Don't know if anyone has had this problem. Searched the forum didn't find anything. My problem is with my X axis I calibrated it. It's good in the positive direction but in the negative direction it's out by about 1/2" more negative. If anyone can help thank you.

Suggest you check that all the mechanical stuff is right, ie; that a pulley is not slipping or something is loose.
RICH

I have checked all the mechanical stuff everything is tight. It cuts small parts very well. It only happens on larger cuts. I'm using threaded rods & nuts.

Could it be that there is some tightness or binding in the threaded rods which is greater in one direction than the other ?

Tweakie.

Even a plain threaded rod with nuts won't have 1/2" backlash.................something is loose.
RICH

I just re-calibrated the X axis again. Not sure if the size of bit I'm using would make a difference, using a .25. I'm measuring from the center of the bit it seems like I'm off now by 1/8". I'm moving 8" x+ back to zero it's good center to center of bit. I've marked a board with  an 8" mark. Going in the negative direction it lands on the outside edge not the center off bit, so I got an extra 1/8" .

May I suggest you use Mach's calibration feature which is available via F6 or the Settings Tab in ach Mill.
RICH

Even a plain threaded rod with nuts won't have 1/2" backlash.................something is loose.
RICH

Rich,

You have to think outside the box.
An axis which is tight in one direction only can loose steps in that direction and then be 1/2" out when it returns, nothing needs to be loose. 

Tweakie.

Tweakie,
The problem is that one can't hear what the motors sound like or have a visual on the movement. Figure that if it can move in one direction
without binding and skipping steps one would assume it can reverse. At 1/2" it's not loosing steps it is "skipping"...play on words but a big difference between the two.
If the steps per unit are correct, a reasonable velocity & acceleration along with proper pulse duration, then after a move the backlash would be rather consistant and thus any
loss in position greater than the backlash becomes a mechanical issue............at least in general. Of course other assumptions are made in the above ........ you know how it goes.
RICH  

I found my problem. It wasn't a mechanical or mach setup problem. It was a dumb ass human error. I put my center point on the peace I was cutting in the wrong place in relation to my model's center point. OK call me a dumb ass. Plus it wasn't a 100% calibrated. It's all good now.

Thanks for all your help