New plane Config issue

[Guest simagic3d]

I've been through probably every param in both .cfg and .air files and can NOT for the life of my figure out what's wrong. I've narrowed whatever it is that's wrong to the .cfg but have no idea why the real world data (contact points, cg, ref, etc) make this plane sink thru the ground, get to the model's center/pivot point, then pop back up to static cg location and rapidly repeats the process (like a feedback loop).The plane, once I can get it off the ground, seems to fly just fine, albeit it needs some fine tuning.When I put FS in slew, the plane will slowly rotate in pitch, tail going down, so I assume this is the stability problem that's causing the ground sink, but have no idea why or how to fix the numbers. I've looked at a similar plane's cps, cg, ref, by pluggin in lights to see such in FS. The only thing about that particular plane was that the designer did NOT place the model's pivot/origin at 25% mac. It was at the center of its bounding box, but the plane was rock steady on the ground and flew great.Attached is the side view of the model with all the important contact and data points listed. Maybe someone can see something majorly wrong right off the top and suggest what value(s) need to be change and where. One major point not shown is CoL. According to Rec 1204 in AirEd, COL is 3" fore of CG, which places it btwn AC (mac/4) and CG. Seems appropriate to me, but maybe it's not. Ideas?http://forums.avsim.net/user_files/57036.gifThanks so much to the assistance in the past!

[Guest simagic3d]

Just got AFSD working and what the heck do I find it saying about CG? That it's 20 FEET BELOW me and 8 feet aft! What the heck!?? My .cfg CLEARLY says 1 ft below datum and 4 ft aft. See... ![WEIGHT_AND_BALANCE]reference_datum_position = 3.34, 0.00, 0.00empty_weight_CG_position = -4.00, 0.00, -1.00CG_forward_limit = 0.0CG_aft_limit = 1.0[contact_points]max_number_of_points = 21static_pitch = 0.0static_cg_height = 2.94point.0 = 1, -1.38, 0.00, -4.00, 3000, 0, 0.33, 40, 0.3, 2, 0.75, 0, 0, 0, 0, 0 // center wheelpoint.1 = 1, -4.53, -2.20, -4.00, 3000, 1, 0.66, 0, 0.3, 2, 0.75, 0, 0, 2, 0, 0 // left wheelpoint.2 = 1, -4.53, 2.20, -4.00, 3000, 2, 0.66, 0, 0.3, 2, 0.75, 0, 0, 3, 0, 0 // right wheelpoint.3 ... point.13I guess the keyword I'm screaming here is... HELP!

[Guest Ron Freimuth]

>The plane, once I can get it off the ground, seems to fly just>fine, albeit it needs some fine tuning.>>When I put FS in slew, the plane will slowly rotate in pitch,>tail going down, so I assume this is the stability problem>that's causing the ground sink, but have no idea why or how to>fix the numbers. I've looked at a similar plane's cps, cg,>ref, by pluggin in lights to see such in FS. The only thing>>Attached is the side view of the model with all the important>contact and data points listed. Maybe someone can see>something majorly wrong right off the top and suggest what>value(s) need to be change and where. One major point not>shown is CoL. According to Rec 1204 in AirEd, COL is 3" fore>of CG, which places it btwn AC (mac/4) and CG. Seems>appropriate to me, but maybe it's not. Ideas? When you set the empty CG in ACM, it appears to use that position for the CG reference. The wheel locations in FS are based on distance from the FS_Reference, not a shifted CG. Set the empty CG to 0.0 in all directions, then see if ACM shows the wheels in a different location. Set the LG, etc. with that 0, 0, 0 empty_CG. Finally, change the empty_cg to the correct location manually in aircraft.cfg after saving from ACM. You don't even have to save from ACM, I don't like the way it formats the lines in aircraft.cfg. So, I experiment in ACM, but don't save. Entering the values I set in ACM manually. Actually, without the pilot in the vehicle, leaving the CG reference at the Visual center (also the center of Lift unless shifted in the AIR file CoL record) pilot sitting well below the wing, the empty CG may work well enough. Setting a station load for the pilot well below the reference datum will move the loaded CG down and may also move it for/aft. You might experiment with moving the main wheels in the direction that keeps the AC from tipping on its tail. I don't think the scrape points worked in FS2K2 so the tail would sink into the ground if the wheels were too far forward. Regardless, if they aren't in the correct location TO rotation will be poor. Ron

[Guest simagic3d]

> When you set the empty CG in ACM, it appears to use that position> for the CG reference. The wheel locations in FS are based on distance> from the FS_Reference, not a shifted CG.What about AirEd and AAM? Do they reference things differently?I don't have ACM.

[Guest Ron Freimuth]

>>> When you set the empty CG in ACM, it appears to use that position>> for the CG reference. The wheel locations in FS are based on >> distance from the FS_Reference, not a shifted CG.>>What about AirEd and AAM? Do they reference things>differently?>I don't have ACM. Oh, you image looked something like ACM displays. But, not identical. Most of the distances are set in aircraft.cfg. Many things in the AIR file have no effect. However, the CoL record will move the Wing's Center of Lift if necessary. I always set 'sweep' to 0.0 and set the "wing_apex_long" position 25% of the MAC ahead of the FS_Reference. Of course, evenything is shifted if the reference_datum is offset. MAC is Area/Span so it's easy to calculate MAC/4 to set the L.E. I've had numerous problems setting LG lines, at times I've moved the MG so the AC will just tip on its tail. That told me where the current longitudinal CG was. The Wing_apex has to be set correctly for the %CG to display correctly. AFSD also figures out the vertical CG offset but I'm not sure just how it calculates it. It's always possible the values displayed aren't right for uncommon offsets, especially when the empty CG isn't at the CoLFS_Reference. I often adjust the vertical distance for the LG by looking at the wheels in the spot view and experimenting with the strut lenghts and compression values. I think the 'length' and static compression is based on empty weight. The struts compress more as load increases. The .mdl has to articulate the distances correctly for realistic strut 'stiffness'. That doesn't seem to be your problem. If the current CG is off vertially the flight charactistics will change. Of course, the longitudinal CG is even more important. All I can suggest is experimenting with the MG location, then place it just a small distance behind the balance point. You could also move the pilot or a test station load to change the weight on the MG. In big jets typically 95% of the AC's weight is on the main gear. Which means the MG isn't far behind the actual CG, but the NG is much farther forward. In fact, I'd think the ratio would be about 20:1 in that above case. To get by for now, you could add a 'tail wheel' in the LG section. But, don't set it on the ground, put it just a bit below where the tail would scrape. Clearly, it shouldn't touch the ground. But if it did at least the tail shouldn't sink any further. Giving time experiment with more adjustmets while watching in the spot view. People have also suggested setting lights at the same locations as the bottom of the wheels are set in aircraft.cfg. That way, you can see if the images of the wheels are really where you set them in aircraft.cfg.Ron

[Guest simagic3d]

> Oh, you image looked something like ACM displays. But, not> identical.That's the model in Lightwave. I model and UV map in it sinceit's SO much easier than gmax. Then I export as 3ds and importto gmax for tagging surfaces, parts, and animations, et. al.> Most of the distances are set in aircraft.cfg. Many things> in the AIR file have no effect. However, the CoL record will> move the Wing's Center of Lift if necessary. The aerodynamics don't seem to be that far off. It flies likeI said but needs tweaks to eliminate the pitch flutter, etc.It just will NOT sit on the ground whatsoever no matter howCG is set. I got CG located with the pilot removed and theweight compensated so that AFSD reports a CG that's very closeto the model's origin/pivot (CoV I think you call it), but FS just seems to ignore the wheel contact points as far as notletting them penetrate the ground, even though FX for dust,sparks, and lights are perfectly positioned.Tell ya what I'll do. Attached is a proxy model, .cfg, and .air.See what it does and maybe it'll be something so simple I'll goaway with that stupid dumbfounded look on my face. But, maybe it'llbe something more. You have definately more exp. with these thingsthat I do, so make yourself laugh and see if you can at least MakeMe Sit Still. It's really small (71K) with nothing but the threefiles).http://forums.avsim.net/user_files/57091.zipAnyone else is welcome to have a laugh, too.

[Guest Douglas K]

I removed contact points 3-13 and reworked the longitudinal and vertical positions, compression and damping of the three main points a little. See below:[contact_points]point.0= 1, 1.38, 0.00, -4.10, 3000, 0, 0.33, 40, 0.125, 2, 0.9, 0, 0, 0, 0, 0 // center wheelpoint.1= 1, -4.83, -2.30, -4.30, 3000, 1, 0.66, 0, 0.25, 2, 0.75, 0, 0, 2, 0, 0 // left wheelpoint.2= 1, -4.83, 2.30, -4.30, 3000, 2, 0.66, 0, 0.25, 2, 0.75, 0, 0, 3, 0, 0 // right wheelIt seems to sit, taxi and stop OK now, but you'll have to fine-tune it.Cm_q Pitch Moment-Pitch Rate (Damping) in REC 1101 was set too high. I lowered it to 17000 as a quick fix for the shaking. It could go lower than that, but that's your call. The damping and the MOI's for all axes are interrelated. If you set the damping high, and the MOI is too low, then the shaking will occur.

[Guest Ron Freimuth]

>>That's the model in Lightwave. I model and UV map in it>since it's SO much easier than gmax. Then I export as 3ds and>import to gmax for tagging surfaces, parts, and animations, et. al.> I checked your site and it appears you fly a real Weedhopper.>The aerodynamics don't seem to be that far off. It flies like>I said but needs tweaks to eliminate the pitch flutter, etc.>It just will NOT sit on the ground whatsoever no matter how>CG is set. I got CG located with the pilot removed and the>weight compensated so that AFSD reports a CG that's very close>to the model's origin/pivot (CoV I think you call it), but FS>just seems to ignore the wheel contact points as far as not>letting them penetrate the ground, even though FX for dust,>sparks, and lights are perfectly positioned. I see this in your aircraft.cfg: [WEIGHT_AND_BALANCE]reference_datum_position = 3.340, 0.000, 0.000empty_weight_CG_position = -3.150, 0.000, 0.15 That puts the empty CG 0.15 ft above the wing. It is clearly below, even without the pilot. Your pilot load is commented out.//station_load.0=170.000000, -4.700000, 0.000000, -2.400000, pilot Without the pilot the high CG won't help ground stability. Since the LG is about -4.1 ft below the reference_dataum. Typical of larger AC. However, with the real CG a couple of ft below the wing reference, the prop thrust will cause the AC to pitch down when it speeds up. Do you notice that in the real Weedhopper? It's better to pitch down than up, since pitch up with throttle could make an AC stall. I never saw this before:[rudder]rudder_rate.0 = 1005 Is it in the FS2K4 SDK? I pasted Douglas' new LG lines in aircraft.cfg as comments. I see he also found a lower Cm_q (-17000/2048) helped. Wing twist and incidence should be set to 0.0 in FS9. Actually, FS9 doesn't use them (unless FSEdit is used), so changing twist to 0 won't make a difference. TBL 404 needs to have CL at AoA=0 appropriate to adjust flight pitch. I assume it's satisfactory as it is.wing_incidence = 0.000wing_twist = -3.000oswald_efficiency_factor = 0.500---- I took a quick look at the AIR file. I see elements from the '1% SS' in it. ;) REC 1101 values appear reasonable. I see Cm_q is set to -27000 =~ -13.0. You might see if half that value improves things. A PA-28 runs about -7.0 (real value). The slope of TBL 472, Pitch Moment vs AoA is rather low. 0.069/0.53 ~ 0.13. Elevator might be too 'strong' for that low slope, you might try cutting 'Lift - Elevator' and 'Pitching Moment - Elevator' to half their current values. Also, the Pitch Trim moment, which likely gives a lot more pitch trim than needed. So, you found a use for the new aileron to rudder coupling set in aircraft.cfg. I imagine it should be adjusted to keep the 'ball' more or less centered in typical turns. I found very light AC (20 to 100 lb) weren't stable (especially in Yaw) in FS2K2. 1200 lb empty is OK. I don't know if FS9 has a problem with an AC your weight. It may have more to do with the low MoI's for small AC. Your values are much higher than the ones appropriate from the small AC I tried. Ron

[Guest simagic3d]

> I checked your site and it appears you fly a real Weedhopper.Yep. Thus the reason I was trying to get FS to balance out with the measures obtained from the real thing, cg, mac, weight, etc. The only thing I have no idea on calculating is all the appropriate coefficients of this and that and the other thing.> [WEIGHT_AND_BALANCE]> reference_datum_position = 3.340, 0.000, 0.000> empty_weight_CG_position = -3.150, 0.000, 0.15> That puts the empty CG 0.15 ft above the wing.But AFSD doesn't report it like that... something like -1.5 or so and the longitude was -0.08. Maybe I reading it all wrong and not understanding what the numbers out of FS via AFSD are, but changing the empty_cg vertical to 0 or slightly negative makes the reported valued go FAR negative, as in -4 or much greater. Same for long.> Your pilot load is commented out.On purpose as you said the pilot station would adjust the dynamic CG after craft initialization, which I assume was causing the -20' reading for CG below and -8 aft from AFSD.> However, with the real CG a couple of ft below the wing > reference, the prop thrust will cause the AC to pitch down> when it speeds up. Do you notice that in the real Weedhopper?Yep. But CG is probably less than 1' down from the boom/water line,not 2 or more.> I never saw this before:> [rudder]> rudder_rate.0 = 1005 > Is it in the FS2K4 SDK?No idea where that came from. I've been copy/pasting parts of one config to another.> I pasted Douglas' new LG lines in aircraft.cfg as comments.And the proxy plane sat still for you too? Interesting.> AIR file. I see elements from the '1% SS' in it. ;)I'm trying all sorts of ways to edit this thing. I assume there really isn't a way to create a new .air from scratch... that one has to use a supplied template or other similar flying craft's .air to apply changes to, right? Wouldn't it be feasible to start from scratch so there isn't any potential uglies from an existing dynamic, like I seems to be getting?And once created in part, couldn't FSedit fill in the gaps using the config file numbers and then the grunged tables be replaced/edited to suit?> So, you found a use for the new aileron to rudder coupling> set in aircraft.cfg. I imagine it should be adjusted to keep> the 'ball' more or less centered in typical turns. Yep. I assume this comes out into the config from the air's Roll Moment due to Rudder in rec 1011. Using invisible ailerons significantly helps the roll rate to reflect what the real plane does in flight with just dihedral and rudder.> I don't know if FS9 has a problem with an AC your weight. Lago's Ultrlight package seems to have their craft set to real world weights and they behave nice and firm on the ground and fly in the air without issues.Thanks for ALL your input and help, Ron and Doug!More questions will inevedibly be forthcoming.

[Guest Ron Freimuth]

>>> I checked your site and it appears you fly a real>Weedhopper.>>Yep. Thus the reason I was trying to get FS to balance out>with the measures obtained from the real thing, cg, mac,>weight, etc. The only thing I have no idea on calculating is>all the appropriate coefficients of this and that and the>other thing. All I can say is much of the AIR file and lines in aircraft.cfg look about right. LG problems have given me lots of trouble in my AC. In part because I'm still not sure of some details. >> [WEIGHT_AND_BALANCE]>> reference_datum_position = 3.340, 0.000, 0.000>> empty_weight_CG_position = -3.150, 0.000, 0.15>>> That puts the empty CG 0.15 ft above the wing.>>But AFSD doesn't report it like that... something like -1.5 or>so and the longitude was -0.08. Maybe I reading it all wrong>and not understanding what the numbers out of FS via AFSD are,>but changing the empty_cg vertical to 0 or slightly negative>makes the reported valued go FAR negative, as in -4 or much>greater. Same for long. I've suggested to Herve' Sors a lot of the AFSD calculations. To give appropriate numbers the wing_apex_xxx locations have to be consistent with how MS calculates CoL, CG, etc. I moved the Wing up or down if appropriate, and also set the CoL vertical offset in the AIR file since I'm not sure the vertical setting in aircraft.cfg works. Hope I'm not doubling the overall effect. I found that when the wing is above the current CG, the AC levels it's wings better. Similar to the Dihederal Effect parameter in REC 1101. While moving it down results in less Dihederal Effect so the parameter in REC 1101 may have to be increased. I think the Dihederal Angle set in aircraft.cfg only moves the average vertical CoL up. Another of the things that we have to verify experimentally to figure out what FS is setting when it loads the AC. Now many of the [aerodynamics] lines have an effect on the ground and flight characteristics. But, many others only have an effect with FSFoobar is used. Finally, tail areas and distances have an effect on FFB, but clearly they only need be approximate for that. FSEdit uses those distances, unfortunately it still sets a lot of inapproprite values in the new records it adds to an AIR file. >> Your pilot load is commented out.>>On purpose as you said the pilot station would adjust the>dynamic CG after craft initialization, which I assume was>causing the -20' reading for CG below and -8 aft from AFSD. Changing the location (one can even change the weight from the FS9 Menu during flight) should result in a change in AFSD's CG offsets. NOTE: at least some of them are in inches, not ft. Ah, MAYBE that's why you thought they are so low and far aft. ! >> However, with the real CG a couple of ft below the wing >> reference, the prop thrust will cause the AC to pitch down>> when it speeds up. Do you notice that in the real>Weedhopper?>>Yep. But CG is probably less than 1' down from the boom/water>line, not 2 or more. The CG spot on your graphics looked about right. Those graphics were very nice, the problem is to be sure FS sees the same things. As I mentioned above, the main effect of vertical CG vs the wing's 'waterline' appears to be in wing leveling. Which should be high in the WeedHopper (for general stability). Note if vertical CG is too high on the ground any vehicle will tend to tip when turning or stopping. >> I never saw this before:>> [rudder]>> rudder_rate.0 = 1005 >> Is it in the FS2K4 SDK?>>No idea where that came from. I've been copy/pasting parts of>one config to another. You might comment it out; I have no idea if '1005' is a reasonable value. Perhaps FS added the line when it saw the aileron-rudder coupling set to 1.0.>> I pasted Douglas' new LG lines in aircraft.cfg as comments.>>And the proxy plane sat still for you too? Interesting. Heck, I didn't have time to set up your AC. In fact, I DL'ed a 31 MB beta of a commercial AC I did the FM on last night and haven't even installed it yet. >> AIR file. I see elements from the '1% SS' in it. ;)>I'm trying all sorts of ways to edit this thing. I assume>there really isn't a way to create a new .air from scratch...>that one has to use a supplied template or other similar>flying craft's .air to apply changes to, right? Wouldn't it>be feasible to start from scratch so there isn't any potential>uglies from an existing dynamic, like I seems to be getting? You may have started with the air file of a similar AC. It did have some interesting curves set. I decided they were probably OK. Many people use Jerry Beckwidth's 1% SS. Milton likes it. However, I'm not sure how it would work on your AC. You could also start from scratch (don't destroy what you have) and use FSEdit. At least I found the new FS9 version fixed some things from the older one. I think FSEdit sets a lot more or less appropriate. However, it still sets things far from appropirate, even though they are easy to calculate. Now I know how to fix many of the poor FSEdit values. So if others look reasonable I tend to think they are appropriate. But, that isn't of much help to those who don't know the main problems with FSEdit. One thing I found was it set the pitch trim value about 5X too high. Making trim too critical. It sets LG drag to zero, but anyone with some familiarity with these parameters could put a reasonable value in. Clearly, it isn't a factor with non-retractable gear. >And once created in part, couldn't FSedit fill in the gaps>using the config file numbers and then the grunged tables be>replaced/edited to suit? I just commented on that. Maybe FSEdit would get the LG, W&B correct and you could go back to your current AIR file. >> So, you found a use for the new aileron to rudder coupling>> set in aircraft.cfg. I imagine it should be adjusted to>keep the 'ball' more or less centered in typical turns. >>Yep. I assume this comes out into the config from the air's>Roll Moment due to Rudder in rec 1011. Using invisible>ailerons significantly helps the roll rate to reflect what the>real plane does in flight with just dihedral and rudder. I just discoverd that changing the Yaw due to Aileron from a small negative value to a positive value improved turns in my jets when the FS HDG/NAV lock was set. This Adverse Yaw is usually small in real AC, but making the ailerons a bit overcompensated appears to aid in turn coordination. >Thanks for ALL your input and help, Ron and Doug!>More questions will inevedibly be forthcoming. BTW, 'southern.edu' isn't SMU, is it? SMU is about 10 miles south of me. Ron.

[Guest Douglas K]

> I pasted Douglas' new LG lines in aircraft.cfg as comments.>And the proxy plane sat still for you too? Interesting.Dean, I doubt they would work if pasted as comments!I was in a rush last night and my reply was too brief. I had to remove ALL the contact points EXCEPT point.0, point.1 and point.2 before the Weedhopper would sit properly. That's what I meant when I said "I removed contact points 3-13". I did this for troubleshooting only. With the contact points section it's best to simplify things and work from there. There appears to be a problem with one or more of the other 11 points, as soon as I've had more coffee I'll look into it.Once again, the contact points section I'm using looks like this:______________________________________________________________________[contact_points]point.0= 1, 1.38, 0.00, -4.10, 3000, 0, 0.33, 40, 0.125, 2, 0.9, 0, 0, 0, 0, 0 // center wheelpoint.1= 1, -4.83, -2.30, -4.30, 3000, 1, 0.66, 0, 0.25, 2, 0.75, 0, 0, 2, 0, 0 // left wheelpoint.2= 1, -4.83, 2.30, -4.30, 3000, 2, 0.66, 0, 0.25, 2, 0.75, 0, 0, 3, 0, 0 // right wheelmax_number_of_points = 21static_pitch = 0.00static_cg_height = 3.95gear_system_type = 4______________________________________________________________________

[Guest Douglas K]

I found the problem contact point. With point.9 removed the Weedhopper's subterranean tendencies will stop. point.9 = 3, 0.94, 0.00, -2.30, 3000, 0, 0.00, 0, 0.0, 0, 0.00, 0, 0, 4, 0, 0 // prop strikeBest Regards,

[simflyerguy 0]

>When I put FS in slew, the plane will slowly rotate in pitch, tail >going down, so I assume this is the stability problem that's >causing the ground sink>Errrmmmm...No, this isnt a problem with Stability.. check your joystick/yoke calibration... this sounds like its the real culprit... the flight model has absolutely nothing to do with SLEW and vice-verse.Andrew

[Guest Ron Freimuth]

>>When I put FS in slew, the plane will slowly rotate in>pitch, tail >going down, so I assume this is the stability>problem that's >causing the ground sink>>>>No, this isnt a problem with Stability.. check your>joystick/yoke calibration... this sounds like its the real>culprit... the flight model has absolutely nothing to do with>SLEW and vice-verse.>Andrew I tend to have that problem. An AC will move sideways in Slew mode since my JS is loose in the mid range. I reduced the JS slew rates in FS9.cfg (FS2004.cfg?). But, increased the altitude rate which is set from the KB so I can slew up faster. I think I also reduced the KB pitch rates. I like to slew on the ground to an up and down nose pitch, then hit 'Y' and let the AC bounce back level. That gives some idea of LG dynamics.Ron

[Guest simagic3d]

> BTW, 'southern.edu' isn't SMU, is it? Nope. More like Southern Adventist University, 20 mileseast of Chattanooga, TN.