This week, I gave my new autopilot its first workout in instrument conditions, climbing out through a low cloud layer in Ottawa, bumping through heavy rain clouds over Vermont and New Hampshire, and flying a localizer approach into Beverly Airport near Boston.

First, I’ll provide a bit of background. The autopilot has four modes as currently configured (without altitude support):

• ST: stabilization mode, which keeps the wings level (or in a bank angle selected by rotating a small knob). In this mode, I can fly a hyper-stabilized plane, much like the over-simplified flight model in Microsoft Flight Simulator. Other autopilots sometimes refer to this as WL (wing-leveler mode).

• HD: heading mode, which keeps the plane flying on a specific heading that I select using a bug on the heading indicator. The bank angle will vary as necessary to maintain the heading; for example, if I put the plane in a yaw using the rudder pedals, the autopilot will lower one wing noticeably to keep the plane flying in a straight line, and if turbulence knocks me off a heading, the AP will turn the plane to correct.

• TRK (LO): VOR-tracking mode, which keeps the VOR CDI centered using a low-pass filter (that means that it reacts slowly, so that it’s not constantly zig-zagging when the VOR scallops, as they often do.

• TRK (HI): localizer-tracking mode, which keeps the VOR CDI centred using more-aggressive corrections, as you’d want on an ILS or LOC approach, where it’s essential to stay right on the centreline.

The stabilization mode (ST) is virtually useless for me – I’ve enabled it only in my initial tests and in pre-flight checks. If I want to steer the plane using the autopilot, it makes much more sense to twirl the heading bug on heading indicator and have the AP follow it.

For enroute, I’ve tried both the heading mode (HD) and the VOR-tracking mode (TRK LO). The trouble with the VOR mode is that VOR needles scallop a lot back and forth, so even with the low-pass filter, the plane still gently weaves during flight. It’s nice because I can just set and forget until VOR passage — no heading corrections required — but especially in IMC, I found the heading mode much more effective. With gentle adjustments, I was able to keep the virtual CDI in my portable GPS perfectly centered, even in cloud and mild turbulence, sometimes going as long as 15 minutes without having to touch the heading bug.

I tried the localizer-tracking mode (TRK HI) for my localizer approach into Beverly. At 10 miles back, it didn’t seem to be doing a good job, sometimes allowing nearly a half deflection on the CDI, so I disengaged the AP before I got too low and flew the rest of the approach by hand (flying an approach in IMC isn’t a good time to troubleshoot). That might just have been an anomaly, though, and I didn’t test it inside the final-approach fix (FAF), when signal might be better — I’ll do a test flight on the ILS approaches in Ottawa in visual conditions to see how it performs there.

I’m learning to deal better with the lack of altitude hold. On my first couple of flights, I was trying not to touch the yoke when it was in HD mode, and instead to correct altitude deviations using the trim. Even though I worked hard to damp out oscillations (anticipating them by trimming before the pitch reversed direction), it was still a challenge in turbulence.

On my Boston flight in IMC, I took a different approach, and tried sharing the yoke with the AP: I pulled or pushed it to control pitch and then trimmed to relieve pressure (the way I normally would), but managed to get used to letting the AP still turn the yoke while I was holding it. I was able to trim more effectively that way, and didn’t have to correct as often, even in turbulence.

The big conclusion, though, is that the AP made a huge difference in fatigue. I can (and have) hand-flown harder and longer flights in instrument conditions, but I’m very tired after I land from the constant second-to-second attention required for the scan and other IFR tasks. The autopilot let me relax a little bit, study the approach plates more carefully, pay more attention to the engine gauges (which can slip out of an IFR scan, when keeping the plane upright is the main concern), and just generally relax, while still keeping an active (but less frenetic) scan.