Accuracy determination & accuracy improvement

I have a JGRO style machine, that I am in the process of taking out the slack, and modifying the structure to remove some of the play & flexure. There are a lot of posts that show mods to do this, but I would appreciate it if someone could post some accuracy figures and if modifications were made, which mods gave the most improvement.

Here's where I stand:

When I cut 2" squares, they are undersize or oversize about .030" to .040" depending on which direction I cut. Right now, I still have the basic JGRO structure, and ther only thing I have done is to re-set all of the bearings to remove the play.

I tooks some static measurements with a dial indicator: I have a .035" flex at the router bit when I apply a 10 pound side load on the bit (pulling toward the 1" pipe that runs parallel to the base). The flex at the gantry wall is about .012". I put some clamps on the pipes to eliminate any looseness that would be hard to eyeball, but there was not any improvement.

I applaud your testing approach. Measure, modify, measure for change. It will pay off. Also try to throw in some stress.

I believe that if you get different results by cutting in different directions, you have backlash (or slop) in the system. This can be caused by any looseness in the leadscrew, leadscrew bearings, linear bearings, structure, etc. You probably will have to tighten it up before it will diminish.

When you cut your squares, also check to make sure that they are square. A larger size would make this more apparent. Check them with a square and also measure the diagonals, they should be the same. Creating level can be done by skim cutting a temporary top.

I don’t have my JGRO anymore, but two things that I worked hard on were the flex of the long pipes, and the leadscrew bearings. I believe that the flex can only be solved with support. Vertical uprights pressing up and in on the pipes will help a lot.

A lot of my backlash was caused by the leadscrew bearings, not the leadscrew nut. With the power off, see if you can move the gantry, etc. You should not be able to.

Leadscrew mounting can help three problems at once: whip, backlash, and motor wear. The trick is to “fix” the screw at the motor end and not let the motor’s bearings take the load while firmly attaching the screw to the frame. It takes 4 nuts, two MDF blocks, and two bearings in a configuration as in the picture. The MDF pieces form a thrust plate. The bearings sit in the recesses and are back to back. The procedure is to loosen the motor coupler, lock the two nuts nearest the motor together, and then tighten the next nut to compress the two bearings together against the thrust plate. Finally tighten the outside nut to lock it all together. Now tighten the motor coupler. Firmly attach this to your machine with nuts.

Steve

Splam- thanks for the input. I did have some backlash in one leadscrew where I had not tightend the anti-backlash nut. I found it by measuring the spindle movement before and after clamping the bearings to the pipe on that axis. (about .005 improvement).

I see three main reasons for lack of accuracy:

-backlash, play or slop between components
-deflection of structure
-X, Y & Z axes not mutually perpendicular

(I am sure there is more, like tool runout, lead screw accuracy, computer problems, interference, lost steps, etc., but the above items are (IMHO) the big hitters)

I have seen a lot of retrofits, some indicating that the problem was solved, and others indicating that a lot of effort went down the tube. What I am trying to do is to find ways to evaluate the problem before implementing a fix.

I would agree with your list.

You may also find that the tolerances change at different locations on the table. The pipes can flex in the middle causing one situation and also the pipes may not be mutually parallel (if there is such a thing) causing a situation on one end that is not the same as the other.

How did you apply your 10 pound side load?

Good luck,
Steve

To apply the 10# load, I hooked a fish scale to the cutter & pulled.

Geometry is measured in pitch, yaw and roll along with straightness & position. You can envision the axis of the machine is the axis of an airplane then pitch it the nose up nose down, roll is the rotation around the axis and yaw is the rotation perpendicular to motion of the axis. Usually the major areas to look first are straightness and position.