Terrific Beta Mustang but it is unstable: FM needs tweaking

[Bahger]

I do not criticise the experts at ED lightly so let me first say that I am very impressed with their Mustang so far.

 

The following observations are based on r/l experience in single-engine propeller aircraft, including tail-draggers. I feel the flight model for the DCS P-51 lacks essential stability in most flight regimes and requires too much constant manipulation of all control surfaces to attain minimal stability in climb, cruise and descent. Here are some specifics:

 

- Takeoff and climbout: Rudder torque is well modelled but very slow application of trim correction prevents smooth transition in yaw stability from climb to level flight.

 

- Trim lacks authority. I think this is by far the most difficult thing to implement in a sim, especially of an a/c like this, because trimming an aircraft is essentially a tactile experience. However, once power input is set and the a/c is at a constant airspeed in level flight, it has to be possible to trim off all control pressures. It isn't, yet, in this simulation. Trim input, especially elevator trim, never "nails" the selected attitude but merely reduces oscillation of the nose. This applies to climb and descent, too; it needs to be much easier, even in a high-powered, torquey a/c like this, to trim for a constant airapeed and attitude, to the point where the pilot can be hands-off.

 

- Power reduction induces way too much rudder departure on descent. In general, the leftward yaw departure on descent is somewhat exaggerated but the most noticeable problem in this regime is leftward massive yaw on reduction of power.

 

As I said, I love the a/c and undertand that, like all a/c, it has to be flown with finesse. However, this is all the more reason to attain minimal stability in all axes and flight regimes and to make sure that the trim tabs in all axes have real authority, or the Mustang is just too fussy and high-maintenenace to fly, let alone to fight. If I were a WW2 squadron commander I would not want to send my pilots out to fight in a fighter that requires such constant management of yaw, roll and -- particularly -- climb stability.

[G-Lock91]

That's one helluvan observation. I understood about 5% of that but it sounded REALLY GOOD :thumbup:

[71st_Mastiff]

Yes I have noticed too. Controls axis's don't translate well; it's in beta. I hope it improves this next patch. I can not get a clean A/C. :joystick:

[flightace37]

I think the trim problem is more an issue of trim accuracy and speed of trim. Perhaps an auto-trim button that we can hold down to cause the "virtual pilot" to trim to maintain existing flight conditions over a period of several seconds would do the trick. I don't think that could really be considered a cheat, but ED could put it under the "cheats" control section if necessary. It would be far more accurate than dialing in everything by hand using the digital controls, and faster to boot. As far as implementation goes, use a "spring" system, where the rate of change is proportional to the offset from the correct trim setting, with an upper bound. When the trim is within epsilon of correct, then just set it to the correct position to make sure it is as precise as possible.

 

Bahger, do you think that some of the torque behaviors are above and beyond what you would expect, given the fact that we're dealing with a much higher performance engine than is typically installed in civilian birds? Would some of the pitch stability problems you're feeling be explained away by a full auxiliary tank? I do know that the behavior improves dramatically as that tank empties and the CG moves forward to the correct position.

 

Have you tried taking off with a full load of 10xHVAR, along with as much fuel as permitted by the gross max T/O figures? While takeoff even with a 100% fuel load is quite doable (I can do it in a straight line with no cross wind), making a successful roll with that weapon load on board has proven to be almost impossible.

 

I wish we could get the opinion of a real P-51D pilot on the feel of the model.

 

I do have yet to try the .2 patch, so there may be improvements there.

Edited May 5, 2012 by flightace37

[Weta43]

I feel the flight model for the DCS P-51 lacks essential stability in most flight regimes and requires too much constant manipulation of all control surfaces to attain minimal stability in climb, cruise and descent.... it has to be possible to trim off all control pressures. It isn't, yet, in this simulation. Trim input, especially elevator trim, never "nails" the selected attitude but merely reduces oscillation of the nose. This applies to climb and descent, too; it needs to be much easier, even in a high-powered, torquey a/c like this, to trim for a constant airapeed and attitude, to the point where the pilot can be hands-off.

Somewhere on this forum (or on the web) you'll find a link to a video by a "Dr Wagner" the owner of a P-51D, where he specifically states that the P-51 never trims to where you can just stop flying and let it fly itself, however I can say that at constant speed, it is possible to trim the plane in game to fly hands off at constant speed, heading and altitude for 10 minutes or so at a time. Remember also that the movement of the trim by the game inputs (& the change from trimming via trim surfaces rather than the whole control surface, is a WIP.

 

Power reduction induces way too much rudder departure on descent. In general, the leftward yaw departure on descent is somewhat exaggerated but the most noticeable problem in this regime is leftward massive yaw on reduction of power.

 

As I said, I love the a/c and undertand that, like all a/c, it has to be flown with finesse. However, this is all the more reason to attain minimal stability in all axes and flight regimes and to make sure that the trim tabs in all axes have real authority, or the Mustang is just too fussy and high-maintenenace to fly, let alone to fight. If I were a WW2 squadron commander I would not want to send my pilots out to fight in a fighter that requires such constant management of yaw, roll and -- particularly -- climb stability.

 

That's what I'd have thought too, but it's not what people that have flown it say. See the attached docs, the first from an active F-16 pilot on the complete lack of yaw stability and need for constant trimming, the second a report by the US Navy on the P-51's non - suitability as a carrier borne aircraft, because of its poor lateral stability, sloppy low speed handling, and dangerous yaw & roll reaction on application of power at low speeds. (edit - sorry missed some of that from the snip)

Edited May 5, 2012 by Weta43

[Yo-Yo]

I must say that massive yaw (and pitch!) reaction is a intrinsic behaviour of the plane powered with the 1500 hp engine for only 22 sq. m of wing area.

 

If you read the WWII pilots' memoirs you will find that very often they say about trim. As far as I remember at least one of them mentioned the holes in the left glove due to constant trim inputs.

By the way, before beta was released I have many hours of constant consultation with our boss, Nick Grey, who has hundreds hours in P-51 and he spent a lot of hours flying DCS P-51 to check the model accuracy.

 

There are two main reasons that r/l pilots first impressions can cause this kind of reaction.

You are right that trim is too slow if you youse buttons or mouse to rotate the knobs. It's a compromise between trim accuracy and and I agree that there is a great difference between r/l tactile feelings and this kind of in-sim manipulations. I found a good way to have fast and precise trim using absolute axis for rudder and elevator trim as there is no reason to trim ailerons. I posted the hint how this axes must be configured to obtain fast and accurate trim. I think it must be in FAQ soon.

 

All in alll - it's only a problem of sim cockpit and not of the model itself.

The same problem is for the rudder pedals. Plastic sim pedals are not so stiff as in RL so the feel of pedals will be different and you have to control it in a different way...

 

The second main general sim problem - is a lack of acceleration info. In RL you get this info instantly via your body. In sim you have to calculate it using visual information that requires much more time and a lot of training.

[Yo-Yo]

And to clear things up: very often "stability" term is misunderstood. This term means A/C ability to maintain certain AoA and AoS (in short-time period) and its flight path slope (fugoid or long-period stability). The first two are compulsory and the last is desired.

 

The special part of short-period stability is dynamic stability. Once initiated oscillation in short-period movement must be dampen.

 

"Stability" does not mean that the plane flies exactly where the pilot wants because trim is a part of "controllability'. Of course if the plane is stable and trimmed properly it will fly in level flight with zero climb.

[SimFreak]

Yo-yo, in a level flight when I push stick aft, there is a slight yaw motion. But when I pull stick (for), there is nothing. Any reason for this?

[Bahger]

@Flightace37: Yes, the more powerful the engine, the more torque and the more of a handful these a/c can be, but basic stability in all axes and in all flight regimes has to be a prerequisite of a good airplane of any size and type. Yes, a powerful fighter or aerobatic propeller a/c with a big engine, big prop and stubby wings will bite you hard if you get too far behind it or take liberties with it but there has to be a baseline level of stability. In real life, an evenly-loaded airplane that's flying level and in trim will return to that condition after the pilot has, say, rolled into a turn and rolled out again, or climbed or descended the a/c, without adjusting trim or power settings. This P-51 is a long, long way from that basic capability and coordinated fight in it is like walking on eggshells. I take your point about different loads, but as long as it is within its published center-of-gravity limts, a heavily, or even asymmetrucally, loaded aircraft should be stable when properly trimmed.

 

@ Weta43, thank you for responding. I'm sure you're right about the elusiveness of the sweet-spot for trimming this beast in real life. However, there must be a way to harmonise the controls in flight in order to attain a greater degree of stability and in a sim without the benefit of tactile trim feedback, it might be necessary to make trim commands faster and more authoritative, or "mastering" the a/c will become a chore. You are absolutely right about the limitations of trimming by control sruface vs. trim tab; these are two different aerodynamic conditions and I'm hoping that the implementation of trim tabs for all control surfaces will make this airplane feel more stable, both in constant flight conditions and when transitioning.

 

Finally, I'm sure you're right about both the "complete lack of yaw stability and need for constant trimming" and the "poor lateral stability, sloppy low speed handling, and dangerous yaw & roll reaction on application of power at low speeds". These characteristics are certainly present and correct in the sim. The problem is that, given how difficult it is to simulate trim realistically, the combination of these limitations with those of the aircraft make it too much of a mule to fly with precision, except at high speeds, when almost any idiot can convince himself he's a pilot. If this Mustang is going to be both authentic and fun to fly, I suspect there will need to be more attention paid to harmonising and modulating all control surface commands, including trim via tabs, the object of which should be to make stable flight in all attitudes not exactly easy but attainable and sustainable.

Edited May 5, 2012 by Bahger

[Yo-Yo]

Yo-yo, in a level flight when I push stick aft, there is a slight yaw motion. But when I pull stick (for), there is nothing. Any reason for this?

 

The plane is less responsive for negative g so you have less angular velocity pitching down. Less velocity - less prop gyro-moment.

[159th_Viper]

Have you tried taking off with a full load of 10xHVAR, along with as much fuel as permitted by the gross max T/O figures? While takeoff even with a 100% fuel load is quite doable (I can do it in a straight line with no cross wind), making a successful roll with that weapon load on board has proven to be almost impossible.....

 

Fuel at 100% with 10 HVAR, cross-wind of 4m/s with a turbulence factor of .6, aircraft % weight 108%:

 

http://forums.eagle.ru/showpost.php?p=1448358&postcount=119

 

Edit: New track added - previous Beta version corrupt on playback.

 

Track: 108% rockets crosswind turbulencetake-off.trk

 

 

 

- Trim lacks authority. I think this is by far the most difficult thing to implement in a sim, especially of an a/c like this, because trimming an aircraft is essentially a tactile experience. However, once power input is set and the a/c is at a constant airspeed in level flight, it has to be possible to trim off all control pressures. It isn't, yet, in this simulation. Trim input, especially elevator trim, never "nails" the selected attitude but merely reduces oscillation of the nose. This applies to climb and descent, too; it needs to be much easier, even in a high-powered, torquey a/c like this, to trim for a constant airapeed and attitude, to the point where the pilot can be hands-off.....

 

Cannot agree.

 

I would submit that it is probably your controller settings that need tweaking. Herewith track where I take off with default 68% fuel load. Take-off at 12H00 and reach altitude at 12H02 where I go hands-free, configure for cruise and dial in my trim tabs.

 

At 12h04 my trim is dialled in and still hands-free, I coast along happily at 46 inHg/27RPM/320mph with very minor trim tab corrections when needed, which was not often. I carry on ambling along, completely hands-free, until 12h10 when I get bored. Track ends shortly thereafter. I cannot see how the aircraft trim can be any more stable: Any more and she'll be on rails.

 

Have a look at the track when you get the chance - I may have missed something/misinterpreted data/behaviour, entirely possible :)

 

Track: Trim.trk

Edited May 5, 2012 by 159th_Viper

[Vibora]

Yo-yo, in a level flight when I push stick aft, there is a slight yaw motion. But when I pull stick (for), there is nothing. Any reason for this?

What you are saying makes no sense, you always push forward and pull aft, at least in this universe, considered euclidean :).

[Vibora]

Fuel at 100% with 10 HVAR, cross-wind of 4m/s with a turbulence factor of .6, aircraft % weight 108%:

 

http://forums.eagle.ru/showpost.php?p=1448358&postcount=119

 

Edit: New track added - previous Beta version corrupt on playback.

 

Track: [ATTACH]65424[/ATTACH]

 

 

 

 

Cannot agree.

 

I would submit that it is probably your controller settings that need tweaking. Herewith track where I take off with default 68% fuel load. Take-off at 12H00 and reach altitude at 12H02 where I go hands-free, configure for cruise and dial in my trim tabs.

 

At 12h04 my trim is dialled in and still hands-free, I coast along happily at 46 inHg/27RPM/320mph with very minor trim tab corrections when needed, which was not often. I carry on ambling along, completely hands-free, until 12h10 when I get bored. Track ends shortly thereafter. I cannot see how the aircraft trim can be any more stable: Any more and she'll be on rails.

 

Have a look at the track when you get the chance - I may have missed something/misinterpreted data/behaviour, entirely possible :)

 

Track: [ATTACH]65423[/ATTACH]

EDIT: Disregard. I could see your trk finally.

Edited May 5, 2012 by Vibora

[HiJack]

I also trim successfully when configured for cruise. Works fine I think. A bit more difficult to be stable during dogfighting and A2G attacks but that must be just practise. I trim ailerons, elevators and rudder and I am using a Saitek X52 Pro stick.

 

(HJ)

Edited May 5, 2012 by HiJack

[Vibora]

Could those ones that consider they are able to trim successfully put here the devices they use as well as the settings?

Devices they use.

If they use HAT switches for trim and in that case for which trimmer.

If they use axis and in that case, which axis and which settings.

 

That could help other people to set their devices and analyze if there is something wrong with some type of devices.

For example, trimming pitch with a HOTAS Warthog is hard as it's too sensitive (each button press should move the trim wheel much less).

[HiJack]

Could those ones that consider they are able to trim successfully put here the devices they use as well as the settings?

Devices they use.

If they use HAT switches for trim and in that case for which trimmer.

If they use axis and in that case, which axis and which settings.

 

That could help other people to set their devices and analyze if there is something wrong with some type of devices.

For example, trimming pitch with a HOTAS Warthog is hard as it's too sensitive (each button press should move the trim wheel much less).

 

I made a guide for trim using the X-52 pro. I still use the same settings for aileron and elevator.

http://forums.eagle.ru/showthread.php?t=82773

 

(HJ)

[HiJack]

Now, I would like to see a track with sharp turns using a free tailwheel with the stick forward. Now do not come and tell me that it should not be used! There must be some sense to that madness also :D Only whay to do a sharp turn is to break one of the wheels with stick in nutreal, that is not correct acording to the manual. The tail wheel should be free when breaking with one of the front wheels.

 

(HJ)

[Bahger]

Could those ones that consider they are able to trim successfully put here the devices they use as well as the settings?

Devices they use.

If they use HAT switches for trim and in that case for which trimmer.

If they use axis and in that case, which axis and which settings.

 

That could help other people to set their devices and analyze if there is something wrong with some type of devices.

For example, trimming pitch with a HOTAS Warthog is hard as it's too sensitive (each button press should move the trim wheel much less).

I use a CH Flightstick and ProThrottle. I use the trim hat (or view hat) on the Flightstick for trim, up/down for elevator trim, left/right for rudder trim. It's possible that each press moves the trim too much, as I am always trimming through the desired value rather than nailing it. I do not have any spare axes so if trim cannot be operated effectively via a hat, my career as a virtual Mustang pilot might be a short one.

[Vibora]

I use a CH Flightstick and ProThrottle. I use the trim hat (or view hat) on the Flightstick for trim, up/down for elevator trim, left/right for rudder trim. It's possible that each press moves the trim too much, as I am always trimming through the desired value rather than nailing it. I do not have any spare axes so if trim cannot be operated effectively via a hat, my career as a virtual Mustang pilot might be a short one.

If you don't have any axis available then you'll have to live with the hat. Though you won't be able to trim perfectly, you can approximate it enough, imagine you are flying in rough air :) I don't think you should desperate because of that. In fact I prefer to use the hat instead of the axis trims, and that's what I usually use. You just need a perfect trim for a long range flight, which we don't use to fly in virtual flights.

In fact, I was asking ED to implement two kinds of trimmer keys, one as it is now and another one much less sensitive, so that you can fine tune the trim. I will talk to them again and ask for it once more.

[lawinge]

The pLane is impossibly unstable in yaw axis. At 200 mph it's possible to get the p51 to yaw back forth at an unbelievable angle with it taking quite a while to settle back down. That is serious instability. I think if an airplane could yaw that far at high speeds it would tear itself apart. I have mainly flown a piper cherokee for my nearly 600 hrs. as a private pilot so i don't really know how a p51 flies, but I believe it would have more stability than cherokee 140 especially at high speeds. That said, the p51 is a lot of fun to fly, although I would never get into it in real life.

[HiJack]

the p51 is a lot of fun to fly, although I would never get into it in real life.

I would! If even just to taxi down the runway :D

[cichlidfan]

I would! If even just to taxi down the runway :D

 

:thumbup::thumbup:

[aaron886]

Bahger I'm going to try tackling your post here, with the understanding that I have not spent THAT much time in DCS:Mustang yet. I'm a CMEL/CFI ASEL with no tailwheel yet but had a few experiences of my own in taildraggers and gliders that are a sort of strange mix of taildragger and unicycle. :D About 15 years in flight sims of all flavors.

 

 

 

I feel the flight model for the DCS P-51 lacks essential stability in most flight regimes

 

We'll start here. When you say stability I assume you're referring to eccentric dynamic stability. When you make an input, it feels like the flight model responds with "bounciness" that you have to cancel out. Like it's too sensitive. Sound right?

 

This may not be far off. In fairness, I don't feel like it's a "touchy" flight model at all, relatively speaking. It feels quite stable in pitch, with a relaxed positive static response. Consider that this is, for its time, a pure-bred fighter aircraft. The control surfaces were adequately large to provide a strong moment at high speeds... it won't fly much like the Piper Cub you may have flown. It's a slick, temperamental fighter... any WWII pilot's memoir will tell you all the fighters of that day had a mean streak if you weren't careful on the stick.

 

Speaking of stick, your particular desktop flight controls will never translate perfectly to a flight model. They don't give you tactile feedback, they're springy and generally too light, and their throw range is often limited by the fact that they're dinky side-sticks. Expect to need to tweak your axis curves carefully! (Very important!) I'd suggest starting with 25% curvature in pitch, 15 in roll, and 30 in yaw... then go from there. Too much and you'll have difficulty making fine control inputs on the steep part of the curve, too little and you'll struggle to make fine adjustments in more neutral control positions.

 

 

- Takeoff and climbout: Rudder torque is well modelled but very slow application of trim correction prevents smooth transition in yaw stability from climb to level flight.

 

Obviously in real life in the P-51 the pilot has a trim wheel he can probably move pretty quickly if he needs to, but that's not the proper use of trim. You've probably been taught this way... trim is designed to relieve control pressures, not replace stick inputs. You should be coordinating the aircraft with the primary flight controls during climb-out, not the trim. You may elect to set a particular takeoff trim setting to help alleviate initial left-turning tendencies after rotation, and maybe even make constant adjustments after takeoff, but that should not stop you from making a smooth transition at all. (Period.)

 

That said, it sure would be nice if we could define a "repetition" rate of an input button, much like it is possible to do in FSX. Very helpful for customizing trim rates.

 

- Trim lacks authority. I think this is by far the most difficult thing to implement in a sim, especially of an a/c like this, because trimming an aircraft is essentially a tactile experience.

 

Absolutely right about it being a tactile experience... but I don't agree with you about insufficient trim authority. I haven't been in a situation where I've needed more than ~1/3 of the total yaw trim available. Certainly not even close to that in pitch.

 

Trim input, especially elevator trim, never "nails" the selected attitude but merely reduces oscillation of the nose. This applies to climb and descent, too; it needs to be much easier, even in a high-powered, torquey a/c like this, to trim for a constant airapeed and attitude, to the point where the pilot can be hands-off.

 

Getting a little confusing here. Trim is not beholden to you the operator. :D It's not going to dampen oscillations for you... it's your job as the pilot to configure the aircraft such that it will maintain its attitude first, THEN trim. Even then, the aircraft is not going to truly fly "hands-off." Even the docile, stable aircraft you've probably flown won't do that. I spend most of my flight time in dinky Cessna and Piper aircraft, the biggest being a twin engine Seminole. None of the above will fly "hands-off" even if perfectly trim. They'll diverge. And they're not a Mustang, with further relaxed stability characteristics. It's on you to make fine pitch, power, and airspeed changes to maintain flight path. Trim just makes the inputs required a little smaller.

 

 

- Power reduction induces way too much rudder departure on descent. In general, the leftward yaw departure on descent is somewhat exaggerated but the most noticeable problem in this regime is leftward massive yaw on reduction of power.

 

On this one, I think you're off-base. (And I think you meant a strong RIGHT yawing component with power reduction) If you're drawing from your experiences with little 150hp taildraggers, you're going to be in trouble on this one. You're talking about a 1500 hp beast spinning a big fat heavy propeller. When you pull the power back you are dramatically changing forces on the aircraft. Aircraft like the Mustang are physically shaped to counteract left-turning tendencies with the expectation that most of the time, the pilot will be at a high power setting. Someone can correct me if I'm wrong, but I bet the Mustang even has its wings at slightly different angles of incidence to counteract torque. This means when that strong torque force isn't there, the airplane will want to yaw right and will require left rudder.

 

 

 

...the Mustang is just too fussy and high-maintenenace to fly, let alone to fight. If I were a WW2 squadron commander I would not want to send my pilots out to fight in a fighter that requires such constant management of yaw, roll and -- particularly -- climb stability.

 

As it should be. These aircraft were a true challenge for the pilot to master and control, even greater to employ to its potential. That's why they're every pilot's dream!

 

I just flew DCS:M 15 minutes ago and I was marveling at how great the flight model feels... absolutely ED's best work in flight modeling yet. Anyway, I understood your post and I don't want it to sound like I'm talking down at you, I just wanted to provide counterpoint because I think you've just had the wrong experience thus far, or maybe your controls aren't set up right. Give it a little time.

[lobo]

By the way, before beta was released I have many hours of constant consultation with our boss, Nick Grey, who has hundreds hours in P-51 and he spent a lot of hours flying DCS P-51 to check the model accuracy.

 

Being an aircraft that TFC owns and P-51 pilot Nick Grey has flown, I would think he would have very high demands for the P-51D flight model and simulation, which should translate into an incredibly realistic pc simulation experience for the DCS P-51D enthusiast. :thumbup:

 

The second main general sim problem - is a lack of acceleration info. In RL you get this info instantly via your body. In sim you have to calculate it using visual information that requires much more time and a lot of training.

 

I think you hit the nail on the head Yo Yo. IMO this is one of the reasons that flying a high fidelity simulation of a high performance tail dragger will always be (to a certain degree) more challenging when sitting in front of a monitor than when flying in RL. Specifically ground ops involving the tail wheel.

 

BTW, awesome simulation as usual Yo Yo. You guys truly rock. :pilotfly:

[EvilBivol-1]

Regarding the powerful torque effects seen when changing power settings at low speeds, I came across this article early in DCS Mustang development, which discusses this issue and the devastating effects in can have when mismanaged:

http://www.avweb.com/news/pelican/pelicans_perch_87_killer_go-arounds_195755-1.html

 

It's a long write-up, so to summarize - the evidence suggests that the pilot came down too hard on landing and abruptly increased power, probably intending to go around. However the burst of power at low speed flipped the aircraft over and landed it upside down, sadly killing the pilot.

 

Another site that talks about the same accident quotes Jeffery Ethell, himself a well known warbird pilot and aviation historian (killed in 1997 flying a P-38 ) saying this:

"...the most challenging part of flying the P-51 is a balked landing go-around. Pushing the throttle forward with gear and flaps down at low speed is just like losing an engine in a high-powered twin. There is a real minimum control speed (Vmc in twins) at which the aircraft will roll over on its back unless power is reduced or airspeed increased."