Is 4.6 tenths per inch too much taper for a lathe?

[erikmannie]

So I just took three hours to do extremely careful turning and measuring.

I used calipers and two different brands of 1-2" micrometers to take a total of 14 measurements at each position A, B and C.

The dimensions written on the workpiece are the average value of my measurements for that position.

The taper is .0019" over 4.1" which is .00046" per inch (4.6 tenths per inch).

I know that my lathe is not level, but is this okay or do I have a problem?

 

[pdentrem]

Do this first.

http://manuals.chudov.com/Rollies-Dads-Method-of-Lathe-Alignment.pdf

Pierre

 

[erikmannie]

Do this first.

http://manuals.chudov.com/Rollies-Dads-Method-of-Lathe-Alignment.pdf

Pierre

Will do. Thanks.

 

[RJSakowski]

The taper is .0019" over 4.1" which is .00046" per inch (4.6 tenths per inch).

I know that my lathe is not level, but is this okay or do I have a problem?

It depends upon the work you are doing. A half a thou per inch is approaching the accuracy limit of most machining operations and would be an acceptable error for many short turning operations. If you intend to make longer shafts to precise diameters, it could be an issue. But then, deflection becomes an issue on longer shafts and your taper would help to reduce deflection issues. When supporting work with the tailstock, the tailstock can be adjusted to provide zero taper.

FWIW, the acceptance tolerance on my G0602 is .0003"/" and it measured at half that on QC inspection.

 

[benmychree]

Tool wear could account for a little of that, but continous turning of a long length should not be used as the measure of accuracy, a test bar should be relieved in the center, leaving a short collar at each end, you then take light cuts on each in succession, measure the results, and adjust leveling accordingly with leveling screws or shims under a leg at either end, as required.
With possible apologies to Rollie's dad, this method has better possibilities of accuracy, as there is no wobbling to deal with. The test bar need not have more than 6" sticking out of the chuck, perhaps depending on the size of the lathe, and should be no smaller than about 1" diameter.

 

[projectnut]

Here's a copy of an inspection sheet that came with my Sheldon MW-56-P lathe:

 

[RJSakowski]

Wobble using the RDM method is immaterial as the averaging used in the method cancels it out. The only requirement is that the test bar be round. Even differences in diameter can be accounted for.

I prefer to use the RDM method as it doesn't require making a new skim cut after every adjustment. Unlike the two collar method, both vertical and horizontal offsets can be measured. There are two measurements made at each position with a dial indicator or a test indicator as opposed to a single micrometer measurement so there is a larger probability of measurement error.

In the end though, the best test is one which reflects the actual use and that is where the two collar test excels. After making my adjustments using the RDM method, I will finish up with the two collar test to verify the work.

A modification of the RDM method can be used to adjust the tailstock and set tapers using the tailstock offset method. To set a taper, the test bar is mounted between centers and the bar is rotated, recording the minimum and maximum indicator readings at two positions a known distance apart. The readings at each position are averaged and the difference between the averages is divided by the distance between the two positions. With a DRO, this becomes almost trivial. In absence of a DRO, a gage block or 1-2-3 block can be used to set an accurate interval.

I also use the RDM method to determine the parallelness of my mill spindle to the column ways. The method is essentially the same as that used on the lathe and again, I can determine offsets in both the x-z and y-z planes.

 

[benmychree]

Wobble using the RDM method is immaterial as the averaging used in the method cancels it out. The only requirement is that the test bar be round. Even differences in diameter can be accounted for.

I prefer to use the RDM method as it doesn't require making a new skim cut after every adjustment. Unlike the two collar method, both vertical and horizontal offsets can be measured. There are two measurements made at each position with a dial indicator or a test indicator as opposed to a single micrometer measurement so there is a larger probability of measurement error.

In the end though, the best test is one which reflects the actual use and that is where the two collar test excels. After making my adjustments using the RDM method, I will finish up with the two collar test to verify the work.

A modification of the RDM method can be used to adjust the tailstock and set tapers using the tailstock offset method. To set a taper, the test bar is mounted between centers and the bar is rotated, recording the minimum and maximum indicator readings at two positions a known distance apart. The readings at each position are averaged and the difference between the averages is divided by the distance between the two positions. With a DRO, this becomes almost trivial. In absence of a DRO, a gage block or 1-2-3 block can be used to set an accurate interval.

I also use the RDM method to determine the parallelness of my mill spindle to the column ways. The method is essentially the same as that used on the lathe and again, I can determine offsets in both the x-z and y-z planes.

After test cuts are made on the two collar specimen, an indicator can be used during adjustments; best of both worlds?

 

[erikmannie]

I am going to have to read up on these methods.

 

[7milesup]

Me too Erik.