Hello,
you can read "reliable and robust estimation of flux, angle, speed and torque across use conditions", "Best speed input into the control System", "high quality feedback signals", ...
but is there a rough estimation of the resolution of the estimated speed? I do understand that the estimation gets more inaccurate for low speed, but if my motor spins with 50 Hz for example, is the estimated speed accurate at <0,1Hz, approx. 0,1Hz, 1Hz or more than 1 Hz?
Thank you,
Carsten
Hi Carsten,
You won't have any problems with the resolution of the speed variable. Since speed is derived as an observer output within FAST, we are only limited by numerical resolution, which is very high.
Please take a look at the speed output plot below taken from the FAST observer operating with a slowly increasing speed command operating around 50 Hz. The red curve shows the velocity response for ADC resolution = 12 bits, voltage and current gain errors are 0.2%, offset voltage = 2mV, and offset current = 2 mA. Offset errors create a speed ripple equal to the motor frequency. Gain errors create a speed ripple equal to 2x the motor frequency.
The green curve shows the velocity response for ADC resolution = 12 bits, but all gain and offset errors are set to zero.
As you can see, the velocity resolution is irrelevant compared to the noise resulting from measurement errors.
Best Regards,
Dave
