SR22 AP Banking way too much

Hey Everyone,

It has been a while since I flew the SR22 and I was hoping some of the issues have been fixed since I flew it last. One of the issues that bothered me a lot was the autopilot banking way too much to either turn or get on a course.

A standard rate turn is 30 degrees which is what the autopilot should commence when a turn is requested. The SR22 typically does around 40 degrees.

A second issue is intercepting and tracking courses. Real autopilots calculate a 45 degree or 30-degree intercept angle depending on how far you are away from the course. I did a test where I manually aligned the plane to a 45-degree intercept course to a track (Track 360, Heading 315) and enabled NAV mode. The plane immediately banked heavily to the left setting a new heading of 280, continued until the track, then attempted to make a hard right to 360, but of course, overshot the track, then continuous left and right turn followed until it was able to get on the track. It was painful to watch.

Anyone else has this issue as well? Did Asobo said anything about this at all?

I haven’t seen excessive bank angle selection on the AP, and I’ve been flying the SR since launch. NXi simply utilizes the existing AP code logic on any aircraft it’s installed, so that’s the PID at work, not the FMS. I’ve had 25 degree bank angles. What might be helpful is that if you’re still flying stock G1000, switch to the NXi. While the AP logic is out of the FMS control, it does play a part by NXi creating more realistic course projection turns instead of the impossibly sharp or lazy wide course changes in the default FMS.

The other piece about the oscillation - this tends to happen if you are at extremely low speeds and or higher angle of attacks. Also, attempting large course change while climbing seems to also trigger this base course weave.

I used to use the working title mod. It is now available in the marketplace so I make sure to have the latest version before I fly. Is there anything else I should use?

Autopilots use an advanced implementation of PID controllers to capture and retain a specific target altitude, heading, rate of descent etc. and respond to changes / deviations.

The software implementation of that is suffering a lot. Even in the C172, the autopilot fails to close on to the target heading, overshoots, then turns the other way etc…
It shouldn’t be such a great challenge to do it right, as those maths and functions have been around for years in computer simulations.

At some point a few months after release they really messed it up (remember the oscillating A320), then corrected it to a degree, but is still lacking.

Nothing else I can think of. Because the Cirrus is Premium / Deluxe, no one’s taken the daunting task of trying to mod it like Dakfly has with the Longitude.

Also, moved to #self-service:aircraft-systems to better focus the topic.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.