All of the grounds on the 1201 are internally connected, so your two power supplies won’t be fully isolated. Back EMF shouldn’t be a problem for your MCU supply, but motor noise could affect your signals and connected power supplies. Even if you had full electrical isolation, however, motor noise could still induce noise in nearby wires. There are a number of steps you can take to reduce the effect of motor noise on your system if it ends up being a problem (solder 0.1 uF capacitors across your motor terminals, route power lines away from signal lines, keep your power/motor leads as short and thick as possible, twist your motor and power leads around each other and/or add a ferrite bead).
Given the connections shown in your diagram, you should not need the extra link between your grounds of your two power supplies. Maybe your connection between the MCU’s ground and the motor driver is not good? Can you show a diagram of the connections you think should work but do not? The MCU and motor driver must share a common ground (Jan talked about this in more detail in another of your threads). There are potentially three ground connections you need to make to the motor driver:
- The motor power ground.
- The logic power ground.
- Ground from your control signal source.
Since you are powering the driver from a regulator on your control signal source (your MCU board), connections 2 gets you connection 3 for free (or vice versa). If your motor power source was the power source for your MCU board, then connection 1 would get you connections 2 and 3 for free, but it’s still good practice to include a separate ground connection between the boards for your digital signals.
By the way, you aren’t actually using the motors shown in your diagram, are you? The TB6612FNG motor driver would be underpowered for those motors.
- Ben