Anyone ever "stretch" ER collets slightly beyond their capacity for use without problems?
i.e. 15MM collet (.5905) for 19/32" (.5937) tooling or even using a 13MM collet (.5118) in for 33/64" (.5156) tooling.
I know that they can be damaged if stretched too far, but does a few thousandths make a difference on most applications since they are open that far without being pushed at all?
Runout desired is .001-.003 max. if it matters.
I wouldn't try it on a smaller collet size - but they come in smaller increments to make up for it anyhow.
As a rule of thumb I try to avoid expanding ER collets more than 1% of the bore size For a 13mm collet that would be 0.13mm =3D 0.005"
I *guess* the collets should be capable of expanding as much as they can collapse. But don't quote me on that... besides, you run into all kinds of problems with tangent points and scrape marks - just ask Cliff & BB =20
What the ER collets absolutely don't like is clamping on the flutes of a drill with 15 degree or less helical twist - You may quote me on that one...!
Yes. You can "stretch" them a little. You'll find out just how far once you stick the collet in the cap. There isn't much room between the OD of the collet, and the ID of the cap (not on the ER-11's or
12's, at least).=20
You can get away with a few thousandths, but be careful how much you increase the OD of the taper. That can cause the collet to not fit properly in the sleeve, AND reduces the number of threads which the cap can grip.
Try not to do it unless it's late at night, and you can't run out to get one that will do the job properly. Do it too often, and you are going to wreck the collet and there goes your accuracy forever, if you are drilling holes. I think it creates chatter, if you are milling
Collets are cheap. Go ahead and stretch it; but...
You'd be better off going to the next larger size and compressing it almost to its minimum. ER collets are noticably MORE accurate when they're squeezed small than when they're stretched big. If runout is an issue, then stretching even a little bit is to be avoided.
Of course, I've thrown away ER collets that gave me runout numbers only a tenth as big as what you're asking about. If you're using ER's in the first place, you ought to be able to get less than .001 TIR no matter what you do to them. If you can't, then throw away the collet, the holder, the nut, and the last guy that used any of the above.
We have a customer that I provided a 17/32" collet for a 33/64" tool... He said it wasn't close enough tolerance and after about a week of trying to convince him, we settled on giving him a 13MM collet because the tooling slipped into the collet without actually stretching it... It was open far enough "stock" to receive it. However, it peaked my curiosity as to how far most people thought or had experience in stretching them...
The "pros" say the collets can close down .039" so I assumed they might be able to open up .005" or so... But didn't feel (and still don't feel) comfortable offering them as a standard option that way.
Runout is .0004 to .0005 on most of the ER collets I've seen for sale with high precision pieces being available at .0002 or less. But... I presumed that making a smaller collet work for a larger tool might somehow affect that because it is making the collet rely on a portion of the collet holder that is higher up the taper than normal, etc.
Oh well... Lots of variables. I guess it boils down to this:
My customer has a collet that works... I have an option for customers who don't understand the .039 rule (ER16+) and runout numbers are good enough for most basic drilling operations that it simply won't matter unless there is damage occuring to the tooling or collet / holder.
Collets are ground as if they were solid tapered bushings. That is, they either haven't been split till after they're ground, or they've been split almost through, with a small solid section left over to be removed after grinding. Either way, the result is the same. The collet's real, proper, precise size is the same as its ground ID when it was solid. A 10mm collet, for example, is exactly right ONLY for a 10mm tool. At any other diameter, it's distorted and sprung just a bit. Sometimes more than just a bit.
After splitting, a collet can be compressed to give reasonably good grip on things smaller than it's nominal ID. The result, if you're using a collet at just a little less than its nominal size, is series of parallel line contacts that are slightly shorter than the collet's overall length. Half of these lines start at the front of the collet and go almost to the back end. Half start at the back and go almost to the front. If the collet is squeezed more than a little, it grabs the tool with two circles of points - one circle at the front and one a the back. This can still give very good concentricity; but it'll slip much more easily than when holding a nominal size tool. It's also worth noting that the contact between the collet and it's holder work the same way. Shortened line contacts or front-and-back circles of point contacts. Traction is lost on BOTH the ID and OD of the collet. And, of course, as the collet is compressed, its outside diameters get smaller, so it fits deeper into the taper of its holder.
If you attempt to use a collet for something bigger than its nominal size, however, then you WILL have problems. First, a collet stretched oversize will need to move outward, away from the taper in the holder. This shortens and distorts the OD contact against the holder, and diminishes the extent to which the holder can make the collet (which is springy) retain its indended shape. The results can't be good. Second, the radius on the ground ID is now smaller than the radius on the tool. Anyone who's ever bored the jaws on a chuck knows what happens in that case. Instead of a single line of contact between the tool and the middle of each of the collet's "jaws", you'll get two lines of contact with each jaw, right at the jaw's edge. If you were holding something soft, and wanted the collet (or the chuck) to bite into what its holding, then that might be attractive. But when what's being held is tool hard, it tends to bite back. The edges of the collet segments themselves get compressed and rolled out of shape, with no results that are consistent or predictable enough to allow good concentricity. Also, instead of lines (or points) of contact that are equally spaced around the tool's OD, as when the collet segements are each touching the tool at only one place, you get pairs of contact points (or lines), and non-equal spacing. This is pretty much a guarantee that the tool will rock and squirm as it's tightened, and will find a final resting place that's not where it ought to be.
When a good collet maker says that a collet will hold diameters from, say,
9mm to 10mm, what he's actually selling you is a 10mm collet, with compression expected and allowed. It's NOT a 9mm collet intended for stretching. If, as you say, your customer had trouble (whole thousandths of trouble??!!) with runout on a 33/64 tool in a 17/32 collet, then the collet wasn't the problem. The real trouble was something like a bad holder, a bad machine setup, or just improper use, assembly, and tightening.
Bottom line, though: Collets can work smaller than nominal. They should never go bigger.
Incidentally, although the .0004" or .0005" runout numbers you mention are pretty standard, it's been my experience that almost any good manufacturer (Lyndex, Nikken, Ericson, etc.) will generally produce their collets to much tighter specs than that. I very, VERY seldom spend the money for the "premium" brands or models. You can buy a whole set of reasonbaly priced ER collets, and maybe, sometimes, find one that can't hold .0002" TIR or better, right out of the box, every single time. (Assuming proper use and cleanliness, of course.) If I happen to find that one of the collets in the set is a lemon, I just toss that one and replace it with another from the same source. Odds are that the replacement will be just fine, and I end up with a whole set of "premium" collets without spending premium dollars. And I know plenty of other people who do the same thing.
I use ER collets on my drill grinders, and it's pretty much expected that a ground drill will have point runout and lip height values that are maybe .0002 or .0003" TIR, all the time, every time. And that's AFTER you've combined the error in the workhead spindle, the error in the ER holder, the error in the collet, the error in the drill's straightness and roundness, the errors in the machine, the heat and pressure from grinding, and everything else you can think of. The collet itself is typically a very, VERY small part of that total .0002" or .0003".
It's good that your customer is finally happy; but I'd be real surprised if the collet actually had anything to do with the source, or the solution, to the problems he complained about.
I agree... What I meant by "simply" is that runout is not an issue on this particular job. He simply didn't like that we were closing a collet within it's range. He thought the close-down was too much (much less than the .039" they specify possible). So... After a week of showing him competitors' sites, suppliers' sites and various charts proving he could use the supplied collet, I simply gave him what he wanted - which was a collet that needed to stretch .004 or so to meet his tool. Actually, the collet was already open that far but was rated for .5118 whereas his tool was actually .5156.
Bottom line... I asked about stretching collets to see if it was a widely used practice. I'll stick to the intended sizes of collets unless a customer demands that I provide something else.
It's just like the customer who refused to take my suggestion of a DA collet chuck when drilling wood in a vertically "UP" position and insisted on a key-type chuck. After a month, he bought the DA collet chuck when his key-type froze up from dust and debris getting into the chuck just like I had warned him.
We got a chuck sale out of it, but it didn't make me feel good in any way, shape or form to sell him.
We simply turn jobs down now when people do stuff like that. I wouldn't turn down a job simply because of the .004" stretch situation, but it is a similar circumstance where the customer wanted one thing and we knew that was right but gave in.
We will probably never hear from the customer looking for the "right" collets, but at least we tried to educate him.
Actually, this is what Dan Murphy, Bryce, me, Beege, etc. tried to explain to Cliff but he either couldn't conceptualize it, or was trying to cover his original boo-boo with his normal M.O. of Cliffasions and misdirection, unsuccessfully I might add. LOL
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.