KC Lee gave me a link showing a technique to improve grounding on scope probes. Thank you!

I immediately try it and found that it does reduce stray coupling from the board.

Unfortunately, the noise is still persists.

I later found that it is actually the high impedance balance leads that cause significant voltage drop every time the bridge transitions.

Next plans:

*Partition analogue and digital grounds

*Provide some power decoupling for the opamp

*Stiffen the supply rails for the synchronous buck converter chip

*Add flyback diodes to catch fast bemf

*Maybe replace the balance leads with lower gauge