Facing Error

.

But as the result is [presumably] still way outside Warco's original certificated performance, I guess there is still scope for fault-finding and rectification.

MichaelG.

My bet's still on headstock misalignment .

Keith.

It's very possible but you have to make the presumption that someone with the presence of mind to measure for concavity in a face would also have checked for taper in their normal turning, especially if it's to such an extent.

The only thing to do is ask SOD:

Dave, have you checked that parts you turn which are held normally in the chuck for taper? If not, that is the very next thing you should do.

It does have the minor complication that a tapered part might be the result of a twisted bed whereas a faced part would be barely affected by bed twist if at all.

Given that on page 1 Dave said he halved the error by locking the topslide (but only Jason noticed) I do wonder if it's just a matter of checking gibs and slides and locking the saddle as well...

Neil

.

Ahem ! ...

Jason prompted that action, but he wasn't the only one to notice the improved result.

MichaelG.

Edited By Michael Gilligan on 15/08/2018 18:21:35

As I suggested it I did notice the results, I'd rather look at the figures 0.23 down to 0.08 which is somewhat better than half as Dave said in subsequent posts. Still about another 30% to chase down though.

Might also be worth checking the faceplate with a straight edge and feeler gauges and if that is good mount it up and run the DTI along that with the base mounted on the cross slide. I'm assuming Dave has not skimmed his.

May not have been a smart thing to do given convex is bad, but - top-slide and saddle locked - I faced off the chuck-backplate in reverse, that is with the tool cutting from outside to inside with the lathe in reverse and the tool-post at the back:

This results in a convex face. It has the same slope as the concave cut when the lathe faces off in the usual way with the motor running forwards, but in the opposite sense.

Putting the plate flat on my milling machine and measuring the error with a DTI in the spindle confirms the slope at about 0.07mm in 60mm.

The error is consistent with the slide not running parallel to the headstock. Need to double-check, but the absence of error when cutting parallel (tbc) suggests that the headstock and ways are true making the cross-slide chief suspect.

Turning cast-iron has left a filthy mess and everything I touch is black. This includes the kitchen floor! When I've cleaned up I'll try and prove it's the cross-slide that's wrong.

Dave

.

I wasn't suggesting for a second that you didn't notice

I simply mentioned that you were not the the only one.

... and Yes, I also paid attention to the figures, as I expect did several others.

MichaelG.

_______Need to double-check, but the absence of error when cutting parallel (tbc) suggests that the headstock and ways are true making the cross-slide chief suspect.

SOD, Do I get the impression you are turning your test bar between centres (as in tbc above) in which case it wont tell you if the headstock is out of square only if the tailstock is in line with the headstock.

Keith

Dave, are you having a job getting any closer to the headstock? Looks like you have the topslide fully extended to the left and quite a lot of tool overhang. If the problem is partly down to movement in the ways then getting the overhang as small as possible will help.

I'll check by turning a 6" of ⌀ 1½" aluminium rod held in the chuck, not between centres. Thanks for the reminder!

Dave

Yes, I've still got the telescopic swarf guard fitted which keeps the saddle well away from the headstock. To reach the backplate I've used a long boring bar and cranked out the top-slide almost to maximum as well.

I don't think the overhang is the cause because I first noticed excessive dishing when the work was held in a 3-jaw allowing the tool-post to get personal. Also, with the top-slide locked as you suggested the error is about the same irrespective of how close to the headstock the tool-post is.

My original measurement was faulty in that a loose top-slide made it look worse whilst measuring on the opposite side of the slide doubled the reading. I'm more confident that the actual error is about 0.07 over 60mm concave. This hasn't caused any real-life issues but I'd like to reduce it if I can.

With luck it's something to do with the gibs. Last time I had an unexplained parallel error it turned out to be swarf inside the chuck. Once I'm satisfied it's the slide, I'll take it off, have a look, and give it a clean.

Ta,

Dave

Edited By SillyOldDuffer on 15/08/2018 19:37:15

Hmm

Back to basics.....

Headstock spindle are the bearings tight with no play? Not much error there.

Headstock aligned to bed? turn a preferably thick say 25mm or greater diameter bar and check for taper? If the error is small the headstock is aligned to the bed with little error.

Facing test turn reveals a large error. Why? The cross slide is not aligned at 90 degrees to the bed.

I would remove the cross slide clean and replace it then carefully reset the gib for a nice firm but not overtight fit.

Do another test facing cut.

Regards John

Edited By John McNamara on 16/08/2018 00:41:47

When you first set the lathe up did you level the bed to ensure there is no twist ? You really need an accurate machinist level for this but even a standard spirit level is better than nothing. I have heard of some that have deliberately twisted the bed to get it to turn parralel not knowing the actual fault was a headstock cocked to one side.or nose up / down . Its a simple check , sit a level on yor cross slide and note the reading then wind the carriageto the other end of the bed and check again - there sould be no deviation on the level .

Once you have this set do a parralell turn test and if it is out of kilter you may need to shim the headstock to correst , this is a common thing with some brands of mini lathe . I also check with a ground mandrel to see if there is any nod on the spindle as this effects it turning parralel

If you can confirm that the bed is not twisted , the headstock is aligned correctly it only leaves one thing - the cross slide is not perpendicular to the spindle .

Looks very much as if XD 351's late entry has won first prize!

Because I'd checked the lathe was turning parallel recently, I assumed that it still was.

After cleaning up (there is still dirt appearing) I:

1. checked the slide was parallel with the spindle spigot: it showed the same error as facing, ie the cross-slide and headstock are not parallel.
2. Inspected the the cross-slide and experimented with the gibs to no avail.
3. To confirm the headstock is true with the bed, I cut a test-bar. This revealed a taper of about 0.1mm over 90mm, which is BAD and a surprise.
4. Checking the geometry confirms that the taper error is dimensionally consistent with the facing error indicating strongly that the head-stock is misaligned. Bit worrying because the headstock isn't adjustable on this type of lathe - it's machined to fit the bed prisms.
5. Checked to confirm the bed is level: it's not! My level is crude, but it shows clearly that the lathe is leaning slightly toward the operator, and that the headstock end is leaning nearly a degree more than the tailstock end. The bed is twisted.

I think I know what's happened. My garage floor has a small high spot under the headstock and I cured the problem after installation by resting the stand on a pad of roofing felt about 6mm thick. The lathe wasn't twisted and disn't cut tapers. It's likely that weight and vibration have crushed the pad over the last few years and the lathe is now resting unevenly on the bump again.

I might see if tapping a wedge under the stand helps but really I need to either fit adjustable machine feet or level that part of the floor properly with some goo. Right pain to get under the lathe because of the amount of junk in the way.

Astonishing the number of times bodges like this have come back to bite me. Ho hum...

Many thanks for all the help, the various suggestions guided me to the target. Fingers crossed the lathe is cured taking the twist out.

Cheers,

Dave

Therein lies the conundrum. If the cross-slide is not perpendicular you would still not see any dial movement by following the cutter's path with a dial gauge, which it seems is what's happening.

What it could be is that the slide or gib is worn or such a poor fit that cutting forces are deflecting the cross-slide where dial gauge plunger is not.

Confusion earlier about this point. True the DTI follows the tool-path BUT it was positioned on the opposite side of the plate to that cut by the tool. In that position the DTI sees the hill caused by the taper and registers twice the error. When positioned on the other side, the DTI parallels the tool-path and all appears well.

Dave

Ah, then the error is doubled as I'm sure someone mentioned previously.

Where are you Dave if you're close to North Kent I could pop by and assess it for you?

Tes, I have a "thing" about lathe beds not being twisted; and Dave says that his is, beacuse the roofing felt packing has settled.

Dare I suggest mounting the lathe on a substantial angle iron frame, (a rectangular frame made of 50 x 50mm angle iron, with flat face upwards and the web facing down). If the lathe is them mounted on four studs as large a diameter as the lathe feet will accept, with the studs secured to the angle iron by nuts above and belw the top face. The lathe is then secured by nuts above and below the feet.

This is the way that my 300Kg lathe sits, (using 1/2 UNF threads) and getting rid of twist was simple.

By adjusting the nuts on the lathe feet, it ought to be possible to remove any twist from the lathe bed, and hopefully return to turning parallel.

If you do not want to move the lathe about the shop, clamp the frame to the concrete floor with Rawlbolts, or something similar, and then mount the lathe and remove any twist from the bed.

Once this is satisfactory, you can go back to checking the alignment of whatever takes your fancy.

Howard

Thanks Howard, I haven't decided what to do about it yet and suggestions are welcome.

Interesting experiment today. I used a crowbar under the stand to tilt the lathe to see if the level showed any sign of the bed untwisting. Not at the time, but retesting after removing the crowbar showed the twist had nearly halved. Cutting another test bar showed the taper has also halved (roughly). I think lifting the lathe moved it slightly on the rough floor and the bed's less twisted in the new position.

Health and Safety note: lathes are top heavy beasts. I think it would be easy to tip mine over with a crowbar! Maybe the improvements should include bolting it down.

Dave