Altering a new backpt

I want to fit a new 3 jaw chuck to my old lathe, the spindle thread is 11/2" x8tpi which is the same as Boxford.

The Boxford backplate which I have just bought screws on fine but the register is about 5thou bigger, so I want to bore the BP out and fit a reducing collar, my first thought was to make it oversize then heat the BP, cool the collar and squeeze them together, but I wonder if Loctite might be a simpler option?

Anyone got any thoughts?

Loctite!

Yep ! Loctite. If it fails then use a more agressive plan. Noel

A sleeve for the spindle-backplate? I run backplates with up to 0.020" clearance and down to 0.0005", they all repeat exactly the same.

Edited By old mart on 31/01/2022 21:36:28

I have back plates on my Boxford that range from 2>10 thou clearance. The backplate should run right up to the face of the spindle. The spindle face should be unmarked & the bottom face of the back plate register that goes up against the spindle face should also be flat. A little oversize on the register is no problem at all. Dont try to fix something that aint broke.

regards

I agree with the above. At least screw it on and check it first (several times). Odds-on it will repeat pretty well.

If I don't get it running true on the register my new 3 jaw will run out, if it was a 4 jaw independent that might be acceptable but not for me on a 3 jaw.

You don't have to get it running "true" on the register, just consistent. IE reapeatable that it sits in the same position everytime you put it on. Often the thread and the flat face of the shoulder do this, without the register being a tight fit.

Then you have to put the backplate on the spindle and machine the backplate in situ so that the part the chuck locates on is running dead true to the spindle. That way, if the backplate is consistent/repeatable, the chuck will always run true. Well, as true as a 3 jaw can. Which is often up to 3 thou out of concentric depending on the position of the jaws etc.

But if you really want to fit a sleeve regardless, you will have to bore the hole in the backplate out and fit a sleeve that is undersize in the bore and bore to final size after it has been fitted. You could make the sleeve about 1 thou oversize on the OD and press it in cold using a large nut and bolt and flat washers to pull it in nice and evenly. Bit of Loctite on there for insurance as well could be used.

Then after boring the sleeve to fit your register size-for-size, you will still have to mount the backplate on the spindle and machine it in situ to fit the new chuck.

Look at this on the lathes.co.uk web site Page 7

Go down to Making a Backplate for a Screwed Spindle Thread

Read the part near the bottom regarding experience with a Boxford

Martin C

Have you machined the backplate insitu to fit your chuck ? If not do so before trying anyting else.

I have never seen a chuck/lathe where you need the register ID of the backplate to be any closer than stated by myself and others. The only way the OD of the spindle register would affect alignment is if it were a press fit with the ID of the BP. Then you would probably be dragging the backplate sideways on the spindle threads. It would be better to check the register faces on the spindle and the BP.

regards

The important parts which affect the fit of a backplate are the squareness of the rear face which contacts the face of the spindle and the threads which are normally 60 or 55 degrees. The tightening up of the threads centre the faceplate and the square end contact does the rest. The so called "register" does nothing at all, except on the rare spindles with a square thread which cannot self centre.

It seems sensible to do nothing until you have tried the fit several times, measuring the run-out each time. Only if the results vary significantly do you need to alter fits etc.

Cheers, Tim

Before skimming the front of the backplate, I would always fit and refit it to the spindle at least 6 times to settle it down.

After a Google I realise this subject has been done to death before on here, with supporters on both sides with out a resolution.

I only asked whether Loctite would be suitable for holding the collar in, and it appears it is

Here is a link to a previous discussion.

https://www.model-engineer.co.uk/forums/postings.asp?th=74991

There might be one way to resolve this once and for all. I am a firm believer that the spindle "register" is not needed. In a previous discussion I was challenged to try reversing my faceplate an seeing how it ran! This flippant challenge was soon resolved with the use of a suitably sized parallel ring spacer between the reversed faceplate and the spindle abutment face( I used the outer bearing shell of an old magneto bearing) It ran and located just as well when reversed as when mounted conventially. If the people who believe in the spindle "register" were to try this for themselves, then measure the run-out both ways and report them hrere, we could resolve it, or at least be able to separate the armchair conjecurists from the engineers.

dave8

Failing to resolve the issue in the past strongly suggests the register isn't important: if registers mattered, there would be firm evidence in their favour, and it seems to be AWOL!

But absence of evidence isn't proof. We need more facts. David Bennett8's valuable experiment indicates registers don't matter, but we don't know yet how Ian's collar performs in practice. Most grateful to Ian if he would post before and after measurements.

Definitely wouldn't bet the farm on it, but my guess is the collar won't make much difference - if the register isn't important there's a 50:50 chance the result will be slightly better or slightly worse. Worth trying though - I might be completely wrong!

Dave

You are happy with the way you do it Dave which is all that matters, I am happy with the way I do it, I don't say you're wrong but it's not my way, so I will fit a reducing collar to locate on the register as Acorn tools intended.

Good for you, but please contribute some feefback with before and after measurements, also on the reversed faceplate experiment.

Dave8

Edited By david bennett 8 on 02/02/2022 11:59:47