The block diagrams in the DRV8703D-Q1 datasheet show the negative terminal of the motor connected at the SP pin. Thus, in the "slow decay" mode ("off" half of a PWM cycle), the freewheeling current flows through the low-side FET and bypasses the current-sense resistor, Rsense. To accommodate this topology, we need to connect both the positive and negative terminals of the motor directly to our control, consuming two pins on the control's connector.
However, we are very constrained in connector pin availability. To conserve connector pins, I would like to connect the negative terminal of the motor directly to ground, external to the control. This saves a pin on our connector, but it means that in the "slow decay" mode ("off" half of a PWM cycle), current will be flowing backwards through Rsense, resulting in a negative voltage at the SP pin.
The datasheet specifies an absolute limit of -0.5 V at the SP pin. In normal operation, the voltage across Rsense in our application will never exceed 30 mV (in either direction), but I can ensure that the -0.5 V limit cannot be exceeded by putting a diode in parallel with Rsense.
As long as the negative voltage at the SP pin does not drop below -0.5 V, is there a problem with this approach? How will the DRV8703D-Q1 react to a negative voltage at the SP pin?