Yes, I was mentionning this specific formula, but I have confused your situation of Orthogonal Blocks CCD with the Orthogonal CCD (without blocks). The formula I mentioned is applicable for Orthogonal CCD, but not appropriate for Orthogonally Blocked CCD.
As an example for a 3-factors Orthogonal CCD, I have :
- 8 factorial runs (denoted with +/- signs only)
- 6 axial runs (denoted with a/A signs)
- 9 centre points (by default, denoted with 000)
Using the formula:

I can find the same axial value than the one displayed by JMP (1,668) for orthogonal CCD.
For orthogonal blocks CCD, the axial value is increased compared to orthogonal CCD to ensure that there is no correlation between blocks and main effects (+ interactions):

If you force the axial value for an orthogonal blocks CCD back to the orthogonal one (or any other value) for a design with blocks (according to the number of centre points), then you'll degrade the performance of this modified design (here with a calculated value of 1,525 (orthogonal CCD) for a 3-factors CCD with 8 factorial runs, 6 centre points, 6 axial runs and 3 blocks):


No matter if Orthogonal or Orthogonal Blocked CCD designs, the number of centre points will impact the axial distance calculation. More infos here Central Composite Design - Orthogonal Axial Values
Unfortunately I don't have the formula to calculate the axial value for orthogonal blocks CCD.
Hope this answer will help you,
Victor GUILLER
"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)