You are using two coils at 90 degrees to to create a spinning magnetic flux so that your magnet (needle) aligns with it yes?

I think the problem is largely the way the flux density from each air cored coil is almost constant (close to and across) the face of each coil. Outside the "projected" tube formed by the coil, the flux varies with distance at a much steeper rate and round about halfway between the two quadrature coils the change in flux per angle turned will be quite linear. I'll try and make a picture that helps. Bear with me.

You are using two coils at 90 degrees to to create a spinning magnetic flux so that your magnet (needle) aligns with it yes?

I think the problem is largely the way the flux density from each air cored coil is almost constant (close to and across) the face of each coil. Outside the "projected" tube formed by the coil, the flux varies with distance at a much steeper rate and round about halfway between the two quadrature coils the change in flux per angle turned will be quite linear: -

Looking at the picture above you can see that the density of the lines of flux at the ends of the electromagnet (solenoid) is fairly constant. This means for a given excitation current in the coil you'd measure near enough constant flux density across the "projected" face of the solenoid. You are then reliant only on the reduction or increase in flux from the non-aligned coil to move the needle. As this coil is now a significant distance from the aligned coil, the flux density it produces is much less.

I guess this is the problem of an air cored motor like you describe. If you looked at a more conventional motor, its stator windings are really close to each other: -

In addition to them being close to each other, the mag field is distributed by the iron core making the problem you describe very minimal.