Always use a machined in place mounting plate for chucks and collet chucks, it's the only way you can get perfect results, as any lathe user will know.

From memory the SL lathe has a M12 or M14 threaded spindle nose so have you made a backplate that screws onto that and then has a register to suit which Arc collet holder?

I can see why using collets is a good idea but for an SL to take ER25 a lot of mass and overhang is inevitable and is going to take away some of the benefits. Even so the run out you mention is more than I would expect but from your description its not clear what is causing the problem. You need to go back to basics and measure the runout in stages starting withe the spindle nose itself.

Ian P

Technically this shouldn't make a difference but I've found sometimes it does. Gently tap the cutter all the way around with a rubber hammer and reposition the spanner that does the nut up on the other side of where you originally did it up to do it up further. The only reason for this working can be that there are tiny differences in taper between the collet and the chuck and/or the nut (assuming everything is spotlessly clean and oiled). Most people's first thoughts would be "cheap collets" but a little gentle tapping and further tightening can seat the cutter exactly where it should be, showing that in fact there is no "fault" with the collet and for some reason the cutter has simply failed to find its perfect position first time.

Andy, you may find this interesting, my experience with ER16 collets on a Unimat 3 fitted with new ER16 spindle.

I started with buying a backplate and ER16 chuck from ARC, the backplate was rubbish, bored and threaded grossly out of true. I started to make my own but as it has to be done on the Unimat itself for concentricity I found that doing it from steel was really beyond the capability of the machine and I felt that aluminium wouldn't be robust enough. (Mind you now I've made a tangential toolholder it might be different.) After having a few tries at making a screw-on one piece chuck (because I don't have a topslide for the Unimat) that were way out of true, I decided on the new spindle approach. As you'll see in my previous post, the spindle taper is pretty good but I found a big difference between nuts – as I've somehow accumulated 5 of them I was able to select one that is probably acceptable.

I probably also need to make sure I'm tightening the nut enough.

He possibly or probably did. We have sold 100s of such backplates, and continue to do so.

Every batch is consistant, except for drilling of fixing holes in certain cases, but not proved to be a big issue. The final turning of registers is left to the machine operator, on his machine. It is difficult to say or prove clearly if the backplate is rubbish. All i know is this: Out of every 100, we get three to four complaints. The rest either seem to be usable for the user or if they have a complaint, it is something which they have failed to raise with us.

If the feeling of complaint is strong, we will just refund and move on, and leave it to the machine operator to get on to finding their own solution.

Ketan at ARC

The problem is likely to be that the nut is only done up 'firm finger tight'.

Even ER16 needs to be done up a good solid spanner tight. Also check the collet is clicked into the nut properly. The collet can't engage the nut properly once constrained by the holder so it needs to be fitted to the nut before inserting the collet into the holder..

Neil

I have pondered during a job ER collet variance in runout when tightening. What I found is not to easy to put in words, but here goes..

The nut thread obviously has some play ( very small on a decent chuck, but must still have some play/clearance!)  when fingered onto the collet body thread, and when tightened up against a collet's top taper, the two threads find a mating medium. I found this medium to vary measurably with each tightening. As a result, while the nut is being tightened, it will pull up against the body thread on the side that the tightening torque is applied – ie, you grab the spanner and pull on the nut, pulling it over in that applied force direction. Since the nut moves ever so lightly in that direction, it's outer collar taper impinges more on one side of the collet than the other, closing that side more initially. This moves the held cutter over a fraction ( 0.005 to 0.008mm in my case). Even when fully tightened, the cutter is not centred properly. This is repeatable, and the rate of applied spanner force easily varies the runout from 0.005, all the way to even 0.01mm.

So, I tried using two spanners, 180deg opposed, trying to apply the same force to each spanner, and what do you know! First attempt gave 0.002mm, second 0.0025, third 0.0017.

It seems the trick is to not pull the nut to one side when tightening, and I get good results every time with my method.

However, it is not easy to do mounted in the machine – My tests were with the collet chuck firmly secured on a vise with Al jaws. Since doing up collets normally requires a spanner holding the chuck body, and another fastening the nut, I find myself short of an arm..

Hope my description makes sense.

Any bright ideas as to a spanner/jig/mechanism to hold the chuck and fasten the nut with a double ended spanner?

edit – 'than', not 'then'….

Edited By Joseph Noci 1 on 09/12/2018 11:19:21

Joseph,

This is just to add to your observation…

Certain clockmakers/watchmakers who we serve as customers have the SIEG C0 which we used to sell. They do this work for a living, as their profession. They would totally overhaul the machine to meet their specific needs… outside the norm of general hobby use, as the accuracy is not designed for use by a watch/clockmaker.

These customers set up their machines for specific jobs.. permanently. They would tell me (they because I heard this from at least three different customers from their profession), that they would only be satisfied with the backplate and ER lathe collet chuck assembly, after they got the finest level of accuracy out of it. To do this, after they got the maximum level of accuracy as it stood by tightening the nut/collet assembly, they would proceed to measure the runout all the way around, marking the points of deviation on the backplate, and introducing bits of paper in such areas, 'as shims' in between the backplate and the chuck, to improve the accuracy further. this works well for them.

This is all provided the backplate is correctly threaded in the first place.

Ketan at ARC.

Edited By Ketan Swali on 09/12/2018 11:45:18

A smidgen of oil helps on the tapered surfaces too, as I think there must be a tendency for the collet to twist as it's tightened.

Martin.

On the subject of Collets,I bought a set years ago,which I use on my mill. The holder has a threaded shank so would be no use for through work,but for the moment that,s not an issue. What I would like to know is this :

How do I tell ER16 from ER25 ? There are 8 in the set from 4 to 16 mm and the holder is MT2 which I fitted with an MT3 sleeve. Have never used them in the lathe,only to hold cutters in the mill.

Just curious.

Hi, this is very true about some people who have worked in jobs that you may be afraid to question their qualifications. I used to work for a big corporate company that had British Standard status, it wasn't quite a blue chip company like 'R R' but it did have some complex machinery. At times I had to correct some other workers who were new to the company, who had all these higher qualifications than myself, but what they lacked was the experience that I had gained working on these machines and the process they were used for. In my current job I have met other workers some of them machinists, who know there job well, but don't have the full fitting skills that I learnt over the years, but on the other hand I have acquired experiences that some of these people have that I never did have before. We are all learning little snippets every day, whether we realise it or not, even if it is by the odd little mishap. I can always remember a saying on the back of a box of matches (can't remember the brand, but there were hundreds of different sayings on the back of their boxes) which was pointed out to me when I first started work, and I quote; "Experience is the best form of education" and I have found that to be true.

I'm wondering if the OP has enough clearance in the corner of his register, as even if the chuck is sitting a few microns off in any one place of a true backplate, it will throw the concentricity out at the collet.

Regards Nick.