Encoder signal to servo drive & controller; split or through drive?

I am setting up my first CNC. I am using a PC with LinuxCNC and Mesa boards 5i25 and 7i77. I am using Fanuc AC servo motors powered by Allen Bradley Ultra series 100 servo drives.

I chose that Mesa board combo because it would allow me to have encoder feedback to the controller (as well as to the drive) for better control.
Now as far as actually wiring it up, I was planning to split the encoder signal at a terminal strip; one branch going to the servo drive (on connector J2), and one branch going to the Mesa board, like so:
http://www.cnczone.com/forums/attach...d=370906&stc=1 <== this is from the AB Ultra 100 manual and is fine to do per the manual

As I read the AB Ultra 100 manual, it describes an encoder output:
http://www.cnczone.com/forums/attach...d=370908&stc=1
So the encoder goes into the drive on J2, pulses divided (or not), and then sent out on J1 to the controller.

Which of these methods is best? Is there any advantage to using the encoder outputs on J1?

How are you interfacing the Fanuc AC commutation signals, as normally the Fanuc uses a proprietary commutation signal?
Al.

Quote:

---

Originally Posted by Al_The_Man

How are you interfacing the Fanuc AC commutation signals, as normally the Fanuc uses a proprietary commutation signal?
Al.

---

I have no idea but it was working previously. When I bought the machine it had Dynapath Delta 20 controls interfaced to these same drive and motors. All I'm changing is the controls. Maybe the previous owner put ordinary encoders on it.

By piggybacking on the encoder lines I would be concerned about introducing noise into the system. I have always used the encoder output from the drive in any system I have put together. Never tried it the other way.

noise is bad

use good shielding and/or ferrites

there is more than one way to skin a cat, smoke em if you got em

In a good motion control card there is no real need for a intelligent drive, and the need for dual loop system.
A simple PWM drive can be used.
There is a member here that made a convertor for Fanuc servo's.
It is usually that or replacing the encoder.
Fanuc 4 bit shown converts to pseudo sine wave.
Al.

Quote:

---

Originally Posted by Al_The_Man

In a good motion control card there is no real need for a intelligent drive, and the need for dual loop system.
A simple PWM drive can be used.
There is a member here that made a convertor for Fanuc servo's.
It is usually that or replacing the encoder.
Fanuc 4 bit shown converts to pseudo sine wave.
Al.

---

Are you saying the 5i25/7i77 is not a good motion control card? Or are you saying I can forego the encoder feedback to the card? Or am I missing the point?

The (intelligent?) AB Ultra 100 Servo drives are already in place and I have not budgeted to replace them. Should I consider replacing them in your opinion?

First I cannot comment on the Mesa as I have not used them, but from utube videos etc they appear to do a good job, I tend to always use ±10vdc analog drives which I believe these card will do.
And with a decent controller where the encoder goes back to the controller card, a simple drive such as the A-M-C, Copley etc can be used, these are simple torque mode (trans-conductance) drives.
This avoids expensive intelligent drives such as the AB which AFAIK requires software to set them up, parameters etc? And which entails tuning two loops.
If you already have them try commissioning them first before replacing.
Maybe the Linux forum may have someone who has used them.
Interesting to know how the Fanuc commutation was handled.
Also are they used as original 3 phase servo motors or BLDC with the AB drives?
Al.