Lathe alignment problem.

[mu38&Bg#]

.002/2 of taper is rather large. Even if bearings were totally shot, I'd expect moving the headstock to change the taper on a turned shaft. I'm uncertain why the tailstock comes into the discussion when trying to sort headstock alignment.

 

[mcostello]

I am hoping that it shows the lathe is actually of turning a straight shaft, and not some other undiscovered problem. It was .012 in 2" when I started. Installing new springs seemed to help but I cannot make it any better.

 

[mcostello]

.002 in 2" is miserable to work with.

 

[goldstar31]

I wonder whether there is a very simple explanation that has yet to be explored. I was re-reading George Thomas in one of his books called the Model Engineers Workshop Manual and was plowing my way through his discourse( s) on alignment. I wonder if tool height or a worn base on a tool holder might be or contribute to cutting a waisted cone of metal.

Again, I was wondering whether tests should be carried out with the additional and problematic top slide instead of a more rigid tool station utilising the boring table.

And then there is the story of 'points per square inch'- the story of those who can scrape and those who ---well, cannot.

Perhaps enough to get on with- methinks?

Norm

 

[Hopper]

.002 in 2" is miserable to work with.

So get two dial indicators and set the headstock spindle parallel to the bed, without that .022" offset, as already discussed. It's really not that hard once you have a second indicator telling you exactly what that headstock is doing when you turn the jacking screws.

You are right that the fact that your lathe will turn a parallel shaft when between centres, after suitable "dialing in" of the tailstock centre indicates the source of the problem is not bed wear or low tool position or loose gibs or anything else like that.

So the problem must be headstock to bed alignment. Buy a second dial indicator and spend some time and care getting it just right.

 

[mu38&Bg#]

I think some photographs of what is being adjust and how things are being inspected would be helpful. Something is being lost in communication. It's simply not possible to change the angle of the headstock relative to the bed and not see a change in taper.

 

[mcostello]

That's what I thought also. The adjustment screws are buried and hard to get to. It will have to wait for a while as I am swamped with work and need the lathe for it. Thanks for the ideas and I will continue this post when time allows.

 

[petertha]

I am hoping that it shows the lathe is actually of turning a straight shaft, and not some other undiscovered problem.

I made this suggestion on another post regarding head stock alignment. Any chance you can get hold of an accurate test bar with taper profile on one side? Its the best 50 bucks I ever spent. For example if you have MT5 spindle, insert a MT5-MT3 adapter sleeve, then the MT3 ended test bar. Now you have temporarily completely eliminated chucks and cutting forces & stock overhang... you can focus on alignment as step-1. Mount the DTI to apron & traverse down the side of bar to detect any headstock yaw rotation in or out. You should also put DTI ball on top of bar, repeat & detect if the headstock is pointing up/down for whatever reason (which will also cut a taper).

But getting back to the headstock rotation, putting a DTI on the end of the bar should be showing you movement as you adjust the jack screws & re-lock down the headstock. Goingto the end of the bar is maximizing this sensitivity. The rotation point is somewhere inside the headstock, but we dont care. We are looking for confirmation of movement direction & amount at some consistent poisition for test purposes. Note that some lathes actually specify a permissible amount of 'point in'ie to operator side of tailstock. I had a somewhat similar issue adjusting my tailstock in/out. Even though I could displace the axis with the 2 adjusting screws to a target zero position with DTI in constant contact with a test bar, turns out the torquing effect of securing down the position clamp bolt could make it wander off zero again. So it was a bit of an iteration. But point is I would have never known this unless I had a direct reading dial giving me constant feedback. That's why I wonder if your headstock hold-down clamping mechanism is possibly somehow over-riding your jack screw setting.

Something is not adding up because the amount of discrepancy you are seeing by cutting a coupon should stick out like a sore thumb on this basic alignment test. Now if the alignment happens to show its within accurate tolerance, then we proceed with work holding & cutting related issues. But cutting stock to ascertain taper is practically a waste of material if the headstock orientation is questionable at this time. Introducing the tailstock over & above this is a recipe for chasing your tail. Maybe the HS is out & TS is in, but how would you ever know?

I think this 'barbell' or 'Rollie's dad method' are the go-to diagnostics where people have 'cast in place' headstocks. In other words they are presuming the factory alignment must be pretty close to begin with. But when you have a lathe with the ability for the HS to be rotated & reset into a yawed position, all bets are off. Its a different machine. You need to nail this down first.

 

[Hopper]

I think some photographs of what is being adjust and how things are being inspected would be helpful. Something is being lost in communication. It's simply not possible to change the angle of the headstock relative to the bed and not see a change in taper.

+1 on pictures. We seem to be going in circles otherwise.

+1 on impossible to rotate headstock without changing taper -- unless the pivot pin is missing, broken or otherwise fubared, or not fitted to this particular model etc. A dial indicator placed each end of the headstock would reveal what is actually going on.

 

[mcostello]

When I can get to it this will be the first thing I will try. I have not thought of up and down as being important in this problem, but I will research this also. I am planning to see if I can pull the pivot pin but it might involve pulling the headstock which is probably in the neighborhood of 300-400 pounds. I don't have a test bar but I do have a collet setup and some ground stock which would work. Nothing makes sense at this time, but I will keep on it. Thanks, Mark.

 

[john_reese]

Mcostello,
You need the pivot pin to remain in place. Just back off the hold down bolts a bit, then go after the jackscrews.
I am not in favor of test bars. You don't know how true they are and a little dirt or a burr in the taper can throw them off. Instead, chuck up a heavy bar. It needs to be chucked the full length of the chuck jaws. Make it as long as you can get away with without chatter during light cuts. Starting about an inch from the end reduce the OD of the bar up to about an inch from the chuck. There are two reasons for reducing the OD: it eliminates the potential for tool wear between the ends of the bar, and it saves time because you don't have to machine the whole length of the bar. Take a cut off the raised bands to true them up. Take a light cut off both raised bands without disturbing the cross slide or compound between cuts. Measure both both diameters. Ideally they will be the same. If not, adjust the jackscrews and repeat the test until the cut diameters are equal.

Use a large diameter bar to reduce the effects of deflection. Thick walled tubing is OK.
Use the longest bar that is practical to increase the sensitivity of your measurement of taper.
No tailstock or steady rest support.
Get the headstock running true before you even think of messing with the tailstock.

 

[mcostello]

What You have suggested I have done many times, it shows the headstock end bigger than the tailstock end. Adjusting from there does no good. I have not used two indicators yet to see what that shows.

 

[goldstar31]

If the bed is not in twist, you have wear in the front shear or wear or mis-adjustment in the saddle.

I hope that you have not been using a damn fool top slide which is another variable that can cause trouble

Cheers

Norm

 

[petertha]

Lets say hypothetically the HS was aligned 100% parallel to the way axis viewed from the top. But you had microscopic grit or burr or casting warp or whatever between the underside of the HS block & the mating face of the lathe bed such that the head stock was pointing net up or down slightly viewed from the side. Even assuming perfect cutting finish, nil cantilever bar deflection under tool load, stress relief, material skin effect... etc. two 'different' barbell diameters would not point you to the correct HS misalignment diagnosis. Worse yet, you 'interpret' this resultant taper cut as a 'typical' deviation in the horizontal plane and now start rotating the HS to compensate. Now you have 2 problems instead of just 1 and they are somewhat masking one another. If you ran a DTI along the top of a test bar, it would indicate this.

There are ways to measure test bars to get a sense of accuracy by rotating them to different marked quadrants, at different length positions, remove & re-insert to different clock position. Counting on a lathe to cut to the +/- degree of center-less grinding operation on a test bar seems like a stretch to me. Never say never with an import, but I would hope it should be as good or better unless a completely botched job. Actually that's not even true. The ground finish allows you to compare diameters all the way down the test bar in opposing orientations using your same best micrometer. So if that checks out the remaining deviation to validate is banana effect which should always hilite the same under spindle rotation. For sure the tapered surfaces have to be spotty clean. But if we are going to second guess that then we may as well include mounting anything to our tapered spindle nose, or dead centers into the tailstock etc. The fact that things are not adding up on this makes me think its something outside the normal box.

 

[mcostello]

This is odd. Nothing makes sense. I do have time to put the collet attachment on and check with a ground bar. Will report back tomorrow night. Another funny thing is when I turn a shaft and it is bigger on the headstock end, then go to boring out a holeit also is bigger on the headstock end. Seems like it is just the way it wants to go.

 

[petertha]

Actually both attributes make sense if HS is pointing out to operator side I think.

 

[goldstar31]

I could be wrong but I have this gut feeling that 'something needs to be attended to in addition to fuddling with the headstock.

I recall tackling a wreck of a lathe and filling the gaps ( which shouldn't really be there) where the badly fitted gib strips didn't quite fit with metal based polyester.

There is a whole library of comments on badly. fitted gibstrips in Chinese/Taiwan lathes.
So these things are not unknown. I think Baz was part of the discussion

Norm

 

[goldstar31]

On a different quest , I was browsing John Moran's excellent Gadgetbuilder.com site and noted again that he had written up how he had corrected his Far Eastern lathe.

I noted that he included a sample test of alignment checks.

There you have it- or not

Norm

 

[ShopShoe]

I just re-read the original post again.

It seems to me that there may be something in the alignment or tracking of the saddle / cross slide / compound / toolholder area.

If it was my lathe, I would probably disassemble those parts and inspect them carefully, following advice from some of the books mentioned earlier. I would use the same reference material to check the bed again with independent measurements, not using the tailstock, headstock, or saddle assembly as a reference.

I would expect to find some things that need attention in all of this, some of which I might be able to address and some I might have to live with. I would be very careful in deciding to do any scraping or grinding, which could end up creating an additional error that would make the general situation worse.

In reassembling, I would put the saddle together, then test alignment without the cross slide. Add cross slide and re-test, then add compound and test. If I was to try cutting a bar for basic testing, I would try it without my usual toolholding setup (IE: remove the QCTP if I'm using one).

I'm rambling, I know, but I had half-a-cent to add.

My gut feeling, which is totally unscientific, is that there is more than one problem in general, as well as a specific problem causing the basic error.

--ShopShoe

 

[goldstar31]

ShopShoe
Your comments are clear enough. . I suspect that there is saddle problems.
In any case, there should be a constant maintenance schedule on the lathe ways and the saddle..
Regards

N