3D SURFACING, YOUR BEST GUESS AT THIS PROB

So, I have successfully surfaced this with a 1/16th ball mill a couple times. All the other times is comes out like this photo attached. Really fuzzy and raspy on the flats.

I am running it at MAX rpm for the mill (11,500. I know 20k or 30k would be better but aint got it.) and at about .0005 IPT chipload and .0035" stepover, attempted with both scallop and raster(parallel) toolpaths. Tried with both 2 and 4 flute endmills and only had success a couple times.

A few questions I have are, given the 3d shapes, how much should I leave for the ball? I've been trying .003 before the final pass but should I do more? .010 or .012? Im stumped... Tried everything.. Maybe more chipload?

It looks like you've chipped the end of your ball endmill.

How far was he tool projecting from the collet?

Well.. We're using a brand new ball mill every single time....

The tool is choked up super close to the collet and its using a REGO PG tool holder with almost zero runout

I would do an experiment.
Anyway it is long story, requires thorough investigation.
I would machine a piece of surface which goes ( always ? ) good, trying to destabilize the table with hand.
It would be easy indication if table is stable. The next - almost the same destabilization attempt - headstock.
It shows if headstock is stable.
And, of course, I would check static tool runout. Must be less than 1 micron.
The next - a bit bigger ball tool tip.
These experiments will provide many information for discussions.