Question On Aligning The Lathe Spindle And Bed Twist

[sanddan]

question one should be:
How is it possible that you get any reading in the first part of the measurement going in on the faceplate?
The indicator is mounted on the compound and should follow the exact same path the cuttingtool followed.
The measurement on the part going out of the faceplate is logical.
Then indicator will follow the path of the cutter going in, even past the centre where the faceplate actually is coming out.

on edit:
after thinking about it the only way to explain the measurements, which are consistent, is that the faceplate gave way under the pressure of the cutter.
More towards the outer rimm , lesser towards the centre.
As a problem also exists while turning in a chuck i suspect the bearings of the spindle need adjusting.

As PM stated in his post "I've done many shafts and parts that had to be perfect, but almost never when facing". This seemed to make sense to me, get the alignment true using the lathe in the area it will do most of it's work. I suspect the wonky readings I got on the faceplate were do to the carriage not being 100% true to the ways over it's full travel. I used a 1.5" dia steel bar I found in my steel storage and got the same amount of taper as the 2.5" aluminum bar so the taper was not due to flex in the chuck from the bigger bar but is due to headstock alignment. I am still trying to decide what to do about the vibration I am getting when the lathe is running. If I level the lathe the loading on the feet isn't even and I get an oscillation that gets the lathe rocking front to back. I can get rid of most of that by balancing the load on the leveling feet but that tends to through off the overall level. I am thinking about making a base for the lathe that spreads out the adjusting feet in hopes that would stabilize it. I don't want to raise the height too much as it is great right where it currently is so I'll have to think on that.

 

[janvanruth]

It does make sense if the spindle axis isn't perpendicular to the cross slide's line of travel (as with a misaligned headstock) AND if he isn't running the DI along the exact path that the cutting tool traveled from the front edge toward the center of the faceplate. It could be the angle of the photo, but that appears to be the case, assuming the cutting tool was on center.

Tom

an indicator that follows the path of the cutter will ALLWAYS show zero movement
even if the headstock and the cross slide are both several degrees off
measurements not taken on the centerline are useless

 

[janvanruth]

As PM stated in his post "I've done many shafts and parts that had to be perfect, but almost never when facing". This seemed to make sense to me, get the alignment true using the lathe in the area it will do most of it's work. I suspect the wonky readings I got on the faceplate were do to the carriage not being 100% true to the ways over it's full travel. I used a 1.5" dia steel bar I found in my steel storage and got the same amount of taper as the 2.5" aluminum bar so the taper was not due to flex in the chuck from the bigger bar but is due to headstock alignment. I am still trying to decide what to do about the vibration I am getting when the lathe is running. If I level the lathe the loading on the feet isn't even and I get an oscillation that gets the lathe rocking front to back. I can get rid of most of that by balancing the load on the leveling feet but that tends to through off the overall level. I am thinking about making a base for the lathe that spreads out the adjusting feet in hopes that would stabilize it. I don't want to raise the height too much as it is great right where it currently is so I'll have to think on that.

A lathe, any lathe, shouldnt vibrate while running not under load.
Without the chuck there is no vibration?
Mount a, known to be good, chuck, and see what happens.
Taking away the cause of the vibration rather than dampening it should be the way to go.

 

[higgite]

an indicator that follows the path of the cutter will ALLWAYS show zero movement
even if the headstock and the cross slide are both several degrees off
measurements not taken on the centerline are useless

Yep, that's pretty much what I said. Or maybe just implied.

Tom

 

[Tozguy]

Once headstock alignment to ways is assured and lathe is more or less level (not critical) the two collar method will show if there is twist in the ways, which is essentially what was done. Use of the word 'taper' might be misleading. A .0015'' larger collar at 10'' from the chuck is not bad in my books. Outboard hub can be expected to be slightly larger due to deflection of the work even when all else is perfect. 4'' to 5'' between collar measurements is enough for 1,5'' bar, longer is inviting more deflection and 'confusion'. Tool and turning method are important to get the least deflection.

OP's approach was correct. Adjust tailstock cabinet feet front and back to remove twist. Adjust to spec. for twist in ways. It is normal that both TS feet do not end up with the same pressure on the floor. Do not try to accomplish anything else by adjusting TS cabinet feet. Recheck after running the lathe for several hours.

 

[4GSR]

I am setting up my new PM1340GT lathe and have a question. I leveled the lathe on the long axis and across the two flat ways. I then installed a 2.5" dia by 12" aluminum bar in the chuck and after truing the bar I made a light cut (.002") at each end (bar only supported in the chuck). Over the 10" span I saw .008" taper with the headstock end being smaller than the tailstock end. I next extended the front adjuster foot on the tailstock end and made repeated light cuts measuring the 2 ends of the bar. I was able to get enough twist out of the bed to reduce the taper to .0015" over the 10" but at that point the rear adjusting foot at the tailstock end wasn't touching the floor anymore. After extending the rear foot until it made contact with the floor I made another cut and confirmed the taper didn't change. Here's the question.

Does this seem like an excessive amount of twist in the bed or more like the headstock isn't aligned properly with the ways? How do I tell which it is or which should be adjusted?
...snip....

First thing first. You did correct thing by leveling your bed to eliminate any twist in the bed. And you did correct thing doing a test cut on a piece of stock which gave you an indication of .008" taper in 10". Stop!
You do not adjust the feet of the machine to remove the taper. This is WRONG! wrong! wrong! Even the best of the best dictate this. But it is wrong! You don't do this on a new lathe! Ever! This goes for the two collar test also. The two collar test is for aligning the tailstock only in my book.

I'm sure in transit from Asia to USA, the alignment was disturbed. What you have to do is re-align the headstock to the bed. If I understand it correct, the Asian lathes have a easy way to adjust the headstock to get it back in alignment. It would be better to use a "known" test bar mounted in the headstock spindle and indicated for near zero accuracy.
Cranking the carriage back and forth, using a tenth reading indicator, check for readings. From the readings recorded, determine how much is adjustment is needed to get the headstock back in alignment. Keep checking and adjusting until you get as close to zero as you can get it. After you do this, set up and take a light cut of about .005-.010" and check the amount of taper you are getting. One word of caution, you are never going to get a perfect zero reading from end to end. Just get as close as you can get it to zero!

Once you do that, then you can move on to the two collar test and check the tailstock alignment to the headstock. Again,make your adjustments to the tailstock, not to the bed!
Sorry for the rant, just had to get it off my chest! Ken

DON'T JACK WITH THE FEET ON THE BED!! The bed needs to stay in alignment all the time, with no twist in the bed!

 

[Tozguy]

Use of machinist levels to determine if there is twist in the ways is inadequate. The only sure way to get reliable twist measurement is with the two collar test. It is important to do the test as specified and have a smaller section in between collars. Must not move cross slide or compound to cut both collars.

My lathe came with an accuracy check list done before it was crated up. It was crated very well with wood but wood works just from humidity changes. The most likely and normal result of the whole transportation and handling issue is some unwanted twist in the ways. Sanddan's progress in reducing twist result from ,008'' to ,0015'' shows that he was on the right track to reducing twist but only if the spindle alignment was assured first. But was the turning test done without moving the cross slide and compound between cuts on the collars?

 

[qualitymachinetools]

Hey Guys, some more useful posts here for sure.

Yes 4gsr is correct on this model of machine, you wan to level the bed first, without even taking a cut. And then adjust the headstock in to the bedways. On a machine where the headstock is mounted direct to the V Ways and no adjustment possible, you can adjust level to get it as close as possible.

But this machine has a bolt-type adjustment on the headstock and is meant to be aligned, that is part of setup. So:

1. Level the lathe, with a level on the cross slide going back and forth (Left to right) to get it close.

2. Adjust headstock to get the taper out

3. Then you can mess around with the tailstock and slight tweaks of the bed level, say you are dead on straight in the first 10" but out at the end of the bed at 40" the tailstock is off, you can tweak the level to get that perfect.

 

[4GSR]

Sandan did not indicate what type of level he used. But regardless, when you induce a "twist" as he did trying to get the lathe to cut straight, something is wrong! You can get a Starrett 99 level to do a good job of leveling IF your are careful with your measurements. I have both a 99 and 199 level, both work well with bed leveling on a small to medium size lathe. Anything bigger, I prefer to use a transit to tram in larger lathes.
But getting back to the two collar method. First of all, you must know the lathe is in perfect alignment when it left the factory before a two collar method can be used. Apparently in Sandan's case, this alignment was disturbed AFTER it left the factory and is indicated in his opening statement. He did try to take the twist out of the bed and obviously he was inducing a twist into the bed trying to eliminate the taper in the test bar he turned. This is where you stop and re-analyze what's going on. He did that. He ask us for advice. Lots of good things have been mentioned.

 

[stupoty]

an indicator that follows the path of the cutter will ALLWAYS show zero movement
even if the headstock and the cross slide are both several degrees off
measurements not taken on the centerline are useless

If the cross slide was a little of and the test indicator was in a different place to the cutting tool I would have thought that some variation might show?

Although I see what your saying so long at the dti and tool were in the same position relative to the cross slide it should be identical (or all zero)

Stuart