Slide 1 Slide 2 Slide 3 Slide 4 Slide 5 Slide 6 Slide 7 Slide 8 Slide 9 Slide 10 Slide 11 Slide 12 Slide 13 Slide 14 Slide 15 Slide 16 Slide 17 Slide 18 Product List
Solution InstaSPIN-BLDC Slide 4
Shown here is a traditional sensorless commutation scheme based upon timing the zero crossings of the back-EMF waveform. So what one will notice in this case is that for every commutation interval, there will be two phases that are driven—one will be driven high and the other one will be driven low. The third phase is not driven and during that time the back-EMF signal is ramping either high, or it is ramping either up or ramping down to be more accurate. So the way that the scheme works is for every undriven phase where one is looking at the back-EMF signal, one wants to measure the point that it crosses zero and that is shown here by the green lines. So what the designer is going to do is take the back-EMF signal when it is crossing zero and run that into a comparator or run it into an A to D converter input and measure the time at which it crosses zero. Then wait until the next commutation interval when another phase’s back-EMF signal crosses zero and take a timestamp on that, as well. Now, assuming that the motor is spinning in a constant velocity, take the first timestamp, subtract it from the second timestamp to get a timed delta and if the back-EMF zero crossings are occurring at the same spot within each commutation interval, one will know the width of each commutation interval because it should be the same as the timed delta that has just been calculated. So next take this timed delta, divide it by two, and then take that value and add it on to the most recently input captured time value. When that is done, set it up as an output compare and then wait for the timer to reach that compare value. Once that happens, it generates an interrupt and then that becomes ones commutation interrupt service routine. So keep doing this over and over, always calculating the delta time between the zero crossings, again, that should correspond to ones commutation zone width, and then always scheduling up a new output compare interval which should, again, if everything is working correctly, it should correspond to the time when one is supposed to commutate the machine.
PTM Published on: 2013-01-24