Attached are 2 Gcodes. The first is the complete code. The second is just a few lines from the first. Something with G17 arc plane arcs that are connected seems to be causing the X and Y program extrema to calculating wrong.

warfboy,
What values are you getting for the extrema?
RICH

Screen shot of program extrema attached. It should be much smaller.

Thanks,
I just wanted to make sure i wasn't doing something strange. I can't fix it and have never seen it before, but, may have a clue on what is happening and will post shortly with some info so Brain can take a look.
RICH

swarfboy,
What program generated the code?

MODIFIED: added a few more questions..
Was any scaling used on the drawing on export?
Was any scale factor used in MACH?

MODIFIED:11/12
What size cutter?
What version of mach?
I am curious about this. Notice that there are crop circles in the code. Yopu would think the extrema would be based on the
part dimensions shown, but it may be a bug that if there crop circles exist the program bases it on something else. Knowing what cutter size would probably help in evaluating the intrepetation of extrema. You could take the view that something is wrong, or ,
on the other hand, giving additional info to pinpoint something saves time.
Your nickle swarfboy, 
RICH

swarfboy,
Attached is a screen shot of the cuts and a CAD drawing of your posted gcode.
RICH

Cutter size doesn't matter.
No scaling was used.
In mach scaling is not changed = 1:1
This code is for cutting in mm but it doesn't matter.
Visualmill 6 Pro generated the code.
Mach backplots it fine as long as IJ Mode is set to absolute in the general config. If you set it to incremental you will get something like you have with the arcs not making sense.
Its a Mach 3.42.15 bug where its not calcing the program extrema properly.
The extrema should be approximately  X min=0, Xmax=13, Ymin=- 7 , Ymax=7 (these are roundabout numbers I just rounded to a whole integer)

I just downloaded the latest dev version .018 which installed as a V.017, and it seems like V.017 fixed the one axis but as you can see something still wrong with the "X".
RICH

This was fixed a few days ago   The lockdown has this in it.. please test and tell me if I missed anything

Thanks
Brian

Brian,
I just downloaded .018 and ran swarf's file. And the X extremes are not correct.
RICH

MOIDIFIED:

Will download V.018  lockdown and try it again.