[Thomas Utley]

All,

I'm writing to solicit opinions on "coupled" motion such as driving the X-axis of a gantry router from both sides of the table. Which of the following is better?

Option A: "Mechanical" coupling (example: both sides of a gantry driven by a rack and pinion system actuated by a single servo/stepper motor via belts/gears/shafts)

Option B: "Electrical" coupling (example: each side of a gantry driven by twin servos/steppers but synced together via control software, with or without closed loop feedback)

Followup question: Which common control software (Mach3, EMC/LinuxCNC, Flashcut, etc.) is better at Option B?

Thanks for your thoughtful comments,

Tom

[Al_The_Man]

To use option B with servo's you would need a control capable of gearing, this is a typical application using a Galil card, but should be possible in EMC or Dynomotion.
Al.

[Thomas Utley]

To use option B with servo's you would need a control capable of gearing, this is a typical application using a Galil card, but should be possible in EMC or Dynomotion.
Al.

Al, I don't follow. What do you mean by "control capable of gearing"? Under either option A or B, you might or might not have gear ratios involved depending upon the desired positional resolution you need to achieve for your particular machine. "Gearing" includes leadscrews/ballscrews, by the way, under my definition.

[Al_The_Man]

Gearing as it pertains to CNC motion controllers is for example, in the case of the Galil controllers I use in the same application, where a Slave axis is geared or synchronized to the encoder of a Master axis and follows it according to a set ratio, which in the case of Galil can be changed on the fly if you need to.
It is essentially electronic gearing.
MotionCode - Galil: We Move The World
Al.
.

[Thomas Utley]

Al, I don't follow. What do you mean by "control capable of gearing"? Under either option A or B, you might or might not have gear ratios involved depending upon the desired positional resolution you need to achieve for your particular machine. "Gearing" includes leadscrews/ballscrews, by the way, under my definition.

OK. I think I follow you now. You're referring to coupling motion between two axes, such as rigid tapping wherein you need the z-axis vertical position to vary at a set ratio with the rotary motion of the spindle to generate threads.

My question was (excuse the pun) geared more toward a single axis driven by either two motors synchronized by software or a single motor with some sort of mechanical means to bridge from one side to the other. Generally speaking, this would be to keep the X-axis square to the Y-axis and avoid "racking" issues when reactive forces away from the centerline are generated during milling.

[BAMCNC.COM]

Well.. I'm no expert here but a driver card can do this without software, I only know this because I've owned a RepRap Prusa, which has 2 Z axis motors, and as far as I knew, software had no clue.

[Al_The_Man]

My question was (excuse the pun) geared more toward a single axis driven by either two motors synchronized by software

That is exactly the example I mentioned by the use of gearing, which is the only way using servo's, one must be the master and the other the slave.

Well.. I'm no expert here but a driver card can do this without software, I only know this because I've owned a RepRap Prusa, which has 2 Z axis motors, and as far as I knew, software had no clue.

It is done using steppers, but not possible to have two servo axis, each with an identical independent command.
Al.

[underthetire]

Mach 3 has the option to slave a motor. Easy check box selection. Industrial controls have the option as well, just a lot more set up.

[Al_The_Man]

How does Mach slave a servo, especially with no feedback to Mach?
Industrial controls do it through the encoder feedback synchronization.
With Galil it is two, two letter commands and it is done.
Al.

[underthetire]

He said stepper, or servo, and it is possible with the Ajax control using servos and Mach with feedback.

[Al_The_Man]

Yes the same as Dynomotion does it with Mach and Galil with gearing, with servo encoder slaving.
Mach is just used as an operator front end.
But it requires a common DSP or equivalent controller, the regular mach step/dir through the P.P. will not do it however.
Al.

[erikjgreen]

I believe Mach can do it with step/direction interface drives, or possibly with closed loop servos, but either way it's not perfect synchronization, it's just "good enough" for most uses, IE it'll avoid racking if nothing gets stuck.

Perfect synchronization of two servo motors is the realm of specialized hardware as you mentioned.

Erik

[Thomas Utley]

I believe Mach can do it with step/direction interface drives, or possibly with closed loop servos, but either way it's not perfect synchronization, it's just "good enough" for most uses, IE it'll avoid racking if nothing gets stuck.

Perfect synchronization of two servo motors is the realm of specialized hardware as you mentioned.

Erik

So then why not simply couple both screws together mechanically with a belt, chain, or rack & pinion and drive with a single motor big enough to handle the accel/torque demand?

[erikjgreen]

So then why not simply couple both screws together mechanically with a belt, chain, or rack & pinion and drive with a single motor big enough to handle the accel/torque demand?

Lots of reasons.

A major one is backlash, followed by simplicity of design, efficiency, cost.

If you need a lot of power, two motors are usually a better bang for your buck overall than one bigger one with mechanical coupling.

[Al_The_Man]

Both methods have been used in the past, it just depends on the mechanics as to whether the single motor mechanically coupled is capable of driving both sides simultaneously without the result of inertia causing a slight lag on acceleration and slight racking being the result.
i.e. Take into account the weight of the Gantry.
The likes of MG Systems used the dual servo approach on some of their heavy gantry plasma tables, using a motor and planetary G.B. on to a R&P.
Al.