Problem with GA Circuits in MSFS

[c912039 112]

Hi,

I seem to be having issues with GA aircraft flying circuits at an airfield used by a large flight school.

The GA aircraft never seem to actually leave the ground and continue to trace the circuit pattern all at ground level.

I have RealTraffic radar showing, as well as LittleNavMap so I can compare the data between what RT believes is the aircraft speed, heading and height, and what PSXT is injecting into MSFS (using LMP and clicking on the simulated GA aircraft to view what PSXT is setting for speed, height etc.

RT definitely shows the aircraft climbing to a cct height of 1000 ft, and yet the PSXT aircraft shows it never leaves the ground.

Any ideas why the PSXT GA aircraft might not be flying at the correct height that RT is showing it at?

For reference, the airport is YPPF, and the UTC date is Dec 16 @ 0100 UTC and thereafter.

 

Regards

[kiek 1,224]

5 hours ago, c912039 said:

Any ideas why the PSXT GA aircraft might not be flying at the correct height that RT is showing it at?

Theer could be a problem with the RT data not showing Vertical speed and/or the GA aircrfat not reaching Vr speed threshold. What GA aircraft type are you talking about.

[c912039 112]

DA40's

[kiek 1,224]

2 minutes ago, c912039 said:

DA40's

That's a light model indeed. I'll have to replay your -well documneted (!)-  situation to see if there is something wrong in the RT data. 

 

[kiek 1,224]

In P3D everything is normal. Now I'm gonna start up MSFS (which takes a lot longer).

Question: Could it be that you are not using a DA4 AI aircraft but a user flyable one from Asobo?

 

[kiek 1,224]

In MSFS it looks pretty normal too, I'm using AIG and FSLTL:

YKH      VH-YKH             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
YTG      VH-YTG             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
YNK      VH-YNK             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
VOZ417   VH-YQR    ADL->SYD VOZ B738 =3= FSLTL_B738_Virgin Australia
FD593    VH-FDK    ADL->PUG PVT PC12 =3= AIGAIM_U-Haul Pilatus PC-12 - N605TQ
MZO      VH-MZO             PVT M20P =7= FSLTL_HTAI_P28A_ZZZZ
YNJ      VH-YNJ             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
YBW      VH-YBW             PVT DA42 =8= AIGAIM_Mokulele Airlines Cessna 208B Grand Caravan - N857MA
VHFVE    VH-FVE    ADL->    PVT PC12 =3= AIGAIM_Metro Aviation Pilatus PC-12 - N852AL
YQZ      VH-YQZ             PVT DA42 =8= AIGAIM_interCaribbean Airways De Havilland DHC 6-300 Twin Otter - VQ-THG
BEV      VH-BEV    ADL->    PVT PC12 =3= AIGAIM_Metro Aviation Pilatus PC-12 - N399AM
USL      VH-USL             PVT C172 =8= AIGAIM_Advanced Air Pilatus PC-12 - Surf Air
YBQ      VH-YBQ             PVT DA42 =8= AIGAIM_Tradewind Aviation Pilatus PC-12 - N590TW
YBT      VH-YBT             PVT DA42 =8= FSLTL_HTAI_P28A_ZZZZ
YTH      VH-YTH             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
EQJ      VH-EQJ             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
PQM      VH-PQM             PVT DA40 =7= FSLTL_HTAI_P28A_ZZZZ
 

Edited December 16, 2022 by kiek

[mike stryczek 23]

I see this phenomenon in  FSX.

It happens when aircraft is on T&Gs sequence when it first takeoff  and flies first circuit it looks great and then after touch and go it stays on the ground and often pops up to 1000 AGL in the middle of downwind to continue correctly.

I have tried to compare to RT information by clicking on the  RT radar and observing climb an attitude.

I also have radar in the SIM and it shows climb information yet the aircraft stays attached to the ground.

It only happens after T&G.

Every fresh takeoff from taxing always looks amazing.

It is not always happening but happens quite often

 

mike

[c912039 112]

I use FSLTL but not AIGAIM objects.

 

I had configured PSXT to look in community folder for aircraft.  PSXT found the FSLTL library, but also the couple of GA 3rd party aircraft Im currently flying.
PSXT was using the JustFlight PA28 Warrior as the model for some of the DA40s. (By the way I didnt experience any performance or frame rate drop when those models were used, I am limiting my frame rates to 30fps but Im also using an old 1080Ti GPU - I was surprised I didnt get fps drop below 30 when these complex models where used instead of out of FSLTL.

After your response, I set up file in regcodes to force all of the DA40 regs that fly out of YPPF to use a specific model, so I could test different configs, so I can better understand what might be happening.

I set up the regcode entry to use FSLTL_HTAI_P28A_ZZZZ for all the DA40 regs belonging to FTA. After forcing PSXT to rebuild the AI_liveries.xml, all the DA40s were represented by that model, however no change was observed in circuit flying.  The DA40s that were on touch & go still did not take off and completed the entire circuit at ground level, despite RT showing their correct altitude, speed and (sometimes questionable) VS.

I also noticed that for the very few instances, they seem not to become airborne at the correct runway point (that RT showed they were lifting off) but instead PSXT had them use the entire runway length PLUS more length extending into the field before takeoff.

The same happens for the aircraft that are actually airborne, on final PSXT has them touch down way before the runway, and yet RT seems to show the correct touchdown location.

For the record, I am using AUSCENE YPPF third party airport scenery.

I have a livery for the FTA DA40s that uses Asobo's DA40 model.  So, thinking that perhaps the FSLTL_HTAI_P28A_ZZZZ model performance config was a mismatch for the DA40's, I experimented by putting in the FTA DA40 livery into community, and modifying my regcode text file for the DA40 reg codes, forced PSXT to use the FTA DA40 livery/Asobo DA40 model.

There was no change to the problems with the aircraft flying circuits. No matter if I used the Asobo DA40 model, or the FSLTL_HTAI_P28A_ZZZZ, neither model in PSXT flew or matched the positional information that RT was providing. (By the way, having all those DA40s depicted with the Asobo model, also didnt appear to cause me to loose any FPS.).

 

So, in summary, even after forcing PSXT to use the simple FSLTL_HTAI_P28A_ZZZZ model, GA aircraft flying circuits dont seem to depict the real life A/C height from RT.

 

The reason I'm so keen for this to work, is using RT & PSXT I can simulate what it was like to try to fly with upwards of 5-6 other aircraft in the circuit at one time when I was flying there (ex student of FTA). Therefore its important that the aircraft fly the patterns exactly matching the data coming from RT, and, if possible be the same size/shape as the real aircraft, and it having to use Asobo's DA40 model for DA40 traffic ends up with too much of a FPS hit, then having the much simpler FSLTL_HTAI_P28A_ZZZZ model is better than nothing.

 

By the way, as a sub-note, Asobo have a number of 'generic' models that they can use when showing other live users.
asobo-aircraft-generic-piston-multiengines,asobo-aircraft-generic-piston-singleengine  etc.
I thought these might be useful simple models to depict GA aircraft where there is no FSLTL or AIGAIM models to match.
The aircraft.cfg probably doesnt contain the info that PSXT might be expecting, but I wonder if there might be some way for PSXT to make use of those simple generic models that Asobo provide, to increase/bolster the number and types of generic models that could be used with PSXT and MSFS.

 

Regards

David

 

 

[c912039 112]

1 hour ago, mike stryczek said:

I see this phenomenon in  FSX.

It happens when aircraft is on T&Gs sequence when it first takeoff  and flies first circuit it looks great and then after touch and go it stays on the ground and often pops up to 1000 AGL in the middle of downwind to continue correctly.

...

It only happens after T&G.

...

It is not always happening but happens quite often

 

Hi Mike,

you might be onto something there. After watching a couple of hours worth of T&G at YPPF (using all sorts of different models for PSXT to use to depict the DA40s), and reading your comments, I believe you have described what I saw, that PSXT seems to have problems with aircraft that are doing T&Gs in the circuit.

In general, aircraft that were taking off for the first time and heading out NW to the training area seem to show takeoff and climb out correctly.

David

[kiek 1,224]

Hi Mike (and others),

That behaviour in FSX we have  discussed in the past, and I thought I had fixed the T&G... 

In MSFS everything is a lot more complicated. Due to bugs in Asobo/MS the AI aircrfat seem to be "glued to the ground". Take-offs in MSFS are not as beautiful (rotate and climb) as in P3D/FSX. There are still independent  autonomous forces that try to keep the aircraft on the ground, while PSXT commands them to pitch up and leave the ground. They should behave as dumb objects fully controlled by PSXT but there is still an "Asobo capatain" playing a role. This may lead to longer takeoffs in the Sim then in reality, and may brake T&G..

But if T&G is worse again in FSX? , I'll have to look into my code.

Cheers,

Nico

[kiek 1,224]

Hi guys,

I think I have found the  problem with T&G. It is caused by a change I made a few weeks ago. Now fixed.

Please download 33.2.1. and try again.

Nico

Edited December 17, 2022 by kiek

[mike stryczek 23]

The attachment to the ground persists in 33.2.1.

I remember sometime ago we’ve had that problem in the previous version and it was resolved

mike

[kiek 1,224]

1 hour ago, mike stryczek said:

The attachment to the ground persists in 33.2.1.

I remember sometime ago we’ve had that problem in the previous version and it was resolved

mike

Yes I know and I reverted to that version.

More after Dec 28.

[mike stryczek 23]

Nico what I meant was, it was long time ago that it was fixed.

 In the last several versions I was observing the ground attachment

Since it’s very obscure and I’m on the old FSX maybe one of very few,  I didn’t wanna bother you about it

 

mike

[kiek 1,224]

Aha, thanks for that extra piece of information.

There are no more then 10 FSX users  indeed...

However, all versions share the same basic code so I prefer that you do "bother" me with problems.

 

EDIT"I think I've really found the problem now, but I will release it Dec 28.

Edited December 18, 2022 by kiek