Originally Posted by
Water pilot
The fact that the AOA sensor uses set screws instead of indexing shows to me that it was never considered to be a part critical to safety. If there was one moment when Boeing "jumped the shark" it was when they promoted a part whose failure was rather innocuous to one that had authority over a large flight control surface. I think we might be going somewhat down the same road with automobiles, the parking radar that beeps when you get too close to a wall can now slam on the brakes at 70mph. One hopes that automotive engineers are more attentive to Murphy's law than aerospace engineers.
It looks to me that the setscrews index the resolver to the balanced assembly of the vane and the counterweight and that the counterweight is indexed to the vane with a flat on the side of the vane shaft. One can see the flat-bottomed groove in the counterweight that receives the vane shaft. Maybe there was a failure of one or more of the screws that hold the item 45, shaft retainer (?), in place. If so, the resolver would still be attached and the vane would mostly follow the oncoming wind. It seems unlikely as the plane was not exposed to a high g load at the time; I would expect it to fail that way on a hard landing or bumping down the runway during take-off.
While it might be the case that the resolver connection could slip, there isn't any resistance except to angular rate of the vane accelerating the core of the resolver to create loads passing through those set screws. It they were completely loose, the resolver would have no torque to turn it and it would report an unchanging value. Resolvers are small and very easy to turn; those loads are small.
In the case of Lion Air, it's possible the inaccurate resolver was indexed 20 degrees off from where it should have been; it remained exactly that amount through both flights. ET302 saw the resolver reading in accordance with aircraft acceleration just like a plumb bob would follow.
One final thing - the gear attachment is not adjusted to change the calibration of the sensor, the body of the resolver is turned to do that after a rough alignment according to the gear teeth, so the gear could be an interference fit on the resolver shaft and the vane shaft with the setscrew being a belt for suspenders approach. The resolver body is indexed with three cleats.