Question:
Why does the display of the actual speed value fluctuate under C0051, although the motor behaviour is alright? Why is the fluctuation of incremental encoders higher than of resolvers or SinCos encoders?
Answer:
The output signal of the rotary transducer can actually deviate in single scanning cycles more or less from the actual value. On average - after several scanning cycles - there will be the actual speed value.
Cycle times of EVS9300 servo inverter:
- Encoder evaluation: 125 µs
- Speed controller: 250 µs
- Position controller: 1000 µs
The deviation's frequencies of the actual speed value are very high in comparison to the time constants of motor and mechanics. Consequently, the machine is like a filter, so that the displayed speed fluctuations have normally no consequences for the drive process.
Analog signals of the resolver are continuously evaluated. In case of incremental encoders a time window of 125 µs counts the pulses according to the number of pulses per revolution including four fold evaluation. In case of SinCos encoder there will be no additional continuous interpolation of the analog voltage level. Practically, the evaluation solely via a frequency (incremental encoder) has negative consequences for a constant speed display, especially in case of low speed.
The display under C0051 always reflects an instantaneous value (unfiltered). After an update time of ca. 1 s in the GDC's monitor a fluctuation of the display becomes visible, this has, however, not this influence on the drive behaviour (e. g. smooth running).
For a constant presentation of the actual speed value a signal filtering, e. g. via a PT 1 element, can be useful. In the control loop a filtering should normally, however, be avoided. They have negative influence on the stability and lower the dynamic response.
Remark:
In case of a master frequency coupling you should prefer the usage of a set value to the signal MCTRL-NACT (actual speed value) due to the mentioned reasons (virtual master).
