how do you measure the fit of a collet?

[rimcanyon]

How do you measure the fit of a collet? e.g. many collets I have seen are slightly sprung when new, so a ½" collet will easily allow a ½" rod to be inserted, sometimes with obvious clearance until the collet is tightened.

Another version of this question might be: How do you measure the grip of a collet as it is tightened, along the bore of the collet, to verify that it is gripping evenly?

 

[TGTool]

It would help to know what collet spec you're talking about since there are significant differences. ER series collet close down parallel, while others such as 5C and R8 do not. This can affect the clamping force, particularly if the piece in the collet is not exactly the same as the nominal collet size.

How do you measure grip? A torque wrench comes to mind as one good measure of grip or resistance to slipping.

 

[rimcanyon]

OK, ER series then. How do you measure the fit?

 

[rimcanyon]

Bill, you miss my point entirely. i.e., suppose I purchase a set of ER collets and a holder. How do I verify the fit of the collet in that holder? Runout of work held in the collet is one measurement, but it does not tell you if the collet's fit to the holder is good. If the fit is good, the grip of the collet on the work (assuming the work is perfectly cylindrical) should be the same along its length. If the fit is poor, the collet will grip the work more at one end.

 

[Bakafish]

You wouldn't be able to directly determine the culprit with this method, but a precision rod mounted in an ER collet and holder that had some error would presumably allow the rod to move out of axial alignment (the collet taper would conform to the holder, but that would create a loss of parallel of the inner bore that holds the rod) so the bore holding the rod would pinch front or rear and have some form of bell mouth. By using a trusted precision grade collet, you could throw shade at the holder for any errors. Having just wasted a ton of time rebuilding a small spindle around a substandard CN holder, I'd suggest paying extra for quality parts if you can and not worry about this sort of uncertainty.

 

[Bakafish]

Downside to that is that the Taiwanese-made ER (two grades offered) are actually rather good. Pick either 6 'tenths" average TIR or TWO "tenths" average TIR and go.

Because.. "super premium" genuine Rego-Fix goods, ex Europe, NEW, and not of uncertain provenance and possible exposure to an STD.. off ePrey... don't just cost ten percent more.

They can cost ten TIMES more.

That said, I am happiest as to both superior quality and value-for-money with my "MariTool" goods, ER and TG.

I got my Rego-Fix ER16 collet set as a part of a surplus 3R EDM drill head that was being auctioned for pennies, I couldn't have afforded them otherwise. My main spindle is also equipped with a Rego-Fix ER25 holder, I'm a big fan but yeah, they are not cheap. I have a good number of Trustco ER25 (high precision grade) collets which are sourced from Taiwan, and agree they seem to be really good for the money. Especially when Amazon Japan messed up on the pricing and I got about 8 of them for under $5 each.

 

[Paolo_MD]

On the top of Bakafish suggestions, I would cover the outside of the collet with sooth, insert a pin very close to the collet nominal size, and tighten the collet in the holder: observe where the sooth transfer to the holder. If you have an equivalent number of "spots" in the front and the back, the fit is fine. If you have more at one end, likely the angles of holder and collet don't match.

Paolo

 

[Bakafish]

On the top of Bakafish suggestions, I would cover the outside of the collet with sooth, insert a pin very close to the collet nominal size, and tighten the collet in the holder: observe where the sooth transfer to the holder. If you have an equivalent number of "spots" in the front and the back, the fit is fine. If you have more at one end, likely the angles of holder and collet don't match.

Paolo

I am not sure what "sooth" is, but two strips of 'plastigauge', one at the front and one at the back of the precision shaft, might show a differential clamping force that would indicate an issue.

 

[Bakafish]

Forsooth! Speaketh he of "soot"?

Otherwise:

noun
archaic
noun: sooth

1. truth.

And "true running" IS what we seek!

Ahhh, soot!

 

[Bakafish]

Not really. Plastigage is FAR too crude for the fits involved and you cannot get it OUT without messing it up and losing the information anyway.

Even harlot-scarlet (see "AOC") lipstick can work better.

Or flash copper plating.

I CAN .. put a ten MILLIONTHS (per division) DI onto such goods. Think uber-high-RPM integral ER snout on a liquid-cooled spindle.

But I do not.

Takes half the dam' DAY even to set it up reliably.
It's Old Skewl "mechanical" (Hamilton). Force exists.

You DO understand a basic noga or equivalant MOVES when you get down to the millionths?
A fully "triangulated" mount has to be implemented.
But only once every many, many years.. to vet 10EE spindle bearings.

Modern factory QC would use "non-contact" metrology.

I've honestly never used plastigauge before, just seen it used and thought it might show something if used differentially. Not saying you could get an actual value out of them, just trying to think of how I'd do it if I didn't have the tools (which I'm assuming is a constraint here.) Honestly the time you would have to take, and the instrumentation required to ensure accurate results would both quickly outway any cost savings of a cheap part compared to just buying 'known good' products.

As far as millionths, I've long accepted the metric system as my mental model (other than for temperature where C is mostly stupid.) I have several sub-micron indicators (Mitutoyo and Mahr), and use Nikken stands for when I need them to be stiff. Triangularization sounds like a better solution, but these stands are built to hold a micron and seem to do so in practice.

 

[Scottl]

Bill, you miss my point entirely. i.e., suppose I purchase a set of ER collets and a holder. How do I verify the fit of the collet in that holder? Runout of work held in the collet is one measurement, but it does not tell you if the collet's fit to the holder is good. If the fit is good, the grip of the collet on the work (assuming the work is perfectly cylindrical) should be the same along its length. If the fit is poor, the collet will grip the work more at one end.

Use blueing to check the fit?

 

[Bakafish]

It isn't like 'the usual' taper-fit.

ER MOVES. A lot! Price yah pay for the wide collapse range.

That's why 'soot' could work better.

Transparent ER holder and collet? No Joy.
By definition, they won't be the ones actually used!

I think the idea isn't to blue the taper itself (since the collet's flexibility will likely allow it to conform to the surface of the taper anyway) but to blue the tooling or internal bore of the collet, as any error in the collet or the taper itself should result in uneven front and rear clamping of that internal bore. Obviously perfect tapers and a bad bore or pin would result in the same exact pattern, so even if you saw something, it isn't conclusive in isolation. I thought plastigage may be able to show such a differential, but I defer to those who've actually worked with the stuff.

 

[Bakafish]

Very interesting! My mental model was that the inner bore just linearly collapses, but I can see now that couldn't really work over a range. I wonder how they machine that internal geometry? Is it just bored and then the cuts are made and the geometry just works, or is the bore machined later with some amount of preload? Or is it actually some more complex lobed profile? Fascinating to think about, would love to see an in depth 'how it's made' on them.

 

[Conrad Hoffman]

Dumb question, but I've never seen it specifically stated. For a 5C or my little 3AT collets, when clamping a nominal size pin, is the taper supposed to have full axial contact with the spindle?

 

[triumph406]

Even harlot-scarlet (see "AOC") lipstick can work better.

Easy there big guy, I like AOC, and if she were to check fit my tool, and leave a ring of harlot-scarlet during the fit check process, well, I'd never wash it off.

 

[triumph406]

Why are we not surprised?

No fear.

Whole assembly would FALL off, soon enough.

With the duct tape and superglue I keep in stock, easily repaired

 

[TGTool]

With the duct tape and superglue I keep in stock, easily repaired

You might keep the name of John Wayne Bobbitt's surgeon handy.

 

[triumph406]

You might keep the name of John Wayne Bobbitt's surgeon handy.

Already on speed dial, just in case.....

 

[FamilyTradition]

Stringline the stairs of the Acropolis. Not straight. Center rise.
Mark One Human eyeball distorts. Crowning make them LOOK straight. Oterwise they would APPEAR to sag.

Imperfect world. Ever' body has to cheat to make it APPEAR 'more nearly perfect'.

Same with working on old houses. My brother was helping me put in some replacement windows a few years ago and going mad trying to get it level and get the windows to open and close right.

Problem is, the house itself ain't level to the ground anymore, and window frames not completely square. Got to just slap the things in and split the differences best you can to make it look good, while still functioning as intended.

Remember, perfect is the enemy of good.

As you have touched up on in later posts in this thread, we can get all scientific with it and try to get things within the millionths - but even when you get to a few tenths, sometimes you don't want to breathe on it for fear of effing it up.

Again, as has been mentioned, very few instances where you need it to be perfect. Otherwise, good is good enough.

Therefore, I usually measure the fit by asking: Does it work for what I am doing and am I getting the desired results? If my runout tolerance is .001 and the collet runs .0005, I won't nitpick that - SEND IT! If my tools are all sharp and and properly supported, but still getting chatter, I figure a re-bore (on the soft collets) or replace is in order. Problem solved in a lot of cases.

 

[Conrad Hoffman]

Cannot. There's a line of enhanced pressure.

Each "petal" is ground to a curve of slightly lesser radius than the forcing taper.

Never noticed how long-running spring collets "lobe" their closers?

That makes perfect sense but I can't see anything but flat when I hold a straightedge up to the taper on our Hardinge 5C collets, even in bright light. If there's a radius it's darn small. I should blue it and see what hits what. OTOH, there are three definite ridges in our spindle that correspond to the slots in the collets. Probably should grind that thing, but I haven't worked up the courage. It can't be good for new collets to be running in an old non-round spindle.