[Deano7/11]

guys
am considering a lathe for retrofit and have been looking at a gsk retrofit kit. the only thing is the feedback loop is closed in the drive only and not the control.

what are the pro's and cons of this?

i assume the control is then effectively 'dumb' and only assumes the servos have made it to the position it has instructed?

what are your thoughts about this?

[Al_The_Man]

In these the loop is closed only to the drive as you say, and they are responsible for detecting errors and pass back to the controller, usually in the form of an e-stop.
Having always worked with commercial CNC I have always preferred to use the systems that close the loop back to the controller.
In my case it has been the in-slot motion cards, one of the many nice advantages is the electronic gearing, for variable gearing 1 or more slave motors off a master.
Another that some of the stand alone drive do not offer is the ability to disable the drive and turn the motor during an E-stop, when the motor is re-enabled the system knows exactly where it is.
Max.

[thebodger]

Definitely close the loop to the control is preferable.
For all the reasons that Al has stated, plus the greater tunablity and troubleshooting.
I have been servicing industrial grade CNC machines for about 40 years, all shapes and sizes, I have never ran across a control that didn't close the loop to the control.
In fact most newer controls feed the encoder to both drive and control.

[Deano7/11]

great replies guys. exactly the info i was looking for.

feedback to the controller it is then.

looks like linux is the way i am heading then.

[thebodger]

I would look at linux for sure.
It's very cost effective -> read free.
However consider a very good value for money add in card from Mesa or Pico.

[mactec54]

Deano7/11

Most good servo systems today, will have the Encoder feed back to the Servo Drive, then the Encoder signal goes from the Drive to the Control, the control is closing the servo loop, what they call a dumb servo drive, does not have the Encoder feed to the Servo Drive, it just goes straight to the control

[Al_The_Man]

The systems I use & prefer do not have an encoder or any kind of feedback to the drive, the encoder returns to the motion controller/trajectory planer only.
The drives are operated in torque (transconductance) mode.
Using mainly Galil Motion cards now.
Al.

[mactec54]

The systems I use & prefer do not have an encoder or any kind of feedback to the drive, the encoder returns to the motion controller/trajectory planer only.
The drives are operated in torque (transconductance) mode.
Using mainly Galil Motion cards now.
Al.

Yes you are using non intelligent Drives, sometimes referred to as dumb drives, these can be easier to set up sometimes

It works just the same, Yaskawa Mitsubishi Etc there Encoders all feed back to the Servo Drives, then the Servo Drive is connected to the Galil or what ever motion control card/ hardware you are using, they use just the same Velocity/Torque mode,the prefered method for cnc control

[Al_The_Man]

, these can be easier to set up sometimes

Not really easier, the PID loop still has to be tuned, just that it is done in the controller and not the drive.
This enables the drive to be a simple transconductance amplifier i.e. no DSp processor required.
Al.

[Larken]

guys
am considering a lathe for retrofit and have been looking at a gsk retrofit kit. the only thing is the feedback loop is closed in the drive only and not the control.
what are the pro's and cons of this?
i assume the control is then effectively 'dumb' and only assumes the servos have made it to the position it has instructed?
what are your thoughts about this?

If you using Mach3 then the loop needs to be closed to the drive , since you can't close it to the control anyway. As far as having Mach3 keep track of the encoder position when the drive is disabled .... there are ways of doing this (in the near future).

Imo there is almost no difference in performance between closing the loop at the drive and closing the loop at the control . Both have pro's and cons.
The PID loop in the smart Drive is just as powerful as the PID loop in the control. Both are trying to keep the machine position the same as the commanded position. With the Step/dir drive the Commanded position is transfered to the drive in real time through the step/dir. A bonus of using the Step/dir drive is it takes a lot of headroom off the control s/w.
But if your machine has a glass scale and there is backlash, then closing the loop to the control has its advantages.
Larry

[Al_The_Man]

There are many instructional videos on the Galil site by Dr Jacon Tal, that outline the theory of PID tuning and other aspects of motion control, they are worth checking out.
Although they cover Galil they general include many universal common aspects of motion control.
Having used Galil for over 30yrs, IMO they are very hard to beat in the way of a PC based motion controller, Galil in particular where you get one card controlling up to 8 axis at 12Mhz encoder F.B.
Al.

[thebodger]

Hello Larry.
Your comment about glass scales and backlash is correct.
However there is usually a limit to the amount of backlash that can exist, and still have the system work.
Usually a very small amount can be tolerated. Less than .001". More than that and the system runs very rough, or flies off into wild oscillations.
Or it may not work at all.
However the latest generation of high accuracy finishing machines my company is buying are equipped with direct drive linear servo motors.
Backlash is not an issue because there is none.
if I'm a bit off topic sorry bout that.

[Al_The_Man]

However there is usually a limit to the amount of backlash that can exist, and still have the system work.
.

One of the Galil instructional video's covers the dual feedback system, scale and motor encoder to combat backlash, the PID loop is divided between the two.
Al.