Spindle runout

[RSG]

Talk of mill wear and spindle accuracy lately has got me looking at mine and I was disappointed to find I have 6 tenths at the bore. Upon inserting a collet and some drill rod I found the total runout was 1.5 thou. down approx 2.5", that's ridiculous! Knowing that my mill is one of these - http://www.grizzly.com/products/8-x-30- ... Feed/G0731 is it worth taking the spindle out and having some work done on it? My thoughts were the rest of the machine likely isn't capable of achieving anything better overall anyway but if it would I'd be interested in doing something.

I'd appreciate some thoughts on this.

[RSG]

I forgot to ask, would buying better collets (than cheap Chinese) help with collet runout?

[John Hasler]

RSG writes:

> Talk of mill wear and spindle accuracy lately has got me looking at mine and I was disappointed to find I have 6 tenths at the bore. Upon inserting a collet and some drill rod I found the total runout was 1.5 thou. down approx 2.5", that's ridiculous!

But you don't know how much of that 1.5 is due to the spindle, how much to the collet, and how much to the rod (better to use precision shaft than drill rod).

Check the inside of the spindle taper again and mark the high spot. Put your collet and rod back in, find that high spot 2.5" down, and mark it on both the collet and the rod. Reposition the collet in the spindle and see if the high spot moves with it. Do the same with the rod in the collet.

[ctwo]

I'm trying to remember the youtube channel of the guy that showed regrinding an R8 spindle bore in place - which would seem to be the best approach since you account for whatever is in place...

Basically, he angled the head so either the top or bottom edge of the taper was parallel to the table travel - using a DTI on the table and moving the table in X to indicate the head angle.

Then he mounted an air grinder to the table, fired up the spindle I think in low, and ran the grinder in-out of the spindle, using the knee to set the grind depth.

I think he ended up with no needle, maybe a width.

heck, the Bar Z guy. maybe it wasn't that good... https://www.youtube.com/watch?v=CFAkb93_V3M

[pete]

How old is the machine, how much work has it done? Sure your spindle bearings are in good shape and the bearing preload is correct? Then just how good is the spindle bore, rust, dings etc. With a stop set up to position your collets location so it can't move back a standard V block can be used to slowly spin each collet with the finger of a dti just inside the collets bore for checking runout. It might be worth checking the collets R8 taper if that's what taper it has. Something seems to be drasticaly out for sure. I thought Bar Z / Shadon HKW did a pretty good job regrinding his R8 taper but the bearings and there adjustments have to be as good as possible before doing that job. If your collets are adding significant runout then it's time to start looking for much better. Better be sitting down though if you start looking at what a place like Hardinge wants for there extra precision ones.

[SteveHGraham]

Possibly a dumb question: how many collets did you try the measurement with?

[spro]

Heck guys, don't be lost in the woods here. Enough of that going on. With no spindle at all, he could measure rise and fall of the table. The main thing is sweep and that goes back to previous information. Before even that, is wear in the knee and the vertical slides of the column. We must consider those vertical ways and their wear. To flatten a table within supposed working distance is not the same as when it is very high or low. The Knee is heavy and supporting all the tables and imparts considerable wear to the column slides, over time.

[ctwo]

spro, I think you were posting in the mill wear thread...

I wonder about checking the collets empty in the mill spindle with just a slight tension of the draw-bar, or a shorter recessed pin so the ID could be indicated right in the spindle.

Is a V-block going to work well enough? I would think the spring to open would be inconsistent enough, but maybe that would work just fine given you would be measuring across the web?

[pete]

Well if your spindle is showing a great deal of runout that seriously throws the exact measurements out. Still not impossible though since that could be subtracted from the tir. But then your dealing with the complication of the fact the collets OD and ID runout may or may not be canceling out or even increasing some of the inaccuracys. Checking those bearings and preload would be my first step. Then possibly regrinding the spindles taper if it's still showing excessive runout. With the spindle at the bare minimum of runout you can then judge just how much the collets are out. Or at least rough check them on a V block. If there's gross errors you then know it's time to buy better.Fwiw I learned after wasting far too much money that it's almost never cost effective to go cheap with work or tool holding. Inaccuracys in either do show up in the part. Shimming and fudging the part or tool position can be done, but it sure gets frustrating to do so every time.

[wlw-19958]

Hi There,

I have some experience in this area. Many years ago, I decided to
rebuild the spindle in my Millrite milling machine. the original bearings
were (I think) class 3 Timken tapered roller bearings (total allowable
runout .0003"). I thought I would do better and got some class 0 bearings
(.0001" total runout allowable) and put them in. This was before I learned
that precision bearings have the "high spot" marked on the inner and outer
races so, I didn't pay any attention to them.

I put the new bearings in (without aligning the "high spots") and the runout
in by spindle was more than before and it got even worse the farther out
from the spindle I measured. My mentor suggested I grind the taper in the
in the spindle (my mill has an R8 taper) by setting the head at the correct
angle and using my tool post grinder to do the grinding.

I carefully measured and did the deed and it improved the runout greatly
but it still suffered the farther out from the spindle I measured. Then, I
read up on precision bearings and learned about the "high spots" on bearings
and the importance of aligning these marks on the spindle.

So, I took my quill apart again (this was several years later mind you) and
pressed off the bearings and re-mounted them with the "high spots" in alignment.
This threw the runout off and I had to re-grind the taper again. Afterwards,
the runout was good at the taper and didn't exaggerate as much as the
measurement was moved father out from the taper (mind you, I am using a
collet and a precision dowel pin to test this so there could be some inaccuracy
in there).

Re-grinding the taper will help but with an R8 taper, I could only grind the bell
taper at the mouth of the R8 taper. I did not re-grind the cylindrical section
at the back of the spindle. I plan to do this some day but I would have to first
cut out the original cylindrical section and press in an slightly undersized
ID sleeve that I can grind to size. This is the only way I can see of guaranteeing
getting the runout "perfect" (or at least, as perfect as I can).

Good Luck!
-Blue Chips-
Webb

[RSG]

Possibly a dumb question: how many collets did you try the measurement with?

Not a dumb question because I just did that very thing with a 3/4" collet and this time it was the same as the spindle at .0006" so it is better than the 3/8" collet for sure. There is no question that the Spindle has runout but the collets seem to exaggerate the problem.

I was looking at some collets on the KBC website "Lyndex" I believe and wonder if they would be more accurate than the Chinese stuff I have now. Since the 3/8" collet is the most common one I use I might just buy one to try it.

[RSG]

Good suggestions so far, thanks for the input.

Some answers to the questions so far, The mill is only 4 years old and I use it about 150 hours a year so it doesn't see much wear. I have a tool post grinder but not sure I am skilled enough to use it for this situation. I would doubt the quality of the bearings are very high. I have no idea what the preload is or should be for them either, I'd have to do some research.

@ John Hasler - you mention re-positioning the collet in the spindle but I'm not sure how you would do that when the collet will only go in one way based on the guide pin.