Shape-based Feedforward Tuning:

Final Adjustments

You can see that at the end of the tuning of Kfff it appears that we have some Kaff error.

	10,000 count move Trapezoidal move:       5e5 accel, decel, vel	Green - Commanded Velocity              Yellow - Commanded Acceleration                    White - Position Error
	Kfff = 500 Kvff = 0 Kaff = 100,000 Kp = 100 Ki = 0 Kd = 2,000 Peak Position Error = 7 counts

After you have adjusted Kaff, Kvff, and Kfff, you may notice some slight errors that were not as evident before. This is particularly normal if you started with a large position error peak. In our example, the original peak position error = 187 counts before we adjusted Kaff.

Now that our peak position error is around 7 counts, some slight errors are now visible. We will now attempt to resolve these few remaining errors and lower our peak position error even more.

First, let's adjust Kaff. Make Kaff = 101,000.

	10,000 count move Trapezoidal move:       5e5 accel, decel, vel	Green - Commanded Velocity              Yellow - Commanded Acceleration                    White - Position Error
	Kfff = 500 Kvff = 0 Kaff = 101,000 Kp = 100 Ki = 0 Kd = 2,000 Peak Position Error = 5 counts

As you can see, the peak position error is a bit lower now.

In order to visually see how much we were able to reduce the position error by using Shape-based Feedforward Tuning, look at the Before and After plots of our data using the same YScale.

Before Kaff = 0 Kvff = 0 Kfff = 0

After Kaff = 101,000 Kvff = 40          Kfff = 500

Previous | Next