Post by volkerboehme on Jun 7, 2009 10:16:54 GMT -5
Hi,
His is a post of FsAviator on the Sim Outhouse Forum. It might be intersting to read for anyone interested in classic aircraft with tailwheel steering. The link to the original page is here:
www.sim-outhouse.com/sohforums/showthread.php?t=15432
Here we go:
***************************************************
The Kaydet was an FS8 release. You may be attempting to use FS8 flight dynamics within FS9. The flight dynamics I supplied within the official FS9 Stearman Kaydet update, very soon after the release of FS9, do not have that problem. They are dated 27th September 2003.
I suggest you download the FS9 update again and re-install only the FS9 aircraft.cfg, Stearman Kaydet.air and PT-17_ref.txt with that filedate.
I don't know what filenames are used for that update on different servers, or which still offer it. They are likely to differ. You need the one whose aircraft.cfg begins;
************************************************** ************************************************
Updated FS2004 flight dynamics for the Stearman N2S-3, PT-13A, PT-17 and PT-18 Kaydet by FSAviator
************************************************** ************************************************
If you have installed additional liveries since the original download you will need to copy them from your current aircraft.cfg so save it as aircraft.bad before installing the FS9 flight dyanmics update (again).
Before adding any liveries to the newly installed cfg make sure that every sim= line for every livery reads;
sim = Stearman Kaydet
FSAviator
*******************************************************
*******************************************************
<<Please, if you are able, provide me with the updates yourself. I'm having a bit of trouble locating them.>>
Unfortunately I was unaware that Dave had issued a 2005 official update to the Stearman with revised MDL and gauges. I now realise that my 2003 FS9 update for my original FS8 FD was replaced on all download sites in 2005 because its content is now incompatible with the current FS9 MDL and gauges. I would therefore be inappropriate to bring my aged 2003 version back into circulation by any means. That precludes solving the problem via simple re-install.
<<..I've got .(max steerable tail wheel angle) set at 40 degrees, and it seems to pivot enough for me. I wish we could give a real 180 castoring wheel, but it just doesn't wanna work with one.>>
I believe most real Stearmans had steerable tail wheels linked to rudder motion, but its not at all the point.
Since there are now two 'why won't my taildragger turn on the ground' threads running side by side, neither of which is coming close to grasping the issues involved, I will try to explain in detail how and where yaw propensity is encoded, and what (sim) pilots must do to induce small turning circles on the ground, in taildraggers with fixed skids or castoring tail wheels, and no brakes, after the flight dynamics contain the necessary code.
Towards the end of this long post I will explain that *steering* is a completely different concept to *yawing* and why failure to grasp and apply the fundamental difference between yawing a vehicle with a rudder and steering a vehicle with a steerable wheel is at the root of all the pointless confusion in these threads.
Many real small taildraggers have only fixed skids or castoring tail wheels and many have no brakes. Even if small taildraggers have a steering mechanism or brakes they are ancillary to the process of yaw control. Wheel steering and brakes are not required to yaw aeroplanes either in real life, or within FS9. Such aircraft are turned by carefully incrementing prop wash across the fully deflected rudder which immediately yaws the tail, even if the thing underneath the fuselage is a fixed tail skid. Even on a soft high friction surface.
Yaw propensity due to propwash is encoded within the air file not the aircraft.cfg. Consequently as everyone has discovered messing about with max tailwheel angles in an aircraft.cfg does not even address the issue. High tail wheel deflection *limits* do not *cause* rapid yaw rates. If the air file has inadequate yaw propensity due to prop wash across the rudder the taildragger will not turn with a realistic turning circle on the ground, regardless of maximum tail wheel angle, and it will not turn realistically anyway if the (sim) pilot fails to generate exactly the *correct* RPM to induce a small turning circle.
I addressed this in detail and most recently within the Ansaldo SVA5/9 release and its 'how to fly the ..' notes last summer thus;
*************************
Ground handling:
This aircraft has no brakes. Hold the stick aft to dig the skid into the surface to slow down. Fish tail left and right with the rudder to scrape the skid (several feet in each direction) across the surface at low speed for maximum braking...... To turn tightly at low speed first create propwash by setting 800 RPM then deflect the rudder fully into the propwash whilst holding the stick forward or neutral to unload the skid. The aircraft will 'rudder round' whilst creeping forward very slowly. On a soft surface it will turn tightly. Remember the rudder will be ineffective with inadequate propwash. If you need to turn to avoid an obstacle during the landing roll or at low IAS you may need to create 800 RPM of propwash to generate an adequate turn rate......
************************
Co-efficient of surface friction is encoded by BGL authors. Most taildraggers do not belong on hard low friction surfaces. A higher rate of yaw is induced on e.g. smooth tarmac, but the propwash needed to induce that yaw causes instantaneous forward acceleration on tarmac; which is unwanted and increases turning circle.
Tyre rolling resistance versus the BGL co-efficient of friction is encoded by FD authors. Small turning circles in taildraggers with no brakes and no wheel steering require both a high friction surface such as grass or dirt *and* high vehicle rolling resistance. Loading and unloading the skid (or castoring tailwheel) with elevator is a key part of the friction control process. (Sim) pilot failure to unload the skid or tailwheel reduces yaw rate and may preclude yaw altogether at RPM values which are consistent with only creeping forward whilst inducing yaw. The ability to load and unload the skid, or a castoring tailwheel, (elevator response to propwash), is also encoded in the air file, not the cfg file.
The Ansaldo SVA 5/9 demonstrates all this better than the Stearman Kaydet because it has a skid and so the Ansaldo makes a better taildragging ground handling trainer, but the sim pilot skills carry across to castoring tailwheels once learned with a skid in the Ansaldo.
If taildraggers within FS9 will not yaw, with the tail unloaded, and with ‘adequate’ propwash (RPM) over the rudder, it is because the ‘yaw propensity due to prop wash’ value in the air file is inadequate. Different air file editors and different versions of the same editor will assign different names to that data field, but they should all identify it as hex location 0x08fe within REC 1101 of the air file and the value can be increased using any air file editor. The elevator response to propwash which may need to be increased to unload the rear contact point to promote yaw on the ground is also within REC 1101 at hex location 0x08c4. Increase the value of 0x08fe slowly until it delivers realistic yaw rates versus *modest* prop wash with the stick held full forward to unload the tail. I would expect the values for most taildraggers to be between 0.015 and 0.03 as encoded within the Ansaldo SVA5.
I suggest anyone interested in understanding this topic begin by testing the ground handling of the Ansaldo (noting that 800 RPM delivers optimal pilot control over skid friction and yaw rate on grass) and by studying its REC 1101 with an air file editor. The Cub (other thread) produces less propwash than an Ansaldo or Kaydet and may need higher (0x8fe yaw propensity) values. It is anyway just a question of altering those air file values empirically using an air file editor until the ground handling matches the Ansaldo, or the developers personal experience of real taildragger ground handling if they have real world experience of that particular aircraft type.
The Ansaldo SVA5/9 can be downloaded for study from Avsim. The file names are Sva5_v01.zip and sva_5fix.zip.
The aircraft.cfg has only limited relevance to this topic. However inadequate tyre rolling resistance also precludes a ‘realistic’ turning circle because the modest and carefully (sim) pilot moderated prop wash needed to yaw the aeroplane in real life then induces excessive and immediate forward acceleration instead of just instantaneous yaw of the carefully unloaded tail.
The friction of skids and rolling resistance of tyres can be increased simply by having many, so within the Ansaldo cfg we see,
point.0 = 1, -16.9 , 0 , -2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0
point.1 = 1, 1.45, -2.9, -5.6 , 1200, 0, 1.15, 0, 0.2, 1.7, 0.8, 0, 0, 2
point.2 = 1, 1.45, 2.9, -5.6 , 1200, 0, 1.15, 0, 0.2, 1.7, 0.8, 0, 0, 3
point.3 = 3, -16.92 , 0 ,-2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0 <<<<<<<
point.4 = 3, -16.92 , 0 ,-2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0 <<<<<<<
point.5 = 3, -16.92 , 0 ,-2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0 <<<<<<<
point.6 = 2, -0.093,-12.6, -2.2, 1200, 0, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0
point.7 = 2, -0.092, 12.6, -2.2, 1200, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0, 0
point.8 = 2, -16.9 , 0, -2.4, 9900, 0, 0, 0, 0, 0, 0, 0, 0, 9, 0, 0
point.9 = 2, 7.0901, 0, -2.77, 1200, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0
This delivers ‘realistic’ dynamic friction braking on soft surfaces via (sim) pilot load and unload with joystick of that contact point in aeroplanes with no brakes. The locked tailwheel (point.0) is present in the dynamics mostly to drive certain animation effects that ‘belong to tailwheels’. We see them when the pilot loads the tail with stick and pulls the invisible tailwheel into ground contact with prop wash (or ground rolling IAS) over up elevator. When we pull the invisible locked tail wheel into contact yaw friction increases allowing fishtail braking.
It is probably better to use a flat helicopter type skid of moderated surface contact length at the desired location to moderate tyre / skid (variable tyre pressure and surface area - such as tundra tyre) friction instead, but cfg values are used to model undercarriage co-efficient of drag, (friction = tyre rolling resistance), not propensity to yaw versus RPM applied by the (sim) pilot. The co-efficient of friction (rolling resistance of the vehicle) is encoded by the FD author, but the applied friction must be controlled by the (sim) pilot with carefully moderated prop wash and carefully moderated elevator deflection, carefully loading and unloading the tail contact point to dig it into the grass/dirt (or not) as required. Sometimes we want the tail contact point to ‘grip’, and sometimes we don’t. Unless we have substantial rolling IAS we have to control its grip with prop wash as well as elevator.
Incorrect developer evaluation of yaw inertia also plays a vital role in this particular process of course, but in tiny aeroplanes it is unlikely to be the primary design time error since even exaggerated values tend to be too low to preclude adequate yaw propensity.
In relation to both current threads it is vital to grasp that this is not just a ‘wrong values encoded in software’ issue as posters to forums always seem to assume. It is very much a handling skill that has to be learned and applied correctly by (sim) pilots. Many sets of flight dynamics are reported as broken in forums when the real problem is just pilot error, (failure to learn a particular skill).
Just as in real life taildraggers with skids and / or no brakes do not belong on hard surfaces and just as in real life there is always a carefully pilot moderated RPM needed to yaw the tail rapidly *without inducing unwanted forward acceleration* versus the co-efficient of friction of the current BGL surface under the contact points. It is also necessary to unload (the friction or rolling resistance of) the contact point we intend to yaw with prop wash (using the joystick to apply down elevator into the carefully moderated prop wash) if we hope to achieve ‘realistic’ turning circles in small tail draggers.
The Ansaldo demonstrates the consequence of (sim) pilot failure to moderate RPM, and (sim) pilot failure to moderate tail loading, to minimise turning circle. The rudder is always fully deflected when we intend to yaw such aircraft on the ground since we will always minimise the RPM and maximise our rudder deflection on any surface type since we always wish to induce maximum yaw rate with minimum vehicle acceleration to minimise turning circle.
Deflecting the rudder to its stop is however just the simplest part of a four part learned skill.
1) Apply full rudder
2) Evaluate surface hardness and friction
3) Moderate RPM (prop wash) accordingly
4) Moderate unloading of tail with elevator accordingly
Realistic tail dragger dynamics will deliver ‘poor’ turning circles on hard surfaces, but skilful (sim) pilots will be able to achieve tight turning circles on soft surfaces. Everyone should evaluate whether they are deploying the required skill set before fiddling with the dynamics of any download.
Altering maximum tail wheel angles in an aircraft.cfg cannot induce yaw propensity at zero IAS, nor will it allow the sim pilot to unload the tail contact point to increase or reduce its rolling and yawing resistance. The aircraft.cfg controls very little of the dynamics. The original posters in both relevant threads may have understood that of course. If there really is a ground handling dynamics error in either aircraft cited by the original posters it can be fixed, but not by fiddling with cfg files. This topic depends almost entirely on developer and consumer understanding of how to encode and moderate real world prop wash effects to induce yaw without inducing forward acceleration. Maximum tail wheel angle is irrelevant. In real life and in FS9 it is often zero (fixed skid).
The reason fixed skids are common on real aircraft designed to operate from soft surfaces is because they give excellent control over load/unload of the forward and side force friction of that contact point on soft surfaces. It is an error to remove that fine control from the simulation of such aircraft and aircraft with steerable tailwheels may need even more skilled control of prop wash, not less. They are more likely to accelerate when the (sim) pilot adds prop wash and may just skid the tailwheel unless the (sim) pilot takes great care to load it adequately and *steer* it very carefully.
That opens up a whole other can of worms since many / most (sim) pilots completely fail to grasp the difference between *steering* a vehicle versus *yawing* a vehicle.
In the real world when we ride a bicycle or motor cycle we do not assume that max steering wheel angle = max turn rate at every vehicle speed. We realise that we need large steerable wheel angles at slow speed, but ever smaller steering angles as speed increases, else we will skid the steerable wheel and lose directional control. Most sim pilots fail to use any common sense when trying to *steer* aeroplanes within a detailed ground handling simulator. They continue to treat vehicles with steerable wheels as though they only had rudders. They over control as though attempting to induce yaw with rudder because they use the same hardware input device in the simulator. Sim pilots must come to terms with the fact that sometimes those pedals or twist grip are a rudder control and sometimes they are a steerable wheel control.
*The difference matters.*
Steering with wheels and yawing with rudder requires different skills and most FS9 users fail to deploy either skill correctly. They lose control of their energy state on the ground even more readily than they do in flight.
That is why we see so many of these daft threads in which people propose one daft change to *maximum* steering angle after another. This nonsense will continue across multiple forums until flight simulation users grasp the difference between steering a vehicle with a steerable wheel, *carefully avoiding significant % rudder deflection to avoid yaw = skid) and the totally different concept of yawing a vehicle with max rudder. If the vehicle has a steerable wheel our job is to *match* the steering angle versus current vehicle speed just as though we were riding a bicycle. Most sim pilots just cannot grasp that and behave as though excessive applied steering angles will not immediately cause and perpetuate skid and increase of turning circle, so they whine that the cfg values are broken.
*They cannot be broken*
70 degree (max steering) lock may be true and required to move away from rest in a tight parking bay, but if any sim pilot subsequently loses control and causes skid it is not encoding the real max steering lock that causes the skid and loss of control. It is sim pilot error and only sim pilot error. If we fall off a bicycle it is not because the designer created a front wheel that can turn to 90 degrees, it is because we were stupid enough to suppose we would not cause loss of control if we applied a large % steering angle at anything above 1 knot.
The vast majority of sim pilots apply far too large a steering wheel angle to the steerable wheel for their current vehicle speed and simply cause the nosewheel or tailwheel to skid over the BGL, every single time they try to turn. They behave as though FS9 is incapable of modelling skid due to pilot error. FS9 actually models skid due to pilot error just fine and most sim pilots skid the steerable wheel everywhere, and almost all of the time, with the steerable wheel turned to wildly excessive lock, skidding over the BGL, delivering little or no steering because the user has induced and perpetuated steerable wheel skid.
Consequently many users gain additional control by halving or quartering the maximum steering lock in the cfg, not because the encoded maximum was wrong, but because reducing the maximum angle reduces their ability to induce loss of control via perpetuated skid with wild oversteer angles versus current vehicle speed.
Many sim pilots avoid inducing and perpetuating skid in any turn only by imposing upon themselves a tiny maximum wheel lock that will be useless at low vehicle speed. Because they refuse to learn how to *steer* and pretend that all aeroplanes are *yawed* they spend their lives meddling with max values in cfgs going from max values that cause loss of control when they stupidly apply them at high speed to lesser max values which prevent small turning cirles at low vehicle speed.
Controlling the turning circle of any vehicle is all about deploying the relevant and applicable skill, not cfg content. It is all about causing or preventing skidding. Skids will skid, and wheels will skid when turned to any inappropriate angle versus current vehicle speed. We simply choose whether to make them grip or not. We all get by perfectly well in the real world without differential braking in real automobiles and with max steering lock angles that would cause skid if deployed at any significant speed (even 10 KTS). We don’t demand that the max steering lock of our car be reduced so we can never skid because we realise we need a large max steering angle to park and unpark.
This topic does not require a grasp of advanced rocket science. It’s just common sense.
Steering is not yawing. Yawing an aeroplane requires maximum rudder deflection *to deliberately induce skid*, but preventing skid of a steerable wheel requires small applied rudder angles even at modest taxi speeds.
In a real automobile we get tactile feedback through the hardware device known as steering wheel that we are oversteering in a turn. MSFS fails to provide that feedback cue and most sim pilots skid the steerable wheel into and around most turns achieving a very poor turn radius, *because they failed to steer*. They apply wheel angles that would only be appropriate if a steerable wheel was some kind of rudder used to induce yaw = skid. A steerable wheel is not a strange variety of rudder. It is the opposite of a rudder. A steerable wheel exists to *avoid* yaw = skid. Large deflections of the steerable wheel do not induce small turning circles. They induce skid of the steerable wheel, and cause large turning circles because they degrade its effectiveness.
Judging the steering angle that induces oversteer in FS9 without tactile feedback is difficult, but it is just common sense. There is no excuse for sim pilots believing and pretending that no one can steer a bicycle unless the front wheel is prevented from adopting a large angle at low speed. Altering the maximum values of steering wheels in aircraft.cfgs is not the solution. It really is time that the whole FS community got a grip on this and stopped pretending that the dynamics are broken because they allow large steerable wheel angles, or that the loss of control many sim pilots experience when turning is due to anything other than skids they have induced by confusing steering with yawing and applying massive oversteer angles to the steerable wheel.
Many sim pilots induce wheel skid even at very low speed by applying massive steering angles.
In a taildragger sometimes we want a contact point to skid and sometimes we don’t. It is *not* the max wheel lock encoded in the cfg that controls the wheel loading or the yaw we can induce. As the Ansaldo demonstrates a fixed skid is no impediment to small turning circles and if you replace it with a castoring tail wheel in the cfg it will just make fine control of the friction of that contact point by the sim pilot more difficult. If you substitute a steerable tail wheel with a small maximum lock all that happens is that it becomes a skid once (often before) it reaches that maximum lock. Everbody has to stop pretending that ground handling of aeroplanes is not all about carefully avoiding or inducing skidding of the ‘third contact point’ by careful moderation of thrust and deployed steerable wheel angles versus current vehicle speed.
Otherwise FS forums will continue to be full of people insisting that it is impossible to ride a real bicycle and telling one another that one particular max angle of the steerable wheel is ‘correct’ or ‘best’ or ‘essential‘. The very idea is nonsense.
Reducing max wheel lock in FD does not prevent loss of directional control at much smaller wheel angles all of which constitute oversteer at most vehicle speeds, and increasing it does nothing to induce yaw when the speed is zero or minimal.
The sorry truth is that feedback of sim pilot oversteer error in MSFS is so poor, that it almost always goes undetected as pilot error by the sim pilot and the dynamics are blamed for what was simple pilot error even though the dynamics detected the pilot error correctly and imposed the correct consequence of the induced skidding of the steerable wheel across the BGL. Directional control is correctly lost. It is lost most easily on hard surfaces.
Once a sim pilot has induced skid the only way to turn the vehicle is yaw by rudder and if the vehicle in question was designed to be steered, not yawed, then the residual yaw induced with the rudder is tiny. It’s no good blaming the dynamics. They are doing their job of evaluating and imposing the consequence of pilot error.
Even 10 degrees of applied steering angle is enough to cause skid of the steerable wheel at very modest speed whether the vehicle in question is a bicycle or an aeroplane. The sorry truth is that many / most sim pilots don’t even attempt to steer aeroplanes defined as having steering. They dumbly attempt to yaw them with massive rudder defection instead (skid the steerable wheel sideways with the rudder) and when they succeed in turning the third contact point into a skid with no steering and lose directional control they claim the dynamics are broken.
Sim pilots must think harder about whether they are trying to steer or yaw the vehicle during simulation. The techniques are opposed techniques. FS9 supports both, but many / most sim pilots just confuse the concept of yawing (applying max rudder angle to deliberately skid the third contact point) with the concept of steering (never applying significant % yaw axis control deflections to always avoid skidding the third contact point) .
FSAviator
*************************************************************
*************************************************************
<<I still can't quite figure out why the Stearman's dynamics simply don't respond to changing the tailwheel to 180>>
They do. You have successfully removed all ability to steer the tailwheel. After you unload and reload the aircraft, you will have full yaw control via the rudder and no steering control via the tailwheel. Without propwash the yaw propensity versus (full) rudder will vary as a function of current IAS. As I explained the issue is not about tailwheels. It is whether the air file is allowing you to substitute propwash for IAS at zero IAS.
In my second reply I indicated the likely value required in REC 1101 field 0x08fe. The value present in the air file you are using is 15 times smaller than the proposed upper bound. Consequently you are unable to induce yaw with propwash. Doing what works in a real Stearman (or Cessna 120) will not work until realistic values are present in the air file.
<<Maybe the friction values to have something to do with it. I'll check into it more later on. >>
Encoded friction (tyre rolling resistance) is only an issue in aeroplanes without differential brakes. You intend to prevent inner wheel rotation with differential braking rendering rolling tyre resistance irrelevant. Since you intend to implement differential brakes you must obviously remove the cfg code;
differential_braking_scale = 0
If after you add differential braking by removing that code the locked inner wheel is seen to be skidding at the applied RPM you have applied excessive RPM. Each BGL friction is very likely to be different to whatever real surfaces you are used to operating on. You just have to accept that you are on a different surface in a different place and moderate RPM accordingly.
The reason it will not turn at low speed in FS9, with or without a castoring tailwheel, is that yaw propensity due to propwash is inadequate, and is nothing to do with tailwheels or brakes, as you keep proving with your tests after you vary them. The air file requires a yaw propensity in 0x08fe which will yaw the tail round with rudder fully deflected at an RPM that will not skid the (locked or unlocked) inside mainwheel on any BGL surface except ice. Once that value is augmented to adequacy the tail will yaw at low speed with a fixed skid and no brakes at all, or any other combination of contact point code and braking in the aircraft.cfg, because those variables only have the potential to limit what the air file induces.
FSAviator
********************************************************************
********************************************************************
<<Is it possible for a tinkerer with limited ability (like myself ) to implement this fix? >>
Yes.
Two different easy to use, but difficult to fully understand and configure, freeware air file editors are available from e.g.,
pagesperso-orange.fr/hsors/fsairfile.html
For this simple task Air Ed 1.52 is your best bet. The page says you need an updated aired.ini, but you won't to simply edit propwash effects. Leave all the files as they are and wherever you install the aired152 folder. Simply double click the supplied AirEd.exe, then navigate to and open the air file you are using. Then find and expand Record 1101.
Look down the long list of FD data fields in REC 1101 for Cn_dT Yaw Moment
What I believe to be the latest official version of the Stearman FD (FS04 PT-17 Ver 1.1.6 Apr 05) has this set to 0.002.
Look in REC 1 of the air file after you open it. If you are using a different air file the solution is 'likely' to be identical, but of course a different (unauthorised?) air file could have many problems.
Consider increasing the value incrementally until you can yaw the tail with fully deflected rudder into 'modest' propwash at very low taxi IAS with the stick held forward to unload the tail. If you overcook the value you will induce wild yaw authority at high thrust making co-ordinated turns in climbing flight and aerobatics harder to control; so only increase the value by just enough to solve the ground handling problem. Try increments of 0.005 and test each time.
If you define a personal key combination in FS9 (options/controls/assignemnts/Reload user aircraft) which causes air file reload you can then use that key combinbation after each edit and test each increment until you have just adequate yaw authority during very slow taxi. There 'should' be a higher value which makes differential braking irrelevant.
As you know the tailwheel animation may not behave as it would in real life whatever you code in the dynamics. Since the real Stearman had tailwheel steering (with disconnect to castor at unknown angle) I suggest that you implement that capability using no more than 45 degrees max in contact point.0 rather than implementing castoring and no tailwheel steering during your tests.
However......
Nothing you do will change the MDL tailwheel presence, so consider making point.0 a type 3 contact point = steel tipped skid.
point.0= 3, -17.3, 0, -6.2, 1200, 0, 0.3, max angle, 0.250, 2.5, 0.85, 0, 0, 0, 0, 0
Test whether defining it as a steel skid with both 0 and then 180 as max angle makes the dynamics behave more like a castoring tailwheel than actually defining the friction of that aft contact point as a rubber tyre.
The real Stearman had main wheel brakes. I still do not know if they could be applied differentially as delivered. That is not a given at the relevant date.
Remember the rear contact point has more friction with both cockpits occupied and yawing will require more RPM causing a larger turning circle. You should test on both grass and dirt, not just hard surfaces. Use MS default BGLs, not third party ones, during all FD ground handling tests.
Finally to all readers, please remember that many reported FD errors do not exist and the real problem is faulty user technique or expectation. Inexperienced tinkering with air files usually does more harm than good. The only reason I replied to this thread was that I believed the original question related to my original September 2003 FS9 update FD for this FS8 aircraft. I do not endorse generic air file tinkering by anyone who does not understand flight dyanamics.
FSAviator
His is a post of FsAviator on the Sim Outhouse Forum. It might be intersting to read for anyone interested in classic aircraft with tailwheel steering. The link to the original page is here:
www.sim-outhouse.com/sohforums/showthread.php?t=15432
Here we go:
***************************************************
The Kaydet was an FS8 release. You may be attempting to use FS8 flight dynamics within FS9. The flight dynamics I supplied within the official FS9 Stearman Kaydet update, very soon after the release of FS9, do not have that problem. They are dated 27th September 2003.
I suggest you download the FS9 update again and re-install only the FS9 aircraft.cfg, Stearman Kaydet.air and PT-17_ref.txt with that filedate.
I don't know what filenames are used for that update on different servers, or which still offer it. They are likely to differ. You need the one whose aircraft.cfg begins;
************************************************** ************************************************
Updated FS2004 flight dynamics for the Stearman N2S-3, PT-13A, PT-17 and PT-18 Kaydet by FSAviator
************************************************** ************************************************
If you have installed additional liveries since the original download you will need to copy them from your current aircraft.cfg so save it as aircraft.bad before installing the FS9 flight dyanmics update (again).
Before adding any liveries to the newly installed cfg make sure that every sim= line for every livery reads;
sim = Stearman Kaydet
FSAviator
*******************************************************
*******************************************************
<<Please, if you are able, provide me with the updates yourself. I'm having a bit of trouble locating them.>>
Unfortunately I was unaware that Dave had issued a 2005 official update to the Stearman with revised MDL and gauges. I now realise that my 2003 FS9 update for my original FS8 FD was replaced on all download sites in 2005 because its content is now incompatible with the current FS9 MDL and gauges. I would therefore be inappropriate to bring my aged 2003 version back into circulation by any means. That precludes solving the problem via simple re-install.
<<..I've got .(max steerable tail wheel angle) set at 40 degrees, and it seems to pivot enough for me. I wish we could give a real 180 castoring wheel, but it just doesn't wanna work with one.>>
I believe most real Stearmans had steerable tail wheels linked to rudder motion, but its not at all the point.
Since there are now two 'why won't my taildragger turn on the ground' threads running side by side, neither of which is coming close to grasping the issues involved, I will try to explain in detail how and where yaw propensity is encoded, and what (sim) pilots must do to induce small turning circles on the ground, in taildraggers with fixed skids or castoring tail wheels, and no brakes, after the flight dynamics contain the necessary code.
Towards the end of this long post I will explain that *steering* is a completely different concept to *yawing* and why failure to grasp and apply the fundamental difference between yawing a vehicle with a rudder and steering a vehicle with a steerable wheel is at the root of all the pointless confusion in these threads.
Many real small taildraggers have only fixed skids or castoring tail wheels and many have no brakes. Even if small taildraggers have a steering mechanism or brakes they are ancillary to the process of yaw control. Wheel steering and brakes are not required to yaw aeroplanes either in real life, or within FS9. Such aircraft are turned by carefully incrementing prop wash across the fully deflected rudder which immediately yaws the tail, even if the thing underneath the fuselage is a fixed tail skid. Even on a soft high friction surface.
Yaw propensity due to propwash is encoded within the air file not the aircraft.cfg. Consequently as everyone has discovered messing about with max tailwheel angles in an aircraft.cfg does not even address the issue. High tail wheel deflection *limits* do not *cause* rapid yaw rates. If the air file has inadequate yaw propensity due to prop wash across the rudder the taildragger will not turn with a realistic turning circle on the ground, regardless of maximum tail wheel angle, and it will not turn realistically anyway if the (sim) pilot fails to generate exactly the *correct* RPM to induce a small turning circle.
I addressed this in detail and most recently within the Ansaldo SVA5/9 release and its 'how to fly the ..' notes last summer thus;
*************************
Ground handling:
This aircraft has no brakes. Hold the stick aft to dig the skid into the surface to slow down. Fish tail left and right with the rudder to scrape the skid (several feet in each direction) across the surface at low speed for maximum braking...... To turn tightly at low speed first create propwash by setting 800 RPM then deflect the rudder fully into the propwash whilst holding the stick forward or neutral to unload the skid. The aircraft will 'rudder round' whilst creeping forward very slowly. On a soft surface it will turn tightly. Remember the rudder will be ineffective with inadequate propwash. If you need to turn to avoid an obstacle during the landing roll or at low IAS you may need to create 800 RPM of propwash to generate an adequate turn rate......
************************
Co-efficient of surface friction is encoded by BGL authors. Most taildraggers do not belong on hard low friction surfaces. A higher rate of yaw is induced on e.g. smooth tarmac, but the propwash needed to induce that yaw causes instantaneous forward acceleration on tarmac; which is unwanted and increases turning circle.
Tyre rolling resistance versus the BGL co-efficient of friction is encoded by FD authors. Small turning circles in taildraggers with no brakes and no wheel steering require both a high friction surface such as grass or dirt *and* high vehicle rolling resistance. Loading and unloading the skid (or castoring tailwheel) with elevator is a key part of the friction control process. (Sim) pilot failure to unload the skid or tailwheel reduces yaw rate and may preclude yaw altogether at RPM values which are consistent with only creeping forward whilst inducing yaw. The ability to load and unload the skid, or a castoring tailwheel, (elevator response to propwash), is also encoded in the air file, not the cfg file.
The Ansaldo SVA 5/9 demonstrates all this better than the Stearman Kaydet because it has a skid and so the Ansaldo makes a better taildragging ground handling trainer, but the sim pilot skills carry across to castoring tailwheels once learned with a skid in the Ansaldo.
If taildraggers within FS9 will not yaw, with the tail unloaded, and with ‘adequate’ propwash (RPM) over the rudder, it is because the ‘yaw propensity due to prop wash’ value in the air file is inadequate. Different air file editors and different versions of the same editor will assign different names to that data field, but they should all identify it as hex location 0x08fe within REC 1101 of the air file and the value can be increased using any air file editor. The elevator response to propwash which may need to be increased to unload the rear contact point to promote yaw on the ground is also within REC 1101 at hex location 0x08c4. Increase the value of 0x08fe slowly until it delivers realistic yaw rates versus *modest* prop wash with the stick held full forward to unload the tail. I would expect the values for most taildraggers to be between 0.015 and 0.03 as encoded within the Ansaldo SVA5.
I suggest anyone interested in understanding this topic begin by testing the ground handling of the Ansaldo (noting that 800 RPM delivers optimal pilot control over skid friction and yaw rate on grass) and by studying its REC 1101 with an air file editor. The Cub (other thread) produces less propwash than an Ansaldo or Kaydet and may need higher (0x8fe yaw propensity) values. It is anyway just a question of altering those air file values empirically using an air file editor until the ground handling matches the Ansaldo, or the developers personal experience of real taildragger ground handling if they have real world experience of that particular aircraft type.
The Ansaldo SVA5/9 can be downloaded for study from Avsim. The file names are Sva5_v01.zip and sva_5fix.zip.
The aircraft.cfg has only limited relevance to this topic. However inadequate tyre rolling resistance also precludes a ‘realistic’ turning circle because the modest and carefully (sim) pilot moderated prop wash needed to yaw the aeroplane in real life then induces excessive and immediate forward acceleration instead of just instantaneous yaw of the carefully unloaded tail.
The friction of skids and rolling resistance of tyres can be increased simply by having many, so within the Ansaldo cfg we see,
point.0 = 1, -16.9 , 0 , -2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0
point.1 = 1, 1.45, -2.9, -5.6 , 1200, 0, 1.15, 0, 0.2, 1.7, 0.8, 0, 0, 2
point.2 = 1, 1.45, 2.9, -5.6 , 1200, 0, 1.15, 0, 0.2, 1.7, 0.8, 0, 0, 3
point.3 = 3, -16.92 , 0 ,-2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0 <<<<<<<
point.4 = 3, -16.92 , 0 ,-2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0 <<<<<<<
point.5 = 3, -16.92 , 0 ,-2.4, 800, 0, 0 , 0, 0.1, 1.7, 0.7, 0, 0, 0 <<<<<<<
point.6 = 2, -0.093,-12.6, -2.2, 1200, 0, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0
point.7 = 2, -0.092, 12.6, -2.2, 1200, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0, 0
point.8 = 2, -16.9 , 0, -2.4, 9900, 0, 0, 0, 0, 0, 0, 0, 0, 9, 0, 0
point.9 = 2, 7.0901, 0, -2.77, 1200, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0
This delivers ‘realistic’ dynamic friction braking on soft surfaces via (sim) pilot load and unload with joystick of that contact point in aeroplanes with no brakes. The locked tailwheel (point.0) is present in the dynamics mostly to drive certain animation effects that ‘belong to tailwheels’. We see them when the pilot loads the tail with stick and pulls the invisible tailwheel into ground contact with prop wash (or ground rolling IAS) over up elevator. When we pull the invisible locked tail wheel into contact yaw friction increases allowing fishtail braking.
It is probably better to use a flat helicopter type skid of moderated surface contact length at the desired location to moderate tyre / skid (variable tyre pressure and surface area - such as tundra tyre) friction instead, but cfg values are used to model undercarriage co-efficient of drag, (friction = tyre rolling resistance), not propensity to yaw versus RPM applied by the (sim) pilot. The co-efficient of friction (rolling resistance of the vehicle) is encoded by the FD author, but the applied friction must be controlled by the (sim) pilot with carefully moderated prop wash and carefully moderated elevator deflection, carefully loading and unloading the tail contact point to dig it into the grass/dirt (or not) as required. Sometimes we want the tail contact point to ‘grip’, and sometimes we don’t. Unless we have substantial rolling IAS we have to control its grip with prop wash as well as elevator.
Incorrect developer evaluation of yaw inertia also plays a vital role in this particular process of course, but in tiny aeroplanes it is unlikely to be the primary design time error since even exaggerated values tend to be too low to preclude adequate yaw propensity.
In relation to both current threads it is vital to grasp that this is not just a ‘wrong values encoded in software’ issue as posters to forums always seem to assume. It is very much a handling skill that has to be learned and applied correctly by (sim) pilots. Many sets of flight dynamics are reported as broken in forums when the real problem is just pilot error, (failure to learn a particular skill).
Just as in real life taildraggers with skids and / or no brakes do not belong on hard surfaces and just as in real life there is always a carefully pilot moderated RPM needed to yaw the tail rapidly *without inducing unwanted forward acceleration* versus the co-efficient of friction of the current BGL surface under the contact points. It is also necessary to unload (the friction or rolling resistance of) the contact point we intend to yaw with prop wash (using the joystick to apply down elevator into the carefully moderated prop wash) if we hope to achieve ‘realistic’ turning circles in small tail draggers.
The Ansaldo demonstrates the consequence of (sim) pilot failure to moderate RPM, and (sim) pilot failure to moderate tail loading, to minimise turning circle. The rudder is always fully deflected when we intend to yaw such aircraft on the ground since we will always minimise the RPM and maximise our rudder deflection on any surface type since we always wish to induce maximum yaw rate with minimum vehicle acceleration to minimise turning circle.
Deflecting the rudder to its stop is however just the simplest part of a four part learned skill.
1) Apply full rudder
2) Evaluate surface hardness and friction
3) Moderate RPM (prop wash) accordingly
4) Moderate unloading of tail with elevator accordingly
Realistic tail dragger dynamics will deliver ‘poor’ turning circles on hard surfaces, but skilful (sim) pilots will be able to achieve tight turning circles on soft surfaces. Everyone should evaluate whether they are deploying the required skill set before fiddling with the dynamics of any download.
Altering maximum tail wheel angles in an aircraft.cfg cannot induce yaw propensity at zero IAS, nor will it allow the sim pilot to unload the tail contact point to increase or reduce its rolling and yawing resistance. The aircraft.cfg controls very little of the dynamics. The original posters in both relevant threads may have understood that of course. If there really is a ground handling dynamics error in either aircraft cited by the original posters it can be fixed, but not by fiddling with cfg files. This topic depends almost entirely on developer and consumer understanding of how to encode and moderate real world prop wash effects to induce yaw without inducing forward acceleration. Maximum tail wheel angle is irrelevant. In real life and in FS9 it is often zero (fixed skid).
The reason fixed skids are common on real aircraft designed to operate from soft surfaces is because they give excellent control over load/unload of the forward and side force friction of that contact point on soft surfaces. It is an error to remove that fine control from the simulation of such aircraft and aircraft with steerable tailwheels may need even more skilled control of prop wash, not less. They are more likely to accelerate when the (sim) pilot adds prop wash and may just skid the tailwheel unless the (sim) pilot takes great care to load it adequately and *steer* it very carefully.
That opens up a whole other can of worms since many / most (sim) pilots completely fail to grasp the difference between *steering* a vehicle versus *yawing* a vehicle.
In the real world when we ride a bicycle or motor cycle we do not assume that max steering wheel angle = max turn rate at every vehicle speed. We realise that we need large steerable wheel angles at slow speed, but ever smaller steering angles as speed increases, else we will skid the steerable wheel and lose directional control. Most sim pilots fail to use any common sense when trying to *steer* aeroplanes within a detailed ground handling simulator. They continue to treat vehicles with steerable wheels as though they only had rudders. They over control as though attempting to induce yaw with rudder because they use the same hardware input device in the simulator. Sim pilots must come to terms with the fact that sometimes those pedals or twist grip are a rudder control and sometimes they are a steerable wheel control.
*The difference matters.*
Steering with wheels and yawing with rudder requires different skills and most FS9 users fail to deploy either skill correctly. They lose control of their energy state on the ground even more readily than they do in flight.
That is why we see so many of these daft threads in which people propose one daft change to *maximum* steering angle after another. This nonsense will continue across multiple forums until flight simulation users grasp the difference between steering a vehicle with a steerable wheel, *carefully avoiding significant % rudder deflection to avoid yaw = skid) and the totally different concept of yawing a vehicle with max rudder. If the vehicle has a steerable wheel our job is to *match* the steering angle versus current vehicle speed just as though we were riding a bicycle. Most sim pilots just cannot grasp that and behave as though excessive applied steering angles will not immediately cause and perpetuate skid and increase of turning circle, so they whine that the cfg values are broken.
*They cannot be broken*
70 degree (max steering) lock may be true and required to move away from rest in a tight parking bay, but if any sim pilot subsequently loses control and causes skid it is not encoding the real max steering lock that causes the skid and loss of control. It is sim pilot error and only sim pilot error. If we fall off a bicycle it is not because the designer created a front wheel that can turn to 90 degrees, it is because we were stupid enough to suppose we would not cause loss of control if we applied a large % steering angle at anything above 1 knot.
The vast majority of sim pilots apply far too large a steering wheel angle to the steerable wheel for their current vehicle speed and simply cause the nosewheel or tailwheel to skid over the BGL, every single time they try to turn. They behave as though FS9 is incapable of modelling skid due to pilot error. FS9 actually models skid due to pilot error just fine and most sim pilots skid the steerable wheel everywhere, and almost all of the time, with the steerable wheel turned to wildly excessive lock, skidding over the BGL, delivering little or no steering because the user has induced and perpetuated steerable wheel skid.
Consequently many users gain additional control by halving or quartering the maximum steering lock in the cfg, not because the encoded maximum was wrong, but because reducing the maximum angle reduces their ability to induce loss of control via perpetuated skid with wild oversteer angles versus current vehicle speed.
Many sim pilots avoid inducing and perpetuating skid in any turn only by imposing upon themselves a tiny maximum wheel lock that will be useless at low vehicle speed. Because they refuse to learn how to *steer* and pretend that all aeroplanes are *yawed* they spend their lives meddling with max values in cfgs going from max values that cause loss of control when they stupidly apply them at high speed to lesser max values which prevent small turning cirles at low vehicle speed.
Controlling the turning circle of any vehicle is all about deploying the relevant and applicable skill, not cfg content. It is all about causing or preventing skidding. Skids will skid, and wheels will skid when turned to any inappropriate angle versus current vehicle speed. We simply choose whether to make them grip or not. We all get by perfectly well in the real world without differential braking in real automobiles and with max steering lock angles that would cause skid if deployed at any significant speed (even 10 KTS). We don’t demand that the max steering lock of our car be reduced so we can never skid because we realise we need a large max steering angle to park and unpark.
This topic does not require a grasp of advanced rocket science. It’s just common sense.
Steering is not yawing. Yawing an aeroplane requires maximum rudder deflection *to deliberately induce skid*, but preventing skid of a steerable wheel requires small applied rudder angles even at modest taxi speeds.
In a real automobile we get tactile feedback through the hardware device known as steering wheel that we are oversteering in a turn. MSFS fails to provide that feedback cue and most sim pilots skid the steerable wheel into and around most turns achieving a very poor turn radius, *because they failed to steer*. They apply wheel angles that would only be appropriate if a steerable wheel was some kind of rudder used to induce yaw = skid. A steerable wheel is not a strange variety of rudder. It is the opposite of a rudder. A steerable wheel exists to *avoid* yaw = skid. Large deflections of the steerable wheel do not induce small turning circles. They induce skid of the steerable wheel, and cause large turning circles because they degrade its effectiveness.
Judging the steering angle that induces oversteer in FS9 without tactile feedback is difficult, but it is just common sense. There is no excuse for sim pilots believing and pretending that no one can steer a bicycle unless the front wheel is prevented from adopting a large angle at low speed. Altering the maximum values of steering wheels in aircraft.cfgs is not the solution. It really is time that the whole FS community got a grip on this and stopped pretending that the dynamics are broken because they allow large steerable wheel angles, or that the loss of control many sim pilots experience when turning is due to anything other than skids they have induced by confusing steering with yawing and applying massive oversteer angles to the steerable wheel.
Many sim pilots induce wheel skid even at very low speed by applying massive steering angles.
In a taildragger sometimes we want a contact point to skid and sometimes we don’t. It is *not* the max wheel lock encoded in the cfg that controls the wheel loading or the yaw we can induce. As the Ansaldo demonstrates a fixed skid is no impediment to small turning circles and if you replace it with a castoring tail wheel in the cfg it will just make fine control of the friction of that contact point by the sim pilot more difficult. If you substitute a steerable tail wheel with a small maximum lock all that happens is that it becomes a skid once (often before) it reaches that maximum lock. Everbody has to stop pretending that ground handling of aeroplanes is not all about carefully avoiding or inducing skidding of the ‘third contact point’ by careful moderation of thrust and deployed steerable wheel angles versus current vehicle speed.
Otherwise FS forums will continue to be full of people insisting that it is impossible to ride a real bicycle and telling one another that one particular max angle of the steerable wheel is ‘correct’ or ‘best’ or ‘essential‘. The very idea is nonsense.
Reducing max wheel lock in FD does not prevent loss of directional control at much smaller wheel angles all of which constitute oversteer at most vehicle speeds, and increasing it does nothing to induce yaw when the speed is zero or minimal.
The sorry truth is that feedback of sim pilot oversteer error in MSFS is so poor, that it almost always goes undetected as pilot error by the sim pilot and the dynamics are blamed for what was simple pilot error even though the dynamics detected the pilot error correctly and imposed the correct consequence of the induced skidding of the steerable wheel across the BGL. Directional control is correctly lost. It is lost most easily on hard surfaces.
Once a sim pilot has induced skid the only way to turn the vehicle is yaw by rudder and if the vehicle in question was designed to be steered, not yawed, then the residual yaw induced with the rudder is tiny. It’s no good blaming the dynamics. They are doing their job of evaluating and imposing the consequence of pilot error.
Even 10 degrees of applied steering angle is enough to cause skid of the steerable wheel at very modest speed whether the vehicle in question is a bicycle or an aeroplane. The sorry truth is that many / most sim pilots don’t even attempt to steer aeroplanes defined as having steering. They dumbly attempt to yaw them with massive rudder defection instead (skid the steerable wheel sideways with the rudder) and when they succeed in turning the third contact point into a skid with no steering and lose directional control they claim the dynamics are broken.
Sim pilots must think harder about whether they are trying to steer or yaw the vehicle during simulation. The techniques are opposed techniques. FS9 supports both, but many / most sim pilots just confuse the concept of yawing (applying max rudder angle to deliberately skid the third contact point) with the concept of steering (never applying significant % yaw axis control deflections to always avoid skidding the third contact point) .
FSAviator
*************************************************************
*************************************************************
<<I still can't quite figure out why the Stearman's dynamics simply don't respond to changing the tailwheel to 180>>
They do. You have successfully removed all ability to steer the tailwheel. After you unload and reload the aircraft, you will have full yaw control via the rudder and no steering control via the tailwheel. Without propwash the yaw propensity versus (full) rudder will vary as a function of current IAS. As I explained the issue is not about tailwheels. It is whether the air file is allowing you to substitute propwash for IAS at zero IAS.
In my second reply I indicated the likely value required in REC 1101 field 0x08fe. The value present in the air file you are using is 15 times smaller than the proposed upper bound. Consequently you are unable to induce yaw with propwash. Doing what works in a real Stearman (or Cessna 120) will not work until realistic values are present in the air file.
<<Maybe the friction values to have something to do with it. I'll check into it more later on. >>
Encoded friction (tyre rolling resistance) is only an issue in aeroplanes without differential brakes. You intend to prevent inner wheel rotation with differential braking rendering rolling tyre resistance irrelevant. Since you intend to implement differential brakes you must obviously remove the cfg code;
differential_braking_scale = 0
If after you add differential braking by removing that code the locked inner wheel is seen to be skidding at the applied RPM you have applied excessive RPM. Each BGL friction is very likely to be different to whatever real surfaces you are used to operating on. You just have to accept that you are on a different surface in a different place and moderate RPM accordingly.
The reason it will not turn at low speed in FS9, with or without a castoring tailwheel, is that yaw propensity due to propwash is inadequate, and is nothing to do with tailwheels or brakes, as you keep proving with your tests after you vary them. The air file requires a yaw propensity in 0x08fe which will yaw the tail round with rudder fully deflected at an RPM that will not skid the (locked or unlocked) inside mainwheel on any BGL surface except ice. Once that value is augmented to adequacy the tail will yaw at low speed with a fixed skid and no brakes at all, or any other combination of contact point code and braking in the aircraft.cfg, because those variables only have the potential to limit what the air file induces.
FSAviator
********************************************************************
********************************************************************
<<Is it possible for a tinkerer with limited ability (like myself ) to implement this fix? >>
Yes.
Two different easy to use, but difficult to fully understand and configure, freeware air file editors are available from e.g.,
pagesperso-orange.fr/hsors/fsairfile.html
For this simple task Air Ed 1.52 is your best bet. The page says you need an updated aired.ini, but you won't to simply edit propwash effects. Leave all the files as they are and wherever you install the aired152 folder. Simply double click the supplied AirEd.exe, then navigate to and open the air file you are using. Then find and expand Record 1101.
Look down the long list of FD data fields in REC 1101 for Cn_dT Yaw Moment
What I believe to be the latest official version of the Stearman FD (FS04 PT-17 Ver 1.1.6 Apr 05) has this set to 0.002.
Look in REC 1 of the air file after you open it. If you are using a different air file the solution is 'likely' to be identical, but of course a different (unauthorised?) air file could have many problems.
Consider increasing the value incrementally until you can yaw the tail with fully deflected rudder into 'modest' propwash at very low taxi IAS with the stick held forward to unload the tail. If you overcook the value you will induce wild yaw authority at high thrust making co-ordinated turns in climbing flight and aerobatics harder to control; so only increase the value by just enough to solve the ground handling problem. Try increments of 0.005 and test each time.
If you define a personal key combination in FS9 (options/controls/assignemnts/Reload user aircraft) which causes air file reload you can then use that key combinbation after each edit and test each increment until you have just adequate yaw authority during very slow taxi. There 'should' be a higher value which makes differential braking irrelevant.
As you know the tailwheel animation may not behave as it would in real life whatever you code in the dynamics. Since the real Stearman had tailwheel steering (with disconnect to castor at unknown angle) I suggest that you implement that capability using no more than 45 degrees max in contact point.0 rather than implementing castoring and no tailwheel steering during your tests.
However......
Nothing you do will change the MDL tailwheel presence, so consider making point.0 a type 3 contact point = steel tipped skid.
point.0= 3, -17.3, 0, -6.2, 1200, 0, 0.3, max angle, 0.250, 2.5, 0.85, 0, 0, 0, 0, 0
Test whether defining it as a steel skid with both 0 and then 180 as max angle makes the dynamics behave more like a castoring tailwheel than actually defining the friction of that aft contact point as a rubber tyre.
The real Stearman had main wheel brakes. I still do not know if they could be applied differentially as delivered. That is not a given at the relevant date.
Remember the rear contact point has more friction with both cockpits occupied and yawing will require more RPM causing a larger turning circle. You should test on both grass and dirt, not just hard surfaces. Use MS default BGLs, not third party ones, during all FD ground handling tests.
Finally to all readers, please remember that many reported FD errors do not exist and the real problem is faulty user technique or expectation. Inexperienced tinkering with air files usually does more harm than good. The only reason I replied to this thread was that I believed the original question related to my original September 2003 FS9 update FD for this FS8 aircraft. I do not endorse generic air file tinkering by anyone who does not understand flight dyanamics.
FSAviator