The customer has a very well equipped 1992 Beech A36 Bonanza. The customer reports an intermittent nav tracking problem with the autopilot. The customer reports, after a period of time the autopilot will make an un commanded left turn. The problem has been occurring for about 1 year and the customer is finally tired of it and wants the problem corrected.

The avionics involved with the problem is a KFC-150 autopilot, Sandel EHSI, and Garmin GNS-530. The problem has been looked at by another shop, but no resolution has been found. The other shop did try a different KFC-150 autopilot controller, but the problem was present with the replacement KFC-150 autopilot controller. The only other data point is that the display on the Sandel is always good, even when the autopilot is acting up.

We either have a bad course datum signal from the Sandel, a bad GNS-530, bad roll servo, or an intermittent wiring problem. Everyone hates intermittent problems, here is how Penn Avionics figured this one out.

The customer left the plane at Penn Avionics for us to repair. We test flew the aircraft. After about a 1 hour flight, the autopilot did in-fact make an un commanded left turn away from the correct course.  After about 15 seconds of confusion, the autopilot did correct itself and re-intercept the desired course. The only thing we could determine during the event is that the course datum output from the Sandel, was still able to drive the autopilot, and the left/right display on the Sandel was always centered even when the autopilot was making the hard left turn. Before we could gather any additional information, the autopilot starting working fine and did not act up again.

We returned to N10 and landed the plane. We had a hunch that maybe the analog left/right outputs of the GNS-530 were the cause of the problem.  The Sandel is on the 429 data bus and that data bus is not what is driving the autopilot. The analog left/right outputs from the 530 drive the autopilot in the nav mode (along with the course datum signal from the Sandel).

We pulled the GNS-530 for a bench test. We mated the 530 up with a Garmin GI-106A CDI on the bench and hooked up an outside antenna. A quick bench checked, showed the left/right was OK, and the GI-106A CDI showed correct left/right deflection.

Not convinced that the 530 was innocent, we put the 530 in our environmental chamber. We had the GI-106A hooked up again with the outside antenna. We entered a waypoint on the 530 and the CDI  displayed a centered course as it should have. We set the chamber to 115 deg F, and waited to see what happened. Sure enough, after about 30 minutes the left/right pointer on the GI-106A started to intermittently pin to the left (along with the up/down pinning to the top). We had found the culprit.

We replaced the main board in the GNS-530, and the left/right was solid. We replaced the 530 in the aircraft and sent the customer on his way.

Copyright 2004 Penn Avionics, Inc rev 12/2004

Penn Avionics, Inc 610-409-0328

www.pennavionics.com