Your cross slide gibs are worn?

OMG is that lathe running while you are dragging the pointer of the dial indicator across the backplate? If so, please stop it at once. It does horrible things to your nice piece of precision measuring equipment.

The opposite sideways forces on the plunger as you move across the centre point are most likely the cause of the needle fluctuations.

Also, your dial indicator needs to be much more solidly mounted for precision testing like this. It looks like you have a full sized dial indicator on one of those little hydraulic arm bases made for a small 1" face finger-type dial test indicator. Flex in the mounting set up, combined with the sideways forces on the dial gauge from the moving backplate mean your gauge is all over the place like a rash.

You need to mount the dial indicator on a solid piece of square bar clamped in the toolpost, so its as rigidly mounted as a cutting tool. Keep overhang to a bare bare minimum. Or at least  use a full sized magnetic base, set up with minimum overhang from base to gauge. Best to mount magnetic base solid on the bed way, not the carriage too.

Then without the lathe running, put the plunger on the face of the backplate at about the 9 o'clock position and slowly rotate the spindle BY HAND. The reading you get on the dial gauge is teh amount of axial runout you have. Repeat process with the dial gauge plunger running on the outside diameter of the backplate to check radial runout, or concentricity.

A further check for flatness of the machined face would be to check it with a straight edge, or with bearing blue on a surface plate (or piece of flat plate glass at a pinch). It should be either dead flat, or slightly concave so the chuck seats on the outer area of the face and will not rock. If the face is convex, adjust up the gibs in your carriage and cross slide as best you can and take another fine cut, with carriage firmly locked. If still convex, you probably have a worn carriage ways on the bottom of the carriage that are allowing the carriage to cant at an angle, or possibly worn cross slide ways. Not much you can do about that in a hurry. Instead, turn a relief of a few thou in the centre section of the face so the chuck bears on a ring around the outer portion of the face.

Running the dial indicator across the face as you did in the video, even with the lathe not running, tells you nothing much really, other than the dial indicator is following the same path the cutting tool did. The face could be concave or convex to billy-o and the dial guage follwing the same path as the toolbit did will read no variation. So don't even bother with that test.

Edited By Hopper on 05/05/2017 23:29:54