Facing bar ends parallel on the lathe.

This arises because I got involved with a project to make a riser for a VMC mill as detailed in another thread . It got me thinking. How does one turn the faces of a bar parallel?

The inimitable doubleboost just turned the piece end-for-end in the 3-jaw. I tried that with with a 4" diameter billet, making sure the workpiece was referenced to the flats on the jaws rather than the grippy bits. It turned out about 0.03mm out of parallel across 100mm. Not good!

I'm wondering if it can be done better by bolting to a faceplate which could be skimmed before attaching the work?

Robin

No need to skim your faceplate unless there is something measurably wrong with it. You could ruin a good plate.

But yes it's more reliable to bolt to faceplate to get faces parallel. Holding in the 3 jaw by a relatively small area can allow movement under load. Or bits of swarf or even duff chuck jaws can throw it out too. And always tap the the job home carefully with a brass hammer to make sure job is seated on all jaws. A 4 jaw would give better grip than a 3.

PS. But the faceplate method can mean you end up with the two ends parallel to each other but out of square with the outer diameter. This happens if the first unmachined face bolted to the faceplate is not square to the OD to start with -- as is common with power hacksaw cuts on stock. Probably for a riser, the OD is non-functional so won't matter if it is not square to the faces.

If you want to get the OD square to both ends and both ends parallel, the best way would be careful set up in the four jaw. Get OD running true at both ends and job seated firmly on all four jaw steps with a brass hammer. Then carefully machine one end. Then either flip it around and do the same again or use the faceplate to do the second face.

Using a tailstock centre would help keep things in place in the chuck too, facing across to leave a small tit in the middle to be removed last and perhaps even relieved a little so load bearing is around the outside.

Edited By Hopper on 08/08/2020 05:12:34

1st to check, is the cross slide actually square to the run of the lathe. Some are not and are not adjustable in any way to make them square. Which is why on some older machines, you need to set the compound to get a flat face.

It's not really a lathe job, there are too many potential errors

I would use a shaper (or if I had one a grinder)

the faces must be parallel, and flat, and at right angles to the diameter

Lathes do round bits and shapers do flat bits

I believe that most lathes are set up a bit convex or concave to start with, so even the facing off isn't really flat which doesn't matter with most jobs but when you need it really spot on you need the right machine

I suppose a decent mill with a fly cutter would also do flat

Edited By Ady1 on 08/08/2020 07:14:53

As you were using the outside jaws and the work was not back against the chuck it would be easy enough to get a dti onto the previously faced now back face of the work to check it for true running. Only tap the work back against the jaws if you know they are totally true.

Regarding hoppers point about ends not parallel that is easily solved by taking a cut off the OD at the same setting when facing the first end.

Edited By JasonB on 08/08/2020 07:16:58

.

Given the purpose of the exercise ... That would seem the obvious choice,

MichaelG.

Have a look on the Stub Mandrel site for Schlessinger's work and then you will know rather than merely believe.

It turned out about 0.03mm out of parallel across 100mm. Not good!

Good grief man thats a tad over one thou - I think its fantastic!

Obviously the person who has the requirement here has a mill, though it might be in pieces. But this is an obvious milling job, it would need a large lathe and chuck to hold it accurately for turning. I had a similar issue facing and drilling a pendulum bob and did that on the mill.

Surely it must be easier for the mill to be reassembled and used rather than a dodgy lathe setup?

Essm,

You beat me to it. I would say that 0.03mm in a 3 jaw chuck is quite good. You could also try checking the jaws with a DTI to see if that is the cause of this small error.

For reference, the jaws on my had it from brand new, top of the range, very expensive 4 jaw chuck is out by about 0.001".

Phil H

Just measured the 100mm dia EN3 flywheel blank I turned the other day in 4 places around the rim no more than 2/10ths deviation which is 0.005mm so I must be doing something right though did not clock the back so maybe I could have done better but job did not need it.

It turned out about 0.03mm out of parallel across 100mm. Not good!

But not too bad for a 3 jaw either ? Close enough to sort with a scraper if you are in a position to measure the error.

Could you not put the component between centres, face one end, reverse the part & face the other ? Should be pretty close to parallel done that way. May need a hole drilling & tapping in a non-working area to take a drive peg. I have a cylinder square at work that appears to have been machined that way, complete with an off-centre tapped hole at each end.

Nigel B.

The riser block can be fitted in one of three positions. Given that Chinese machines often have “errors” built into them it might actually be and advantage to have a slightly less than square riser?!

Turn it between centres and both faces will be parallel with each other.

Face the tailstock end first, remove part from lathe, refit back between centres with faced end now located at headstock. Face the tailstock end. Job done.

Martin.

Martin you came in while I was writing,

I would simply rough out somewhere near size and centre each end with a protected, or recessed centre. Drill and tap at one of the mounting holes on each face for a driver peg. Then mount the work between centres to take a light skim down each face and over the O/D if desired.

This does assume that the lathe is already set-up correctly.

Regards

Gray,

Edited By Graham Meek on 08/08/2020 11:26:09

Agreed, Graham.

If a fairly large central relief can be made at each end, which leaves a face of sufficient size for its intended purpose, then a much smaller area will need facing and the concave/convex facing issue becomes almost irrelevant.

Martin.

Oops. I missed that. Yes you are not going to get much better than that on the average lathe.

.

I am genuinely surprised to see ‘the book’ available there

... ‘though of course that is not all of Schlesinger’s work.

MichaelG.

.

Edit: But time has passed since this discussion:

https://www.model-engineer.co.uk/forums/postings.asp?th=132159

Edited By Michael Gilligan on 08/08/2020 14:03:01