Went from SV32 to SV42 servos for both pitch and roll servos. The SV32's that I removed had a lot of slop between the output shaft and shaft bushing. Not shear screw slop but main shaft slop. The SV42's are much tighter. I had earlier production run servos for the SV32's, don't know if Dynon changed anything here or not. I suspect the slop in the output shaft was causing some of the performance issues that I was having before.
The pitch servo I have set at 85% torque, sensitivity 15, VS at 600 ft/min. The roll servo is at 60% torque, 7 sensitivity, and 40 deg bank angle max. I have not finished playing around with the roll servo settings yet and MAY go back to a NEW SV32 servo with a shaft that's not so sloppy. The SV42 does however give me more setup flexibility so at the end of the day may just stay with the SV42.
Today while testing in smooth air, the altitude would hold within 10'-20', and the roll would hold within a degree or two. It does GPS and Nav tracking very well too. So....this system is performing extremely well. The next test will be in turbulent air. I suspect I may have to increase the torque on both servos a little in turbulent air. FWIW in the Bonanza that I had until last summer, its high dollar AP (King KFC-200) performed in a similar fashion to the Dynon (i.e. 10'-20' altitude deviation, +/- 1-2 degrees heading deviation. The King did have auto trim which was nice, but then again the Dynon system is about 10% of the price of a King system. In turbulent air, the King was +/- 50'-100', and about +/- 4-5 deg heading deviation. It will be interesting to see what the Dynon will do in similar conditions.
Dynon, you have produced a great system. After system tweaks my Dynon autopilot is performing solidly. Thanks!!