Originally Posted by
pasta
I never really thought that through, but it's blindingly obvious now you point it out.
How does your Pa's compass stop the needle dipping without inhibiting its rotation?
This is the early '50s prototype. The magnet is under the pointer and atop a long thin and very light shaft which runs through a jewel bearing in a plate near the top and on down to another at the bottom. To put it simply, the needle just can't dip and any sideways load on the bearings is too small to matter.
It was designed for use in gliders - early units were loosely mounted so that they could be tilted to match the bank angle in thermals. Later units got an external gimbal. The advantage over a floating ball type was the very low inertia, and they were pretty popular until the fluxgate arrived.