How good is a good faceplate?

[Robin GrahamParticipant @robingraham42208]

I recently asked about turning the ends of a round bar parallel on the lathe. At some point using the faceplate was mentioned, and I took to the idea thinking that I could skim the plate to make sure it was perpendicular to the lathe axis. Hopper said:

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

This set me thinking. What is a good plate? So I made some measurements, rotating the spindle:

Results (in degrees, microns) were :

0, 0

90, -5

180, + 20

270, +22

so it's about 25 microns or a thou off over 120mm from the lathe axis.

Coming from nowhere in this engineering thing, but enjoying every nanometre of the journey, my question is whether this is reasonable for a 'jobbing' lathe of Eastern extraction. And, of course, if it can be improved upon.

Robin.

[Robin GrahamParticipant @robingraham42208]

[martin haysomParticipant @martinhaysom48469]

plenty good enough for me but it all depends what you plan to do

[Simon Williams 3Participant @simonwilliams3]

"Reasonable"?

Huh! I'd be mighty chuffed if I'd measured that.

For "reasonable" read "pretty blooming excellent!

Rgds Simon

PS But (if I remember right) that wasn't Hopper's concern.  To make a flat face at right angles to the axis of rotation you need the saddle to move at right angles to the face plate.  So this is checked by passing the dial indicator across the face plate surface, rather that rotating the face plate wit the DTI stationary.

Correct me if I'm wrong anyone?

Edit – PS added to treat Premature Send Button Syndrome

 

Edited By Simon Williams 3 on 15/08/2020 00:34:24

[Mike PooleParticipant @mikepoole82104]

I suppose the question is whether if you skimmed it would it be any better and if the next time you mounted it would it be the same? Management of our hopes and expectations needs to be tempered with reality.

Mike

[Jeff DaymanParticipant @jeffdayman43397]

take it off , put it back on, and re-clock it,  3 times. I'll bet it is a little different every time, at the 20 micron-ish scale.

and it does not matter at all, for most ME work.

If you skim it enough times, the whole face plate will disappear.

Make some parts, have some fun. Life's to short to be chasing rainbows.

Edited By Jeff Dayman on 15/08/2020 01:26:55

[HopperParticipant @hopper]

I agree with all the above sentiments. As a beginner, on a lathe of unknown quality, with unknown wear, you will struggle to improve on the one thou runout and quite possibly make it worse. Plus there is the flatness issue, even more difficult to achieve than the runout due to wear on the carriage way surfaces and cross slide.

Certainly try the suggested take the faceplate off and clean the threads and register faces etc and retest again, several times. You might be surprised at the variation.

The layer of rust on the faceplate (or is that exaggerated in the pic?) could make that much difference alone.

Headstock bearing adjustment and wear could make that much difference too.

If you just want to have some fun, bolt a disc of metal to the faceplate and take some test cuts across it and see how it turns out. When you can get that perfect, unscrew the faceplate and clean the threads and put it back on and see if its still perfect. When it repeatedly is, you can think about facing the face plate.

To measure flatness, you need to either use a straight edge and feelers across the face. Or you mount a dial indicator on the cross slide and run it across the BACK half of the machined face, ie between the centre line and the REAR of the machined face. If you just run on the front half, the dial indictor just follows the same path as the tool and reads 0-0 regardless of any angle on the face. But the back half is angled the opposite way. So the true variation is half the reading over that surface. Flatness across the faceplate should be either perfectly flat ie 0-0 reading or very slightly concave, probably within a thou or so for home hobby work on an old machine. This is because a job will sit firmly on a concave surface but will rock on a convex one.

Personally, I would leave it as is and carry on. If you want to machine the ends of your large round bar any closer to parallel than one thou, use the suggested method of between centres and face both ends in one set up.

Then you need to work on your ability to measure less than one thou with a large micrometer. (Digital calipers really aren't accurate enough.)

 

Edited By Hopper on 15/08/2020 02:22:18

[jimmy bParticipant @jimmyb]

I skim a face plate when I first mount it.

Jim

[Michael GilliganParticipant @michaelgilligan61133]

Robin,

May I put things into perspective from a different viewpoint ?

Your total indicator reading is less than 30 microns; which is a value that you may recognise … It’s the standard thickness to which petrological samples are prepared, to enable analysis with the polarizing microscope [many substances being transparent at that thickness]. It is certainly non-zero, and non-trivial in some disciplines, but it is small in the context of of the faceplate as a bolting surface.

Remember that the lathe will face the job “flat” regardless of those errors, and the major issue is lack of repeatability: According to Kinematic theory, the workpiece makes firm contact at three points, so the ‘terrain” means that you will get variation in tilt, according to the location of that ‘tripod’.

You might enjoy calculating those tilts for a hypothetically perfect workpiece attached to the rugged terrain of your faceplate.

MichaelG.

.

[*]

”flat” being a special case of “conical” … and generally set to be very slightly concave.

[larry phelan 1Participant @larryphelan1]

Let sleeping dogs lie !!!

[John HinkleyParticipant @johnhinkley26699]

I would wager that bolting a workpiece to the faceplate could introduce "stress distortion" (I just made that up) leading to greater run-out than you've measured on the faceplate alone.

I'd ignore it and worry a bit less – I'm completely grey-haired already.

John

[Nigel McBurney 1Participant @nigelmcburney1]

I would expect most ordinary lathes after some years years use to have some run out,expect a thou. Faceplates are easily distorted when jobs which are not flat and stiffer than the faceplate are bolted down too tight,a faceplate which has a permanent distortion has to be skimmed. To get a job machined very parallel,get a subplate of steel or alloy,mount this on the face plate skim it true then mount the work on the subplate using clamps and tapped holes in the sublate,alternatively a round sublate can easily mounted in a 4 jaw chuck . only snag with subplates on larger jobs is that the length of gap is reduced.For some skimming jobs such as brake discs I have the the slotted table removed from an old rotary table,this gets mounted in the 3 jaw and gets a couple of thou skimmed off when ever I use it,eventuall it will get too thin but not in my lifetime. A large old brake disc could be used as a subplate,mounted in the 3 or 4 jaw chuck.

[JasonBModerator @jasonb]

Well mine was skimmed the first time I used it and a second time due to some surface rust on one side but going by the comments here the old Chinese machine ain't doing too bad

Edited By JasonB on 15/08/2020 10:21:23

[Neil LickfoldParticipant @neillickfold44316]

As long as the faces are always very clean, and you don't get accidental stuff between when plates or chucks are changed, it will stay the same. It is interesting to see how little pressure is required to move the outer edge of the face plate buy a few hundredths of a mm in or out. When we have very large objects on a face plate on a cylindrical grinder, we only stop it for the minimum time to measure only. Otherwise it keeps rotating slowly to prevent a distortion in the assembly. Heavy objects out from the face take very little to move or influence.

[old martParticipant @oldmart]

Your plate is probably fine for 99% of applications, it will distort every time work is attached to it anyway.

I have just made one to fit the Atlas lathe we are doing up at the museum. I used the backplate of one of the chucks which came with the lathe. The chuck is scrap anyway. That got turned down on the Smart & Brown to 3" od ready for bolting to a faceplate which I had bought cheap on ebay. The 9" faceplate was cheap because it had a 1" X 10 thread and I was the only person interested in it. The faceplate was bored through 2" diameter and a 3" recess was made in the back for the backlpate to sit in. Using the rotary table on the mill, I drilled, tapped, and counterbored for 8 off 6mm SHCS and the halves were assembled with 620 Loctite. On the Atlas, it runs 0.002" tir on the face, and 0.001" radially. It may get a skin eventually when the Atlas gets powered up.

[Robin GrahamParticipant @robingraham42208]

Thanks for replies. It's reassuring to know that my plate is OK, and I take the point that if I mounted it again it would measure up differently. I'm envious of Jason's precision though!

As things turned out the job was turned and faced one end in the 3-jaw, then reversed and set in the 4-jaw to finish:

Slightly scary swinging a 24kg 160mm diameter billet at 700rpm (there was a time-served RR engineer present, he said it was OK!). Anyhow, it worked and no people or machines died.

It's been an instructive project on several fronts. I'm learning I hope!

Robin.

PS the 700 rpm stuff was with tailstock support before anyone has a heart attack.

R.

Edited By Robin Graham on 18/08/2020 23:27:30

Edited By Robin Graham on 18/08/2020 23:35:47

[HopperParticipant @hopper]

Well done. Good to see a result and thanks for posting a result. So many dont bother.

One point with the reading in Jason's video is that it is usually recommended not to run a dial indicator on a shaft or job rotating under power at speed like that. Apart from wearing flat spots on the plunger end, the side loading on the plunger combined with the speed of change in reading can result in the plunger binding and massively under-reading the run out. Recommended practice in most camps is to rotate the spindle slowly by hand to get the most accurate reading. I'm sure Jason would have done that too but the video does not mention it so could be a bit misleading.

[Author]

Posts