Re-Grind 5C Collet throat

[jjtgrinder]

I got my new 5C collet chuck mounted to the G4003G and it is showing .0005 TIR indicating on the inside surface of the chuck nose (taper that closes the collet). I put a collet in and a accurate test bar (protruding 5 ") , indicates .0006 TIR at the nose of the chuck. Then indicated 4" out from the nose, .004 TIR! Yipes!

I took out the test bar and closed the collet completely. Indicated the "throat" of the collet just inside the edge.
.002 - .003 TIR. Indicated along the bottom surface of the collet, more irregularities. Full disclosure, these are an "off-shore" inexpensive collet set.

I am thinking of putting the tool post grinder to work and lightly regrind the "throats" of the larger collets.

Any thoughts on this? Has anyone tried this before? What suggestions do you have?

 

[jbolt]

There are a number of areas where errors can occur with a 5C

Test bar fit in the collet? If not a good fit it could get skewed as the nose of the collet pinches at the front.
Fit of the tail end of the collet in the chuck?
Concentricity of the threads in the tail end of the collet.
Concentricity of the chuck threads that capture the collet.
Bore alignment of the chuck to the spindle.
Debris between the mating surfaces.

I have a 3/4" Hardinge Special Accuracy collet (.0002" TIR) and a precision ground bar for checking holders, chucks, closers, drawtubes etc.

I generally don't use the 5C collets for work that sticks out more than two diameters. If I do I will support the end in a live center.

To grind a collet you would need to preload the jaws of the collet. Not sure I would bother with that. If the chuck is well aligned and concentric I would look at a better quality collet for high precision work.

 

[mksj]

Second what Jbolt indicated. You did not specify the chuck and how it is mounted to the lathe. If it is a direct mount or a Set-Tru. Typical Chinese 5C chucks direct mounts will be out 0.002-0.006" measured on the internal nose where the collet mounts. I returned two Chinese 5C chucks before I went to a Set-Tru type 5C chuck. The TIR of a decent set of 5C collets is 0.0006" or better. So at most with a test bar you are looking at around 0.001" or just slightly better. Dial it in with a Set-Tru, switch collets and you will be chasing your tail with the numbers. Since 5C collets only clamp at the tip of the collet, the clamping surfaces are is small, and the smaller the collet size the shorter the clamping length in the collet. It is typical to get some skewing depending on how they clamp and the clamping mechanism. At 4" out I get about 0.002" TIR with a Bison 5C set-tru zeroed at the collet nose and using a 3/4" ground bar.

Thoughts, I would not try to regrind the chuck collet mating surfaces. Even a high quality non Set-Tru 5C chucks spec. a run out of 0.0006" at the nose or better if you are lucky, yours it better than most.

 

[chips&more]

If you have not done so already. I would take that import 5C chuck apart and thoroughly clean it and look for irregularities. Then put it back together and try your test findings again. If still not correct. I would take a billet of material mounted in a chuck. Turn the billet into a 5C mandrel and DO NOT REMOVE from chuck. Mount your new 5C chuck on the mandrel and grind the back end of the new 5C chuck. Then try your test findings again…Good Luck…Dave.

PS: you can make just about any chuck into a Set-Tru type if you use your noodle.

 

[Chipper5783]

Hi JJ, lots of good ideas here (I did a similar thing on my chucks that Dave chips&more is recommending with quite good results - except I don't have a grinder so I just took a skim cut). I would be very, very hesitant to grind the collet bore. I added a 5C collet chuck about a year ago (tired and used from the auction). I took it apart and cleaned & lube'd it. I bought a back plate from Shars and fitted the chuck (first of all fitted the plate to my D1-4 per recommendations elsewhere on this site, match marked the orientation, then trimmed it for the chuck, decided to make it a snug fit on the register - instead of a "tap true") . The result was reasonable, about 0.001" TIR. I have found that if I play around with how it is tightened to the spindle, I can get the 0.0005" finger dial doesn't wiggle on a 0.5" ground dowel both close to the collet and an inch out (I didn't try it further out - I'll try it next time I have things set up).

My point is I suggest you put your effort into how the chuck is mounted. Even a POS chuck and low end collets ought to be significantly better than what you are getting. I suspect the issue is not the chuck or even the collets. You can play with the mounting details several times (a little cut here, a skim there and so on) without really hurting the chuck itself. Perhaps buy or borrow a good collet (or take one of your's somewhere you can get it checked out) to remove one variable.

You did not say what type of mount you had. The beauty of mounting plates is you can custom fit your own spindle tools and get quite good results.

Let us know how you make out. David

 

[jjtgrinder]

Thanks for the reply's.

My chuck is a Chinese set-tru type from Lathe-Master. I bought a D1-5 spindle chuck adapter from Grizzly and machined it (on the spindle) to accept the chuck back plate. That all worked out with no appreciable run-out on the face the chuck back-plate mounted to. I'm sure there's some TIR from the "stack-up". I'm happy with the results. I feel like I can get a little improvement work doing the recommended cleaning and adjusting.

My biggest problem are the cheap chinese collets! I have decided any collet grinding is folly.