#2485 Static pressure changes while fgfs window has no focus

[Huntley Palmer]

This shows as instability in c172p Autopilot, Vertical speed - Elevator PID; after changing windows focus and returning to the fgfs window the property: /environment/static-pressure-inhg can spike while the fgfs scenery 'jumps' a few frames. This property is differentiated to yield target ROC for the plane which, in turn dips, climbs and sometimes recovers when fgfs window regains focus. Wondering, too, if this explains how some users observe apparent violent IAS changes blamed on METAR ?

[legoboyvdlp]

Wondering, too, if this explains how some users observe apparent violent IAS changes blamed on METAR ?

That is a separate issue, primarily to do with lack of data provided by METAR. I investigated this issue and made some mitigating changes about a year ago which reduced the instances of this.

Are you using advanced weather or basic weather?

[Huntley Palmer]

Basic, explain if you think there's a difference, please.
I should add, I'm scoping the output of static pressure from static.cxx

PPS Advanced waether makes no diff that I can see,

[Huntley Palmer]

Here's a screencap of static.cxx' output at systems/static/pressure-inhg: I timed mousing to another window at output == 22.30xxx . Before that point the pressure drops steadily ( 600fpm ) but the numbers freeze and then recover while the LMB is held .

[legoboyvdlp]

Advanced weather has interpolation routines while basic weather doesn't seem to have those' that's the main difference. The fact it happens with Basic suggests its probably something in C++ not working correctly.

It's actually somewhat as if the environment code stops executing while the window is out of focus..?

[Huntley Palmer]

Thanks ! Of course, the weather determines the pressure ( duh hoy)

Looking at the PID, it's trying to double integrate, target ROC driving Elevators and I'm not seeing much success so far in slugging the differentiator and PID gains, with static pressure low passed sufficiently the PID oscillates.
I'll see what separating the integration with a 'target pitch' stage in between does for stability.

Tks again !

[Anonymous]

This could be related to the time-interpolation of the weather not working in live METAR mode: https://sourceforge.net/p/flightgear/codetickets/2636/