Excentric collet

[Tim StevensParticipant @timstevens64731]

I have an ER25 12mm collet which runs excentric. The bore of the holder is true to about 0.005mm – pretty good, I guess. But the collet carries tools which run out by 0.03mm – measured on a clean shank of a new cutter.

My question is – am I likely to get better than this, and have you a recommendation for make or supplier?

Regards, Tim Stevens

PS: is that really how I should spell excentric?

Edited By Tim Stevens on 22/04/2016 14:59:24

[Tim StevensParticipant @timstevens64731]

How much is too much?

[Andy HoldawayParticipant @andyholdaway]

Tim, I've no idea about your collet, but I think it should be eccentric!

Andy

[Raymond AndersonParticipant @raymondanderson34407]

Collets are usually measured for run out at 50mm from the collet face. Normal accuracy ones are usually around 10 / 20 microns, higher accuracy ones have figures around 6 microns or better. So yes, you can get better, and the better you want the more expensive they are. Rohm, Regofix and Royal are all available in Standard accuracy and higher accuracy, as are many others.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Tim Stevens on 22/04/2016 14:58:10:

My question is – am I likely to get better than this, and have you a recommendation for make or supplier?

Regards, Tim Stevens

.

Tim,

As I mentioned on another recent thread:

Here, for Lottery winners, is the Regofix catalogue.

They invented what is now known as the ER collet system, and enshrined in DIN standards.

Enjoy !!

MichaelG.

[Thor 🇳🇴Participant @thor]

Hi Tim,

I have both ER-16 and ER-32 collets and collet chucks. A runout of 0.03mm is more than I get on my cheap Asian made collets, the ones with least runout that I have are Vertex brand. I agree with Raymond and MichaelG about their brand recommendations, they are outside my budget unfortunately.

Thor

[Tim StevensParticipant @timstevens64731]

My measurement was taken just outside the collet – I was concerned that the tool itself might be off-set as it was cutting intermittently. Clearly the collet was at least partly to blame.

It seems to me, though, that the most important dimension to check for accuracy when milling is 'at the cutting end of the tool' which of course can vary with the tool size and extension. It is quite likely that some collets can be checked as 'within a micron' at the bore, or at 50mm, and still hold the tool cock-eyed.

I shall now see what can be done in the shopping dept – I note though that no-one has suggested a supplier whose reputation is better than others … ?

Thanks for the info

Regards, Tim

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Tim Stevens on 22/04/2016 17:51:34:

I note though that no-one has suggested a supplier whose reputation is better than others … ?

.

Regofix

[Jon GibbsParticipant @jongibbs59756]

I don't know whether this helps but the internal surfaces of the closing nut might also be expected to have an effect on the run-out of something held in the collet.

Have you checked the internal taper on the inside of the closing nut for swarf and muck?

Some information here… **LINK**

HTH

Jon

[Swarf, Mostly!Participant @swarfmostly]

Hi there, Tim,

You don't say how you're fitting the collet into the chuck – the correct way is to 'click' the collet into the closing nut first, then engage the nut with the thread on the chuck a turn or two, insert the tool and tighten. (Having cleaned everything first, as others have said.)

If you insert the collet into the chuck while still separate from the closing nut and then screw on the closing nut and tighten, the nut will bear on the collet asymmetrically and throw it out of alignment. Well maybe 'throw' isn't the right word, perhaps I should have written 'bias'.

Best regards,

Swarf, Mostly!

[Raymond AndersonParticipant @raymondanderson34407]

Tim, Just been checking, and for your 12mmØ collet you should be seeing NO MORE than 20microns runout at 50mm from collet face [ the figures for a higher accuracy collet the 12 mmØ should be 8 microns or better.] The runout will vary at different Ø. Also, when collets are measured the gauge pin is slightly below the max holding Ø for the collet, so for 12 mm Ø the gauge pin would be around 11.995 /997 The more a collet collapses the worse the runout will be.

I have full sets of ER25 ,32 and 40 an they are a dolly mixture of Rohm, Kennametal, Regofix and some Chinese [I think ] and none have ever had any more than 20microns in fact, most are a good bit better than that and all are only standard accuracy.

The only ultra precise one I have is Royal ER 40 at 20mm Ø, in order to grip a highly accurate Carbide bar that I had made in order for me to fit a collet chuck to its back plate with the highest accuracy. That is the only time I have ever had need of a ultra precise collet. Forgive me if I sound like im teaching granny to suck eggs, that is not my intention, but you must ensure that the whole system is "clean " ie not even the slightest spec of dirt in the collet holder , collet, and nut. I always use a medium hard nylon brush to clean the collet slits every time I use them. Its so very easy for a tiny spec to stick in one of the slits then the accuracy goes haywire. Your runout at the bore of the holder is good, its when the axis of the spindle, and the axis of the collet are not parallel to each other [ thats the one when the runout increases the further out you go].

Cheers.

[norm nortonParticipant @normnorton75434]

Interesting subject. I was discussing ER collet errors today with someone at the Spalding Show. I note the comment above by Raymond " the more the collet collapses the worse the runout.."

I have a set of ER25 metric collets, unbranded so they must be from 'budget land', which work OK when the job fits the size of the collet before tightening. If I have to tighten more than 0.5mm on diameter then the runout is worse than gripping in the 3 jaw chuck. It makes a nonsense of the ER system 'adjustability'.

As a result I have purchased some ER25 imperial collets so that when I grip imperial bar material it is reasonably well centred.

Do the expensive/branded metric collets close down better and still offer reasonable concentricity? How much does one have to pay to get 20um at 50mm, with moderate closure onto the work?

Norm

Edited By norm norton on 23/04/2016 18:10:49

[Anonymous]

Posted by norm norton on 23/04/2016 18:09:52:

I have a set of ER25 metric collets, unbranded so they must be from 'budget land', which work OK when the job fits the size of the collet before tightening. If I have to tighten more than 0.5mm on diameter then the runout is worse than gripping in the 3 jaw chuck. It makes a nonsense of the ER system 'adjustability'.

That's because the ER collet system was designed to hold tooling. So, for instance a 5-6mm collet is intended to accurately hold a nominal 6mm shank tool. If you close it down to 5.5mm, or even 5mm, then you're on your own regarding runout.

Andrew

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Andrew Johnston on 23/04/2016 18:42:41:

Posted by norm norton on 23/04/2016 18:09:52:

I have a set of ER25 metric collets, unbranded so they must be from 'budget land', which work OK when the job fits the size of the collet before tightening. If I have to tighten more than 0.5mm on diameter then the runout is worse than gripping in the 3 jaw chuck. It makes a nonsense of the ER system 'adjustability'.

That's because the ER collet system was designed to hold tooling. So, for instance a 5-6mm collet is intended to accurately hold a nominal 6mm shank tool. If you close it down to 5.5mm, or even 5mm, then you're on your own regarding runout.

Andrew

.

Andrew,

I hesitate to challenge you, but: Have a look at the statements on p_3 of this PDF [document page 2]

… Are you saying that "over the entire clamping range" [quoted below] actually refers to the "range" of the system, rather than the "range" of each collet ?

… I don't know the answer; and, if you do, I would be grateful for clarification.

[quote]

Special Advantages

T.I.R. ≤ 0.0004" (10 μm) over the entire clamping range 

[/quote]

MichaelG.

 

Edited By Michael Gilligan on 23/04/2016 20:09:37

[Raymond AndersonParticipant @raymondanderson34407]

Hi Michael, Im with Andrew on the original use of ER collets being for tool holding. All the lathes ive seen in industry [ either in the flesh or on you tube ] If the work holding involves collets as opposed to power chucks then they all seem to use any type of collet other than ER ie, 3C, 5C B42 ect. The ER's only appear on the rotary turret where they are used for "live" tooling or static ie, drill bit. The design of the ER system certainly does not lend it's self to production workholding. I use them for workholding where possible but im only a hobbyist . There are certainly better collet systems for lathe workholding.

I think the quote from Regofix would mean across the entire system, rather than across a single collet range. I did a few random tests last night to check, and all the collets I tested [ 6 in total ] from 4 makers Rohm 2 Kennametal1 Regofix2 and I think a Chinese 1 and all showed an increase in runout if clamped down at the lower end of the range as opposed to at the upper end of the range. That of course is only my tests, and I am going to fire off an E-mail to Regofix to see what they say. It is certainly an interesting subject.

Once I hear back from Regofix I will post the answer on here.

Cheers..

[Neil LickfoldParticipant @neillickfold44316]

Tim, if you buy the high precision Regofix collets you will be very happy. Please also note, that the inner taper is required to be running true as well as the outer thread of the holder. I have best results when using the Regofix nuts with the Regofix collets. I only buy their high precision collets. Other brands actually make a higher precision collet from what Regofix offers, but at about 3 to 5 times the price depending on collet size. They are best used on the nominal sizes. For that reason I am not using any inch series cutters, 1/8, 3/16, 1/4 inch. Now that I trued the inner taper on my router , I have a runout equal to the error of the spindle so is less than 0.01mm even at 40mm from the collet face.

The ER collets are best when holding onto a work piece that fits the entire length of the collet engagement. In the mix of collets that came with my 3d router, the 5mm unkown brand is just as accurate as the 3,4,5,6,8mm RegoFix collets I brought. The unnamed brand 3mm , 4mm, 6mm,8mm run out up to 0.15mm. Just shocking.

Neil

[norm nortonParticipant @normnorton75434]

Posted by Andrew Johnston on 23/04/2016 18:42:41:

That's because the ER collet system was designed to hold tooling. So, for instance a 5-6mm collet is intended to accurately hold a nominal 6mm shank tool. If you close it down to 5.5mm, or even 5mm, then you're on your own regarding runout.

Andrew, I fully accept your comments that the ER system was designed for tooling; but that doesn't stop us being creative with our use of our available equipment.

What I don't understand is why a 6mm ER collet is advertised and labelled as 5-6mm, but many of us are finding it only holds accurately at 6mm. Is it because we have bought the cheaper ones? Is the TIR spec applicable over the range 5-6mm, or just at the 6mm nominal size?

I note Neil's comment above that a whole set of his metric collets are "just shocking". Do I now need to buy a full set of PGMS bar to test all my cheap collets, or do I throw them all in the bin and buy a Regofix set?

Norm

[Anonymous]

Posted by norm norton on 24/04/2016 09:58:08:

Andrew, I fully accept your comments that the ER system was designed for tooling; but that doesn't stop us being creative with our use of our available equipment.

Of course it doesn't, just that something used outside of it's design intent may not work quite as expected.

It will be interesting to see what Regofix say in reply to Raymond.

Andrew

[Raymond AndersonParticipant @raymondanderson34407]

Thats the E-mail sent to Regofix I also sent one to Royal [USA] and Rohm ,It will be interesting to see what the makers say.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Raymond Anderson on 24/04/2016 07:35:00:

Hi Michael, Im with Andrew on the original use of ER collets being for tool holding. All the lathes ive seen in industry [ either in the flesh or on you tube ] If the work holding involves collets as opposed to power chucks then they all seem to use any type of collet other than ER …

.

Raymond,

I understand and accept that … and, in fact, do not currently use ER for anything.

[which is actually the underlying reason for my interest in the system!]

I also understand that they are only intended for holding truly cylindrical shanks.

Perhaps I asked my question badly, but the point is: If Andrew's statement is correct, then [Contrary to what many would have us believe] the ER system would be more limiting than conventional collets !!

Perhaps another reference to the Regofix document is relevant:

See the blue box on p_4 of the PDF, regarding 'oversize' shamks

A very interesting subject.

MichaelG.

.

P.S. … I have a good selection of 'watchmakers' collets, but I do regret being persuaded [on this forum] to sell my Burnerd Multisize chuck and collets, instead of finding a lathe to take it. 

Edited By Michael Gilligan on 24/04/2016 11:21:02

[Raymond AndersonParticipant @raymondanderson34407]

Hi Norm, The TIR doesn't go totally haywire if the collet is clamped at the lower end of its range,it's just not as good as at it's nominal size. The same is true of every collet system even a true workholding system like any of the "C " style for example. Even though the "C " system has a very limited collapse range compared to ER, it is still at it's most accurate at a nominal size. Where the ER system really scores is in the toolholding dept, as most Carbide end mills have a shank Ø that equals the cutter Ø [ unlike HSS ] and when you add Intermediate shank sizes into the mix you would need a mountain of collets . but not so with the ER system. And yes, the better ones would have better TIR even when clamped at their lower end compared to cheaper ones. but still not as good as their nominal size.

I like ER collets for workholding [almost as good as soft jaws ] and I use them where size permits after all most folks here [me included] are hobbyists, were I a production turner then I wouldn't want to use ER for workholding. no way, no how.

[Raymond AndersonParticipant @raymondanderson34407]

Hi Michael, just noticed your post I will go have a look at the document. and yes It is an interesting discussion.

Cheers.

[Yngvar FParticipant @yngvarf]

When I first got ER collets I found they had terrible runout.

Took a while to figure out the colled should be "clicked" into the nut.

[Raymond AndersonParticipant @raymondanderson34407]

Hi Michael, Just read the document and although I was aware of the NO oversize shanks It is still the grey area as regards the part "over the entire clamping range " but hopefully the makers I mentioned will respond and there will no longer be a "grey area" Re your Burnerd Multi size chuck, although Iv'e never had or used one the brother has, and he was saying that it is a good system , very versatile and robust. if that is the same system as you had then there are square and hexagonal collets also. Pity you sold it, although we can all be wise after the fact.

Cheers.

[Ex contributorParticipant @mgnbuk]

Some time ago I made the mistake of buying a SOBA branded Myford nose mounting ER25 collet chuck from one of the Harrogate exhibitions. I left it too long after purchase to try the thing on the lathe, to find that it didn't fit, to send it back & this thread prompted me to ave another look at it yesterday. lt looks nicely finished, but would not screw fully onto the spindle nose. A couple of hours lapping with fine paste opened up the apparently tapered bore such that it now screws up to the spindle location shoulder without binding. Applying a dial guage to the collet taper shows 0.07mm runout – I guess that I need a toolpost grinder now to true up the collet taper in-situ.

Non of this has any bearing on the original query but, despite the runout, I mounted a ground 12mm pin in a used 12-11 collet (ex industry) using the SOBA supplied closing nut. Runout approx. 20mm from the collet measured 0.15mm. A new Chinese 12-11 collet was tried, with similar results. I had an ex-industry German Zuern closing nut to hand – a bit grubby, but a squirt of brake cleaner got it looking presentable. Using the same Chinese collet & same pin with the Zuern nut gave a runout of 0.075mm – almost identical to the collet bore runout on the chuck. So it would appear that the nut can have a large influence on the gripped tool runout, all other factors being equal. I didn't have anything accurately ground 11mm diameter to hand to see what effect closing the collet down to it's minimum diameter had on runout.

The SOBA nut does not, on the face of it, appear "badly" made, with the main difference being the German nut has a ground thread and collet closing face where the SOBA is just turned & blacked.

Interesting to see in the Regofix catalogue linked to earlier that the range includes 0.5mm increments as well as 1mm – I wonder if the more restricted operating range collets perform better for runout than the "normal" range when closed down ?

Nigel B

[Author]

Posts