GRT autopilot gain sanity check (8A)

[skirting_virga]

I had a long XC flight yesterday to fiddle with the autopilot. The builder had roll gain set near the default, 100%, which I found was ok in smooth air but exacerbated even the smallest amount of turbulence.


Starting with the servo response test, I had to work roll gain down to 7%, and roll gain damping very high (600-1100%) to get the system to behave. The result seems to be that it is able to intercept and track a course in even bumpy afternoon thermals. It could be a little more proactive, but it's honestly pretty good compared to before.

Does this make any sense? I know the 8 is lively in roll but I wasn't expecting to have to turn it down so low. Does anyone have GRT gain settings from a 7 or 8 to share?

Pitch is acceptable for now with lowish settings and I'm not motivated to further optimize it yet.

 

[clam]

GRT equipment in RV-8. My roll gain is set at 84% and damping at 100%. What I have found to be more influential, is the XTK gain. My AP roll performance was satisfactory in all modes, except LNAV. When tracking an LNAV track, especially on an approach, the roll (course) correction was continuous large amplitude corrections, near 25 deg bank angle corrections all the way down final. I don't recall exactly what the XTK gain is set now, but it is pretty low. It is not 7% kinda low, but I believe down in the low 80/high 70 range.

There are a few other super smart GRT guys on here that will likely add more color and accuracy for you.

I'll add this little bit of experience too, although probably not related based on your low gain settings ... but, I had a pax who fell asleep on a long xctry. I believe that during a course change, his leg had fallen against the stick keeping lateral pressure on it. The roll servo was trying to program in the turn, and torque built up enough to shear the screw. I never got a TRIM alert that I recall on that event. But the roll response then acted unsatisfactorily, and after some troubleshooting, I found the shear screw sheared. Replaced and all was well.

 

[rv8ch]

I don't recall my numbers but they are not far off the recommendations from GRT for RVs - I know the main guy there has a RV-6.

One thing I have seen is that the "180 degree get out of weather emergency button" didn't behave like I expected, and I'd suggest that you test this in good weather with a safety pilot before relying on this feature in an inadvertent VFR into IMC scenario.

The recommendation from GRT - which I have not yet tested - is:

"You can limit the bank angle for all autopilot operations in A/P Maintenance. As for the loss of altitude, the EFIS will use the normal altitude hold response to react to an altitude change caused by a turn. You could try increasing the Altitude Hold gain in General Setup. If it's 100%, try something like 150% or 200% to see what the effect is."

 

[skirting_virga]

The roll servo was trying to program in the turn, and torque built up enough to shear the screw. I never got a TRIM alert that I recall on that event. But the roll response then acted unsatisfactorily, and after some troubleshooting, I found the shear screw sheared. Replaced and all was well.

This is insightful - no idea if it applies to my situation but it raises the idea that I should inspect the servo.

I will note that in the servo response test, the external control loop isn't involved. I did tweak xtk and heading gains a little but most of my time was spent with the roll gain and damping.

My bank angle limit is 20 degrees at the moment

 

[skirting_virga]

Does anyone else have any insight as to why my plane seems to require 7% roll gain and high roll gain damping to eliminate roll oscillations (even in servo response test mode)? Something must be off...

 

[BH1166]

Been behind GRT AP for 800 hours in 6 yrs. All my settings are 100 or 200. I hand fly in really rough air as I don’t want to shear the pin grabbing stick in a tense moment. Flew today, pretty calm air, AP via heading and enav rock solid. Form up a question with data/settings and submit to [email protected] they see it all. I fly 6a…different but I got no settings like yours.

 

[NelsonME58]

I’m hoping an RV-9A owner with an installed GRT A/P may be monitoring this thread. I’ve attached photos of my pitch servo installation and was hoping someone could help me determine if it is correct. The install was completed by a maintenance shop hired by the previous owner. I suspect the rod end bearing installation on the servo control arm may be attached to the wrong side resulting in the angled installation. Any advice is greatly appreciated.

 

[skirting_virga]

I’m hoping an RV-9A owner with an installed GRT A/P may be monitoring this thread. I’ve attached photos of my pitch servo installation and was hoping someone could help me determine if it is correct. The install was completed by a maintenance shop hired by the previous owner. I suspect the rod end bearing installation on the servo control arm may be attached to the wrong side resulting in the angled installation. Any advice is greatly appreciated.

Great idea - can any 7/8/9 owners post pictures of their GRT roll servo install? I want to see where the pivot is on the bellcrank and the servo arm to figure out if mine is set up in a "typical" fashion or not. Given that the pitch servo does just fine, I'm thinking this could be the issue...

 

[rv8ch]

I’m hoping an RV-9A owner with an installed GRT A/P may be monitoring this thread. I’ve attached photos of my pitch servo installation and was hoping someone could help me determine if it is correct. The install was completed by a maintenance shop hired by the previous owner. I suspect the rod end bearing installation on the servo control arm may be attached to the wrong side resulting in the angled installation. Any advice is greatly appreciated.

I have the GRT servos in my RV-8, and I mainly used the Dynon instructions to install them. The Dynon instructions are very good, and the servos are very similar, physically. I also see they now have some videos.

dynon autopilot servo installation - Google Search

 

[skirting_virga]

This is a photo of my roll servo. The dynon instruction appears to specify 2". Mine is about 1-5/8". This might just be the problem.

 

[NelsonME58]

This is a photo of my roll servo. The dynon instruction appears to specify 2". Mine is about 1-5/8". This might just be the problem.

I’m experiencing too little torque in the roll axis even with the roll servo torque setting set to maximum of 15. I assume this is the “problem” you are experiencing as well? I am planning to move my servo arm attach point to the center position in order to increase torque. Haven’t measured the position of the rod end attached to the aileron bell crank yet. Here is a couple photos of my installation….