minimize the chattering of a linear stage driven by a voice coil actuator(VCA)

[iyearning]

I have a linear stage driven by a voice coil actuator (VCA), a capacitive probe is used to measure the movement of the stage. A picture is attached to show the basic mechanical design of the stage. Even though the resolution of the VCA is 0.5 um, the overall resolution of the stage is not that good. When I use controllers to achieve position controller of the stage, the stage is always chattering around the set value. It can only achieve 5 um in open loop and around 10 um in PID control. Is there any method to improve the performance by modifying the controlling VI? 

I am using NI 6363 to give voltage signal to a LAA amplifier and the amplifier will convert the voltage signal to current signal and drive the VCA. It seems that the output of the LAA amplifier is not constant and it is the LAA amplifier that generate those chattering. The command voltage given to the amplifier is extremely constant with error less than 2mv, which will cause only 0.5 um displacement of the stage by theory. Such kind of motion is similar to the resolution of the sensor but it is only 1/10 of the actual motion noise which is on the scale of 5 um.

The stage has a natural frequency of 30 Hz, which is determined by the stiffness of the springs and the mass of the moving stage. When I use FFT to analysis the movement of the stage, I found that the main noise comes from 30 Hz and 50 Hz (both for open lood and close loop PID). I also use FFT to analysis the command voltage given to the VCA in close loop PID control, the peak frequencies are 10 Hz, 30 Hz, 40 Hz, 50 Hz, 70Hz, 90Hz....

Do I need to use some filter between the amplifier and VCA so that it can reduce the noise without influence the accuracy? Or I can reduce such noise just by using some noise block in the VI program?
Thanks
best

[FuManWho]

Hi iyearning,

It seems that your 6363 is sending a clean signal, so changing your VI may not help because all it would do would be to change the response from your 6363, which does not seem to be the issue. A filter seems like a good idea between your LAA and your VCA as it seems that there is some noise being introduced from your LAA somehow. One thing I have to ask is if you are properly shielding your cables as the magnetic field from your VCA might be causing noise as well.

[iyearning]

Hi Paolo,

It seem the influence from the magnetic field is not a big issue in the position control of a voice coil stage(VCS). I attached some picture of the commecial design of similar VCS and all these stages do not shield the cable connecting the VCA and controller. Anyway, I will try this method to see if it can make the result better.

BTW, for the filter, could you please recommend some model for me? As the current noise is just around the level of 1 mA, I think I may need a very precise filter.

Thanks

best

[iyearning]

The pictures I attached are from the website of H2W. Those stages all acheive 1 micron resolution which is limited by the resolution of the linear encoder. The hardware of those stages are quite similar to my stage and the encoder in my stage have 1 micron resolutin too.

[Highland_Controls]

The command signal to the VCA should be a current signal, and as such, has no position control abilities.  An open loop command should result in a constant current, and that would then be a constant force.  If you are using a capacitive probe for feedback, then I am assuming that is an analog signal.  Analog position feedback is always noisy.  What is the resolution of the capacitive probe?  What about the resolution of the analog input?

[iyearning]

Dear Highland_Controls

You are right. The command signal to VCA is a current signal and results in a force output of VCA. As the VCA is fixed to a flexible bearing with stiffness of k, so it will result a displacement which will be sensed by cap probe. The cap probe do measure analog signal and it has a resolution of 0.3 um.

For the analog input of the VCA, we do not get the specification talking about its resolution. The analog current is from the amplifier. The spec sheet just tell us the linearity is around 6%. I attached the spec sheet of the amplifier for your reference.

Thanks

best

[Highland_Controls]

Analog is horrible for tight position feedback.  Post the spec sheet for the probe.  How many bits resolution is your analog input?

[iyearning]

Hi Highland_Controls

I made a mistake. I am not sure if the position feedback using analog signal or not. It just deliver an analog output signal to the NI DAQ system. The resolution of the cap probe is remarkable, you can find the spec of the sensor CPL190. The RMS resolution of my device is 0.03 um.

I am still contacting the vendor about the resolution of the amplifier which is not listed on the sepc sheet.

Thanks

best