around us. With basic motion used to try and trick our inner ear into believing what we see.

However It doesn't matter how good the Sim is it cannot replicate motion completely. Some Pilots actually never handle the motion and can even get sick from it...
( your inner ear doesn't agree with your eyes )

However having said all of that, Sims are an invaluable tool in learning systems knowledge and procedures.


If you can handle the Sim you could handle the Aircraft, isn't that what's it supposed to be about ?

[b737NGyyc]

The most effective training method I have employed over the years with students who have issues over controlling the simulator on V1 cuts was a simple rhyme Squeeze, Freeze, 12 Degrees. This pitch initial attitude is applicable to Boeing 737 600 through 800 series.

Squeeze - the rudder to parallel and not necessarily regain centreline. If recognition is prompt enough the drift from centreline should be minimal and trying to correct back to centreline in the short timeframe between V1 and VR IMHO causes more issues of over controlling than it solves.

Freeze - Once the yaw is controlled and the aircraft is paralleling the centreline, freeze that rudder input. At rotation some aileron will be required to maintain wings level during initial climb out and at this time the aircraft may briefly cross-controlled. With the rudder input frozen it is much easier to finely adjust the aileron input to maintain wings level. This input should never exceed 5 degrees.

12 Degrees - Rotate smoothly and slowly (Boeing recommends 1½ - 2 degrees per second as opposed to the normal 3 degrees per second) to an initial target pitch of 12 degrees (10 degrees on the B737-200) using ADI/PFD raw data initially and not FD cues. Once the gear is retracted, with a positive rate of climb, and pitch is stabilized at at an attitude that results in V2 (don't chase the airspeed, allow it to stabilize and then make corrections) FD commands can be followed and fine adjustments can be made to rudder and aileron inputs to accomplish co-ordinated flight.

From my experience most over controlling occurs as a result of far too aggressive rudder inputs both on the runway and in the air and if treated in the same fashion as PIO on short final (freeze all inputs momentarily then make adjustments in a co-ordinated manner) V1 cuts in the simulator can be transformed from a very stressful event to a non event in fairly short order.

Once the basics have been mastered, then other complications such as weights, density altitude and varying CofG can be introduced.

[Rick777]

We shouldn't forget that the reason for sim training to successfully handle the real thing anytime and anywhere under the worst of circumstances in the airplane. I always thought sims were a little harder to fly than airplanes, but that is OK. If I can fly the sim then the plane is a piece of cake. If a guy can't fly the sim then I sure wouldn't want him in a plane where failure is not an option.

[ZFT]

It should be recognised that the only difference between a Level C and a Level D FFS is with Sound and Vibration (Motion buffet) data. The aero data standards they are built and certified to and the Visual system requirement should be the same.

[Skyglider]

"In another life we trained on the real B737 because a suitable simulator was unavailable. Training was on a long runway and included simulated engine failure at V1 conducted by pulling back one throttle and continuing the take-off. The aircraft was easy to control with no undue yawing or rolling."

"In 2006 I had the CP pull an engine to idle on me at V1 during base training. Very easy to control for real (A320). "

"On a different plane years ago i flew the OPC in the real plane because there was no sim available in the world, V1 cut, SE approach and go-around lot easier than sim."

"I've had the pleasure of Sims in the 737 747 and 777 and then base training in the Aircraft where we did multiple circuits with failures. I can tell you without a shadow of a doubt that in all cases the Aircraft was easier to fly"

Can it be that the real aircraft(s) was very light during base training? & 1 eng. pulled to idle, so still producing some thrust, not a real eng. fail. In the sim did you try reproduce the exact same conditions? surely it's easier in the sim with 1 eng. pulled to idle & light weight VS the "standard" V1 cut with the pax & fuel weight in normal day to day operation?

[BARKINGMAD]

In a nutshell, sims are mostly crap for reproducing handling qualities.

Great machine for practising for the awful day "it" happens, but apalling for giving you accurate "feel" for the plane.

Treat the sim like a gigantic expensive "Compass Test" for qualifying to fly the real aircraft and don't blame yourself for the not-so-accurate performance.

Interesting to note the varied causes for Airline total losses over the last half century, yet we're still obsessed with the "V1 cut" and other manoeuvres in the LPC/OPC menu rather than assessing some of the other HF decisions/actions losing lives and hulls.

I once asked why the roll control of a certain once popular aircraft could not be made more realistic and the answer was cost-the CAA would charge too much to test fly and approve the new more authentic feel, which could have been bolted/programmed in with minimal expense.

Perhaps we could ask WHO is actually test-flying and certifying these otherwise very faithful simulators, how much ACTUAL poling time do they have on type, their relationship to the manufacturers of both sim and aircraft and the commercial urgency to get the beast signed off and earning money?

Try allowing a low hours line-trained F/O to handle the real aircraft at 41,000 feet and observe the phugoids!

I dread being in the bog, trousers around my feet when the rookie has to control the trim runaway or A/P disconnect with NO experience of the "bungee" control effect at cruise altitude.

As I ricochet off the floor and the ceiling attempting to return to the flight deck I will curse those in the regulatory, supervisory and training empires who regard it as a MORTAL SIN to permit crews to hand-fly the aircraft to top of climb.

I would argue that pilots who have not REGULARLY hand flown the aircraft, at all corners of the normal operating spectrum, are not qualified so to do in the event of automatics failures. If to qualify I have to hand fly at flare height, then why not at cruise altitude?

Is there any connection here with contrbutory causes relating to AF447?:ugh:

[A37575]

In a nutshell, sims are mostly crap for reproducing handling qualities.


If you are serious about that (and I am sure you were jesting) why not write to the various simulator manufacturers and offer them your ideas why they are "mostly crap" and suggest a better solution?

[FullWings]

With all the simulators I've ever used, it was immediately obvious from the word go that you were in a sim not in a real aircraft, purely from the handling and feel. That said, I don't think it matters as you're training technique and procedure, not finesse - in fact, it can be counter-productive to concentrate too much on minutiae that may be artefacts of the sim rather than characteristics of the real airframe.

I remember alternating between two supposedly identical (certified to same level) sims during a conversion course and one always gave you a perfect touchdown no matter what you did while the other felt like a crash every time. The instructors were aware of this so were just looking at technique and assured you that it would be OK in the real thing.

[captjns]

I think the simulator is more realistic the jet itself:E

[Sunny Boyle]

For all type rated pilots who have physically FLOWN the aircraft, almost all KNOW that the simulator can NEVER replicate the precise handling characteristics of the actual aircraft.

However for some simulator instructors/checkers who have NEVER EVER flown the actual aircraft but get type rated only on the simulator ( like some Alteon geriatrics up in Korea ), the sim is 100% accurate! They make life very difficult for checkees with their anal retentive ways and intransigence.

[scotbill]

Big fan engines rarely have an instant cessation of thrust with failure - more like a progressive rundown.
Many years of TRE work showed me that some simulators react badly to over-controlling. That was particularly evident where pilots were converting from aircraft that needed an agricultural amount of boot input with critical engine failure. Some pilots also rely on feel much more than pure instrument readings.

The best piece of advice I ever heard on the subject came from a much-respected trainer who said.
"Merely apply sufficient rudder to stop the scenery from rotating round the aircraft!"

It works for simulators (assuming they have visuals) just as well as the aircraft