Y-axis phantom backlash

[Toyhacker]

Hello all,
I have a Prolight 2000 benchmill with newly reground ball screws. In cutting a test piece I've noticed that the x-axis seems dead-on but the y-axis seems to be cutting about .003 under. When I put an indicator on it I don't see any backlash (at least not .003 worth). I also don't see the indicator move when I push and pull on the table. Can anyone think of another source for the error?

Thanks.

[TXFred]

Is the column flexing?

[Toyhacker]

Good question. The column itself seems to be a piece of granite and it's pretty rigid. However, the milling head rides up and down on a pair of guide rods and they could be worn. If I pull on the chuck, I can deflect the dial indicator about .001. I'll see if I can tighten the slides (there's a set screw adjustment) to eliminate that.

Having said that, I'm cutting machinable wax with a 1/8" dia x 1" long flute end mill and I'm taking a .015 finishing cut. It doesn't seem like that would put much side load on the head. Still it's a good place to start.

Thanks

[Toyhacker]

So I've measured the table travel of the Prolight 2000 in the x direction (because it's easier to set up), moving the table six inches to the left and back in one inch increments. The results are:
Commanded Move______Measured Move
0.000_________________0
1.000_________________1.000
2.000_________________2.0025
3.000 ________________3.0030
4.000_________________4.0035
5.000_________________5.0045
6.000_________________6.0060
5.000________________ 5.0050
4.000_________________4.0040
3.000_________________3.0035
2.000_________________2.0030
1.000_________________1.0005
0.000_________________0

The other possibility would be the belts. I replaced them both when I reinstalled the ball screws after they were reground. It's hard to see how that would matter but I'm open to suggestions at this time.

Regards

[SCzEngrgGroup]

So I've measured the table travel of the Prolight 2000 in the x direction (because it's easier to set up), moving the table six inches to the left and back in one inch increments. The results are:
Commanded Move______Measured Move
0.000_________________0
1.000_________________1.000
2.000_________________2.0025
3.000 ________________3.0030
4.000_________________4.0035
5.000_________________5.0045
6.000_________________6.0060
5.000________________ 5.0050
4.000_________________4.0040
3.000_________________3.0035
2.000_________________2.0030
1.000_________________1.0005
0.000_________________0

The other possibility would be the belts. I replaced them both when I reinstalled the ball screws after they were reground. It's hard to see how that would matter but I'm open to suggestions at this time.

Regards

It would be difficult for the belt itself to cause a problem, and the pulleys would create a cyclic error, that varied depending on the angular position of the pulley. What you're seeing almost certainly indicates rather poor accuracy of your screws. The good news is, it appears to be a monotonic error (increases linearly with position), at least over that range of motion, which can be greatly mitigated by tuning the steps/unit in Mach3.

Regards,
Ray L.

[Toyhacker]

Mach3 is a possibiity, but I'd like to try to get up and running with what I've got before I make any more "upgrades".

Does anyone know if the original PLM2000 software contains a file to correct for the ball screw error? I've looked through it several times, but nothing jumps out.

Regards