I am using a 3 phase BLDC motor with current rating of 20 A and voltage rating of 18V. At the moment zero cross detection scheme has been employed in the floating phase for position detection of the rotor while two phases are energized at a time. In the start, motor is commutated for a predefined interval until a specific speed is reached. After that, the motor is commutated normally using BEMF detection method. The problem is sometimes commutation works fine and sometimes not. In other words it is not reliable. For example, i have to play a lot with the parameters so that commutation works fine until smooth BEMF signals are acquired. Still then, the results are not reliable.
How to proceed in this situation? any ideas or references would be appreciated!
The "no-brainer" is to use 3 hall sensors to get position updates every 120 degrees. sin/cos angular senors would even allow for a continuous position assessment.
Your current scheme seems to be a "sensorless" one. As you already found out, the BEMF acquisition is somewhat tricky. I know, TI (ti.com) had (maybe still has) some optimized solution for this. Might require some searching on their website, but I'd expect you'll come up with some documentation/application note(s) and alike.
The problem is I must stick to the "sensorless" scheme. I have gone through a lot of applications notes and documentations. But couldn't find a concrete reliable solution in case forced commutation fails.
Where forced commutation fails and BEMF voltage is still too low, there is no sensorless "solution". You may apply max. power and may look for overcurrent conditions once the motor is assumed to rotate, but basically there is little you can do at lowest speeds - beyond keeping the rpm ramp-up low.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing