Need Help!- Cylindricity

[plifter365]

Looking at a 18i-T controller

3 machines- one of which is producing a wave in cylindricity.

The machine makes a ball screw, so it interpulates rotation (C-Axis) and translation (Z-Axis) linearly. X-Axis defines diameter.

I end up with a wave in my cylindricity of the ball groove. At first glance it looks just like ovality from runout. But all bearings, workholding, and setups have been checked, and rechecked. Ballscrews and such also replaced. All with no effect on the pattern. I am trying to find out if there is anything that could cause the X-Axis not to hold position, or to dither. We are looking for an amplitute of 15micron. I just want to make sure that i have not missed anything in the controller. All parameters were checked against the other machine.

Let me know if you have any ideas. Thanks

[memoryman]

Interesting. Just to make sure: the amplitude is ~.0006"?
Does the amplitude vary with the rotation angle? Or with the thread pitch?.

[plifter365]

Yup 6 tenths....it the amplitude changes some with thread pitch, 1-2 micron. But the timing peak to peak is still 180 degrees of rotation, regardless of pitch. I am basically making an oval if you looked the the trace of the ball from the end. Interestingly enough the wave pattern is a sharp drop, then slow climb to next peak, followed by another sharp drop and so on.

[memoryman]

so, if you were to plot the diameter deviation vs the rotation angle, it would look like a saw tooth?

[Dan Fritz]

Sounds like some kind of axis compensation function is turned on. It could also be an axis "correction" of the interpolated position after a motion is complete. For example, if you make an interpolated arc with G02/G03 and the end point does not lie exactly on the programed arc, then there will be a slight "correction" when the arc finishes in one axis. You will be able to see this correction on the machine position display, however.

Are you programing one long Z move with one multiple-rotation C move in the same block, or are you programing a single C axis rotation with a Z motion multiple times? A single rotation programmed multiple times might cause trouble if you're using any kind of tool wear compensation at the same time.

I would try to set any tool radius or X-Z compensation offset to zero and try a test cut. If the problem goes away, then you may have to program one long Z motion with multiple C rotations to prevent the cutter comp from doing anything funny at the end of each block.

There are also many compensation functions built into the control (like "Beam Sag" comp on mills). These comp functions work a bit like pitch error comp, and you WON'T see any deviation in the axis position on the machine position display.

If one of those comp functions have a value that affects the X axis as the C or Z axis is moved, then you may get the kind of funny X motions that you are getting. Look in the parameter 5700 area for anything that looks like it shouldn't be there. Remember that the C axis position goes up to 360 degrees, then suddenly goes back to 0, then climbs to 360 again. This coincides with the X deviation that you're seeing, so a "straightness compensation" that ties X and C together would cause this.

Hope this helps a bit.

[tc429]

not sure if it could appear as x deviation, nor even sure if its got a separate encoder- but might be worth a look at the spindle encoder if its got one - loose pulley, or dirt in one of the belt teeth? I had one that was badly worn (like .015 wobble on the shaft!) causing some lunging/tool chipping issues...unlikely, but thought I'd throw it out there...
Tim

[underthetire]

I think the first thing to do is to narrow down your search. Put an indicator on x and see if it is actually moving during the cut. Put a test bar in the spindle, check run out, we use a inspection ball in a fixture, witch allows us to check end play and rotation. Eleminate all the mechanical problems first. Then dive in to the control. I run in to a lot of weird problems like this that are usually mechanical, but not all the time. Some of our machines are accurate to .00001 or less, and it takes a huge amount of time to narrow it down.