Historically tuning has been a bit science, a bit art, and possibly a little luck. I know hardcore engineers will say it’s all science – which is true – if you knew all the variables in the system. Unfortunately, whenever I’m helping a customer, I hear the comment “about this much” when asking for details about a system. Not anyone’s fault, but it makes computation virtually impossible and the job still needs to be done!
Auto-tune has gotten better, but Load Observer takes it to a new level. Surprisingly, so few people use it and it’s as easy as toggling the “ON” button. The Load Observer feature manages the load resistance and low frequency dynamics in a system. It eliminates the 10:1 inertia ratio guideline and effectively manages material deformation, such as compliance, in the system.
When the load observer function is enabled, it operates at the drive update rate (NOT the coarse update rate, we are talking microseconds) while the machine runs. During machine operation, the load observer function dynamically estimates the torque required to move the mechanical load and provides the torque required to cancel its effect. The result to the drive: the motor appears as though it is unloaded, which is easy to control. In fact, the load ratio should be zero in the Logix Designer interface when using the Load Observer feature. For this reason, it is recommended to program the motion instructions in actual units, since the default settings for the dynamic limits are very high for an unloaded motor. (Don’t use percent of maximum, unless they are reduced significantly… in fact, just never use percent of maximum, just my opinion).
In effect, it’s our easy button for most applications. The goal is performance AND stability. Technically there is a fine line between the maximum performance and the minimum stability.
I have successfully implemented it in several applications, and here are the highlights. These may not reflect any Rockwell prescribed practice and are only my personal “observations”.
2. Load observer seems to work best with ratios that are on the higher side. Those say below 3:1, the effects are minimal in its default state, unless there is compliance. You’ve already done your homework with an excellent inertia ratio.
3. It works with induction motors with one caveat: you must know the inertia of the motor, or at the very least have an estimate. When entering in custom data, the inertia is not necessary to get the motor to spin, but it is for the observer to function (ref fig 2). I suppose that the observer cannot make a motor appear as 1:1 if the value it’s trying to reach is zero perhaps a div/0 error (again I do not know for sure, this commentary is derived from empirical data, meaning I just tried it!) Most NEMA motors have inertia units in WK^2 which is the same as LB-FT^2
4. Make use of it for vertical applications (especially belted), where maintenance such as belt tension is prescribed, but rarely followed. Also, it can save tooling, due to lack of the auto-tune process and incorrectly set fault actions