Mach2 motor tuning problem???

[kombayotch]

I tuned my motors and with discrete jogging, its right on. I tell it to move 100 thou, it move 100 thou and so on. Put the dial indicator to it and its great.

However, when I run G-code programs, the scale is off...

For example, if I try:

G00 X0 Y0
G01 X1.5

It doesn't go to X of +1.5, it goes to X of +2... but the DRO says its at +1.5.

Ok, this has me baffled... did I miss something in the documentation here?

Similar Threads:

• Problem with speeds, feeds, and motor tuning
• New Machine Build- Mach3 , Config, Motor Tuning problem ( FILM OF MOTOR)
• motor tuning problem
• Backlash or motor tuning problem? How do I test for backlash?
• Little puzzled on Motor tuning Mach2

[derekj308]

Hi kombayotch
Look for any G51 blocks in your program and also check the on screen scaling factor for your controller (Mach2/3?). You could also add a G50 block in which resets scaling factors back to 1. Your scaling factor is either 1.333' or 0.75 (I couldn't work out exactly what you meant travelled) so go hunting for those values in your program and your on screen scaling values. Are your other axis behaving the same way?
Cheers
derekj308

[ravin]

Hi kombayotch
HOW R U? DEREKJ is right also
but i will sugess u that check pitch of the ball secrew
i think if all is ok scalling is off and u should check the pitch in actual of the ball secrew and what u mension in the controler
one thing more if u r using pullys then re check retio of pullys of moter
Cheers

Ravin.

[kombayotch]

No G50 or G51 blocks. This is the entire program:

G00 X0 Y0
G01 X1.5
M02

Essentially, scalling is 1 when I move each axis manually (jog -> discrete), but not when I run G-code... even as simple as that above.

It is the same for both the X and the Y axis. I haven't checked the Z.

There are no pulleys, its direct drive through some Heli-cal couplings.

Other than G50 and G51, is there any other way the scale can get set for G-code but not for jogging?

[ger21]

Set it up to be accurate when you run the code, not when jogging. Trying to set up a machine by making .01 moves is probably not the best way to do it. Your screw pitch may be off, and even small amounts of backlash could throw you off when making such a small move.

[kombayotch]

Actually, I jog it different distances 0.1", 1"... several inches when tuning it. It should be acurrate for both the code and manual operations should it not?

[ger21]

If you jog it 1", does the DRO say 1", and it measures 1" exactly? If the DRO matches the distance you jog, then it should when you run the g-code. Is the DRO correct when you run the g-code.

The DRO should always match your exact position. If it doesn't, then you may be possibly losing (or gaining) steps.

[kombayotch]

DRO matches when I jog. It matches what the results SHOULD BE when I run code, but not what they actually are. Its quite puzzling...

[ger21]

If you jog 10inches does the DRO match? If the DRO doesn't match at any time, either the steps per unit is incorrect, or your losing stepsIt could be that your parallel port is not outputting enough voltage?

[kombayotch]

The DRO ALWAYS matches during jogging. It just when running code that it doesn't match the actual movement. The part that is confusing is that I'm not loosing steps, I'm GAINING them somehow???

If I tell the machine to jog 1.5": It move exactly 1.5" and the DRO says it moved 1.5"

If I use the code with a center drill in the chuck:

G00 X0 Y0 ;go to the start point
G01 Z-0.1 ;lower the tool and make a small dimple
G00 Z0 ;raise the tool
G01 X1.5 ;move the tool 1.5"
G01 Z-0.1 ;lower the tool and make another small dimple
G00 Z0 ;raise the tool
M02

The DRO says that it is at X 1.5", but its actually moved 2". I'm not gaining distance because I'm taking the diameter of the bit as part of the measurement. I'm measure the center-to-center distance between the two dimples.

[victor1946]

Hi,
Have you checked with Artsoft. I'm not sure that Mach2 uses the same software sequence to Jog versus running gcode. Have you tried using the MDI mode? They would probably be very interested in your problem.
Victor