Z longitudinal acceleration provide strange data

[litinoveweedle]

Hi,

when programming AP mode using as input offset $3070 (Z (longitudinal, or forward/backward) acceleration in ft/sec/sec) I've found strange behaviour of this offset:

If aircraft is on ground, parking brake set and not moving, this offset oscillate around values 0.57 - 0.58ft/s/s

Such value is not a low deviation - comparing to take-off run or intensive braking when value of this offset is around 10-11ft/s/s

Only in case I shut down both engines, this offset will came close to 0.

Values are from FSinterrogate from Flight1 ATR72 plane, FSUIPC version 4.60a

Could I please get some explanation of this behaviour? Only one possible coming to my mind is vibration from engines.

Thank for any help.

BR

Litin

[litinoveweedle]

Some more informations:

- X (lateral, or left/right) and Y (vertical, or up/down) acceleration are NOT affected - when aircraft is parked, then both these are close to 0 as supposed to be

- Used simulator is FSX Acceleration

Litin

[herve_sors]

Same observation in both FS9 and FSX. Without knowing how FS calculates acceleration vectors, I will not rely on those values while on ground where different forces apply (idle engine thrust, brake force, ground friction). Whatever it may be it seems fluctuations of body-axe longitudinal acceleration (3070) mainly depend on brake action (fixed small positive value when brakes are ON, small variations around 0 when brakes are released and aircraft is not moving). No more idea by now.

[litinoveweedle]

Thank you for reply.

I can override output when aircraft is on ground and parking brake is set, that is not an issue.

But problem could be if I can't rely on this offset in flight. I am writing complex autopilot using PID controllers. I would like to use Z acceleration for IAS mode by pitch. This is chain of several controllers first PI setting required acceleration to get to proper IAS and second PID comparing this required acceleration against actual acceleration (this FSUIPC offset) and giving required pitch as its output.

Has anyone used this offset for such in-flight calculations?

There is still possibility to calculate Z acceleration as change in IAS over given time, which is coming to my mind more often now. :smile:

BR

Litin

[herve_sors]

Tests I performed with AFSD indicate body-axe accelerations are ok while in flight. Most advanced aerodynamical calculations (CD,CL,etc) I perform are partly based on them and I didn't find any noticeable discrepancy. I also recalculate flight path-based accelerations from them without major problems. So you should have a try

[litinoveweedle]

Thank you very much for your help, I will definitely test AP function based on Z acceleration. If I found any strange behaviour in flight, I will reopen this topic.

BR

Litin