Mach3 - weird coodinate shift with inch/mm gcode change

[Clive S]

Everything is setup to mm.

home your machine then zero the X then in the MDI put g0x100 f500 and check if the machine moves 100mm

[CnCAbc]

home your machine then zero the X then in the MDI put g0x100 f500 and check if the machine moves 100mm

If I put g0x100 f500, my machine will run out of its limits. But if I put g0x10 f500 it will move 100mm (10cm) perfectly.

Thanks

[Clive S]

If I put g0x100 f500, my machine will run out of its limits. But if I put g0x10 f500 it will move 100mm (10cm) perfectly.

Well then I think that the steps per are wrong on the drives

[CnCAbc]

Well then I think that the steps per are wrong on the drives

I set them using the "Set steps per unit" feature, under the settings tab. It's like they're X10.
My driver are set to 1/8 (1600 pluse/Rev) .
I am using Tr8*8-2p (4 starts). And nema 23 @ 1.8°

[hanermo2]

The co-ord stuff and scaling is sometimes iffy.
Mostly, it is not wrong, necessarily.

If the machine was set up with a certain abs. position, homed, then you use a different offset g54xx, it is unclear which units should be used.

Since the work co-ords are not tied to a scalar, how is the sw to know what is/was the correct offset base in global terms.
If a work co-ord is offset say 10,0 mm, and I switch to imperial, should the offset base adjust by /25.4. Why ?

Likewise, should the machine global values adjust, or not ?
Why ?

If the home switch in z is offset 10.000- mm in z, metric, and I switch to imperial, should the home-pos value adjust ?
If I read in the value into a parameter say #3000, should that number adjust ?
Why ?

I am NOT saying it is right, logical, correct.
What I am pointing out is that it is very complex with no clear right answers, mostly.

The UCxxx, and the chinese offline controllers, are mostly much less complex and don´t support a lot of the features in gcode.

[magicniner]

Good call Clive!
I rarely think of the basic machine set up being at fault! :D

[CnCAbc]

Good call Clive!
I rarely think of the basic machine set up being at fault! :D

I got it figured out! It was an extra decimal point. i.e step per was set 2000.xxx instead of 200.xxx
I guess, I assumed the step per start at 1600 and goes up.

Appreciate your help, everyone!

Now the step per is resolved, I need help to figure out why everything except square, rectangle, and straight lines come up deformed.

The shapes below were supposed to be 20mm circle, and 20x20x20mm triangle.

[Clive S]

Now the step per is resolved, I need help to figure out why everything except square, rectangle, and straight lines come up deformed.

Could be missing steps ie. acceleration too high or voltage to the steppers too low. Or even excessive backlash. Or couplings loose on the motors.

Glad you got the steps per right now.

[CnCAbc]

Could be missing steps ie. acceleration too high or voltage to the steppers too low. Or even excessive backlash. Or couplings loose on the motors.

Glad you got the steps per right now.

How to sort them out?
The only things I'm sure is not the issue, are the voltage to steppers and coupling, the seems/feel tight/fine.