QTR-Quadtone RIP

The 'minimum increase between steps" that Paul refers to is apparently quite large. After numerous trials to get the "Linearize Data" routine to take my curves, I gave up. 
Instead, I print without linearization, measure the 21 step tablet, plug resulting L* values into excel increasing as %K increases (or highest L value to lowest), then type those values into the Curve at "LINEARIZE=" step, save the curve and install it, then reprint, remeasure, and import  L* values into Excel to measure linearity. Never fails to give me a linear curve. 

If I see that the non-linearized graph of L* values has large bumps or irregular internal slopes, I do have to go back and adjust the ink values. It doesn't have to be straight, but it should be regular (read this to mean without erratic slopes). My setup requires that I use much less LK than most of the sample curves (or the Calibration curve) suggest, so don't hesitate to experiment with the Gray_values for your intermediate blacks.

Good luck.

Jim