[hrhoward]

Hello,
Have a Kiwa with the 0-M series.
The problem is this: From powerup to powerup it will not home to the same location. Usually about .200" off. As long as the control is not turned off, it will return to home at the same location.
Any ideas?

Thanks,

Heath

[Al_The_Man]

If the home posn.moves back and forth .2 between power ups, I would guess the marker is too close to the striker when it moves back off to look for the marker and register.
Al.

[hrhoward]

Thanks for the quick reply Al,
Don't know if it matters but I forgot to mention this,
It happens on both X & Y but not Z.
X-Y are controlled by the same axis controller. (A168B-2200-0200-05A)

[txcncman]

Pull the covers and clean everything under the table.

[tc429]

Sounds like grid setting isn't matched to the ballscrew- or maybe inch/metric ( parameter called 'ini' bit in parameter 1). On the grid mismatch, it's a rather lengthy description to how Fanuc uses the one rev marker/ grid offsets, I will see if I can find the sheet I typed up for our maint guys...I type slow


Anyways, to verify, need a little more info:

Didn't say if 0-a/b/c/d version...
If unsure, what are the big numbers on the screen at boot up? 453-whatever/462-whatever, etc... That can identify the control version easily.

Do you know the ballscrew pitch? You can just scale it, on Asian machines it will be a nominal MM dimension, usually 6/8/10 mm per rev.

Is the ballscrew directly coupled to the servo?

What are the last 3 digits of the servo part number?

What's in parameter 1
What's the values in parameters 4-7
What's the values in parameters 100-104
If it has a "svprm' soft key on the bottom of the parameter page, what's the values for each axis(page down to each) for AMR, n/m, and reference counter

If the grid IS mismatched, it's easy to find out:
Home machine- dont worry about the error, just home it.
Jog the machine over to touch a indicator, and zero it out.
Power the machine off without moving the slide
Power up, without going towards indicator, go directly home, put a mark on the slide.
Jog back to exact same indicator position you were at
Power down
Power up
Jog/ mpg at a moderate rate TOWARDS the indicator (has to be over 128 counts of following error in the 800 diagnostics to log the one rev) about half inch, then go home, remark slide if different.

If not different, your grid settings match your screw

If your mark don't line up, try again, and if it always repeats if first jogged in the same direction, always jumps to the other mark if first jogged the other way, then for sure, the grid parameters are wrong.


We had a few old 3M-A Enshu mills that had grid errors from the factory years go...it happens. More often though, someone put parameters in for a machine with different screws or encoders.
Fanuc is kinda odd in their homing method for incremental encoders...there's A/B quadrature pulses at 2000/2500/3000 pulses per rev, and a Z pulse once per rev. Fanuc only "sees" the very first Z pulse received at over 128 counts of following error, from that point on, "grid" pulses are generated by counting the quadrature pulses...when the count reaches a number set in the "reference counter" parameter, it GENERATES a fake grid pulse. There is a " grid offset" parameter(508-511) that adds or subtracts a fixed count from that first Z pulse, this is used to shift the grid pulse ideally half a grid from the RELEASE point of the home switch- this way you are as far as possible from misreading the switch.

This stuff is in the zero maintenance book, but the jinglish makes little sense...THE purpose of grid offset is to get the switch release point half a rev from the grid points to minimize chances of skipping home. The timing diagram in the book does show this, but I've yet to see anyone that understood it by reading the book... We dicked around with that Enshu long enough to figure it out...it's simple, but they give a description that is really hard to comprehend( for my little brain anyways)

If your grid count is 5000 and your pitch is 10 mm, you get a grid pulse 2 times per rev- not equal but it divides out as a integer and Z pulse will ALWAYS fall on a grid location... If your grid is 3000 it will not fall in the same place twice, depends on number of revs from home where it read the initial Z pulse- make sense? Starting from same physical position, then reading first Z in either direction will tell you if grid don't match one rev Z position ( or integer divisor of it).