View Full Version : Holding .001 with a Collet Stop?

12-29-2008, 12:22 AM
I'm working on a 40 pc run of parts needing to hold +/- .001, so I got the idea of using a collet stop to accurately set the part for each step. But like so many of my ideas .... it ain't working! :D

It's a cheap ENCO collet stop which consists of noting more than a threaded rod, so I stepped that up with a sleeve over it for a wide flat surface to set the part against. I'm also making sure there's no burrs on the part seating area and giving the pocket a shot of air before setting the part in. I'm using a production closer and tapping with a brass hammer as I put tension on the collet. If there's something else to do seat a part accurately I haven't thought of it yet ....

The particular's are 7/8" dia Stainless 303 at 1.345" long. I'm only holding a 1/2" in the collet because I'll eventually need to turn .820 of the length down to .625. 1/2" Brazed carbide bit under a mister with suds in it on an 11" Logan ..... and I'm getting a beautiful finish. Just not always where I want it. :rolleyes:

The interface between the ENCO threaded rod and the sleeve I made might be under suspicion. The threaded rod is 3/8" with a roughly beveled end and the whole thing isn't exactly centered. I put the hole in my sleeve with a 1/2" drill bit so we have a rough bevel against the cone the drill left. I dunno.

For all I know it's not possible to reliably hold .001 with a collet stop. So here I am. Hat in hand once again. Any advice?


Paul Alciatore
12-29-2008, 12:57 AM
The thing about collet stops is that most of them screw into the back of the collet and position the part in reference to the COLLET, not to the lathe (spindle/headstock). When the collet is tightened, it is being drawn back into the taper in the adapter/closer. The position it stops at will depend on the diameter of the part being held. Any deviation in the diameter of the part will show up as a deviation in the axial location of that part. Even worse, the deviation is amplified: that is, a 0.001" difference in the part's diameter will be increased to several thousanths of difference in the position due to the taper's angle. So, you loose big time. Even if the parts are the same diameter, any difference in the tightening torque on the collet will still translate to a positional difference. Collets are great at centering but poor at axial location.

If you want good axial positional accuracy, the collet stop would need to be mounted to the spindle itself or a fixed part of the collet closer, not the collet. Otherwise, to hold 0.001" accuracy, you need to zero your position for each part you mount. CNC would make this easy. Mount part, touch, zero, go. I am sure there must be other solutions/workarounds for this problem, but I am not experienced enough to know them all.

One thing I have done with part runs like this is to use a three jaw with zero set centering ability. Once is is centered on the first part, the rest will generally center to +/- 0.001". Now I position the carriage at a reference point and use the tool tip to position the part while I tighten the three jaw. Oh, I have a QC tool holder which also has good positional accuracy when tools are changed. Besides, it is really the tool more than any part of the lathe, that the part needs to be positioned in reference to.

12-29-2008, 01:13 AM
Another issue is wether you are using a handwheel colllet closer or a lever type. Lever types are more accurate as they are repeatable in drawbar pull. With a handwheel you can never guarantee it will be the same.

You really need a dead length collet. Hardinge does make a dead length collet set up for 5C.

Got a surface grinder? You could grind them to length.

12-29-2008, 01:34 AM
Mount part, touch, zero, go.

It's that "touch" part I've been trying to avoid Paul. :o It's hard thing for a FNG to learn, but if I have to get better at it ... I will.

And Thank You for the explanation on collet axial challenges. +/- .003 seems to be completely doable but if .001 is beatin' a dead horse it's time to be moving on.

Lever type on the collet closer Macona and since I'm using 5C, what's a dead length collet? It did occur to me that I could cut a 1/2" depth into an emergency collet, and if I do I'm gonna order 2 so I'll have one for the next time.


Frank Ford
12-29-2008, 03:12 AM
The dead length collet is a fairly sophisticated rig - worth a Google search to get more info. It will hold close to absolute length with a reasonable difference in diameter.


Mark Hockett
12-29-2008, 04:26 AM
If money is no object here is a very cool collet nose that holds to collet stationary and pushes the nose forward to clamp.

noah katz
12-29-2008, 04:34 AM
Very educational post, Paul, thanks.

"CNC would make this easy. Mount part, touch, zero, go."

Maybe a good excuse to get a DRO?

12-29-2008, 04:51 AM
DRO or CNC is not going to help you unless you can chuck the material consistently.

With a deal length collet the part of the collet that holds the work rests against the face of the collet closer. The draw tube pulls the collet body back against the collet chuck squeezing the inner portion of the collet. Since it is resting against the face of the chuck there is no lateral movement of the actual workholding portion of the collet.

I see these NOS from hardinge quite often on ebay. They are sold in different parts. A master body that attaches to the drawtube and interchangeable jaw sections.

12-29-2008, 05:20 AM
If you play it right, a DRO could certainly help. Just get it close, then as Noah said, touch-off on the end (or better yet, face it) and set that as "zero". Work all your features from that established zero and go...

Without a DRO, I would do similarly. Set each piece in place using the existing stop, which should get you within a few thou anyway. Then just face each piece to your established "zero" (probably with a DI on the bed?). Than again, just work from that zero once you get there. You could even do the same thing with a hard stop set and using the compound for Z. Just gotta work within your tools...

Mark McGrath
12-29-2008, 05:34 AM
I would machine a set of soft jaws to hold the part.Done correctly you will achieve the tolerance easily.

12-29-2008, 06:03 AM
Good point. That's a prefect solution that should be dead on every time. And with that many to do, the soft jaw overhead is offset, particularly if this is a second op or something that can't be faced after insertion. But facing after setting is generally how I pick up the zero...

12-29-2008, 06:58 AM
Soft jaws are the way to go. They have been around since for-ever and were used when collets might not have been available for what-ever reason. Further, they can extend out to diameters that are impossible with normal collets.

They can handle ANY size within their limits and DON'T have the size restrictions of R8 and C5 and similar collets. And its a true "one size fits all" "gripper" that is as true and concentric to very high orders of accuracy. And you can counter-bore it for use as a stop. But for a good accurate "stop" a good quality chuck with no wear in the jaw "ways" is required.

BUT to ensure that the soft jaws do their job with optimum accuracy, the machining should be done with a bit of "right-sized" material gripped at the back of the jaws so constrain any unwanted movement during "forming" that will detract from the finished job.

The bored hole in the soft jaws should be a neat sliding fit on the job to be machined. The jaws need to be (re?) machined each time they are mounted.

Machining one end of a job with or without soft jaws is up to the machinist. But if the job is the be "end-for-ended" and held in soft jaws the concentricity between the two ends tuned separately will be very accurate indeed as will the depth setting in the counter-bore (if needed/used).

I am surprised that many more don't use soft jaws. It has only been mentioned very occasionally until very recently.

12-29-2008, 07:15 AM
I had thought soft jaws as well. But we dont know if he even has a chuck with 2 piece jaws.

12-29-2008, 07:28 AM
It would have to have been cheaper than the collet closer.

Given that it is for a "paying/paid" job, it might be a question of not what it costs to buy, but rather what it might cost not to buy it.

Anyone who is stuck with the limitations of R8 and C5 collets (ER are OK) will soon appreciate having a soft-collet chuck (but keep it for soft jaw work). But beyond the limits of any collet the "soft-jaw" option for extent to the limits of the chuck as well as its ability to grip very accurately both on the outside as well as inside diameters makes them hard to beat. They are superb for "disk" work as well.

Perhaps its too "old school" and not "hi-tech" enough like collets.

Yeah - well.

12-29-2008, 08:31 AM
You guys are as always ... amazing. I DO have some soft jaws. I DON'T know if there's enough of them left to run a 7/8" part, :eek: , but I'll be finding out as soon as I can. :D

I did do some work on my "touch" and went 2 fer 2, but it's very time consuming. Just plugging in a part would be a lot faster.

Thanx again.


12-29-2008, 09:22 AM
I used a collet stop that was just an adaptor screwed into the back of the collet with a chunk of a bolt threaded into the center of the it. As long as you use care in tightening the collet you should get decent repeatablilty. Learn to use the tools you have at hand, their limitations and their features. The nice accurate pieces we are used to having were made on less accurate machines first. learn how to make the tools you have in hand do the job that needs done, and later you can have to tools that "do it right" when you can afford them. If the OD of the parts is consistant you should be ablr to bring the collet to the same point in the spindle and get good results. Get a piece or 2 of scrap and practice getting the feel of things and then give it a go.

J Tiers
12-29-2008, 09:48 AM
An expanding stop could be made to fit the drawtube. The drawtube is referenced to the back of the spindle, fairly directly, and might be sufficient if two-piece jaws are not available.

The two piece soft jaw chucks tend to have very large and clumsy jaws.... so many collet type jobs might not be a 'good fit" rather literally.

12-29-2008, 10:13 AM
Great ideas floating around here. I like the bit about loading the part, facing it off and continuing without letting go of the part. As soon as the part is removed from a setup it becomes difficult to pick back up on zero for either length or T.I.R.

I am probably repeating a bit of another post here but, at a place I once worked at, we used turrets on our lathes. One of the turret positions had a metal bar as a stop. Pull the saddle to a work stop on the lathe, pull the part against the turret stop and hold against the stop hard, close the collet while still pulling the part against the stop, check stop to part contact again (feeler, shim, light etc), continue machining.

I dont know your exact setup but I have a bar that I put down the spindle of my lathe. I put pressure on the bar, pushing the part against my stop. When I close the collet I try very hard to keep the pressure the same. This is also important. I ran about 20 parts wrong by mistakenly thinking that once the part was against the stop and the collet was snug, tightening the collet a bit more to keep the part from flying out would be beneficial. This extra tightening pulled the collet and the part back into the spindle more helping me machine a bad part. Never forgot that one.

The key to this is as others here have already said. The parts change in diameter and keeps the collet stop from repeating beyond a given amount. Find another reference surface to touch off of.


Mark McGrath
12-29-2008, 11:54 AM
Tiffie,got to disagree with you on this.
"BUT to ensure that the soft jaws do their job with optimum accuracy, the machining should be done with a bit of "right-sized" material gripped at the back of the jaws"

You need to load the jaw to take all the play out and this can only be done with a boring ring or similar setup.Gripping material at the back of the jaws only guarantees getting the play out the scroll.You also need to get it out the slides.
Bore soft jaws with a ring and get concentricity,accuracy within 0.0001"`s and repeatability.Once you get accustomed to working with soft jaws and build up a stack of machining aids to speed up boring you will use them most of the time without even thinking about it.

Mark McGrath
12-29-2008, 11:59 AM
This job would be easy in a capstan/turret lathe if you had one.Feed bar out enough to make one part.Turn OD off turret.Face length with tool in front toolpost on cross slide,part off with rear toolpost in cross slide.
Note.Front and rear cross slide tools are set up the correct distance apart to suit the length of the job.The saddle is never moved.

12-29-2008, 12:34 PM
Any reason why you can't just mill them to length? Just cuz the part is round doesnt mean they have to be faced in the lathe.

12-29-2008, 01:00 PM
A couple of thoughts (though I do like softjaws if they work):

First, do the touch off on the face of the collet right where it meets the work. The collet is moving but the relationship of the internal collet stop to that collet face is fixed.

Second, having done that, you next need to know how far from that touch to go to get your desired length. I made up a quick and dirty compound DRO for my lathe since I don't have a full sized DRO. All it took was a cheap set of digital calipers and about 30 minutes of fooling around:


It just clamps in place on the lathe with the compound at right angles to the cross slide. Now you've got a way to do those lengths precisely that actually goes very fast. It would be nothing to do 40 parts.



12-29-2008, 03:15 PM
Hardinge dead length collets are the way we do most of our length-critical work.

Another way is to set up a through spindle stop rod mounted in bearings to rotate with the spindle. The stop rod mounting is secured rigidly to the back of the lathe. We bought this setup from Dunham company for our CNC lathes.

For second op dead length work on shouldered parts Hardinge has a cap-like piece that fits over the spindle front to seat the parts against. This is by far the simplest method, and could be easily made to fit most any lathe. But, it only works on parts with a shoulder to seat against. Hard to describe exactly, maybe somebody has a Hardinge spindle tooling catalog showing it.

Mark Hockett
12-29-2008, 04:29 PM
An expanding stop could be made to fit the draw tube. The draw tube is referenced to the back of the spindle, fairly directly, and might be sufficient if two-piece jaws are not available.

The draw tube is connected to the collet and moves with the collet so I don't think this will work. On 5C collets the draw tube screws on the OD of the collet and a collet stop screws in to the ID so they are both moving the same.
If the is any variation in the stock diameter it will cause the collet to close at different depths.

Spin Doctor
12-29-2008, 06:43 PM
If the parts already have one turned diameter then you have a shoulder to locate gainst. That is when this type of stop comes in handy. They are discussed in Hardinges Workholding Manual. Basically they set against the spindle nose and when the collet draws in it draws the part up against the stop ring


I made my own as they are dead easy to make. One thing about the 5C is for every .001" difference in diameter you get .003" difference in length.

J Tiers
12-29-2008, 10:12 PM
The draw tube is connected to the collet and moves with the collet so I don't think this will work. On 5C collets the draw tube screws on the OD of the collet and a collet stop screws in to the ID so they are both moving the same.
If the is any variation in the stock diameter it will cause the collet to close at different depths.


That just HAS to be WRONG.

At least for a non-lever closer.....

The COLLET will move back, but by golly, if the drawtube is moving in any way other than spinning with the spindle, there is something VERY wrong.

It is pulling the collet back BECAUSE it is solidly connected, under pressure, to the back of the spindle.

Doesn't help pntrbl very much, because I see that in his second post he did say lever type.... I missed that, and went back searching because that was the only way your post made any sense.

But for any regular handwheel type, the drawtube will offer a good reference.

Mark Hockett
12-29-2008, 10:48 PM
pntrbl stated that he had a lever type collet closer. I have two lever type and one pneumatic type, which is basically a lever type with a pneumatic cylinder on the end, and the tubes go the full length of the spindle and are pulled from the end of the spindle either by the lever or piston action. There are bearings on the end that allow the tube to spin with the spindle while the lever or piston stays stationary.

If he was using a hand wheel type he could install an expanding stop in the spindle bore which would remain stationary regardless of collet movement.

A.K. Boomer
12-30-2008, 11:16 AM
I didnt read through everything so if I repeat its just that,
what iv found useful is if its a part without a bore then I focus on the center as there are "nipples" with just about any lathe piece, make sure the the receptor in the stop has "nipple relief" -- just a little detent in the center of the stop, Just as critical - even though your trying to avoid dead center - try to stay as close to center with the stop piece as possible, use the collet for alignment as thats what it was designed for, dont try to align the piece with the outer parameter of the piece and the stop as most stops deviate from center when tightened and any deviation in the piece can also be compounded on contact, it also increases surface area and therefore increases the chance of chip and or burr error --- so make appropriate stop inserts for particular jobs.

Now for the "touch" --- Even with the best of set-ups,deburring, and cleaning there is going to be deviation.
I would rather use a chucker lever and keep the machine running when changing the pieces out (if the pieces are small)
By slightly loading the piece with a finger at the appropriate RPM you will remove the high and low errors as the piece seeks an equilibrium between the two states of flux, when you feel like you have achieved this little window of opportunity (your finger should detect very little movement in the piece) then mildly and incrementally apply pressure to the clamping lever till the part starts spinning on its own, This practice is more than just what it sounds as its also a great way to "seat" each individual piece to the stop, it reduces chip and burr errors...
Its no guarantee as with lots of production pieces is still bound to net you a few "bad eggs" ---- but - its about the closest thing iv found to work this good and still hold .001" .

12-30-2008, 04:06 PM
Thanx again for all the responses. I got myself set up in some soft jaws and we're crankin' 'em out. I'm finding tenth's of parallelism issues, maybe as much as .0005, but who cares about that? :D

Boomer, what's a chucker lever? I googled it but all I got was some bozo describing a turret lathe with pics of tailstock equipped lathes. :confused:


12-30-2008, 04:44 PM
Lever operated collet closer...