[How do I?] Troubleshooting and Reducing Total Indicated Run out in a New Collet Chuck

[Shiseiji]

Hi all. First TIA in recognition of your time and experience. This is my first backing plate. May be too much information not needed, but I've tried to cover everything I found on the subject trying to educate myself to minimize questions.

I'm working with a Bostar 5C collet chuck for a Logan 820 10" lathe. The spindle is 1-1/2 8 tpi with a .25" un-threaded portion of the spindle often called the register and a .75" threaded part. The .25" vs .1875" un-threaded common on a Souhbend often trips people up when they buy an eBay chuck. FWIW, I measured my Logan 4 jaw and the un-threaded portion has a diameter of 1.513." You can argue with a 4 jaw any un-threaded portion is irrelevant but .013" is a long way from .002" Ops, mine is bigger than that.

Comment. A personal opinion, after reading as many forum posts as I could find, about the so called "register." I suspect over the years "shoulder" dropped of what should be the "register shoulder" leading to much confusion over how the backing plate is made square to the spindle. Just my $.02 .

First I checked the spindle shoulder. Dead flat (yeah!).
After my attempts at an internal close fit mandrel and spacer turned into several entries into the "box of blunders" I chose to get precision spacers from McMaster-Carr to cover the un-threaded portion of the spindle.
I covered the un-threaded part of the spindle with the spacers, did a skim cut on the headstock side of the backing plate. Dead flat.
Reversed the backing plate and did a skim cut on the non-register portion of the backing plate. Again dead flat.

When I mounted the chuck on the backing plate, it would not pull tight. In retrospect, I think I should have tried freezing the backing plate. Oh well, next time.
To the point, trying to skim cut the centering shoulder, I ended up cutting a taper. When I cleaned that up, I have an interface gap of .0025 to .003." Having read of purposely not having an interference of "0" and tapping the backing and chuck into alignment, I figured I could probably live with this.

Except the chuck TIR is not what I expected from a collet chuck and to me is pretty is bad. Worse than the AT & 3C adapters and collets I intended to replace because they have ~ .002 to .003" TIR.

I marked each of the bolt holes and checked the TIR rotating the chuck through each position starting at bolt position #1. Tapping on the chuck made zero differance.
1: +.0065 to -.005" or .015 TIR.
2 : -.0075" TIR
3:+.006 to -.005 or .011 TIR.

If this isn't a "as good as it gets" situation, I'm open to any ideas including re-cutting the shoulder.

Final question. Any harm in reducing the length of the headstock side of the backing plate to reduce over hang?

TIA

Ron

 

[Alexander McGilton]

Shot in the dark without photos.
Have you champfered your edges where they meet radius? I could imagine that the collet chuck is riding on an edge and not on a flat surface.

 

[Shiseiji]

Hi Alexander, thanks for your interest. I checked for full contact between the backing plate and chuck with Prussian Blue. Sorry I neglected to mention. In fact all contact surfaces with the exception of the backing plate "shoulder" diameter that fits into the chuck have been checked with Prussian Blue. I did blue that too but we know we don't have a good fit there.

Ron

 

[srfallsallot]

I have a Logan820. I purchased an ER32 for it. It screws on to the 1-1/2x8 threads. I love it. It takes maximum advantage of the small bore.

 

[Shiseiji]

I have a Logan820. I purchased an ER32 for it. It screws on to the 1-1/2x8 threads. I love it. It takes maximum advantage of the small bore.

The 5C also fits on the 10EE and my Hardinge Indexer. I'm trying to avoid having to invest in, for me, a 5th set of collets.

Ron

 

[Nutfarmer]

I gave up on an inexpensive , import 5C chuck. After going through 2 of them the best I could come up with a TIR on my 5C chuck that was worse than a 3 jaw chuck. Many times import tooling is hit or miss. I found better accuracy by using a draw tube and a spindle adapter for a 5C collet. You might make the collet chuck work by turning where the chuck registers on the backing plate a few thousand under size. Then lightly tighten up the bolts holding the chuck to the backing plate while adjusting the chuck for runout , like you would a true set Buck Chuck. After the run out is adjusted out the bolts can be tightened down.

 

[machPete99]

You may want to add some clearance around the register/shoulder, and around the mounting bolts.
That way you can pump it around to indicate it in.
Even if the chuck is running perfect, you may want to be able to adjust for the collets or the parts being out.

 

[Shiseiji]

I gave up on an inexpensive , import 5C chuck. After going through 2 of them . . . might make the collet chuck work by turning where the chuck registers on the backing plate a few thousand under size. . . ..

Thanks for sharing your experience. Certainly won't hurt and if it doesn't work probably means nothing will. Thanks again.

 

[Shiseiji]

You may want to add some clearance around the register/shoulder, and around the mounting bolts.
That way you can pump it around to indicate it in.
Even if the chuck is running perfect, you may want to be able to adjust for the collets or the parts being out.

You have the second vote. As I said to Nutfarmer, Certainly won't hurt and if it doesn't work probably means nothing will. Thanks.

Ron

 

[RJSakowski]

There are three advantages to using a 5C collet. It grips rounds more securely than a 3 jaw chuck, it is fairly easy to change out work, and it has presumably better concentricity than a 3 jaw chuck. Add to that, if you have the right collet, you can run square stock.

However, as pointed out above, the inexpensive 5C chucks often lack the concentricity. The lack of concentricity can arise from multiple sources. The register of the adapter plate can be wrong, the chuck socket fit can be wrong, the collet can have concentricity issues, Correcting runout will require identifying what the source is.

I would start by checking to see if I can reliably remove and reseat the the chuck. This will determine if there is an issue with the register to the lathe spindle. If the runout remains constant in magnitude and orientation over multiple installs, then I would move on to checking the runout of the backing plate to chuck register. That good, the issue is in the collet socket or the collet. Check the socket for runout at two positions to determine both radial and angular runout. keep in mind that the effect of angular runout is is magnified the further away from the chuck you are so measured runout of a few tenths in the socket can result in significant measured radial runout at an inch or two from the collet. If everything looks good, then the next suspect is the collet itself. This can be evaluated by checking multiple collets and/or by checking in different collet fixtures. Mark everything with a Sharpie so you can determine if direction is consistent.