First of all I confess that I don't have the data for CDG 23, but I think the equation t = 13.5 / (i-1) seems to be too simplified. The equations I derived from the data for CDG 12 and CDG 14 differs very much from this.
1) CDG 12 (Inverse)
t = 111.82 i ^ -0.796
2 <= i <= 20 , i is multiple of the setting.
2) CDG 14 (Extremely Inverse)
t = 73.82 i ^ -2.229
2 <= i <= 6
For CDG 14, the characteristic curves go upto 40 times the setting, but the equations hold good only for the initial portion of the curve, upto 6 times the setting. Curve Fitting techniques failed to find an equation for the remaining part. But the equation shown above is very accurate upto 6 times setting
Once again, I admit that I don't have the data for CDG 23. If some one can send me the curves, I will try to find an equation.
Expecting comments,
Sajith Kumar
sajkumar@yahoo.com