Direct axis control with FSUIPC has unwanted nonlinearity in MSFS

[Alexey]

Hi!

I have a custom FFB yoke and try to setup MSFS with FSUIPC7 in the same way I had in P3D with FSUIPC5/6.

* I don't have joystick device for my yoke. Instead, I directly set axis values through offsets 0x0BB2, 0x0BB6.

* For P3D it works with DirectAxesToCalibs=All set in .ini file

* I don't need any calibration/nonlinearity (the yoke has full range already).

In MSFS this approach generally works, but comparing virtual yoke and real yoke motion I see that virtual MSFS yoke doesn't move linearly comparing to real one. There is a sensitivity curve introduced somewhere which makes it more sensitive at center.

It works same way with and without FSUIPC calibration enabled (with FSUIPC calibration set to linear). I can compensate it via FSUIPC at the expense of movement accuracy (inverse curve), but that's undesirable scenario.

Note: I cannot set or see calibration curves for ailerons and elevator in MSFS because I don't have joystick device for my hardware.

Is there a way to have linear motion in this case? Am I missing something in FSUIPC or MSFS configuration?

Thanks!

 

Version: FSUIPC 7.3.1.1

Environment: MSFS 1.27.21.0, Win10

 

[John Dowson]

9 hours ago, Alexey said:

* For P3D it works with DirectAxesToCalibs=All set in .ini file

* I don't need any calibration/nonlinearity (the yoke has full range already).

But this doesn't make sense...if you have DirectAxesToCalibs set, then values written to the offsets will go through FSUIPC calibration, and so you must calibrate. From the FSUIPC Advance User guide:

Quote

DirectAxesToCalibs: Setting this to ‘Yes’ makes FSUIPC assume that any direct writes to FSUIPC’s offsets for
rudder, aileron and elevator are from a hardware driver and are really meant as axis inputs. FSUIPC directs the values to
its Joystick Calibrations section, where you should then calibrate the inputs exactly as you would for a normal flight
control. The only hardware known to me which benefits form this is the one for the Aerosoft GA28R console. Do not
set this option if you use any sophisticated panels or external programs with their own autopilots, as it is possible that
they route their control values the same way. FSUIPC cannot distinguish the source.
To also allow for brakes to be sent to the Joystick Calibration, this value can be set to ‘All’.

If you don't want calibration, maybe try removing that ini parameter.

Otherwise, has this always been an issue, or is this a new issue since the SU10 release? I ask as there are various issues with calibration since SU10.

If the issue persists, could you add logging for offsets 0x0BB2 and 0x0BB6 (using FSUIPC's offset logging facilities), activate logging for Axes Controls and Extras and produce a log file showing your issue and attach that (zip if large) as well as your FSUIPC7.ini file.

Thanks and regards,

John

[Alexey]

1 hour ago, John Dowson said:

But this doesn't make sense...if you have DirectAxesToCalibs set, then values written to the offsets will go through FSUIPC calibration, and so you must calibrate

Yes, agree. This is legacy stuff I blindly moved from P3D (it was needed specifically for A2A aircrafts).

Set DirectAxesToCalibs back to No, but the behavior is the same - sim yoke movements not linear. With the obvious difference that I now don't see the axes in FSUIPC calibration dialog - but I don't need anyways.

 

Quote

1 hour ago, John Dowson said:

Otherwise, has this always been an issue, or is this a new issue since the SU10 release? I ask as there are various issues with calibration since SU10.

 

Can't say, because this SU10 is when I first tried. Maybe I can try with SU11 beta.

Attached please find FSUIPC7.ini and log file (I removed name & email). Log shows values coming from WideFS (2nd pc), but that I hope irrelevant to the issue.

Log contents left me puzzled: what I see in values, looks fairly linear:

59625 25784 Monitor IPC:0BB6 (S16) = 0
 59640 25784 SimRead: 0BB6="AILERON POSITION"
            FLT64: 0
 

60984 25784 Monitor IPC:0BB6 (S16) = 8216
    61000 25784 SimRead: 0BB6="AILERON POSITION"
            FLT64: 0.501464961099
 

  62328 25784 Monitor IPC:0BB6 (S16) = 15957
    62344 25784 SimRead: 0BB6="AILERON POSITION"
            FLT64: 0.973938106511

 

But look at the screenshot (also attached). The title bar shows current 0x0BB6 value ˜8000, so my expectation for yoke rotation angle to be about 45 deg, but in MSFS it just started moving from straight position. Same behavior for three airplanes I tested.

So no idea.. Maybe it's just visualization issue.

 

FSUIPC7.ini FSUIPC7.log.zip

[John Dowson]

15 hours ago, Alexey said:

Maybe it's just visualization issue.

Yes, I believe this is just a visualization issue that has been around since msfs was released. See this post about the issue on the MSFS/Asobo forums:

https://forums.flightsimulator.com/t/movement-of-on-screen-yoke-does-not-match-real-controller/418391/18

Regards,

John

P.S. The same issue was also reported in FSX - see https://www.flightsim.com/vbfs/showthread.php?263439-The-movements-of-the-saitek-yoke-aren´t-truely-representated-in-fsx

Edited October 15, 2022 by John Dowson
PS added