I have a bit of a stumper. I brought my plane up to an avionics shop to do an installation of a dual G5 system connected to my M20J’s Century 21 Autopilot. When I got the plane back I found that the autopilot in DG mode could not hold a heading and would hunt with rather extreme bank angles around a heading point. Namely, it would go into steep turns (+/-45 degrees+) left and right as it crossed the heading bug in HDG mode and I would have to take over the controls by force and disable autopilot to prevent this undesirable operation. NAV mode in GPSS also shows the same behavior. The Attitude indicator had to be retained (probably for this very reason). On the ground, when the engine was running and the attitude indicator was level, engaging the autopilot where the HDG=HDG bug and HDG mode enabled, the roll motor would hold ailerons level. When the engine shuts down, the AI often settles in a left banking position. When this happens, the ailerons respond by rolling to the right, so it’s clear that the autopilot is responding appropriately to the attitude input, and therefore the attitude connection, which was unmodified on connector CD176 seems correctly connected. Before the G5 upgrade the autopilot worked fine with the original 52D54M gyro.
I made sure that the jumper connecting pins 13 and 14 (DG select pin) on connector CD-194 was removed since the 52D54M gyro was replaced by a G5 HSI as per Century Flight System’s recommendation. The G5 install manual incorrectly targets different jumper pins on CD-185, which does not exist on a C21 autopilot. This makes me question the reliability of the G5 install manual, but that’s another issue.
It appears that on connector CD175, someone took a grey wire out (perhaps it is the HDG signal that is on pin 17 but I don’t know) and wired it to a non-shielded 22 AWG wire. I would probably prefer that it be shielded but I’m sure it’s not the issue as it’s probably heavily filtered. Pin 2, however, is a blue wire that is called course select and it is also supposed to be connected to the GAD29B. Both signals need to be connected and shielded to the GAD29B as per Garmin manuals, though I wonder if there are potential erratas as the information provided is incorrect. CD175 is left hanging and while the grey wire (suspected to be Pin 17, HDG signal) is connected, Pin 2 is left floating. When powered on, the blue wire pin 2 measures 4.825V using a Fluke 87 IV meter wrt airframe ground. Pulling the GTN750 and GAD29B circuit breakers did not change this voltage measurement, making me question if this signal is connected to anything of interest.
I’m supposed to bring the plane next weekend in to the avionics shop for a fix, but if anyone has experience working with the C21 and G5s / GAD29B and can chime in on potential issues, it would be helpful because the shop is trying to do the work on it over the weekend while I wait (it’s about 100 miles away from me).
I know that the CIII autopilot has radio couplers, etc, but this connection is direct. While there is a GPSS scaling factor, there is no option for HDG bug scaling factor so I wonder if the voltage outputs going to the Century 21 are correct. I also have heard that the C21 is supposed to be limited to +/-29.5 degrees of bank in spite of any inputs on HDG or CRS which is a bit strange too. Any ideas?
I did notice this: https://www.beechtalk.com/forums/viewtopic.php?f=21&t=71222 . It does seem the attitude input is suspect because it should be clamped to max deflection, but on the ground it seems to respond, so I'm not 100 percent sure, but perhaps as one poster suggests the signals are intermittent after being relocated from one location to another.
Thank you!