help with drill chuck

[fishfrnzy]

After several years of fighing holes drilled off center I am hoping to solve the problem of holes off center in lathe tailstock drill chuck.

The facts:
1. cant seem to line up tailstock when drill chuck is mounted.
2. round bar in drill chuck indicates increasing reading as I move away from tailstock.
3. indicator reading with bar in lathe chuck and drill chuck tightened on round bar. 2 different morse taper adapters show .020-.030 runout at end of adapter as it and chucks are all rotated together.
4. Chuck is Jacobs 33B with 3/8 24 tpi. Came from garage sale box.

would it be right to assume threaded portion of chuck is bent internally?

Can I just bore out to true up and rethread to 1/2 20 ?

Am I overlooking something else I should check first?

Help please

[oldtiffie]

Adaptor (male) thread

A similar result might be due to the thread on the adaptor (the "male" thread ("bolt") that mates with the female/"nut") if defective might give the same result.

[macona]

How old of a machine is it? Sound like typical wear on the slideway of the base of the tailstock. When they wear the front wears faster then the back so it slowly nose dives over time. I had to shim .005 on the front of mine to get it straight again.

[fishfrnzy]

Yes there is some bed wear .003-007 over length of bed. T/S has been shimmed.

checked T/S ram with indicator on crosslide and all was within .002 on side and top measurement. SB 10L lathe if that helps. I can see the end of the drill wobbling when drilling.

FYI if I indicated with 1/2" rd bar in drill chuck and rotate while pressed loosely into T/S ram I get about .010 2" from jaws so this leads me to believe either jaws or threaded end.

Checked with 2 different adapters with same result.

Thanks

[darryl]

When you mounted the chuck to the rod held in the lathe, did the body of the chuck wobble? My guess is it did, the same as the arbor did. I don't see the likelyhood of the interal threads on the chuck being out, but possibly there's a chip embedded in the body under one of the jaws- maybe a bunch of stuff in there. Can you take that chuck apart?

[fishfrnzy]

the chuck itself wobbled within .002

[pcarpenter]

First thing I would do is figure out whether its the chuck, the arbor, or the tailstock being off one direction or the other. Given that most tailstocks are intentionally adjustable for offset and that this adjustment seems to often allow some out of parallel movement, I would think it is first most suspect.

I would get a known very round ground tool bit and clamp the drill chuck on it and then put it in your lathe spindle and indicate the threaded hole. Indicate the tool bit first if you want to find the error it adds and where its located. Depending on the taper and the adaptors you have, you may even be able to insert the tailstock arbor in a good adaptor sleeve and put that directly in your lathe spindle (removing spindle chuck error from the equation).

Then thread the arbor in and indicate the arbor/chuck/tool bit combination up close to the drill chuck (reading off the fatter end of the tapered arbor). This takes the tailstock out of the mix and shows you what part of your error (there will be *some* from each source...you are trying to find the major source of runout) is from the stuff plugged into the tailstock and what is from the tailstock itself.

Tailstock ways are often (usually?) separate from the carriage ways. Shimming a tailstock fixes way wear issues *only* at the one spot you measured and shimmed. Usually the wear on the tailstock ways is tapered....often increasing as you get closer to the lathe spindle....from sliding the tailstock back and forth. A height gauge could be used to measure to the top of the tailstock spindle from an unworn portion of the ways at various locations. This might be the narrow top surface of a v-shaped way.....which is rarely a wear surface and should therefore still be straight.

I don't think that threaded arbors and chucks are likely to be as accurate as those with a taper on each end. And no...you probably won't be able to drill out and re-thread that *very* hard chuck body....and likely not so accurately.

I have one or two cheap import drill chucks and MT to JT arbors that indicate only a few thousandths runout....that may be a cheap solution if this is what is at play.

Paul

P.S. I just re-read your original post and see you checked the arbor. Its important to remember that runout may be measured in thousandths, but anything concentric like an MT arbor in a tailstock bore is really inducing *angular* error....meaning it gets worse the further from the point of initial error. If your arbor/drill chuck/drill bit combination were long enough, .040 would translate to 10" out a (really) long way from the tailstock :-)

[small.planes]

Have you stripped and cleaned the chuck? it could be jaws rather than the threaded end. Usually the threaded end has a flat register to butt up against a shoulder on the arbor.
Stripping a jacobs chuck is dead easy. There are detailed instructions on the web somewhere, Ill see if I can find them.

Found it: clicky

You could buy new jaws at one stage, but its probably not worth it...

Dave

[darryl]

I suppose another way you could check it is to turn a stub with a shoulder that the chuck will be a snug fit onto. It won't need threads, just enough meat to register with the back of the chuck, and a little inside the hole, short of the threads. There's usually a short section of hole before the threads that's clean. Place the chuck onto that, and use a bolt with washer, etc- whatever it takes to hold the chuck to that stub temporarily.

This should have the body of the chuck running very true. If it doesn't, there's a piece of junk mounted in your lathe, throw it away. If it does run true, then either it can't mount to your arbors properly, or the jaws can't hold anything concentric with the axis of the body.

Next step would be to mount an arbor so it runs true- maybe an adapter would make this possible. In any event, screw the chuck to that and see if the body runs true. I think you've said that it does within .002- is that a wobble or do both ends of the body deviate the same amount? If wobble, the threads are damaged- could be as simple as a chip being sort of welded into the threads somewhere, in which case it behaves the same with either good arbor, or the threads have been bent out of shape by a bad crash or an overzealous tightening at some point.

If there's cam action but no wobble, meaning both ends of the body are out the same at the same points, then anything mounted in the chuck should show the same pattern, ei cam action but not wobble. In other words, the drill bit or whatever remains parallel to the axis and doesn't precess. If there's precession, the jaws are probably shot and/or the slides that they run in. If there's no weldements in the slides, I'd say the jaws are shot, probably from overtightening or a crash at some point.

The collar which turns to loosen or tighten a drill bit is irrelevant as to any wobble it may have, and is not part of the body that I refer to.

I'm not trying to supercede anybody's knowledge of drill chucks or procedures for testing- just presenting another way that I might approach this problem.

There's another possibility, and that's that the body could be bent. There's not a lot of meat between the holes that the jaws slide in, so if one part got stretched, the back of the body could be running true, but the front wobbles. If that's the case, well- use that chuck simply to become familiar with how they come apart, then toss it.