Our organisation uses SM23165D motors to precision locate and hold a diagnostic device - we hope to microns typ (0.005 - 0.01mm) control level. Brakes are required on the system as the diagnostic element is driven by a ballscrew which will backdrive under constant load when the power is removed.
During tests with another system I observed a lag between the power off and brake sufficient to cause a motion jump of 0.1mm. Not very much - but a lot to us.
Furthermore the 0.1mm jump occurs each time the power could cut off - in the same direction. So there could be a cumulative build up of error which could be ultimately very large without the operators knowing it (unless they return to the home position after they know there is a power out).
Do you have data regarding the lag between power off and brake applied in terms of angle on the motor (I can then confirm this is the same linear 'jump' seen on the ballscrew)
You are using the motor in default ail safe mode. This mode just applies power to the brake at all times unless power is totally removed or the motor faults.
This is known as BRKSRV or brake servo.
Alternate would be BRKTRJ or Brake Trajectory.
In this mode, once that command (BRKTRJ) is issues, the processor will apply the brake (De-energize) any time there is no active trajectory in progress.
This will work best when running in Position Mode. Under this state, the processor will energize and release the brake when a Go command is issues and then when destination is reached, it will de-energize and engage the brake automatically. This will occur for all moves. The result is that thew motor will already have its brake engaged before power is lost assuming power was not lost while moving.
so bottom line, issue BRKTRJ at boot up or the top of your downloaded program and give it a try.