The Artificial Horizon — Page 177, Lesson 211

We're now into the heart of the air-driven artificial horizon's inner workings. I want to walk you through the correction system that keeps the gyro vertical, and then the one big weakness this design has — acceleration error. First, the correction system. We have four pendulous vanes, and they're arranged as two opposing pairs fixed to a common spindle. So vanes A and C form one pair, and vanes B and D form another pair on their own spindle. The key idea is that these vanes hang under gravity, and they control airflow through slots in the rotor housing. Here's the principle: if the rotor axis wanders from the vertical, one vane in a pair will hang clear of its slot, leaving that slot open to unrestricted airflow, while the opposite slot gets completely obstructed by its vane. That creates an unbalanced airflow, and this unbalanced airflow precesses the gyro, correcting the tilt and returning the gyro axis to the vertical. Let me make that concrete with the geometry. Suppose the gyro has wandered from the vertical. Vanes A and C are unaffected — they stay half covering their slots. But vanes B and D, on their common spindle, hang down. That means slot B is now closed, and slot D is wide open. A strong jet of air exits through D, and by Newton's third law, that jet produces an equal and opposite reaction, which we call 'R', acting on the gyro. Now here's the crucial part — precession. This reaction R is precessed through 90° in the direction of rotor spin, which is anticlockwise when viewed from the top. So the actual restoring force acts in the direction we call 'P', and that's what restores the gyro axis back to the vertical. The reaction doesn't act where you'd intuitively expect it — it's shifted 90° around by the gyroscopic effect. Now, this whole system depends on the pendulous vanes being affected by the earth's gravity. And that's exactly where the weakness lies. The vanes will respond to any acceleration, not just gravity. So when an aircraft accelerates in a level attitude — like during the take-off run — we get a false indication. Specifically, a false nose-up, right wing down, or climbing right hand turn indication. Let me break that into the two components. First, pitch error. During acceleration, the lateral vanes lag — they swing back towards the pilot. That opens the starboard slot and closes the port slot. This produces a reaction 'R' acting to port. By the rule of precession, the effect on the gyro is as if the direction of R had been moved 90° in the direction of rotor spin, which is anticlockwise. So the gyro gets precessed out of vertical, with the base moving backwards towards the pilot. That movement is transmitted through the guide pin and the horizon bar arm, which brings the horizon bar below the gull-wing — giving you a nose-up indication. So during acceleration, the instrument falsely tells you the nose is up. The roll error is different in origin. That one comes from the inertia of the bottom-heavy rotor housing. The housing is deliberately bottom-heavy for its own reasons, but under acceleration that inertia causes a false roll indication — the right wing down part of that false climbing right hand turn. So the takeaway here is this: the air-driven artificial horizon is self-correcting through those pendulous vanes and precession, but because it relies on gravity sensing, it's fooled by linear acceleration. That's why you see those false indications during the take-off run. The pitch error comes from the lateral vanes lagging, and the roll error comes from the inertia of the bottom-heavy rotor housing.