I'm actually using wave drive profile (only a single phase on)
Aha, then you've already cracked the biggest problem. Now that's out of the way, let's return to your original question.
We need to find the phase current of the motor, and that's not straightforward. From the website spec, your motor has a phase resistance of 1.7 ohms, rated current of 1.7A and a recommended drive voltage of 12V. This means it is not suitable for direct drive with an L293D because Ohm's Law does not fit these values. Such a motor is intended to be used with a chopping drive. We can get around this by adding series resistance to the phase connections to limit the current, but it does raise the question of how you've managed to run your rig at 12V without killing your L293D, because the L293D data sheet shows a peak current maximum of 1.2A, and Ohm's Law shows your motor phase to have been taking over 3A.
So, have you killed a bunch of shields? Can you measure and report the actual phase resistance? Have you done anything else to limit the phase current?
The good news is, if we can solve the current control issues, it's looking much more likely that you can use AAs.
One last thing for the calculations, (assuming we get the current issues resolved): you energise the motor, wait some time t1 (which may be zero), advance the phase excitation to the new position, wait some time t2, then de-energise the motor. What are the times t1 and t2?