OK,.
Now it's up to you to prod around with the power on, according to previous suggestions.
What we need to do is to identify which half of which circuit (there are 4 coils) is on permanently.
Now, it's just occurred to me that because you've 2 identical circuits, each of which is always on, on it's left side, or otherwise on its right side, that we have a danger of confusion. In each circuit, we will find one half that is permanently energised, one correctly so, and on in failure mode. (Unless both have failed) However, being aware of that, we shouldn't need to engage too many more brain cells.
You need to be very cautious at this point (and thereby hangs a tale a couple of weeks ago in my own motor controller for a 16mm ng engine!!) not to connect the Base of any transistor directly to its corresponding power rail (the +ve rail for npn; -ve rail for pnp) lest you invoke the alternative name for a transistor, "The fastest fuse on 3 legs"!
My reason for saying that is that when diagnosing the faults in my own controller, I was prodding around with a couple of leads, either to force a transistor to come on by injecting Base current through a suitable resistor, or forcing it to shut off by a base-emitter short.
Another thought has occurred, and not wishing to worry you, because the motor is stalled in one position, the current through the coils, no longer being commutated, might rise to a damaging level and be causing further mayhem.