ER32 Collet Chuck advise please.

Hi guys, I'm after a little bit of advise please.

I bought an ER32 collet set some months ago for my mill, and thought it would be good to use them in the lathe also (Myford ML7)

I bought a 4" collet chuck and blank back plate from Arc Euro, and I've machined a shoulder on the back plate to be a tight fit on the recess in the back of the chuck. I then used a transfer punch to mark the through holes in the back plate. This is all bolted together with 3 M8 socket cap bolts.

Now, when I put a dial indicator against the flat area behind the collet nut, it's true to less that 1thou, but, when I fit a collet and some ground rod (silver steel) it is out by 15+ thou!!

So my question is, how accurate are collets? I assumed they would be almost spot on every time. Should I machine the shoulder on the back plate smaller by 5-10thou and the set the dial on the ground bar stock and tap the chuck till it's true and then nip up the M8 back plate bolts?

Thanks for any advise, cheers, Rob

Do the M8 bolts have a clearance hole to stop them pulling the collet chuck? I can't believe it's the collet or even the chuck itself.

If you put the indicator on the taper on the inside of the collet chuck how does that read?

If thats good you are installing the collet into the nut before fitting?

I have seen some american sites where people have not done that and had run out

ISTR the runout spec is something like 5 or 10 microns. I have only ever had one ER 32 collet on which I could measure more than 1/10 thou runout.

How did you set it up to machine the register? The outside of the chuck (behind the collet nut is not necessarily true. I would have set a silver steel bar to run true in the four jaw or between centres and then used a collet in the collet chuck to mount it on that bar for turning. Using the biggest collet in your set.

Yest, make sure the bolt holes have adequate clearance.

Russell.

Rob,

Are you assembling the clap nut / collet correctly?

You must pre-mount the collet into the nut ensuring the groove in the collet is engaged with the crescent shaped extractor ring. Them put the collet complete with its nut into the chuck . If you simple put the collet in the chuck and do up the nut no collet however precise and/or expensive will run true.

Have a look at the Arc website catalogue which gives precise instructions.

Another test when a collet seems 'out' is to mark the high spot on both collet and mounting and then loosen, rotate collet, retry, repeat a few times. The high would follow the part that is off, or be random if it is due to technique.

As Ian has said, check Total Indicator Reading on the internal taper where the collet locates. And work either way from there.

Backwards to the backplate setup if you get TIR errors.

Forwards to the collet runout checks if the taper runs true.

If you go down the morse taper collet route don't overtighten it or you'll have the devil of a job getting it out. Don't ask how I know.

Edited By duncan webster on 10/01/2016 18:54:36

Hey guys, thanks so much for your comments. I'll try and answer the questions.

The M8 bolts have a clearence hole of 8.3mm, and none of them (3) bound as I put them in by hand.

Yes I have been putting the collet into the nut first with a faint smear of light oil on the collet surfaces.

No, I have not put the dial gauge into the taper, but on the smooth area at the end of the outside threaded area. As it is all of ground finish, I'd assumed that it was all machined in one operation, so should read true there. I will move the dial gauge to the taper and see how it looks then.

Thanks again for your input, I'll report back once I get it looked at again.

cheers, Rob.

Russell, the collet chuck comes fully machined/ground with its reses machined into the back side. The back plate however is blank, but with the correct shoulder and thread for my Myford.

So, I removed my 4 jaw chuck (my 3 jaw is badly worn, so only use the 4 jaw), and mounted the back plate and turned a shoulder onto it fit the chuck. I then added a witness mark and used a transfer punch to mark the holes into the back plate.

If it's out in the taper, I'll take a few thou off the shoulder and then tap it true before tightening the bolts fully.

Would you all suggest thread lock on these bolts?

Hi Rob

I would suggest you try to work out the cause a bit more before taking the location diameter down.

Like the other members I would be surprised if the chuck was that far out, but if it was I'm sure it could be exchanged.

Are you sure the spindle nose location is repeatable ? If the back plate register bore is not a good fit because of either the spindle nose being worn or skimmed at some point or the back plate accuracy, you won't ever get a repeatable run out and the 4 jaw would hide any such problem.

Perhaps clock the backplate location diameter or the outside diameter if you turned that at the same time.

If that's OK then are there any sillies such as location boss too high (ie chuck not sitting on the right area, which I assume, not having seen the chuck, is on the face with the largest diameter).

Do the holes have any burrs that touch the mating surface.

Good luck - Will

You have an ER32 set up for your mill have you done the same check on that set up it would help identify if it was the collets or the chuck.

Thanks again guys, as soon as I get a chance to get back into the workshop I'll take these tips on board and see how it looks.

Good call re checking the collets run true in my verticle mill.

I know this is late in the day, but I’ve just watch a YouTube vid dealing with this subject...the problem turned out to be the collet chuck nut...different ones gave completely different readings..

Thanks John, tbh I never got to the bottom of it, so carried on with the 4 jaw chuck. Good shout re the nuts, I’ll order a new one and try.
Do you have a link to the YT vid you watched?

Rob,

I have several of these ER32 chucks, although mine are mounted on 'loose' registers.

But as a suggestion - slacken off the backplate bolts sufficient to still hold the chuck on the backplate (but not tightened up) and take another reading (from the edge of the collet taper). You may find the problem goes away (or improves). If so, you may have a problem with either your 'bolts' (cap-screws?) or holes being slightly less than perfect. It takes very little to cause distortion when you really torque things up - which by the way shouldn't be necessary on a well fitted register (I don't have mine really tight even on registers that have some slight play in them).

Regards,

IanT

The loose register trick is exactly what I did with the er25 adaptor on a lathe backplate. Only a small ammount is required, about 0.004", 0.1mm and the normal screw clearance will allow exact lining up if it is necessary. I have bought a small quantity of higher quality collets in the most used sizes, Cutwel and Vertex. Remember, if the collet is not perfect, the fine adjustment may have to be carried out each time the workpiece is changed, as there is no way of indexing the collets in this system.

The wobble that you got is very large, did you face off the backplate at the same time as producing the register? Is the depth of the register less than the depth of the recess in the collet plate? Have you checked the running of the taper inside the collet plate, you will need a lever indicator to do so, the plunger dti's will have difficulty reaching inside. Poor quality collets may have up to 0.003" runout which is bad.

Edited By old mart on 03/06/2020 17:33:47

Thanks guys, I’ll take another look at it, as I’d like to use it more often. If need by I will buy a couple of brand name collets from someone like Cutwell and see how they measure up.

cheets all.

Check the inside of the collet, there's another thread on here last week, turned out to be burrs from cutting the slits

I have not got one myself but if you get a Stevenson's collet block you can use ER collets in a 4 jaw chuck. That way you can adjust runout to suit your needs and use ER collets at the same time.

Martin C