Dialing in a 3 Jaw Chuck

[old mart]

A number in the Fibonacci sequence would be preferred such as 21 rather then 20.

A thread lead of 0.04761904761904761904761904761905 has advantages over a lead of .050

That number is likely only an approximation, it is probably a recurring decimal.

[mattthemuppet]

Another option on old mart's train of thought is to have the adjustment set screws parallel to the back of the chuck, threading into the "boss" on the front of the backplate, with the Allen key end just touching the inside rim of the back of the chuck. You'd have the thickness of the set screw from the thread ID to the inside of the hex to work with. The use a ball end mill or grinder to create a groove in the backplate for the Allen key/ hex wrench to go. Dimensions will be pretty tight, but it's an idea.

[The Metal Butcher]

Guys... it's not that deep. Regular fine thread is fine. You futs with it for 5 minutes, lock it down, and maybe do it again 10 years later.

[dian]

consider differential screws in you designs. (would get a bit complicated.)

[J Tiers]

A number in the Fibonacci sequence would be preferred such as 21 rather then 20.

A thread lead of 0.04761904761904761904761904761905 has advantages over a lead of .050

yes, those will definitely bother the excessively rule-driven among us.

[Bented]

A number in the Fibonacci sequence would be preferred such as 21 rather then 20.

A thread lead of 0.04761904761904761904761904761905 has advantages over a lead of .050

[J Tiers]

The pitch gives the advance per turn (360 deg). Your eccentric gives full travel in a half turn, 180 deg.

So comparing on the basis of advance per turn, or alternately, on thou per degree of rotation, makes sense to compare sensitivity of the adjustment.

Having a bit more adjustment is good, I do not think 0.010" or 0.020" is too much. It's not always for the chuck, although that is the main issue, of course.

[old mart]

If you have a three jaw chuck that is more than 0.004" runout then it wants either some serious work or downgrading to a welding fixture. The movement should be as small as you can get away with, a chuck with 0.003" runout is common, and if the adjustment was 0.006", it would be more than enough. My guess at 0.020" eccentricity may be too big.
JT, I don't understand where the thread pitch comes into the equation, could you explain?

[J Tiers]

A throw of 0.020 would be 0.020 per half turn.

a 24 tpi thread is equal to that, and anything finer such as 32 tpi would do better. Total movement is unlimited, of course. Finer threads would be possible, and if the backplate is steel, would be practical.

[old mart]

Any design thoughts are very welcome at this stage, there are so many variables. As long as the fitting is very good, the eccentrics would only have to have a throw of about 0.020", that would make them quite sensitive.

[J Tiers]

Not sure how the eccentrics would be more sensitive than a fine screw, while still having good range, but it is an interesting thought.

[old mart]

Yes, the idea of eccentric ends is well worth considering, they could be a bit larger than the grub screws as there is likely to be more room inboard of the register to fit them. The finer details of exactly how the eccentric would be made aren't there yet, but they have to be fairly stiff to turn, and 4 are better than 3.

I could definately fit the eccentrics to the 160mm chuck backplate, but there would probably not be room with the 5" and 100mm as I have made them locking on the threaded spindle and that uses up the space. The 160mm has not been modified.

[Galaxie]

This is what I meant by the use of grubscrews in post #77. The screws are only for adjusting the chuck with the holding screws finger tight. The main screws must be tight before using the chuck.

That seems like it could work, although the angle on the cone-points will also want to push the chuck off the backplate (toward the tailstock). That could cause run out on the chuck face if you're not mindful of it.

Another thought I had is instead of cone-points, maybe there is a way to use screws with eccentric turned tips? The eccentrics would push on the inside of the register diameter. In theory, you could adjust just one or two eccentric screws to push the chuck over and then re-tighten the backplate screws.

[old mart]

This is what I meant by the use of grubscrews in post #77. The screws are only for adjusting the chuck with the holding screws finger tight. The main screws must be tight before using the chuck.

[old mart]

I must put my mind to a design which allows screw adjustment of the chuck on backplate, but as there are so many variables, it is difficult. The set tru type of chuck are always longer than a standard type which has disadvantages on a lathe with a short bed. One possible design might be 4 grubscrews spaced at 90 degrees with 60 degree ends threaded through the backplate to hit the step in the back of the chuck where the register is. The maximum allowable movement would never need to be more than 0.006" from dead centre.

Here are three chucks of 100mm, 5" and 160mm. The 100mm has 6 front mount 6mm SHCS. The 5" has 3 front and 3 rear SHCS. The 160mm has 6 studs lock washers and nuts on the rear. As the backplate was slightly undersize, the studs are stepped 10mm/8mm (there were some titanium Westland Lynx gearbox housing studs lying about, NOS, that type were replaced with steel as they were not up to the job).