How to solve a 3-Jaw run out

[Bamban]

I am about to get rid of my 7x10 HF mini and acquire a totally rebuilt Atlas 618. The Atlas all checked out, the spindle itself is running true, outside and inside (based on the video the seller sent), but installing the 3J and indicate on a chucked up true round piece, the indicated run out is .012. Is there a simple fix for this? The seller is more than willing to do any test for me, is there anything I could ask him to do for a fix and check the run out again. Or should I just take the plunge and try to remedy the problem, like get another chuck, maybe. I really like that little Atlas as a replacement for my Chinese mini, except the Chinese runs true with the 3J.

Thought?

TIA

 

[Pat of TN]

Re: How to solve a 3J run out

I have a Craftsman-branded Atlas 6x18. If things are in good shape, you can't go wrong. It's a nice little lathe.

I would have the fellow put the indicator on the OD of the chuck, just to see. But unfortunately, three jaw chucks (especially older/worn ones) are not good for repeatability. Brand-new chucks are only guaranteed to .003 TIR, usually. I

t may be in your best interest to invest in either a new three jaw or, even better, a four-jaw independent chuck & backplate if you decide to purchase the machine. You could get a bit of wiggle room because of this on the price. I got a four-jaw chuck and backplate from Little Machine Shop a few months back for my lathe, and haven't looked back.

About the only way to "fix" a three jaw is to regrind the jaws, which is quite a dirty, literally gritty task. And unfortunately, a lot of the problem is more due to a worn scroll than the jaw teeth.

Don't be afraid of the time it takes to dial in a four-jaw, either. With practice, it becomes quick. Before long, you can get it within 20 or 30 thou TIR by eye then dial it in to whatever you need lickety split. Plus, your TIR is dependent only on you and your indicator!

 

[schor]

Re: How to solve a 3J run out

Dont even worry about the chuck, it is replaceable. I would have him checkout the runout on the spindle itself. Chucks can be ground, spindles are a bit more work.

Put a true (as true as can be had) mt2 arbor into the spindle on the headstock and measure that. Then you should know if things are good enough to warrant the purchase.

 

[Bamban]

Re: How to solve a 3J run out

Thank you all for the input. Short of buying a new chuck, is there a shop out there that grinds jaws whose services are reasonably priced?

 

[Pat of TN]

Re: How to solve a 3J run out

I don't know about that - in a way I doubt it - but it's not all that difficult to do yourself. You just have to rig up some kind of toolpost grinder - which can be done with a pencil grinder or die grinder on a smaller chuck, with a shop-made tool post for it. Then you turn and bore a thick ring to put on the ID jaws that allows you to open them enough to let the grinding wheel in.

Chuck up on the ring and try your best to indicate it - the best you can do is tap it with a hammer on the high spot. When you get it decently true, fire up the grinder and grind the OD jaws until they're ground fully... you can then test it with a dowel pin or drill rod, or just turn a short length, cut it off, and grip it.

But unfortunately, re-grinding a three-jaw chuck doesn't provide much satisfaction for long. A buddy of mine reground the jaws on his 8" three jaw. The first piece he gripped was within .002 or so. The next piece was within .015, I think?

Considering the time and mess involved - abrasive grit is bad for a lathe - I'd just either try to work with the existing chuck for now (after all, pieces done in a single set-up are as true as can be unless something moves), or look into a new, preferably four-jaw independent, chuck... they can be had from Little Machine Shop for about $110 with the backplate.

 

[wa5cab]

If you put a ring around the outside of the jaws (the surfaces that would be used to grip the ID of a hollow workpiece) and open the jaws to grip the ring, you are loading the jaws the wrong way. Any clearance between the T-slots in the chuck body and the keys on the sides of the jaws will allow the jaws to cock. And after grinding the inner surfaces, you will end up with what's known as "bell mouth". The fixture you use to tighten the jaws against must work when you are turning the chuck wrench in the same direction as you would to tighten on a piece of solid round. For examples of chuck jaw grinding fixtures see Downloads.

Robert D.

 

[David Kirtley]

3-jaw chucks are not really very accurate under the best conditions and are for convenience. Even when you re-grind or lap the jaws of a 3-jaw true, they will only be true at the one position of the jaws for size you used. You can get rid of bell shape in the jaws but that is about it. You will be chasing your tail beyond that. You want accuracy, go to the 4-jaw for big stuff and a collet for the small stuff.

Yes, they have the 3 jaws with the extra adjustment but basically, they are working it like a 4-jaw.

 

[Kernbigo]

use 3 gage blocks or if you don't have gage blocks use key stock, and clamp them on the 45 deg. to load the jaws for grinding, don't use a ring wrong way to load.

 

[John Hasler]

use 3 gage blocks or if you don't have gage blocks use key stock, and clamp them on the 45 deg. to load the jaws for grinding, don't use a ring wrong way to load.

Could you explain? I understand that I should not use a ring on the outside of the jaws but I don't know what you mean about 45 deg.

 

[Hawkeye]

The chuck jaws have the sides angled so that they can close more toward the centre. Actually, on a 3-jaw the angle will be 60[SUP]o[/SUP]. If you put a piece of stock in the gap between two adjacent jaws, against the angled faces (repeat for all three), you will put outward pressure on the jaws while leaving the centre faces free to grind.