Guys sorry to be posting again but I'm going crazy here,

OK here is the issue, I set up perfectly round bar verticly on my table and got the centre with a co-ax indicator I then moved above the workpiece and did a few G0 moves and jogged about a bit I then pressed go to home assuming it would return perfectly to the center I then lowered the Z axis with the preset co-ax and tested it, the y axis was  .01 out of the first reading but the x axis was .1 out I then got the center again and repeated again this time i got .05 out in the y and .11 in the x , I have tried a few more times and it seems to change around within these parameters

I have set my steps via a 1-2-3 block ( and gauge ) and got them to within .01

I have also dropped the acceleration by half but this did not help

I'm kinda thinking backlash but whats throwing me off is that it constantly changes and considering i spend £1000 on ballscresw I cant see how it would be this much

I have made sure the tension on the pulleys and belts is good

The nema steppers are new ( everything is new )

If anyone has any suggestions that would be great

Failing this im going to strap my head to the table and do a Z move right threw my brain

Cheers

What you are doing would only work if you had very accurate homing switches

I suggest you find your zero THEN ZERO THE AXES in G54 or higher then run around testing then GOTO ZERO  ie G00 X0 Y0 from MDI

Then test for accuracy.

Thanks Phil,

I will try that , I have no homing or limit switches on my machine but surely that wouldn't make a difference as its not referencing with machine coordinates before going to home ( by referencing I mean mean moving to the  machine coordinates first then moving to home )?

Guys in my first post I say I press go to home button what i actually mean is I press go to zero button

For some reason I always call it the go to home button

Anyway sorry for any confusion

Are you measuring in MM or inches ?

Have you got a link to the ballscrews and nuts you bought ?

Hi Phil i think its backlash so i got the vice on the mill and put in a 1-2-3 block i then got my 3D tester and zeroed off to the edge I then set the mach dros to zero and then set the jog steps to .01mm per click and moved in the opposite direction ,, it took 6-7 clicks before the dial moved to the first .01 graduation, after this every click measured .01mm i then went in the opposite direction again and it again took 6-7 steps before i had movement on the dial (again after it started moving it was .01 mm with every click of the mouse ) I'm having the same issue on the z axis and y axis

The ball-screws were an optional extra with the machine ( wabeco f1210e   www.wabeco-remscheid.de )

I have also checked this against another dial gauge i have and am getting the same readings

I would be right in saying this is backlash and that is seems pretty high for £1000 ball screws??

I have just seen this, in your last PM said you thought it was cured.

Ok Backlash can come from bearings, mounts, pulleys etc as well as the screws so check them as well. First thing I would do is mount a clock (DTI) to the bearing housing and have the tip onto the end of the ballscrew, step jog one way then the other and see if the needle moves, if it does then you have looseness in the bearings.
Also check for slackness in the pulleys on both motors and screws.
Flex in mounts of ballscrews can also give you the backlash but you should be able to figure out a way of mounting the clock to check for that as well.
Hood

Also check gibs are snug.
Hood

Hi Hood I will check that, I also thought it was ok also till i did another batch and I was getting errors again,

Either way the backlash seems huge for a new machine

That seems far to much for a ball screw to me.
However ... you should also check for noise on the encoder lines (unlikely?) and for any backlash in the bearings at the two ends of the ball screws. The latter can really throw you off.

Cheers