[LeeWay]

I have done this a few different ways now, but have never really been satisfied with the results. I started out cutting steps into aluminum jaws and we know they will not last too long without resurfacing and recutting.
Now all I use is mild steel jaws for steps and use aluminum ones only for profile jaws or jig plates.

Now on the Haas, I want to try it a different way then I have in the past. I will be cutting all 4 jaws in a 4" double vise. I think clamping 123 blocks into the jaws and then modeling this up may give me my best chance for accuracy.

In the past I have just kept installing all new jaws on the same face of the vise. Rinse and repeat just jogging an axis. That takes no consideration for variation across the vise or table, so could show in the parts too. Certainly not a very accurate solution.

Any thoughts, tips or tricks?

[kvom]

I generally the steps on both jaws at the same time, so that the jaws are clamped closer together than the diameter of the tool. Assuming a good tram the steps are flat and parallel to the X axis. One issue is that if there is an radius on the flute ends the corner of the steps can be rounded causing inaccuracy in the depth of the stock. One way to insure against this is to make a final pass with a dovetail cutter to undercut the bottom of the slot slightly.

I don't know any fool proof method to compensate for the moving jaw, but I doubt the variation in a good Kurt vise is more than the tolerance of my mill. Obviously you could face the stock as the first op.

[LeeWay]

I used the 123 block method and modeled it up in Fusion for the GMT vise. I think the results were great. Repeatable for new jaws of the same thickness.
I will do similar on the Orange vise that arrived yesterday.