Unwanted Taper ?

[Cogsy]

I wouldn't think chuck run out would be an issue on a turned piece unless it's an absolute junk chuck and the bore through the jaws is crooked. Now if a test bar is simply being chucked up and measured that's a different story, but the way I read the original post it was after taking cuts so 'normal' run out disappears.

 

[Danuzzo]

Thanks for the posts. Tried it again this morning and was repeating minimal taper. Actually, the ends are about the same; but, some spots away from the ends are about .00003" to .0004" off. what would be the explanation for this?

Now I am working on aligning the tailstock. Back to working on it.

 

[goldstar31]

I wouldn't think chuck run out would be an issue on a turned piece unless it's an absolute junk chuck and the bore through the jaws is crooked. Now if a test bar is simply being chucked up and measured that's a different story, but the way I read the original post it was after taking cuts so 'normal' run out disappears.

You may be right but the only way that parallelism can be achieved in a worn or mis-aligned lathes is with a boring bar- and used 'between centres'. Well, this is the way that I have machined to accommodate parallel round PGMS bars in say 'Quorn' Tool and cutter grinder. Professor Chaddock and George H Thomas in the MoModel Engineers Workshop Manual report. As a mere amateur, I have to agree.

A ga in, taking what is said to be 'a normal cut' if the lathe is worn or incorrectly aligned. Mathematically, my experience suggests that wear or mis-alignment can cause an error of THREE times the measured error/wear.
This, is Pi but it gets worse when the actual wear is about 6 inches from the spindle and or the chuck jaws.
This is why a lot of so called measurement is done now with laser alignment. Earlier it was done with a wire and clock gauges, running between centers and clocked using the carriage.

Clears throat a bit and would mention a failed machine tool factory-- and obviously worthless share certificates.
Actually there were 3 but - the rst of the shares in 'other factories' are still a success.

However, I am 'Connolly Man' but note that others may not share my views.

Digressing further, I have 'special equipment' and several of my lathes over the years have been mounted - without twist, using cheap studding, penny washers and cheap nuts. My surface plate was nothing more than a piece of plate glass and I scrapped with a worn out old file using engineers blue. Initially, I got 'near' using strips of cigarette rolling papers. Years before I made wartime models with used single bladed safety razor blades honed on the inside of glass tumblers.

My experiences- but my I found my late wife's ruler yesterday. It was a 6" wooden one and despite having a list of degrees working down from a Fellowship, that is what she still used from when she was- a school girl.
I have Dad's folding 2 foot brass rule. My son has still my slide rule. Such is Life.
I digressed- sorry

Normani

 

[awake]

Chuck run-out is immaterial once a piece is turned. At that point, the results depend on the alignment of the spindle to the bed, and vice versa.

Danuzzo, look up "Rollie's Dad's" method - much better than taking a continuous cut.

 

[goldstar31]

Thanks for the posts. Tried it again this morning and was repeating minimal taper. Actually, the ends are about the same; but, some spots away from the ends are about .00003" to .0004" off. what would be the explanation for this?

Any number of reasons, but the most obvious( to me) is 'Flex' from any numbers of reasons.

This, as you will find- is a very big and detailed subject.

 

[Danuzzo]

Got the tailstock lined up; but, still have not used it.

I tried turning the 1" mild steel about 2.5" long. No significant taper. I am pleased with the current result.

Thanks for all the input.

 

[goldstar31]

Got the tailstock lined up; but, still have not used it.

I tried turning the 1" mild steel about 2.5" long. No significant taper. I am pleased with the current result.

Thanks for all the input.

Respectfully, you should have a projection of 6" and as my good friend Frazer mentions, cutting two rings and then compare results.
FYI I had several Myfords over the years I stripped the headstocks off the lathes and had them 'Blancharded' to get a 'reference'. So far no one has mentioned this practice. I then scraped a piece of mild steel as long as the lathes and proceeded to scrape the number 1 shear which wears 6 inches from the chuck end. I then scraped and blued the saddle etc so close that little strips of cigarette papers were held.
When all was saisfactory, I could hold a Number 2 Morse. taper test bar which when clocked near the spindle end and then 6" from the free end only deflected when tested with a tenths of an inch dial gauge by less than half a thous.

This is where you should be going. Connoly talks about tests made to 'references' and never mentions these these 'itty bitty funny practices' in more than 500 pages of text.

If you have to remove 'metal' you scrape to a minimum of 25 'spots' in a square inch. This is what the old restorers normally worked to whilst blueing and scraping 'cycles'

 

[Tim1974]

what brand and type of Micrometer are you useing ? i with i could reliably get to .00005

 

[Danuzzo]

what brand and type of Micrometer are you useing ? i with i could reliably get to .00005

I have an old Brown & Sharpe that I have to estimate the 10ths, and I do take more than one reading. In addition, I also use a Harbor Freight digital that does go to tenths, and again, I must take more than one reading to make sure I am getting consistent results. I have never actually tested either mic to confirm they are accurate, as far as the actual measurement; but, they appear to be accurate at least as far as checking the difference of the diameter from one spot to another.

 

[goldstar31]

Check your decimals, they appear pretty dismal

 

[Danuzzo]

Check your decimals, they appear pretty dismal

Whoops. I can't measure in .00000 (5 digits past the decimal). So, anywhere I stated .00000, I mean't 10ths, as in .0000 (4 digits past the decimal). I wish I could get tolerances in the .00000 range .

 

[goldstar31]

Whoops. I can't measure in .00000 (5 digits past the decimal). So, anywhere I stated .00000, I mean't 10ths, as in .0000 (4 digits past the decimal). I wish I could get tolerances in the .00000 range .

Try Statistical Quality Control from McGraw-Hill and if it is no good you can always keep it to prop the broken leg on the chair!

 

[Danuzzo]

Still having fits over trying to figure out why I am getting a smaller diameter on the aluminum in the center of the piece. The two ends are pretty darn close. Now, i have variance of up to .00065" for a distance of about 1" around the center of the cut. Tried different tool angle on 2 different HSS bits , speeds, carbide insert tool. Appears to be in the same area all the time.

Help!

 

[goldstar31]

0.00065". Come on-- digitum extractum If you are still playing about with a cob end, and getting unexpected answers- heaven help you. Project out a decent 6 inches or more and let us know what the result is.

 

[TonyM]

Still having fits over trying to figure out why I am getting a smaller diameter on the aluminum in the center of the piece. The two ends are pretty darn close.

That suggests the headstock/testpiece is pointing up or down in relation to the bed. If this is the case a longer test piece would prove the theory. The low point would remain in the same position distanced from the chuck as this would be the point where the tool is on centre of the test piece. With a longer test piece the diameter would increase the further away from the chuck you get. The tool needs to be perfectly sharp, large diameter piece and very small cuts to minimise deflection.

 

[goldstar31]

Deleted

 

[TonyM]

I don't know if there is any point in progressing this but lowering the tool height should change the position of the small diameter. If you lower the tool and the small diameter moves away from the chuck the head is pointing down. If the lowpoint moves nearer the chuck the head is pointing up. It just sounds like an alignment problem. The head is bolted rigid and the tail has bowed either down under its own weight and that of the saddle and tailstock. Or up if the table is not flat. When installing a lathe you really need to level the bed and the spindle with a high resolution engineers level.
Without one of those remove the tailstock and wind the saddle as near the chuck as possible. Take a couple of pieces of thick felt (4-5mm) the size of the mounts and soak them in shellack. Place them in position where the feet of the lathe go and gently lower the lathe onto the pads leave to dry out. No need for bolting down and the resultant possibility of distortion. They used to be available on a roll for industrial equipment but I cant remember what they were called.

 

[goldstar31]

Deleted

 

[Badhippie]

What about wear in the ways has this been checked. If you are worried about a taper in the .0000 then you should be worried about wear in the ways. Another possibility is all the backlash taken up when the cut is being made

 

[Danuzzo]

No wear in the ways. Brand new.