ER32 COLLET SETS

[Dan_BParticipant @dan_b]

Hi,

I would like to buy a new set of ER32 collets but i am struggling to find just what i'm looking for. I know this is quite a regular topic but i'm really looking for information about what is available on the market now.

I will initially be using them on a lathe but plan to expand my workshop with a milling machine in around 12 months time. I already own a Chinese set bought from eBay and have been quite unlucky in that the quality is terrible – some collets have as much as 0.13mm runout.

I use a bolt on chuck bought from arceurotrade, it has 0.01mm runout in the taper.

I'm inclined to go for a precision set because apart from price i can see no other downsides and i may be tempted to buy a better chuck or maybe a taper chuck option in the future.

I like to buy tools of good brand name but haven't found much on offer in the UK when searching for collets – i suspect most on the market are Chinese sets of varying quality. I accept that these can still be very good but it doesn't give me the same confidence as a good brand name.

These are the options i have looked at so far:

Cutwel 0.01mm unbranded set – Good price and accuracy given in spec. (£86)

Cutwel 0.005mm unbranded precision set – A lot more expensive and non branded (£334)

arceurotrade – Price is OK and decent retailer but no accuracy given in spec (£97.75)

Vertex Standard Set – £209 for a standard precision set

Vertex precision set – £252 – i think i would pay the extra for precision if i went with Vertex, are they a decent make though?

Zoro Indexa set – £145 – no details of accuracy in the spec. Who are Indexa, i cannot find information on the internet?

Gloster Tooling Precision set – £160 for a precision set – this company seem to have decent feedback around this forum. Another unbranded set though but good price.

And that's about all i could find that was of interest. D&J workholding have sets that would be of interest but don't advertise prices.

Are there any other decent retailers that i missed? I think Gloster tooling are favorite so far but I am tempted to the Vertex precision set.

Any other suggestions to consider would be gratefully received.

Thanks

[Dan_BParticipant @dan_b]

[Martin ConnellyParticipant @martinconnelly55370]

First question with ER collets is always are you fitting them properly by snapping them into the eccentric ring in the nut before fitting into the holder?

The second question is are you tightening them up sufficiently and not with a cheap hook wrench as is often supplied in a kit?

Third question is are they free of burrs and debris in the gaps, if not then deburr them.

ER nut wrench example better than a hook wrench

Ball bearing nut for better clamping also reported by some to improve runout. The more tab on this page also shows how to fit the collet in the nut, you may know this already, apologies if that is the case.

These are the main reasons collets have excessive runout.

Martin C

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[IanTParticipant @iant]

Hi Dan,

I guess I could just dive in and ask questions about how you are measuring your run-out and suggest possible ways to counter it. But to summarise, your chuck has 0.4 thou run out and your (worst) collet has just over 5 thou – but what is the best one like? Would it be worth just buying replacements for the really bad ones or simply sending the bad ones back and asking for replacements (or a refund)?

Can I also ask a [probably stupid] question? What accuracy do you actually need? What do you plan to use them for? If you managed to get a collet with (say) a one thou (or less) run out, would that be good enough for the work you do?

My ER collets are not perfect by any means but then I didn't spend big bucks on them either. Most of the time I can live with their deficiencies , mainly because I know they are not perfect and plan the work accordingly.

However, I am not a watchmaker..

Regards,

IanT

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[SandgrounderParticipant @sandgrounder]

I've just bought some from here, very good service and the ones I got were the 0.015mm runout series which will do fine for me.

https://www.shop-apt.co.uk/collet-sets.html

Edited By Sandgrounder on 15/06/2020 06:21:31

[not done it yetParticipant @notdoneityet]

Apparently, cutter life is much reduced as run-out increases – important for CNC production machines. How much that might affect our machines at far less loading and speeds, I don’t know.

There will be a trade-off between initial outlay and replacement cutters (possibly for a looong time). Those with cutter grinders may accept the extra time resharpening, so several factors to take into account, I suppose.

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[JasonBModerator @jasonb]

In the home workshop I expect hand feeding, approx feeds and varying chip loads, less than optimum cooling and swarf removal, etc will do as much if not more to reduce tool life as running a CNC machine with a slightly off collet where all other perameters as set to optimum.

Also tool in the home workshop is likely to be fitted and removed many times which would average out which cutting edge does the work and wears rather than in an industrial situation where tool stays in a holder for it's life and just the same flute will get the wear.

As with most things you get what you pay for and it's upto the individual to decide what they need and what they are willing to pay for it.

[Neil LickfoldParticipant @neillickfold44316]

My best collets are the regofix high precision ones for my ER32 collet chuck. The best nuts are the Regofix ones as well. As far as tightening goes, the regofix have a special coating on them, called the Hi Torque series. I use a bearing one at work as they wont spring the cost for a regofix brand one.

It is important that the thread of the chuck is concentric to the inner taper. The cheap nuts I bought are junk. Need to be set up and have the taper of the nut made concentric to the threads. Another job that is easier said than done.

I did get some er11 collets off ebay that were rated AA quality and said that they were within 0.008mm for tir and they are definitely better than that.

There was another company from Taiwan , making an ER system and the nut has a slot in it for a special spanner that allows the nut to be done up or undone and extract the collet. No spanner flats etc on the nut and no eccentric retention piece as well. The threads were ground and the nut taper ground as well.

They also work very well. Tried finding them to take a photo but are in the moving lost stuff at the moment.

[Dan_BParticipant @dan_b]

Thanks for all of the responses – i will try and go through the points raised:

First question with ER collets is always are you fitting them properly by snapping them into the eccentric ring in the nut before fitting into the holder? – Yes I am definitely snapping them into the nut. This is another issue actually, some collets slip in easy, others are really difficult.

The second question is are you tightening them up sufficiently and not with a cheap hook wrench as is often supplied in a kit? – I wouldn’t like to tighten them any more than I am and I do have a good spanner with multi points.

Third question is are they free of burrs and debris in the gaps, if not then deburr them. – I picked one bad one that I needed for a job, deburred and cleaned it with absolutely no improvement. 0.13 is quite a lot of improvement to find.

I guess I could just dive in and ask questions about how you are measuring your run-out and suggest possible ways to counter it. But to summarise, your chuck has 0.4 thou run out and your (worst) collet has just over 5 thou – but what is the best one like? Would it be worth just buying replacements for the really bad ones or simply sending the bad ones back and asking for replacements (or a refund)? I measure run-out with a 0.01mm Mitutoyo finger clock mounted off the bed way on a mag stand. I have had these collets since last year and the eBay listing made no mention of accuracy – no chance of return now. Some are OK but not many.

Can I also ask a [probably stupid] question? What accuracy do you actually need? What do you plan to use them for? If you managed to get a collet with (say) a one thou (or less) run out, would that be good enough for the work you do? – I only use them on my lathe at the moment but plan on getting a Mill next year. What I have now isn’t good enough, at 0.13mm an endmill doesn’t sound good when cutting and I might as well hold my work in a 3 jaw. My feeling is ‘buy cheap, buy twice’. I don’t want to make that thrice!

I had really been looking for advice on where I could buy a good quality set from a decent manufacturer for around £250. I hadn’t spotted that D&J sell the Fahrion collets and so far I think this is now my favourite option. I will contact them for a price. I also found engineering supplies sell Fahrion collets individually at reasonable price, it would cost £233 for a set of 18. I would have been interested in Rego fix but haven’t found a UK supplier.

Thanks for the advice given.

Dan

[JasonBModerator @jasonb]

Posted by Daniel Bird 1 on 15/06/2020 10:10:43:

First question with ER collets is always are you fitting them properly by snapping them into the eccentric ring in the nut before fitting into the holder? – Yes I am definitely snapping them into the nut. This is another issue actually, some collets slip in easy, others are really difficult.

One of the problems with very cheap "sets" is they can be made up of collets from mixed sources that may have failed test at various stages so one factory's collets may fit easier than anothers. Also the smaller the size the stiffer the collet is so you will find say a 16mm collet snaps in easier than a 2mm one.

[Tony Pratt 1Participant @tonypratt1]

You need to go for a named brand from an industrial supplier which quotes a TIR if you want to guarantee a decent product & of course you will pay a hefty premium. The offerings from Ebay, Banggood, etc etc. are a complete crap shoot, I have used both 'premium grade' & 'not premium grade' & the difference is immediately obvious in fit, form & function.

Tony

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[not done it yetParticipant @notdoneityet]

If you buy premium collets, do give them the respect they deserve.

I watched a recent youtube video where a short tapered piece was hopefully held in an ER collet, but came loose and spun. Likely the collet was toast because of that. Another was tightening on a short workpiece without a same sized piece behind it. That, too, would destroy the precision of an expensive collet. Possibly both those were cheap, or freebies from a Chinese supplier, so they were no worse than originally 🙂, so it did not matter too much.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Daniel Bird 1 on 14/06/2020 21:16:26:

I would like to buy a new set of ER32 collets … i'm really looking for information about what is available on the market

I use a bolt on chuck bought from arceurotrade, it has 0.01mm runout in the taper.

inclined to go for a precision set because apart from price i can see no other downsides…

I like to buy tools of good brand name but …

…

Any other suggestions to consider would be gratefully received.

Thanks

If money is no object, approach an Industrial supplier and buy top-end collets. As the best are about £40 each, all the sets listed by Daniel are "cheap", but that doesn't mean they're rubbish. Look closely at what a £40 collet delivers. They're costly because they're balanced to run at CNC/Carbide speeds, about 40,000 rpm. Is it worth buying one for an ArcEuro chuck in a GH600 lathe? No – waste of money!

If precision is essential, necessary to consider the whole tool-chain. A precision chuck is needed to get the best out of precision collets, and a precision machine is needed to get the best out of a precision chuck. The cost rises rapidly, is it justified?

As Model Engineering is a hobby, chaps are free to spend their dosh as they wish. Nothing wrong with filling a workshop with 'quality' tools and enjoying them. It's not what real engineers do though! To them 'quality' is a red-rag to a bull, because it too often means buying on hearsay and emotion rather than cold logic. Adverts, history and personal opinion are inferior to taking a balanced view. Real engineers satisfy requirements; choices aren't based on quality, rather they target 'Fit for Purpose' and 'Value for Money'. Buying by Brand comes unstuck for two reasons: it wastes money whenever cheaper would do the job and it fails when better is needed. Nuts and bolts are an example; DIY Store fasteners are OK for domestic purposes but a poor choice for weight bearing, car brakes, maritime and – god forbid – aircraft. And only a dingbat would buy aerospace fasteners to patch up an old garden gate! Best to engage brain and ask what tools are for!

Hobby lathes are what they are. Buying a GH600 means a choice has already been made. It's a general purpose lathe good for work in the 0.02mm (1 thou) region. With care the operator can do better, but it takes time and effort. For what I do, 0.02mm is almost always good enough. I have a lathe collet chuck and a budget ER32 collet set (about £120) bought from Warco. I've not bothered measuring collet run-out; instead I cut the chuck's register slightly loose so work can be centred more-or-less exactly by tapping the backplate with a mallet before nipping up the bolts. Works well, but precision gear would save lots of time if I did a lot of high accuracy collet work. I don't!

In practice, 75% of turning in my workshop is done with an inexpensive 3-jaw chuck (0.05mm run-out 50mm out from the jaws.) 4-jaw about 20% of the time because work can be re-centred accurately in it. Collets rarely when jobs benefit from quick resetting, and/or is transferred between lathe and mill. That's just me – a clockmaker would favour collets much more highly.

Bear in mind that the old guys did excellent work on worn or indifferent machines without calibrated dials & micrometers. Spring calipers, chalk, wobblers and fitting are effective down to about a thou, the disadvantage being the time and skill required. Bees knees tools save time rather than enable new marvels. There's a good chance an ordinary hobbyist doesn't need top-end tools unless making something special like one of Barrie Lever's high-performance engines. Wanting to own nice tools is a different matter, but be honest: is the money spent for engineering or emotional reasons!

Arc Euro are good at matching value for money tools to value for money machines. Buying from them (and other thoughtful suppliers) avoids the worst of ebay and filters out the tempting expensive stuff. Owners of ex-industrial machines might be justified in going up-market, but so much depends on how their machine is used. Pootling in sheds and my breed of light hobby engineering is a world apart from a professional 'time is money' jobbing workshop. An advantage of buying upmarket products is they should be reliably specificied; but eyes wide open – they won't be cheap. Always consider if money can be spent more effectively elsewhere.

Newcomers often jump into the deep-end by stripping down machines and measuring stuff like run-out. Both risky and confusing without a certain level of experience, not least because operator error causes much grief early on. I suggest finding problems by cutting metal at first. Put the machine through it's paces. If it performs 'well-enough', don't fret! Look for causes and ask for advice only if it misbehaves. At the moment Daniel proposes spending up to £332 replacing collets when it's possible his cheap set is good enough. My advice, get used to the lathe and it's accessories and decide what it's for before parting with cash.

Dave

[Anonymous]

Posted by SillyOldDuffer on 15/06/2020 12:21:07:

Bear in mind that the old guys did excellent work on worn or indifferent machines without calibrated dials & micrometers.

It would be interesting to know if anyone has actually measured an older model to see how accurate it is, as opposed to parts made to fit.

Andrew

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[JasonBModerator @jasonb]

Out of interest what sort of run out does a typical Clarkson holder give?

Andrew, I could measure my old Stuart 10V made many moons ago on a Unimat 3 with very little equipment, I doubt many of the sizes are that close to what was on the drawing but I do know it will work if I blow into the inlet, though only for a short while before I go blue prior to bursting.

[Dan_BParticipant @dan_b]

Some good points made and i will take on board what has been said.

My original set are not worth messing about with – 0.13mm is a big error to work with and really defeats most of the point of using a collet.

I could narrow down what i need a little more though:

• Ideally i would like collets that give me a total run-out of less than 0.03mm on my machine.
• I don't want to waste money buying collets that will disappoint me.

I do like arceurotrade as a trader and have so far been happy with the tools i have bought. If the consensus is that the set arc sell would tick my boxes then i will listen to the advice and go with them.

Dan

[JasonBModerator @jasonb]

Why not clock all the collets that you have and keep any good ones, then get a couple from the likes of ARC in most used sizes and see if they meet your expectations, if so get some more if not pay for a higher spec ( I do know ARC's spec but not at liberty to say) but by way of compensation will post this again which is good enough for my own personal use.

Also remember that all quoted figures are a MAX runout collets could be better and quoted figures are at nominal size. 0.002mm collets are also generally non collapsable so not suitable for work holding.

DTI mmoved both sides of CTR to ensure it is actually measuring the crest of the shank. ARC collet & bearing nut, ARC or Chronos chuck (not sure) and X3 Sieg

[old martParticipant @oldmart]

I have also had similar experiences with the cheap sets. One point to make is with a mill, rather than a lathe, it will normally be cutters held in the collets. So you do not need a full set, just the sizes which match the common cutter sizes. I have bought good quality collets from Cutwel and also have Vertex collets. If you are also going to use the collets with a lathe, then a full set would be worth getting, as the variations of workpiece size will cover the full range.

I used to have a Clarkson Autolock small size which runs from 6mm to 16mm, plus the imperial equivalents, and there were two nosepieces to choose from. It was well worn, but I could get 0.001" tir with the best nosepiece and 0.002"tir with the other. Since then, I have been lucky to have got hold of two Osborn Titanics in new condition to use with the threaded shank cutters. They both are better than 0.001" tir with all sizes of cutter.

There was a recent thread on collets and is was mentioned that careful deburring of the bad ones along the cut lines inside and out could improve some, so it is worth a try.

Edited By old mart on 15/06/2020 14:05:31

[Clive FosterParticipant @clivefoster55965]

Jason

Did some quick'n dirty measurements of end mill runout in my Posilock "Clarkson" end mill holder in the spindle of my ageing Bridgeport and compared the results with native collets.

TIR with the end mill inserted by hand was 2 thou / 0.05 mm measured with a lever type dial gauge for both 1/4" and 3/4" cutters. I imagine things would be a little better if done after a cut had been taken and the cutter auto-locked more securely into the collet.

For comparison an R8 collet gives around 1 thou / 0.02 mm TIR.

Measurements taken perhaps 3/16" – 4 mm from the collet. Basically as close as was easily practical without bashing the indicator.

According to my notes my spindle is significantly better than 0.5 thou TIR measured off the collet closing taper.

So probably pretty good. But the method was by no means inspection grade so I'd be unsurprised if a proper job showed approaching 50% error. Unlikely to be that bad as I'm usually pretty good at such things but doing it right, with proper error bars and component verification is at least a days work.

Clive

Clive

[Author]

Posts