As long as your marks are inside of the outer pinch rollers it should be able to pick them up.
When you set your print in the cutter and let it register the width do you move the cutter head over to the
right crop mark?
It seems to work best if you set the head so the blade holder is inside the little corner mark and not over it.
Once the head is in place don't reset the origin (if you are doing that), just send the plot file from the computer...
That's interesting. My Flexi/FC7000 always moves to the origin and then starts a diagonal scan from there. Thus if I set the blade holder in the crotch of the lower right registration mark and do not set the origin to the point as well, the plotter will move to the original origin and commence its scan from there, never location the first mark.
The larger the marks the more the feed compensation on the printer and plotter can vary and the plotter will still find all four marks. On small images it's not a big consideration but when you start trying to cut panels over 48-60 inches a larger mark can make the difference between finding the mark or not.
When all else fails you can print your own marks on the corners of a rectangle of known dimensions and use the plotter's built in automatic or manual mark sensing. Or you can print bomb sight marks and use the built-in light pointer and locate them manually. The latter never, as in ever, fails.
If your contour cutting is off in any direction you probably need to go through the marker sensor adjustment outlined in excruciating detail in the plotter manual. The sets the logical distance between the mark senor gear and the blade tip. If the actual distance between these two items is not exactly the same as the logical distance [that being what the plotter thinks it is] the contour cuts will be off in either or both the X and Y directions. If the contour cut is off but the error is not consistent but is cumulative then the problem, or at least part of it, is in the feed and carriage movement compensation or, shudder, some sort of material or belt slippage. The latter cause is most unlikely.