7/11/2018 9:38 AM | |
Joined: 10/11/2006 Last visit: 9/2/2024 Posts: 5902 Rating:
|
Hello, for me the ripple of +/-2 rpm seems to be quite normal. Because the traced r61 is the unsmoothed raw value which comes from the encoder evaluation directly. At an encoder with sin/cos- tracks you have maximum / minimums, zero crossings and so on. To get the information, if the ripples comes from the encoder, mechanic, gears or the motor (for example pole, notches), you can do a trace at 60 rpm (=1Hz = 1rev/s). So in that trace you can isolate the orders of the oscillation by a FFT of the r61 (time limited trace with 1000ms and a sampling time of 125µs). Further you can do a trace of the r479 (actual position). The SINAMICS has a very powerful trace tool (it is possible to trace in the current controller cycle and also math. Function are possible). So here you will see the position deviation (+/-100 the encoder has a pulse number of 512 and a fine interpolation of 2^p418 = 2048 – so one revolution is 512 * 2048 = 1048576; so 100 means approx. 0,034° in my example). Following the FFT of the signal. You see a maximum at 2048Hz (4 x 512) and also at 512Hz, 1024Hz, .. The low harmonics (which comes typically from the mechanic or motor) has a lower order – for example 18 motor notches. In my case they are factor 6 lower! And in my example I haven´t a high inertia which do a smoothing and I haven´t a 1FT7 (which has very low torque ripples). So in other words: It would be absolute permissible to do a smoothing of 1ms (which smooth the encoder parts in the encoder signal, but still not smooth the parts of the mechanic). For that I would use the monitoring functions. So you can parametrize p2153=1ms and trace the r2169. So you can compare r61 and r2169 in one trace at the same scale. Don´t do a smoothing with p1441 in that case, because it would have an influence in the control loop. And a smoothing time in the control loop creates a dead time could have more negative effects (reduction of the dynamic). Attachmentr61.zip (499 Downloads) |
With best regards, |
|
This contribution was helpful to
1 thankful Users |
7/12/2018 3:01 AM | |
Joined: 3/14/2017 Last visit: 8/12/2021 Posts: 3 Rating:
|
Thank you all too much. Actually, i went to a Siemens branch last month, siemens engineer just show me the r63 value of two type of motor whitch were drived by the same system, the encoder resolution of 1FT motor is 2048(see photo 1FT), and the 1FK is 512(photo 1FK7), the r61_1FT is very stable, 1000rpm ± 0.15, and r61_1FK7 is 1000rpm ± 2. what the siemens engineer could confirm is r61 is not the real value of motor speed, and r61 was effected by encoder resolution, but he could not tell me what the real ripple is. the attachment is Scout configuration named 20180619. Finally, acording to your reply above, the real ripple may be 0.034°/r, at the same time , i can understand how the value came out. At last, I will do some test according to your guide above. later i will report my test to you all. Attachment20180619.zip (463 Downloads) |
Follow us on