Swiss Lathes - straight Parts

Check alignment of the bush and collet.

Check the tension of the bush when dry then when it runs the oil feed will be more effective!

If necessary re lap the bush.

make sure the bush is not bell mouthed.

use a fixed bush if you can.

depends on how long the part is and what diameter, can you use a centre?

tool height, too high and work tends to push off.

sure theres more but it's late and I want to watch the film!

Have fun

what are your workpeice dimensions. length, dia etc. more info please.

Is the part long? Could you use a travelling steady to support the component?

All good ideas, but if it's still not straight, cheat the program. that is if you have a .001" taper, then you start with your X(start) program your next line to Zx.xxxX(start+.001) Then your machine will trace straight... I call it faking out the machine.

Thank you for everybody's input and I'm sorry that I forgot to explain the situation better.

The parts I'm having problems with are between 4.0" to 14.0" long and ranging from a diameter of 0.1875" to 1.0". Most of the parts have more than one diameter.

In some cases there are extreme diameter changes on one part, from 0.75" dia. step down

to 0.20" dia. over a 8.0" length.

Is the machine cam operated or CNC? What is the runout you are getting?

Dear guest,

Your question has tricked everyone to looking at the machine as the source of the problem, and not residual stresses in the product which may be causing the departure from straightness.

the product is straight to you because the surface is put into compression by the drawing and straighteneing at the mill, as well as by straighteners before the grinder if stock is ground to you; as the surface layers are removed, by your machining process, the compressive stresses (think girdle)also are removed, allowing the remaining stresses with in the bar to deflect its unsupported mass, especially over the extended length... if the thermal condition/mechanical properties of your raw materials is specified to you, then you can't get a stress relief, which would normally help in this situation on regular carbon and alloy steel applications.

milo

That's a tough problem, if you were turning on a capstan I'd say use a roller box but I'm not sure that's an option on a swiss lathe.

Think most of the problems occur due to an ill adjusted or worn bush, then again on some of your parts, there is alot sticking out! Could you use a pnumatic steady? these can be used on cam or CNC machines if you know the aux codes (was it M12,M13 normally the free ones after collet open M11, collet close M10! I'm going back over 10 years now so I may be wrong!)

Thanks again for all your help. It surly helps me to looking at this situation from difference angles.

I have some more feedback for everybody's brainstorming.

The type of machine we are mainly using is a STAR SV-32J.

The bar stock diameter is centerless ground to a tolerance of +/- 0.0005" and is solution annealed.

We did use preheat treated material in some applications and experience better results, but this is not always a option.

Problem example: a part 8.0" long with a 0.317" diameter and a shoulder 0.50" long with a 0.62" diameter has a TIR of up to 0.030" measured on the 0.317" diameter.

The bar stock diameter is 0.75"

It looks like the part is getting bend during the lathe operation.

Camber of barstock, lets say is 1/16" max deviation in 10 feet. (per ASTM A108- tightest straightness for cold finished bar, yes, i know your ground bar is better)

The equation for this is Length of bar squared divided by 1600 equals deviation from straight. ; For a ten foot bar, Lets see 10 Times 10 diivided by 1600 is .0625", thats a sixteenth. Thats the max deviation; the TIR is double that or .1250"

Ok that works out. now lets do it for one foot- (Its not a simple divide by ten thing, its a geometric relationshiop described by the term above, I donot have the time to show the proof via the pythagorean theorem.) Lets see what it works out to for one foot:

For one foot, the calculation is 1 times 1 divided by 1600 or .000625" Max deviation; the TIR would be double that, or .00125"

For 8 inches, it would be ((8/12)^2 )/1600 or .0002777" Max deviation; times two is .0005554" TIR. THIS is what you should be getting! However, That is one fifty fourth of your actual experience of .030"

That is too far off for me to conjecture and provide remote diagnostics, but i would say that at this magnitude, and with the stock removal that you have described, I am even more convinced than ever that there is a residual stress phenomenon involved. We used to see it all the time milling the flats for butterfly plates for carburetor shafts. mill the flats and watch shaft bow.

The numbers tell me that this is unlikely to be a simple misadjustment/misalignment issue. 54 times theoretical is a whole bunch of error.

FYI, the TIR for a 5 foot bar would be .03125"-

The Star is a great machine, i respectfully submit that I doubt that your problem is machine. MY 3 minutes of calculations.

milo

what! dude, really. what!

Really.

Its a geometric relationship, not arithmetic.

For simple camber, the bar is described as an arc, with maximum departure from straightness at the midpoint between two centers. The centers are at the denominator for the straightness specified, in my example 10 feet .

The construction of two right triangles from the center of the circle of that arc is done by drawing a radius from the center of the circle to each of the two supporting centers gives us the hypotenuse of each triangle; The line from the center of the circle to the point of maximum departure at that point is also the radius; the distance between it and the theoretical flat is "h" or camber, in my example 1/16" for 10 feet).

Subtracting "h" from this center radius gives us the long leg of the both triangles. one half the distance between centers on the theoretical flat or straight line gives the final leg. you can now use pythagorean theorem to solve triangles for any length between centers for a bar of the same curvature (camber). determining the new "h" or max deviation for the new given length; because most of us use indicators to describe lack of straightness, our callouts ar doublethe actual deviation so, we need to double 'h" to get to TIR; multiply "h" by 2=TIR.

I handled straightness complaints for a cold finish bar mill that I later ran, and there were so many people that thought you just took the maximum deviation (1/16") and divided by the length of the bar (10 feet) and it would give you straightness per foot. It was amazing ! So much for the LEARNING vs TEACHING of geometry in ourschools. But it gave me a part time 'gig' teaching straightness, camber, and TIR every year at many customers and trade association national meetings.

Its a geometric relationship, not arithmetic.

milo" I sure hope my words are clear, its easier with a sketch."

Uhh, I get it but then again I really need to see it.

Just give me a CMM and a Mazak. Or hell for that part a manual and a good inspection lab. I'm all set.

cr3

How are you inspecting it? Straightness is tuff to inspect to if your inspection setup is wrong.

first, lay all the inspection reports side by side, are all the parts wrong by the same amount? Does the bend start all at the same location?

Next is it a CMM? if so, how are they finding center or establishing a datum?

lastly, is the tick I like to use (but you have to be sneaky, if the inspector finds out, your toast), send in the same part a second time and see if the result match, if they don't you have an inspection problem.

Good luck

I guess the fact that see improved results from heat treated bar shows that there is some residual stress even though the bar is solution annealed. This stress often reveals itself as soon as you remove the skin from the bar. The other possibility is that your feed rates are too high, this can cause the cutting tool to push the bar away. You have front & back tool-posts, could you have your cutting tool in the front post & a steady in the back post?

Hi

this looks very much like residual stress.

To get rid of this - if thermal treatment at sufficient temperature and time is not an option - you can stretch the rod (before you start machining) to 5 % plastic deformation.

This is done with the thick aluminum plates that are milled to result in lightweight airplane parts. Only with this treatment these plates can be cut and then milled.

Else the deformation on partial stress relief (as stated in another post above) will distort your part.

You can make a test if your just machined part has too high residual stress:

etch it down with acid or electrolytic, if the part is not deforming significantly then the residual stress is ok.

Have success

RHABE

Wow.

what a bunch of mess.

Your problem is not uncommon. You are going to have to evaluate your op sequencing as to where you can resolve this.

2 possibilities that I didn't see listed. 1 Leave material long and finish between centers.

You mentioned you have a ground stock to +/- .0005", and I believe treated. have the critical dimension ground to finish.

I won't ask you to elaborate about your inspection methods.

Problem example: a part 8.0" long with a 0.317" diameter and a shoulder 0.50" long with a 0.62" diameter has a TIR of up to 0.030" measured on the 0.317" diameter.

The bar stock diameter is 0.75"

I mean I can sit here and ball park stuff all day. Turn the .62 down and leave it heavy. finish the 317 dim and flip the part chucking on it and finish the .620 , it should now be concentric the .317.

I assume you are between centers.

Are you checking run out prior to finish pass?

It's late and I'm tired.

what kind of specs do you have? GD&T all around? If so elaborate.

If you are between centers what is your tailstock psi?

Yawn , yes son, I'll read you a story.

Gotta go, my full time job is calling.

Give us some feed back and let us know what is or isn't working for you.

Take care.

The star is a sliding headstock machine. so he has Z 1 on headstock and Another Z (Z2) on backworking. No tailstock. six axis is standard. Calling this a lathe is like calling a maserati "transportation."

The alignment on this machine (unless wrecked) is like jewelry.

54 X theoretical departure from straight is As you say, a bunch of mess.

Its likely the stresses in the material

milo