Comments on: Linear servo motor sensorless homing tutorial

By: Tero

Tero — Fri, 17 Oct 2014 16:49:54 +0000

4 MHz is a default limit but on TODO list is a plan to make it adaptive based on frequency actual demand. Hardware supports up to 30 MHz count rate. Also due to higher resolution encoders becoming better available, also resolution support needs to be increased in standard FW. So someday that should be no problem, but until then, such high rez encoders need customized FW.

By: TonyT

TonyT — Tue, 30 Sep 2014 17:08:06 +0000

That’s what I’ve heard,however, here’s a story:
Many years ago, we built a large (~500mm) system with air bearings over ground granite, with a precision Swiss ball screw with flexures to drive it, and dual encoder feedback. We bought a couple of the subsystems from other suppliers, including the Galil-based motion control system from Vendor I.

Vendor I said that same thing, however, in the system they actually delivered, the linear load encoder had significantly higher resolution than the motor’s rotary encoder. We might have had some issues with the system, but motor instability wasn’t one of them. Of course, our system didn’t have any significant backlash….

By: Jason

Jason — Mon, 29 Sep 2014 18:42:07 +0000

That’s exciting! How would Argon/Ion work with very high res incremental encoders, like 1,380,000 counts per pole pair in a linear motor? Is max input and hose rate still 4Mhz?

Regards,

Jason

By: walter

walter — Sun, 28 Sep 2014 00:06:17 +0000

Thank you.

By: Tero

Tero — Fri, 26 Sep 2014 11:07:35 +0000

Good to know! 500 LPI is quite low but I hope they start making higher res too.

Yesterday I received components for the first linear encoder prototype. It has less than 1 µm resolution.

By: Jason Jack

Jason Jack — Thu, 25 Sep 2014 22:06:28 +0000

Hi Tero,

U.S. Digital have low cost linear encoders. Currently their range has a maximum resolution of 500 lines per inch, but they are developing a higher resolution encoder. No idea what the price on it will be, but their other products are pretty reasonably priced.

Regards,

Jason

By: Tero

Tero — Thu, 25 Sep 2014 19:40:03 +0000

At the moment we should have everything we need for dual-loop development. But thanks for helping hand! I appreciate that!

The statement makes sense. At least having motor FB higher resolution than linear FB is preferred than the other way around. I don’t think system would become easily unstable but it might “hunt” more for final position.

By: walter

walter — Thu, 25 Sep 2014 14:34:14 +0000

Do you need one for testing purposes? (if so, what resolution and length would be most useful). Also, I found this comment on the internet and I wanted to know if this statement is true:

“It is necessary for the motor encoder to have a greater effective resolution then the load encoder to make sure that the controller will see commanded movements of the motor well before the backlash is taken up on the load. Otherwise, the controller could get into a mode where it is building up too much error before it sees any response from the mechanics(load). This can lead to a slight oscillation around the reference position, or even cause the system to become unstable if the backlash is large enough.”

Do you agree with that?

By: Tero

Tero — Wed, 24 Sep 2014 10:23:14 +0000

Yes! On software side, it’s been developed with Ion and finally the outcome should be transferable to Argon (not 1:1 but helps a lot to have it in Ion). We’re also thinking to produce low cost & high resolution linear encoders as there seems to be absence of them. Or do anyone know where to find new linear encoders around €100?

By: walter

walter — Sat, 20 Sep 2014 02:11:43 +0000

Tero, have you done any work on Argon dual feedback?