I don't own that model but I can give you a few hints based on how I calibrate the models I support:
Work on one direction only at a time.
Use an arrow for the test shape so that you KNOW you have the printout on the mat the correct direction. Thus, if the cut arrow is pointing a different direction from the printed arrow on the paper, then you know your printout needs to be placed differently on the mat.
Record the calibration numbers before each test and number your tests. Write the numbers on the printout, as well as in your notebook or wherever you're recording these tests. After each test, write down how far off the cut line is from the printed line. For example, I might write down, the cut line is 2 mm above the print line. And I have a cheat sheet that tells me whether I need to increase or decrease X by 2 mm based on whether the cut line is above or below the print line. This is another thing that will vary depending on the brand and model.
Don't focus on the head of the arrow... that's just to make sure the printout is correct on the mat. Judge the cut based on the horizontal and vertical lines on the arrow only.
Once you have one direction perfected, work on the other. After you get the calibration done, write down your numbers for safe keeping. It's not unusual for the calibration to revert to default values if you install an update.
You might want to also then test four arrows spread out on your printout to see how accurately all four will cut. This will not be related to calibration but to things such as cut speed and if you're cutting into the mat. Accuracy in a PNC depends on a more gentle approach.
If you continue to find that you're making large changes to the numbers but still getting the same results, it could be a bug in SCAL. I dont' recall seeing anyone with a Redsail model post about PNC here, thus you might be the guinea pig.
Good luck! And post back if you have other questions or issues.
