Poor 'Standard Atmosphere' Found in FS9

[Guest Tom Goodrick]

When I test aircraft for which I am developing flight dynamics files, I like to do the testing in 'Standard Atmosphere' Conditions. I have just found that the weather condition you get when you select the calm, standard condition is badly non-standard. I began suspecting something when comparing speeds between FS02 and FS9. I had tested the atmosphere in FS95 and found it very bad for jets. In FS98 it was fixed and very close to 'Standard'. In FS2000 and FS2002 I just spot-checked a couple of cases and found no problem. But today I ran a test with the Learjet 45 and found lots of errors. I calculate the density ratio as the square of the KIAS/KTAS ratio. Here are some examples:Alt ft______KIAS___KTAS___Dens Rat___Std Rate20,000 ft___245____326____0.565______0.53330,000 ft___247____385.5__0.411______0.37440,000 ft___208____397____0.274______0.24550,000 ft___172____414____0.173______0.152For the same indicated airspeed, the true airspeed would be in error up to 6%. At 8,000 ft there was an anomoly resulting in a 5% error although most speed values below 20,000 ft were within just a few %. In all cases the aircraft is slower than it should be.I find this development very dissapointing. I know that when we fly with Real Weather, we experience non-standard atmospheres. But, for test perposes when trying to design to a spec, it is very nice to have a good standard atmosphere. This is a step backward for MS.

[Guest JeanLuc_]

And we never stop reading posts about "poor" simulated aircraft flight models...Thank you for the heads up! I hope it will help some people on the forums that only in an ideal world, things are perfect!

[Guest Tom Goodrick]

I need to make one correction. The anomoly at 8,000 ft that I noted in my Learjet 45 test did not appear when I went up through those altitudes in the Mooney Bravo. That data is below. The error is more on the order of 1% which I find acceptable.Alt FT____KIAS___KTAS____Dens Rat____Std Rat6000______124____135_____0.844_______0.8367000______124____138_____0.807_______0.8118000______124____139_____0.796_______0.7869000______123____141_____0.761_______0.76210000_____131____152_____0.743_______0.738I should mention that the Bravo matched a cruise spec at 10,000 ft of 190 KTAS on 20.4 gph (89% power). 188 KTAS was the spec published in Jan 1998 AOPA PILOT. The Bravo took off at max gross weight.I need to recheck those high-altitude density ratios in an aircraft with good digital airspeed indicators. I'll do that soon.

[Guest Mark McG]

Tom,Although not having mcuh knowledge of the workings of an FDE I have noticed similar problems.I mainly fly the PIC 767, and have been using it undocked in FS2004.This aircraft is rock solid in FS2002, but in FS2004 I encounter the problems described in the thread below:http://forums.avsim.net/dcboard.php?az=sho...=4264&mode=fullHerve Sors has also identified a problem with dynamic pressure and has posted this on Pete Dowsons forum:http://forums.simflight.com/viewtopic.php?...b59bda9549928c9PIC/FS2004 Nil Weather Loaded Test Results:FL 370 M0.80 CruiseTAT:-27CSAT:-56CIAS:264 KnotsTAS:458 KnotsG/S:459 KnotsRunning Weather I encounter the problems described above and a 500KG per engine increase in Fuel Burn at cruise alt.Hope this helps.

[Guest Tom Goodrick]

I have run another test, this one with new digital gauges for ambient density, ambient temperature and ambient pressure. The results are shown below.FS9 ATMOSPHERE ANALYSIS |<<<<<<>>>>>| |<<<<>>>>>>>>>>| ALTITUDEDENSITY TEMP PRES DENSITY TEMP PRES0 0.00237 59.0 29.91 0.00238 59 29.9210000 0.00175 23.3 20.57 0.00175 23.3 20.5720000 0.00126 -12.5 13.73 0.00127 -12.3 13.7430000 0.00089 -48.0 8.90 0.00089 -48 8.8840000 0.00058 -69.8 5.56 0.00058 -67 5.5450000 0.00036 -69.8 3.44 0.00036 -67 3.44 ObservedFROM Correct PercentALTITUDEKIAS KTAS TAS/IAS StandardKTAS Error0 154.5 153.6 0.99417 1.00000 154.5 -0.5810000 229.1 261.6 1.14186 1.16374 266.6 -1.8820000 237.1 315.1 1.32898 1.37148 325.2 -3.1030000 278.1 426.4 1.53326 1.63185 453.8 -6.0440000 203.6 387.5 1.90324 2.02144 411.6 -5.8550000 178.4 426.3 2.38957 2.56580 457.7 -6.87These reults show that the FS9 ATMOSPHERE IS PERFECTLY FINE!The problem is MICROSOFT SCREWED UP THE AIRSPEED CALCULATIONS!!!!I put my new xml gauges on a panel that pops up when I tap W while flying the Learjet 45. I flew UP to each of these altitudes from the runway at Imperial, CA, where I established my Flight Test Base. (I hope the nearby Airporter Inn within still serves those great Margueritas.) I have a test corridor from KIPL to E63 (Gila Bend). (Sorry about the format. It looks great on the editing page.)

[Guest Mark McG]

Hi Tom,I reflew my FS2004 PIC flight in FS2002 and noticed the TAT & SAT were the same in both sims with nil weather loaded.FS2002IAS - 261KIASG/S - 457KIASTAS - 457KIASFuel Flow - 1800Kgs per engineFS2004IAS - 264KIASG/S - 461KIASTAS - 460KIASFuel Flow - 2100Kgs per engineBoth Flights were at FL370 M0.80Not really sure if this helps you or not, but I would have expected to see the same values in FS2004 as in FS2002, not a 1-2% increase in IAS/TAS/G/S and nearly a 20% increase in Fuel FLow.

[Guest Ron Freimuth]

Tom, Note there is a new line in aircraft.cfg:[airspeed_indicators]airspeed_indicator.0 = 1.16, -22.0 At least on some FS9 AIRCRAFT.CFG files. For example, the C208. That could account for some changes in "IAS". I have not experimented with the numbers yet. Though, my tests on my FS2K2 C208 FD, copied to FS9 and 'upgraded' appeared to display the same IAS. Further, I could check TAS with the new GPS, though not as accurately as with a good test gauge. I wish my XML test gauges worked in FS9. I took the 'image' line out and one still worked in FS2K2, maybe it will work in FS9 now. I also ripped off your XML code (posted in another forum) for 'slope'. However, I coded asin (VS/TAS), where VS and TAS were both set for ft/sec. Flight slope reads the same as in AFSD in FS2K2. I may convert it to the reciprocal, 1/slope = L/D if thrust is zero and AC is in a stable descent.-RAF

[Guest Tom Goodrick]

When trying to match a cruise condition (with a spec or between two sims), it is best to match fuel flow. This is directly proportional to power output for the engine(s) whether jet or piston.

[Guest Tom Goodrick]

I would caution against using slope to get L/D because the results can be horendously erroneous. That relation holds only in unaccelerated flight which we seldom see.My XML code has transferred with no trouble. Here is a typical example:(A:Circuit general panel on, bool)(G:Var2) 0 == (A:Avionics master switch, bool) &&%((A:AMBIENT TEMPERATURE, farenheit))%!4.1f!%Ambient Temp FI do not know where this basic code structure originated as I copied an example given me by our mutual friend BobS. But it works nicely for a variety of digital gauges.Your statement about the new line for IAS in the aircraft.cfg file is very intriguing. I will check into it. The 6% error in airspeed (discrepency between IAS and TAS) at high altitude (observed with the Learjet above 40,000 ft) is very disconcerting.I feel we now have three significant problems in FS9: the incidence problem discussed above, the CG problem and now the TAS/IAS problem.I know Microsoft is not concerned about this aspect of sim use, but I am very concerned that our older aircraft, which we have nursed from version to version, may never fly correctly in FS9. It is clear to me that keeping customers happy is the last thing MS thinks about. I am beginning to think MS should remove the word "Simulator" and just call this the "Flight Game 2004."

[Guest Ron Freimuth]

>I would caution against using slope to get L/D because the>results can be horendously erroneous. That relation holds only>in unaccelerated flight which we seldom see. To some degree, Herve's AFSD corrects for accelerations. Anyway, the parameters blank if not dependable. >>My XML code has transferred with no trouble. Here is a typical>example:>>> > (A:Circuit general panel on, bool)> (G:Var2) 0 == (A:Avionics master switch, bool)>&&> ................ My gauges were based on XML code in a commercial product I'm working on. All the strings are at the end, the calculations are done in blocks at the top. I don't understand all of the structure, but got it working in FS2K2. I'll add new formulas to check various ratios, etc. as I desire. Just to get a better idea some relationships during flight. XML modules coupled as they should in my few XML files. I set the 'reload gauges' tooltip code to reinitialize my constants.>I do not know where this basic code structure originated as I>copied an example given me by our mutual friend BobS. But it>works nicely for a variety of digital gauges. I'll probably put some of my gauges on Herve Sor's site. Once I get embarassing structures, etc fixed. I'd also want to simply them.>Your statement about the new line for IAS in the aircraft.cfg>file is very intriguing. I will check into it. The 6% error in>airspeed (discrepency between IAS and TAS) at high altitude>(observed with the Learjet above 40,000 ft) is very>disconcerting. The new indicator for airspeed block in aircraft.cfg wasn't in many FS9 aircraft.cfg files. I copied it from the standard C310 aircraft.cfg to my version. I didn't see it in imported files that got new lines added. Nor did it appear to relate to the two IAS parameters near the top of REC 1101. Smaller AC have a -10 kt offset to IAS to emulate pitot tube pressure errors at increasing AoA's. But, the C310's don't. I checked a 727 I imported into FS9. I remember IAS ~ 251 kts at Mach 0.745 at FL 350. I saw 253 IAS. I do have JWB.gau and a turbine test gauge working on my Lear panel; which I use for testing jets. So I knew Mach was 0.745. The real 727 has virtually no CAS error at 250 kts, and that is how I've adjusted the second IAS parameter in the AIR file. Otherwise, IAS would be several knots off at Mach 0.745. >I feel we now have three significant problems in FS9: the>incidence problem discussed above, the CG problem and now the>TAS/IAS problem. I came up with a new problem: Unfair details possed by beta testers, but not by others. ;) I noted CG was just right in my 727. JWB gauge showed 24.5%. I took advantage of the ability to change payloads in flight and found it got less stable with CG = 40%. It should be. CG at 5% also made static pitch stability too high, making it harder to trim. I set lateral and longitudinal stations at +/- 200 ft from the CG. As already mentioned, no changes in MOI's were detectable. They were last seen in FS2K. I did find something new: Elevator control is cut way down when pitch attitude control is enabled in the autopilot. I think that is probably desirable. As I already knew, ALT and Pitch control are much more stable in FS9. I've also adjusted some of the new autopilot lines in aircraft.cfg. However, 'maximum_pitch' has no effect. In FS2K one could set max and mininum pitches the autopilot would set. The Yaw Damper doesn't even appear to add Cn_r, Yaw Damping. All I can see is it attenuates rudder movement. The 'yaw_damper_gain' only appears to be a BOOL. '0' turns it Off, most anything higher turns it On. I'm not sure the 'throttle_rate' has any effect, either. SPD control is still slow to respond. And useless during turns with signficant winds. I've also mentioned that 'prop effect on rudder', which worked in FS2K2, is now dead in FS9. 'prop effect on elevator' died going from FS2K to 2K2. There are a lot of taildraggers in FS9, CoF. But, you can't even push the tail up with a blast from the prop. Quite unrealistic. Yes, I flew the J-3 for a while. I agree with what you said: it seems to be pretty good once in the air. I checked Pete's forum, Herve found that 'dynamic pressure' wasn't mapped correctly in FSUIPC 3.03. Not surprising, that is an arcane parameter for most, and I think only Herve' used it. Ron

[Guest jcomm]

> There are a lot of taildraggers in FS9, CoF. But, you can't>even push the tail up with a blast from the prop. Quite>unrealistic.>Ron, I did perform a few tests from the default Meigs situation, using the J-3, and looking from the outside view the tail does lift up when we give it a sudden and intense burst of prop... Of course it is not as realistic as what one gets in x-plane, where the amount of up movement varies depending on intensity of the throttle "burst" as well as wind, density, etc....> Yes, I flew the J-3 for a while. I agree with what you>said: it seems to be pretty good once in the air.>You're certainly talking about the normal flight envelope because as soon as you try the simplest possible power off stall it will do the most weird things pretending to be an UFO, or the like ;-)...Regarding rolling friction (longitudinal and lateral components), either they did something about it and it now feels better to me because my airplane no longer sideslips during the takeoff or landing run. As a matter of fact, provided one decreases the sensibility of the yaw axis, takeoffs and landings under crosswinds is now more realistic IMHO. Again, nothing as good as what you can get in x-plane ;-)

[Guest Ron Freimuth]

>> There are a lot of taildraggers in FS9, CoF. But, you>can't>>even push the tail up with a blast from the prop. Quite>>unrealistic.>>>>Ron, I did perform a few tests from the default Meigs>situation, using the J-3, and looking from the outside view>the tail does lift up when we give it a sudden and intense>burst of prop... Of course it is not as realistic as what one>gets in x-plane, where the amount of up movement varies>depending on intensity of the throttle "burst" as well as>wind, density, etc.... It sure doesn't for me. I tested many AC and non show a bit of "prop effect on tail". My C208 does in FS2K2, since I set the prop on rudder so I could steer it on the water without water rudders. But, not in FS9! I know it works in CFS2, and probably in FS2K. Further, without a prop blast to 'steer' an AC like the J3, it has very little steering until one gets to 30 kts or so, where the 'Normal' air flow is getting significant. I think this is why some AC are so hard to keep straight on the runway. I tried a P51D in both FS2K2 and FS2K4. At least the rudder had effect at low speeds; the prop blast made it effective. However, there is no 'prop effect on elevator' in FS2K2, which means I couldn't get the tail up as I should have. In FS2K4, the P51D was much worse on the runway. It would go in circles and I couldn't get it straight. I had to be very careful to not let it get started in a circle. I figure that's becaues 'P-Factor' 'helical effect' and 'torque' are probably all effective (I know P-Factor is back in FS9), but without the other prop on tail effects, taildragger AC become very hard to control at low speeds. Of course, I have all the prop sliders to max. Which I assume, is 'normal'. Seems the FS9 C172SP was also hard to keep straight on the runway. That should depend on the CG, since the traction from the nose wheel depends on the force on it. -RAF

[Guest jcomm]

Ron,what I was referring to is that when I give it sudden throttle, the J-3 lifts the tail a bit. The problem is it's allways the same bit, no matter how intense my burst of throttle is...I would like to take this chance to ask you for some clarification on something I never really understood very well from the semantics point-of-view: What do "prop effect on rudder/elevator mean"? - Does it mean that MSFS used to simulate the effects of propwash on the rudder and elevator control surfaces such that there would actually be a displacement on their otherwise neutral positions? Or does it simply satnd for the effect on the ALL of the tail/horizontal surfaces (ruuder+fin)/(horizontal stab+elevator)?In fact it appears that MS has "stollen" propwash effects on the fin+rudder surface. Maybe they will answer something like:- "We have removed those variables from the FDE simply because we are not interested only in the tail or horizontal surface, but rather in the airplane as a whole, and since we model prop p-factor, torque and slipstream effects, there was no need for it..." Hehe... ;-)In the limit, we will all be looking after each and every former variable by 2006/fs10... The answer from MS? - "We are aware that there are no longer any variables in the FDE. That was done on purpose because we do not want to modell any of the aircraft components by itself. We see the all aircraft as a single point in space, and we lookup an encrypted table provided to us by the aircraft manufacturer and indexed by.... In a word... FDE tweakers, go find yourselves another job..." ----> Just me kidding B.G. don't really take this seriously :-)

[Guest Tom Goodrick]

I think those airspeed lines in the aircraft.cfg are related to calibration to give a properly erroneus Indicated airspeed from the Calibrated airspeed. This was mentioned by FSAviator in another thread. But such calibration usually is neeeded only at very low speeds where the aircraft angle of attack (and the angle of attack of the pitot tube) is high. It should not pertain to jet flight.I am starting to wonder about the accuracy of Mach measurements. I will add that to my Air Data display.Incidently, I tried putting my Air Data display into the King Air 350 using the same method I used with the Learjet 45 and it would not work. I could not make it visible. There seems to something screwed up with MS's new Mini Panel system of which they make a big deal in the Aircraft folder under "default" (a perfectly obtuse identifier).

[Guest Tom Goodrick]

I made some tests of Mach number today with the Learjet using my air data package to which I added Mach number to 3 decimals. At several points from 30,000 ft to 41,000 ft I compared the observed Mach number with the value calculated from true airspeed and temperature. They matched exactly in all cases. Thus, if you fly using Mach number as a guide, you will get the proper true airspeed. This is indeed what you should do from 30,000 ft up.True airspeed in a sim is the direct result of the integrations of the accelerations. That gives it as good a level of accuracy as the sim can deliver. Indicated airspeed, which must be used by pilots below 30,000 ft for proper control of the aircraft, is a calculated quantity from the true airspeed given the density ratio. But I am beginning to think that MS is corretly using mach number to introduce a deviation into the IAS value which could occur in the pitot system due to Mach effects. Maybe this was explained in one of the Learjet documents I failed to read.