The incident

I flew through some light snow showers on my way to Kingston with my daughter this morning, so I turned on the pitot heat just before joining the circuit to make sure the pitot blade was clear. At the end of the downwind leg I slowed the engine, reduced power, dropped flaps, verified 70-80 knot airspeed, turned a tight base over the icy water of Lake Ontario, then looked again at the airspeed indicator (ASI).

35 knots. Way below stall speed.

But the plane was flying fine. The nose wasn’t high, the controls weren’t mushy, the stall buzzer wasn’t blaring, the wings weren’t buffeting, and most importantly, the ice floes weren’t spinning and getting larger in the windshield. I gently pushed the nose down enough to speed up 5-10 knots, but still the needle didn’t move. I checked the altimeter and it was behaving properly, showing a slow descent towards field elevation. That meant a pitot failure.

The trickiest part was the turn to final, almost immediately after the failure, when I’d barely had time to process it — it’s easy to lose airspeed in a turn, even with a functioning ASI. After that, it was pretty much a normal approach and landing (no point declaring an emergency when the runway is less than a minute away). The ASI flickered back to life on short final to show that I was 5-10 knots above my normal approach speed. It froze again at some point during the flare and landing (I don’t look at the panel once I’m past the airport fence), then gradually climbed to 90 knots as I taxied in to park the plane.

The aftermath and resolution

I called an AME (mechanic) at the airport, tested the pitot system by blowing gently into it (no joy), then went out for lunch so that I wouldn’t stay around fretting. Three hours later, the AME hadn’t had time to get to the plane yet, and the ASI still wasn’t responding to the blow test, so I decided to try something else (with the AME’s blessing): I started the plane, turned on the pitot heat, then did a high-speed taxi down the 5,000 ft runway.

The needle climbed again during slow taxi, then dropped at the start of my high-speed run, then climbed up again — then, suddenly, at the very end, it started responding normally. Since there was no other traffic, I turned around and did the same thing the other way, and this time, the needle responded normally the whole way. I taxied around, did pre-takeoff checks, then went back to the runway for a real takeoff roll, prepared to abort halfway if the ASI wasn’t behaving — no problem at all, all the way home (though my mode C encoder started acting up, because there’s a law of physics that at least one thing always has to be broken on an airplane).

The analysis

There must have been some snow or ice near the opening of my pitot blade. Turning on the heat partly melted it and let it get into the (pin-sized) hole, and the water blocked the pitot line, possibly as slush or even a tiny ice crystal. My high-speed taxis, combined with the pitot heat, forced the blockage the rest of the way through the line and cleared it.

Pitot heat on was a good idea, but turning it on just before joining the circuit wasn’t. Lesson: make as few configuration changes as possible when you’re close to landing — if something’s already working, why mess with it? If I’d turned on the pitot heat 10 or 15 minutes earlier, I would have had the ASI failure at 5,500 ft, where it was no risk at all, instead of in the most dangerous possible phase of flight, and it would have worked itself out before I had time to land anywhere. Since I hadn’t turned it on earlier, I shouldn’t have turned it on at all.

In the end, no harm, no cost, and a little bit of extra confidence that I can handle a plane by feel when the ASI fails, at least in VMC.