Pros and Cons of the ER collet system

Certainly a well thought out piece of kit. very impressed. Neil , How are the blanks held for slitting ? obviously the ones that you slit from both ends have to be machined first then slit . The ones that are only slit at the front are they slit whilst still in the raw state, ie a bar before they are machined. The shrink fit holders you made will certainly provide a massive grip,. especially for use in a router. Excellent. work.

Ps, What is the make of lathe you use and does it have live tooling.

cheers.

Chris, You say you left this forum 18 months ago because of NEGATIVE posts about myfords and myford owners from one of the moderators, hmmm,,, Now I wonder who that could have been ?. I have a very good idea so we will just call him Mr Hypocrisy for now. or Mr Bumptious.[ either or ] any one of the 2 would be spot on.

Ps Any resemblance to any person living or dead, is purely incidental.

Edited By Raymond Anderson on 18/09/2016 18:36:14

Oh I'm pleased that has come up on this thread.
I'm 99.9% sure that's me.

But for the record I stated that Myfords are over rated and are living on a name. A point of view I still hold. A Boxford for example is a far better lathe and can be bought for half the price.

Why am I glad it's been brought up on this thread ?

Well there is no difference me saying what I do about Myford's and someone else saying there are better systems than ER's.

If there is a difference then it's  what I have said is better is actually cheaper, whereas what is being bandied about, on the ER's, for better is actually unaffordable to most of us here.

Edited By John Stevenson on 18/09/2016 20:20:49

I just created a fairly accurate model of one of my Er32 MT3 milling chucks and ran FEA on it. If I had not decided to change some of the loads I would have been able to post a photo of the load distribution but it will have to wait for the study to run which will take at least 45 mins. It does predict a clamping distribution that matches the picture of the collet with blue on (which is good as it is an expensive piece of software).

Mark

Ok, here is the predicted stress distribution for a 20mm collet (er32). You can see the stress is concentrated in 3 "rings". I have added a plot of the displacement (at exaggerated scale) which is also interesting

Mark

Very impressive work, and an interesting result, Mark

... The Location of the main pressure ring, just behind the groove, in particular.

... Is this effectively the 'pivot' when the cutter is loaded laterally, I wonder ?

I am a little worried though, by the displacement of the segements at 9 o'clock.

MichaelG.

Raymond, I put a piece inside the collet that is about 0.01mm smaller than the bored collet. I just cut them slowly with a slitting saw on the cuter grinder, so not milled. Then debur them. I also make adapters from the ER40 to ER11 for holding 1mm drills etc. I also use them in the collet chuck when holding small parts, less than 6mm diameter. I wish that I had live tooling but it just has a 12 station VDI 50 turret, 6 outer and 6 inner.

Neil

The irregular displacement is a result of the displacement scale - the analysis was set to "automatic scale" . If I set it at scale 1 you would not see very much.

The maths going on in the background is quite complex (and a bit beyond me to be honest), the collet is subdivided into "elements" and this forms a "mesh" The shape of the mesh can be quite complex and often looks to be arbitrary. For this analysis I had it set to refine the mesh automatically so it refines it close to intricate features and leaves it coarse in plain areas with little in the way of features (slots small rads etc).

The load was set at 5 KN tension on the outside diameter of the chuck body and inside face of the closing nut to approximate the screw force from tightening the nut up. I chose alloy steels for the chuck and collet assy and carbide for the cutter shank. The whole assembly was provided with an "ideal" fixed anchor on the face of the MT3 shank (it is a requirement that something is fixed in simulation space).

I will attach an image of the whole thing in a moment and will try and load the word document report generated by SolidWorks.

Mark

Assy as promised....

i tried saving the report as a pdf but there does not seem to be any way to up-load it, sorry.

Mark

Mark - one critical challenge with setting up an FE analysis is setting up the loads and constraints. In this case if I understand correctly you have applied loads to the "outside diameter of the chuck body and inside face of the closing nut to approximate the screw force from tightening the nut up". The difficulty is that as the load is applied, the force is transmitted through whatever surfaces happen to touch the 2 conical bores. I'm not sure where the forces were applied on the main taper but the distortion although exaggerated here would affect where the forces were applied and also the distortion resulting. Same is true for the movement of the collet segments against the bar, as clearly they would be constrained by the (line?!!) contacts with the bar surface.

The FEA you get with mid-range CAD is remarkably powerful given that it comes free with the installation (SW, Inventor, F360 etc) but you probably need some serious armchair tool such as Ansys to allow for the factors here.

Murray

Murray, this is not the free version of the analysis you get bundled with basic SW, I pay for this add-on (and it is a painfull experiance).

The analysis was done as an assembly not discrete parts although it was done at "ideal" sizes - all diameters started out co-radial etc.

Mark

I've used the thermal FEA add-on for SW and I could understand the pain if I had to cough up for it. Mind you, Ansys would take the pain to another level.

I imagine it would need to know the modulus of the materials as well as the coefficient of friction between the 3 parts to model what happens. I expect you could set the nut and body to have very high moduli (ie very rigid) and the collet to have a realistic value for simplicity(?) given that it is essentially a spring. For a ground and hardened body and collet the coefficient of friction will be fairly low (0.2?).

Rather you than me though!

Murray, my training does not extend to that level of analysis software - you would need to be a simulation engineer for that. I chose to ignore the friction for this as it was just to get an insight - i could add the friction contact in the system but there did not seem much point as I would need to know about surface finish, material properties lubrication etc. I did set the materials for a reasonable approximation to the real stuff but again, it is hard to know exactly what the real thing is made from and what HT was done.

Mark

.

Mark,

Depending on the content and size of the report; I could convert it to jpeg quite easily.

eMail it to me & I will have a go.

MichaelG.

Here is Mark's FE Report, as a set of 14 images

**LINK**

http://www.model-engineer.co.uk/albums/member_album.asp?a=42201

MichaelG.

Thanks Michael G for putting up Mark's pdf. If I was almost "lost" at page 1 by the time I got to 14 I was completely lost !! This is the kind of tests that real engineers would understand and the type of tests that yield valuable data for those that understand them. Well done to Mark for taking the time and trouble to perform the FEA.

Mark, I notice the bulge in the collet, so that must be why there is only point contact [not even line contact ] and why the collet does not grip at all in the middle, only at the ends. and thus why the test bar was only showing contact at the ends. That may be why they are unsuited to High feed High performance milling. I have a stand alone FEA for Siemens Solidedge but I should NEVER have bought it because it is far too complex for me to understand, so in effect it was a shed load of dough [even with a good discount ] wasted . Ahh well we live and learn. and often at great expense. So Mark well done with the FEA. Thanks.

Neil That looks very much like a SAUTER VDI turret on the lathe is that the case?. Also what make is the lathe DMG Mori , Mazak, Doosan or some other. The ER 40 to ER 11 reducers are very neat , good thinking.

CHEERS.

Converting to jpg has inevitably reduced image quality, so; I have added a couple of details of the mesh.

.

MichaelG.

The lathe is a 19 year old Nakamura lathe ,71mm spindle bore, with their LucBei software on a Fanuc control. I also have made ER16 to ER11 for my cnc router at home. I really like the ER system, especially the ease at making special sized collets. On the smaller ones, I ruf them out and drill/bore /ream the holes and for the final finishing, do that between centres on the hole that was created to finish the outer front 30deg and rear 8 deg tapers. I do this on the collet sizes smaller than 4mm. I always slit them last if they are collets.

Neil

Edited By Neil Lickfold on 19/09/2016 20:24:02