Acorn6!!!???

[Gary Campbell]

Yes, I like the look of those additional encoder inputs! I'd like to think they will enable closed loop position control at some stage. Perhaps not immediately but.....

AcornSIX is a step & direction (position) controller. Don't hold your breath.........

[cnckeith]

its possible in the future you can use scales for position.

[suntravel]

Yes, I like the look of those additional encoder inputs! I'd like to think they will enable closed loop position control at some stage. Perhaps not immediately but.....

AcornSIX is a step & direction (position) controller. Don't hold your breath.........

Closed loop with scales can work also with steppers, but with crappy mechanics it is like pearls before swine

And if the mechanics are solid and precise you don't really need scales for <0,01mm tolerance.

Uwe

[rk9268vc]

If you can use scales i wonder if you could make a macro that uses the scales to automatically figure out backlash comp

[rk9268vc]

And if the mechanics are solid and precise you don't really need scales for <0,01mm tolerance.

if you have 0.0005" of backlash and want to make a part with tighter tolerance than that, then scales would be awesome.
Same reason you use scales on a manual bridgeport

or if you are converting an old worn out machine, or manual machine, scales would be godsend

[tblough]

If you can use scales i wonder if you could make a macro that uses the scales to automatically figure out backlash comp

You don't use backlash compensation with scales.

[Black Forest]

If you can use scales i wonder if you could make a macro that uses the scales to automatically figure out backlash comp

You don't use backlash compensation with scales.

I might be wrong but I think he is referring to mapping the ballscrews automatically rather than move a .5" note the reading on the scale, move another .5" map the reading on the scale, rinse and repeat.

[suntravel]

And if the mechanics are solid and precise you don't really need scales for <0,01mm tolerance.

if you have 0.0005" of backlash and want to make a part with tighter tolerance than that, then scales would be awesome.
Same reason you use scales on a manual bridgeport

or if you are converting an old worn out machine, or manual machine, scales would be godsend

I am using DRO on manual machines for faster higher resolution reading and for the mat functions.

Be sure linear scales are not god send healers for worn out machines
Most professional CNCs using closed loop from the encoder on the servo, not from scales.
That works fine if the machine is geometrical ok and preloaded guides an ball screws are used.
If it is worn out, get your hands dirty and fix it

Also think about what happens if the machine made from steel rises 30° in temperature and the scales made from aluminum also....

'Uwe

[tblough]

Most glass scales have a CTE in the 10-6 range and when you are working towards the tolerances a scale equipped machine offers, you probable have those machines in a temperature controlled environment.

When using scales, you don't use screw mapping or backlash compensation.

[Muzzer]

Yes, it's a common urban myth that you can compensate for backlash using closed loop control and everything magically becomes perfect. Try driving a car with half a turn of backlash at the steering wheel along a winding hilly track and see how you (the closed loop controller) get on. With a nonlinear plant model containing significant hysteresis, you have your work cut out. As Uwe says, the only sensible solution is to tackle the mechanical issue at source.

My other CNC lathe has closed loop position control using 5um glass scales and LinuxCNC. In practice, you achieve perhaps "only" 10um tolerances but that's a transformation compared to how it was as a manual machine and almost certainly better than it would be with just the closed loop servos. Some of the inaccuracies are due not just to backlash but effects such as springiness in the drivetrain. You tighten up the gibs too much to reduce slop (one contributor to backlash) and end up with a different kind of lost movement. There is some improvement possible by closing the position loop, so if position encoder feedback is available within Acorn6, a controller function might be of value in some applications.