Autotune Problem
Recently I ran into a problem autotuning a drive. Mechanically, the application was a large accumulator with a very massive festoon. We used an excellent drive running in vector control with an encoder. We were concerned that a rotating autotune would run past the end of travel for the accumulator. Therefore a non-rotating autotune was used.
The result was that autotune was never completed. The software displayed “Measuring Stator Resistance” for a long, long 30 minutes.
With the mechanical brake released and hands hovering over the E-Stop buttons, we performed the rotating autotune. This was completed with all sorts of encouraging messages from the drive software. Supposedly we had a bandwidth or response of 5 radians/second. On running the drive, we noticed a considerable lag in response.
Manually checking the speed regulator revealed an actual response of 0.5 radians/second. That would take 4 to 5 seconds to reach the desired speed. Far too slow for an accumulator designed to take up slack web when the winder stops. By manually tuning the drive, 5 radians/second was achieved.
Lessons learned:
1-Autotune is better than it was 20 years ago, but is not foolproof.
2-Most drives do not verify the result of autotune. You must check bandwidth or response manually.
Retired Manufacturing Engineer - Specializing in Process & Automation
8 个月Over the years tuning Vector drives, I have had bad results with rotating auto tuning algorithms. First bad experience was with a turret winder with Siemens 6SE-70 AC vector drives. Speed loop gain was 5 times too high. Second time was with Powerflex 755 AC vector drive on machine planing 2x4 lumber. The planing head was heavy so mechanical inertia was very high. After rotating auto tune the drive would fault & not run Have found that static tune works, does not cause problems & very rarely fails.
Troubleshooter, Teacher, Web-Handler
8 个月As a mechanical engineer, the last autotune I was (in)directly involved in was a decade ago. 5 identical digital printing presses with state of the art Mitsubishi servo drives. Every autotune gave different results, most unrunnable. Once a runnable configuration was found, it was copied to the other 4. None worked. Each had to be retuned :(