What's new
What's new

Problem with runout

IndyMachinist

Plastic
Joined
Dec 20, 2011
Location
Indianapolis, IN
Turning a shaft roughly 8" long with 8 different diameters. I put the part between centers and check T.I.R. On the diameters closet to the centers I am .0002 T.I.R and on the other I am .0003 T.I.R. On the diameters in between I am as much as .004 T.I.R.. Arrrrgggh

Process, turning one side with jaws bored to stock material size. Then turning it around and locking up a diameter that is 2 inches long with a precision ground, hardened collet for that specific diameter.

I previously ran 450 parts that were great. Literally all I did was shut the machines off and came back to them a few days later later. No changes in tooling, speeds, feeds, nothing.

Diameters are great as for holding size and staying round.

Very confused.
 
Any precision shafts I do gets all roughed to about .030". Then for finishing I go back with a rough tool to within .005 and then a finishing tool to hold all the tenths diameters. Gotta take great care in properly dialing the part Stress Free before putting a center in, I often true up new centers with single point, gotta be sure the centers are true, not too much pressure from the tailstock, heat expansion, so on and on. Some materials are of course worse than others.
 
are you drilling center holes in each end?

IF not:

sounds like you have a chuck that's not running true to the spindle. if your first side chuck is establishing the diameter that you are chucking on for the second side, then your problem would most likely be on the second side chuck. make sure it's indicated in as close as you can get it. some collet chucks don't have adequate chip evacuation and as you cut towards your chuck the chips find their way in between the collet and the housing which will cause all kinds of fun. you could also have a broken bolt on the drawnut. if it was me i would take the collet chuck apart and clean/indicate everything in and see what happens. if you do all that and still don't find anything wrong you may have a sprung collet. that's a hard one to tell for certain. maybe check your chuck pressure also on the second side is not too high or not too low.
 
Yes, I am drilling centers in both ends. Have indicated dead nuts zero on both center drills to rule that out. Also we pulled the collet out and cleaned real well. We are now at worst about .0025 T.I.R but still not quite there as .002 is the runout tolerance. So I think that was part of the problem. Material is 1144 Fatigue Proof so it is not to bad to run. Using Tungaloy chip breakers and holding 30's on the micro.
 
few more things you can try: i'm assuming you checked the tools to make sure they are in good shape.
1. is to indicate the part in as it's clamped on the second side. you will need to have enough material that is machined from the first side hanging out of the collet to indicate on. indicate the part to zero not the chuck.
2. slow your center drill feedrate down. if your using a #16 center i would run it @ rpm of 1500 and feed around .006".
3. check to see if you have any face runout on the end of the part or any shoulder, especially the second side locating face on the part.
4. check face runout on the second side locate inside the collet chuck.
5. try a new collet.
6. don't swallow so much of the part in the collet. if you are chucking around 1/2 way down the part. try chuckin 3/4's of the way away. this will help reduce runout by clamping closer to the first side center drill.
7. most likely not an option but, you could face and center the part first then turn the part complete in one turn operation using a face driver and tailstock. if you had good runout before there shouldn't be too many reasons why it's bad now.
 
Since the problem is in the middle of the shaft, I would highly suspect bowing to be the issue. A dull tool will push it, as will a variety of other things like tool engagement angle (pressure angle). If you are using an insert with a lead angle, might want to try one with a nearly square face to keep the pressure angle more toward the chuck and not toward the part.
 
I was taught that the only way to achieve the best possible runout was to finish shafts etc between centres, which - given the headstock centre is true, the spindle bearings in good conditioon and tailstock centre is true, can only reliably result in concentric dia's etc etc etc.

To date, that teaching, has held true for me.

Collets etc are at best, only a close approximation.
 
I'd have to say you were very lucky to get 450 good ones without any problem.

Drilling both centers prior to a chucking operation is a mistake for super accurate work, unless you are actually going to turn between centers. The reason is that you have no idea if the centers are pointing at one another right down the spindle axis, because you will have disturbed the part after the first hole is drilled. Yet, you are relying on the part OD to be perfectly concentric with both centers, even in the first roughing operation, because you are chucking on it.

If the part is stiff enough to rough without centers to begin with, then rough both ends. If it is not stiff enough, then you can drill the first center, but you should do this without moving the part for the roughing. If the part creeps in the chuck, then immediately, the concentricity of the first center hole is compromised and will need to be checked before using it as a finishing reference.

So with a stiff part, set up to finish the first end, and drill the first center hole at that time.

Then, you need to flip the part and chuck the first finished end (dialing in perfectly on a surface near to the chuck jaws), near the end, and run one finished surface on a steady while you drill the second center hole. That is your only real hope of establishing 3 concentric circles without assuming anything about any part of the setup, unless you can dial a shoulder face from the first finishing operation, after turning the piece end for end. Then turn the final end to finish.

But perhaps buying and using a driving center would be far and away the best way for large volume production, as it is almost brain dead easy by comparison to fooling around with every piece as I outlined above.
 
Chiming in on this...

since I have lots of experience running shafts various ways. Therefore, lots of problems along the way as well. LOL

Process, turning one side with jaws bored to stock material size. Then turning it around and locking up a diameter that is 2 inches long with a precision ground, hardened collet for that specific diameter.

So assuming here...

You are turning this part on the first op to beyond the last shoulder from this end. Centering this end at conclusion of this op. (Good)

You are then chucking on first op turn in a good collet chuck and turning back to meet the first turn. Centering this end at conclusion of this op. (Also good)

Your runout issue is likely in the second op. Here are some of the things that can (and will at some point) go wrong with this setup.

1. Too much over hang, too low chucking pressure and/or the hard, smooth collet can cause/allow part to slightly move during the roughing process. This typically causes tail wag back in the chuck mostly. Check this by roughing the part, re-chucking and fininshing the part.

2. Slight collet face runout/collet interference with shoulder fillet radius can/will cause the part to tail wag back inside and/or outside the chuck. (Jaw lift can have the same effect.) The longer the first op of the part, the more prevalent this problem becomes. Referring back to the start of my #1, you cannot expect to always turn the short end first.

3. Collet I.D. runout causes the whole first op turn to runout. (No brainer)

Anything else I could think to note here would just be variances of these causes with the same effects. Chips in the collet chuck would create at least one of those causes/effects for example.

With any of these mentioned above, the runout will assuredly appear in the middle of the part and mostly where the two turns meet. The only time the runout is mostly near the ends in this type of shaft turning is when the centers are not machined correctly. Always put centers in after finish turning and dwell them a little to be sure they are true and round.

Material stresses are not IMO, the cause of this runout. I could only see this happening if you were chucking against a much larger diameter flange type shoulder that would then be roughed/finished on the other side on second op. Inherent stresses relieved there could then cause the dreaded tail wagging.

Hope this was helpful...

--Gary
 
I was taught that the only way to achieve the best possible runout was to finish shafts etc between centres, which - given the headstock centre is true, the spindle bearings in good conditioon and tailstock centre is true, can only reliably result in concentric dia's etc etc etc.

To date, that teaching, has held true for me.

Collets etc are at best, only a close approximation.

+1 one for this.

But check your tailstock pressure, I used to do some long thin parts in graphite and found the amount of bowing I'd get depended on the tailstock pressure more than anything else

Boris
 
bumps on your seat

Yes, I am drilling centers in both ends. Have indicated dead nuts zero on both center drills to rule that out. Also we pulled the collet out and cleaned real well. We are now at worst about .0025 T.I.R but still not quite there as .002 is the runout tolerance. So I think that was part of the problem. Material is 1144 Fatigue Proof so it is not to bad to run. Using Tungaloy chip breakers and holding 30's on the micro.

Get a stone and clean any imperfections off the collet bearing and tapered surface. If there was dirt and chips between the spindle and the collet there is probably indentations or displacement that occurred on your seats and bearing surfaces. Who used the compressed air?! Try a new collet since some collets can fatigue after much use and some "hardened and ground" collets can be suspect- Hardened and ground to what...?

I like to use center to center myself since I can recut the spindle center if I use soft material or even regrind on the machine. I like to make the drivers for any special jobs. If nothing has changed in your process except dirt in the spindle I would concentrate on the chance that you have developed bumps on your seating surfaces. Then check if you have increased tailstock pressure also (in an attempt to get better runout).
 
thoughts from an inexperienced amateur

Yes, I am drilling centers in both ends. Have indicated dead nuts zero on both center drills to rule that out. Also we pulled the collet out and cleaned real well. We are now at worst about .0025 T.I.R but still not quite there as .002 is the runout tolerance. So I think that was part of the problem. Material is 1144 Fatigue Proof so it is not to bad to run. Using Tungaloy chip breakers and holding 30's on the micro.

450 good parts, change nothing, then it starts going pear shape?

to me, provided you have cheched the obvious places a powder-sized piece of swarf could have lodged, when you leave a machine a while, there is ALWAYS something that changes, and that is , with ghe various heating up and cooling doqn that goes on, lubrication in the headstock and slides has plenty of opportunity to drip, settle, smear - whatever. These effects could perhaps effect TIR a thou or more on diameter. I there any effect pn surface finish visible between the gpod and later batches - that might provide a clue.
 
Are the drilled centers a print call out, or just allowed for support while machining? The way I would approach this in grinding would either be between dead centers (non rotational) Than inspect between center's. Or to drill a center in one end only, and hit every dia. in one chucking, while supported at the center drilled end, then cut off. Concentricity would than be checked in a roller type concentricity gauge. Could any of this apply to machining?
 
Darned good posts here;

I like "HuFlung's" post #9 in general.

If you were mostly using up the .002" TIR on the first 450 then almost anything including somebody's mood could make things change for the worse.

"StressProof", "FatigueProof" and "ETD 180" all can have some non-uniform grain flow and with longer shafts & multiple diameters often need pushed around after machining or grinding. That's another art to itself.

A bent spindle on a turning machine will cause this no matter if it's centered or chucked unless the center is a "bell type center". A bell center will be better in the shaft middle but still would have some runout at the headstock. (I haven't seen much of this in my travels, but it happens). You can prove a bent spindle or crooked center by bluing a regular center, rotating the part between centers & looking to see if the blue is gone at the bottom & top of the center hole. Machining the center or boring soft jaws will fix this.

Speed & balance can cause a part to move from its static true center and stay there while at speed. (I think the chance of this here are nill). Too much tailstock pressure will cause a similar deal by forcing the shaft into a complex shape "think "C" or "S" shape here", once upset from a true cylinder the shape will remain while rotating. The latter is proved by placing a light near the rotating shaft and observing the shadow on a shadow board or mirror some distance behind the shaft. If the shaft is in a complex shape you will see a hard shadow line and a softer "ghost line". The more the distance to the shadow board and the closer the light is to the shaft the greater the resolution. The ability of a shaft to resist this is determined by a length to diameter ratio times force (radial or axial) applied. It is easier than most people think to force the "whole machine" out of shape with a screw.

Working to .002" TIR with multi flips & ops is not all that easy even with good machinery & material.

Good luck

Matt
 
Is there any possibility somebody did something to the machine between the first 450 parts and these last? An accident with a forklift of something. A lot of times people won't admit for fear of consiquences.
 
problem with run out-shalfs

Is there any possibility somebody did something to the machine between the first 450 parts and these last? An accident with a forklift of something. A lot of times people won't admit for fear of consiquences.

at one time working with very high speed shalfs we were required to pre machine to .030 of finish then most INPORTANTLY- stress relive by heat treat all in an up right position when normalized quinched.
then machine and then grind. this way we could always hold to .0001 to .00015 TIR with out any problems. yes it takes longer but look at the total cost if bad. also all ran between centers and were also strighten after stress relived-heat treat. these were shafts that were running in the 30,000 rpm range for textiles.
 








 
Back
Top