Hi Grotto! It sounds like there is some damage of swarf in the system somewhere. You have to strip the chuck give it a clean and check for damage. If you remove the chuck from the back plate you will find some screws three will hold the key bevel gears in and three will hold the back plate on to hold the scroll gear in place remove all but label the plates and gears with a dot from a centre punch to make sure that they are replaced in the same orientation. Clean the back plate and screw on to the spindle and clock it to make sure it runs true. If it dosnt unscrew and check for damage. You can shim the face to true if nessesary. Rebuild chuck and assemble on to back plate. Put a dowel about 20mm dia in to chuck and check for run out. If it dosnt run true you have to clock the chuck and maybe adjust the location diamiter.

David

David and Michael are probably on the right track, but I'd add that I think you can expect the 3-jaw fit on the screwed spindle nose to be more exacting than it is for the 4-jaw even if everything's in good nick. That's the way it was on my Speed 10 – I had to align the 3-jaw more accurately when starting the thread to screw it on, or it would bind on the parallel bore in the backplate before the threads even met. The 4-jaw was a good deal easier.

The accuracy of that location is obviously less significant on the 4-jaw where the user and not the chuck is setting the workpiece centre of rotation.

I guess I'd start by looking at the bore in the backplate for bruising.

The above have it sussed, one way or another.

The back plate threads are likely not anywhere near a precision fit – they are there simply to hold the chuuck on the spindle threads and so do not need to be a really high precision job.

The chuck will be aligned by the registers between the spindle nose and back plate. One axial contact surface and at least one radial contact area.

If there are two radial contact areas, the chuck will be thrown off-line by the affected surface. It does not need any more than a slight dink or tiny piece of trapped swarf to seriously affect the alignment.

Once you are sure the mounting thread is OK, have you ever dismantled the chuck and cleaned it?

Neil

Put a dial gauge on the OD of the chuck, and then against the face. You'll have to pull the jaws out and turn it carefully so you don't catch the slots. If neither run out then you have a worn scroll or bell-mouthed chuck slots. If the OD is true but the face runs out it could be bell-mouthing of the face. If the OD is running out but the face is true then it could be it's not registered to the back plate properly, in which case slacken the bolts and tap it into truth with a soft faced hammer.

If the OD AND the face run out then mark them to ascertain where the high points are. If they are opposite each other on diameter remove the backplate and clock the register face. If that runs out check the rear face for dings and if it has none run a skim over the register face and bolt the chuck back up, true it radially as above.

To check for bell-mouthed slots remove the chuck jaws and turn the chuck so a slot is horizontal. Now mount a DTI on the cross slide and set the probe against the slot and wind the cross slide in. The DTI should not move.

Dear,e.

It seems that your chuck back registering area has a tiny protrusion may be by an object while keeping aside. Its like a single knurl which makes the chuck slightly off set on face. The register or the nicely polished neck of the spindle has to be spot on mating with the female area of the chuck. If all that is not the trouble, swarf inside the chuck scroll, or the jaws assembled in different numbering to the embossed numerals on the chuck body.

If all fail, what you can do is hold a circular ring on the chuck but locking by opening outward. Then use pencil like die grinder mounted on the tool post and slowly do an internal grinding of the chuck jaws.

usually after long use, the jaws develop end plays with reference to the face as there will be wear on the T slot on the chuck body. If its so, and you cannot do anything else then pls go for a new chuck. Any way any brand new chuck is out by a few thou, may be .01, /.02 mm

There is a lathe called Cazenuve made in France which has a separate chuck holding plate that has a protruding ring, and this protrusion has 4 threaded holes and 4 threaded bolts with square holes to adjust , that which shift the chuck by a few thous to correct out of roundness. The three jaw chuck is held to this back plate by alan bolts sunk in behind the plate, and holding the 3 jaw chuck from behind at a 90% torque so it is free to shift to allow correction.

I'd guess there's a bruise that's stopping the backplate locating properly against its datum face, and tilting it. My Myford was a Speed 10, and in any case I can't remember whether the bottom of the backplate register bore or the back face of the backplate contacted first, but it could well be that engagement that's being compromised.

The runout of a straight parallel bar should vary with distance from chuck face if that's what's going on.

As already mentioned, your faceplate requires truing up, a difficult job on a lathe without power cross feed. The difficulty is in keeping a constant smooth feed while facing.

I have already mentioned that the register has little or no affect on the repeatability of a screw on chuck.

If the faceplate is trued up, then the backplate could be attached to it, adjusted for concentricity of the female register, and the threads very lightly skimmed. Then before taking anything down, the rear face of the backplate skimmed true. After that is done, the backplate needs to be screwed up tight and backed off several times to settle it down, before checking the runout of the chuck mounting face and the concentricity of the chuck register. Both may require truing up.

.

The fact that you wre able to do that, apparently without effort [otherwise you would presumably have mentioned it] seems to confirm that the thread is as it should be … That is one of the reasons why I suggested doing that first.

You have now eliminated the thread from the list of suspects; without modifying it.

I think it's most likely that there is some barely measureable blemish on [or in] the parallel register area of the spindle nose [or the backplate]. Your indicator tests are looking good; but they cannot tell you everything, because the probe is only making point-contact [ish]

It's probably time to start applying some blue, and checking that register very thoroughly.

I hope it goes well

MichaelG.

From reading thru the posts, The backplate thread and register appear to be OK but have you checked if the backplate is running true when attached to the spindle ?

If the backplate checks out OK, the problem must be in the chuck and if cleaned and no swarf in the scroll, etc, then the jaws or scroll is worn and the jaws need a re-ground.

Paul.