Autopilot overcorrection in heading mode

[shane]

OK, I am a bit lazy here as the answer is probably somewhere in the documents.
Have 180 with autopilot.
Altitude hold is great, in fact it is brilliant.
Heading hold not working, aircraft turns towards heading bug, goes past it, turns back with increasing bank, goes past again, turns back again this time with high bank angle etc till I loose nerve and disconnect.
How best to start the trouble shooting?
Shane

 

[DBRV10]

First thing to do is make sure your compass calibration has been done correctly, get your up to date inclination etc (refer the manual) and perform a compass calibration.

This will most likely be the problem. Next thing to play with is the roll axis sensitivity. you may need to increase it one or two points.

DB

 

[dynonsupport]

In addition to the above, have you tried the new bank angle mode in 5.3? This has improved the AP significantly for some customers.

 

[Johnkerr]

I initiated a similar thread this morning on the installation area. Having the identical problem with a 10A, 5.3 version. EDC newly calibrated.

 

[rvevz]

Hi all. My unit is the 180 with ap. I have a similar problem which started about 3 weeks ago. On a flight of just over 1 hour the ap started going off course and overcorrecting itself several times. I was baffled so I turned the ap off and back on , which fixed it , or so I thought. There was no problem on the return flight on the following day. However on a 5 hour flight of 1 week ago it started again without warning and the overcorrecting was much worse , up to say 100 degree either way. Over the flight it happened about 5 times resulting in me having to turn it off each time to correct it. But before turning it off a couple of times , I switched from Nan to Coarse and to Track , which made no difference. I also checked that my Garmin 296 was receiving satellite info , which showed all was OK in that department.
I have had my system in operation for 6 months and the inclination setup is correct.
Until the flight of 3 weeks ago it had been working fine.
Because the problem is intermittent and happens without warning , it's probably going to be difficult ( even if it doesn't happen on the next few flights ) to establish if the problem has corrected itself or if it is only sleeping , at the ready to go wandering at some point when it's least expected.
I can't begin to know how to fix it.
Allen B

 

[dynonsupport]

Some generic advice for all:

First, it's much easier for us to track autopilot performance issues if you email support at dynonavionics dot com, since that back and forth is captured in our tracking system. Next, some general questions and advice that might help you narrow the problem down:

1) First, turn on datalogging at 1 update per second, with the log at boot option turned on. This will give us data to look at to help you figure out what's going on.
2) Is the servo slipping? This is annunciated by the axis's mode label going to a yellow background highlighting instead of the normal black background. If the servo is slipping, it isn't controlling the airplane. Momentary slips during turbulence are normal though.
3) Have you tried experimenting with the sensitivity settings? The generic advice is to GRADUALLY turn it up until the aircraft isn't wallowing. If it's getting twitchy, you've gone too far.
4) Is the problem more prevalent in particular directions? Either N,S or E,W? This could indicate either a calibration problem or an issue with the EDC-D10A magnetometer itself. To see if it's a compass issue, see if performance improves when you're in GPS TRK mode instead of HDG mode.
5) Make sure you're running 5.3, and try the alternate control mode in roll by setting turn rate to 0, and then bank angle to something reasonable (say 25 degrees).

 

[DBRV10]

Have to back up DS post here, 5.3 and a well calibrated compass, plus get the torques right and the heading mode is pretty stable. GPS NAV mode is better.

DB

 

[lindsayj]

Another input on the over correct problem. Tonight in the hangar, I tried to find a null in HDG mode with the bug. With airplane pointed east, and the bug centered, the stick moves to the left. Move the bug a degree or two right and the stick will stop moving left and then start right. Move the bug back half way to the left to try to find a null and you can't.  Stick moves in increments up to a quarter of an inch. Take the sensitivity from 25 to 15 and the stick moves in quarter inch increments but does not respond to the bug as fast as it does at 25. But there is still no null. My guess is that this is why you see sloppy heading controi in bumpy air in HDG mode. Is there anything I can do to adjust something to get a Null when the bug is superimposed on the heading marker? This is in an RV-8A with D10A, HS34, 2 servos, and Build 5.3. Should I copy this to support@dynon.com? Thanks in advance...Jack L.

 

[dynonsupport]

Without getting into the nitty gritty - when you're on the ground, that's normal behavior.

 

[lindsayj]

Without getting into the nitty gritty - when you're on the ground, that's normal behavior.

For me, this said it all, because if the stick wobbles on the ground, it's going to do the same thing in the air, from my view. So, here's what I did: First, I adjusted the plane of my magnetometer to be exactly the same as the EFIS, not just within a degree but dead nuts on, exactly the same. Then, I carefully calibrated the servos and ran the test in sequence. Guess what. I GOT NULL in the hangar. Then I adjusted the Mag Inclination and Sensitivity to the airport Lat/Lon instead of the zip code and ran the mag calibration in a location on the airport that had a compass rose many years ago and was at least 1500 feet from any buildings. The results are impressive. HDG Hold is stable as a rock in smooth air and much better than it was in bumps. In smooth air it tracks a WAAS localizer dead nuts on in GPSS, and the ALT Hold would smoothly control a glideslope if we could control VS. The problem seems to be solved. I give lots of credit to Jakej who encouraged me by unsolicited email to seek perfect alignment of the magnetometer. There's nothing new here. Dynon gives you the requirements but you need to pay attention and take dead aim and try to do better than Dynon says, in order to get the best results, I think. I feel much better about all this now...Jack L