I'm a noob, and I'm refurbishing a 1980's Lagunmatic 110 CNC Knee mill. I had about .005" play in the X-Axis, So I decided to check the ballscrew. Unthreaded the ballscrew from the ballnut and spilled the balls all inside the knee and across the shop. From the balls I did recover, it seems the nut was loaded about 50/50 with .123" and .125" balls (measured via chinese digital caliper). Even with that rough instrument I could tell they were not in spec. variance as much as .005" I replaced the balls with some G25 chrome steel balls (100% 0.1250") I got on Amazon (BC Precision brand). Getting ready to put the ballscrew back in but I can't help thinking maybe this deserves better balls? After all this is a very expensive, very high precision ground ballscrew. It seems the better grade balls are all ceramic though. For some reason that seems like a bad idea. I don't know what I'm talking about, but I don't think I want balls in there that are harder than the metal of the nut and of the screw. I would rather have the balls wear than the nut/screw. And the balls that came out were definitely ferrous/magentic. That's why I went with chrome steel. I couldn't find any balls of higher accuracy in steel.

1. should I consider putting ceramic balls in?
2. are there better steel balls I should be using? If so, where do I get them?
3. am I a dummy for putting all same-size balls in, when there were (I think, not 100% conclusive) standard balls and spacer balls installed originally?
4. since I didn't have the benefit of counting exactly how many balls came out (some were lost), is it sufficient that I put in as many as I could? Or is it possible I have gaps between balls that should be taken up by adding more balls somehow?