Looks good .. did this improve backlash?
Short answer, no. That pic was a before and I have better contact now and smoother motion it seems but backlash is still .0025 in the x and .002 in the y. Z has always been spot on at around .0005. I need to get some wrenches and check the bearings.
It looked fairly decent from what I remember, but it's been months since I last had it out. When sighting along the length of the X gib and if your head is adjusted just right, some daylight is visible between the gib and dovetail surfaces and I'm thinking that this might not be a good sign. My parts come out good enough, though, so I haven't explored it yet. Blueing the gib and checking fit that way like luv2ride did sounds like a good idea, though.
Mike
Continuing with the backlash madness, here is another puzzling question:
Do you see more backlash in one direction than on the other? In other words, do you get a different backlash distance measurement when the motor is moving on one direction versus the other?
This sounds kind of ridiculous to me, but it is what I am observing. At least that is what it looks to me...
If this is not as ludicrous as what it sounds, what is the cause? Mechanical component non linearities?
I document my CNC Experience at CNC Dude's Youtube channel. Check it out!
If this is the case then sooner or later your table will fall off one end.(chair)
Phil
heh heh heh, I did say it sounds crazy! Luckily, though, it is about a tenth of a tenth. As a result, I calculate that it would require 2,160,000 actuations from side to side to go from referenced to touching the other limit switch. If I were to do such actuation per second on a non stop basis, it would take 25 days. Considering I don't think I have put a whole day of use into my most used machine, I think I am in good shape.
I guess this is why it is so important to reference the machine each day ;-)
I document my CNC Experience at CNC Dude's Youtube channel. Check it out!
Hmmm, I was indicating my vise today, and thought I would push and pull on the table in the Y direction and see what I got 2-1/2 thou. Ugh....... Didn't have the stomach to check Z or X.
Luckily, I'm just learning the machine and the CAD/CAM process so haven't needed high accuracy so far.
I guess adjustments are in my future. :-(
I posted this on the 'Huge amount of backlash' thread. But for averyone just folling this thread I thought this might be useful:
'I finally threw down and got a new X-axis ball screw. It was a process of elimination, but the fact that I was able to 'push' the table in the X direction about a thou and a half, pointed to the ball screw. All other possibilities had been eliminated: bearing preload, saddle movement.
Results were better than expected:
Old backlash/lost motion: .002
With the new screw in and the six 4mm screws torqued to about 6Nm ( not much ), the gib set loose, backlash was .0001-.0002! Yes that's right, that's tenths! I relaxed the pre-load just a hair, and tightened up the gib; backlash is now about .0003 - .0004. I can live with that!
Adam'
You have me wondering if we can repack the nuts with slightly larger balls...
I'm fighting .002 lost on the x and y both and tearing into the machine again.
Naw. These are high precision, best to replace it with a new one. But make sure you have eliminated bearing preload and saddle movement. A good procedure is ( start with X becuase it's the easiest ):
1) disconnect the table, take out the gib, and slide it way to the left to reveal the ball nut yoke, and snug-up the screws. These are small screws ( M4, M6 ) so they don't take a lot of torque. The re-connect the table and gib. This is a good opportunity to clean up the ways and saddle cavity form all the chips that have gotten in there.
2) mark the position of the Y axis gib screws, then tighten up the Y-axis gib to eliminate sadle movement form side to side.
3) take off the X-axis motor and stick the shank of a 1/2 inch drill into the coupler and tighten, then put a pair of vise grips on the drill so you can move the screw by hand. Then check the feel of the preload on the bearings. There should be some slight drag.
4) With the X-axis gib loose and the 1/2 inch drill and vise grips in place, put a DTI on the spindle and see how much you can turn the ball screw before the DTO registers movement on the table. 1degree of slop will transalte to .00055" of lost motion.
My bet is you have can turn the screw 3+ degrees before seeing anything on the DTI. If that's the case the ball nut is most likely fried.
For the $650 for a new one, it's well worth it. It takes about 4 hours working slowly, so it's not that bad, and it's something you won't have to do again for a long time, if you keep everything well lubricated.
That is a good method to reduce/eliminate the other variables involved. I haven't locked the y gibs and will try that before I do anything crazy.
I do have access to a whole shop of tools and like to make things. If I go as far as to order a new unit I will grind some new spacers and see if I can separate the ball nuts a thou or so more.
It's worth a shot! Let us know how it come out, but it's also possible the screw itself is scored.
So I tore my machine to pieces. The spacers look fairly easy for a machinist with some equipment. I don't have to make anything after pulling out the .001 shim and getting a tight fit. The screws holding the two ballnuts together were loose also. I would suggest anyone that is already checking tightness of the mounting hardware take the screw out far enough to check these screws too. I'm adding blue loctite to the screws to prevent this later.
Leadscrew out on the bench. Check the flat screw but don't over tighten, I don't trust them. They were both loose enough to wiggle to ballnuts.
Spacer removed for measurements. I added a .001 shim between the spacer and nut and reassembled, all was nice and snug.