Runout on ER25Collet??

I had this - lathe spindle good, collet chuck good, backplate bad.

I then had the good fortune to be given a good quality "tenths" DTI, so then worked with a sloppy backplate onto the lathe spindle and "tuned" out the error. Nuisance if you need to constantly swap out the collet chuck (no good if you need to make money out of your workshop), but for a low cycle hobbyist like me, not a material issue.

Steve

I've had this problem and, if you google around you'll find some discussions with Jim Guthrie about it.

I swapped my first Chronos one for a new one which was no better, then returned that and got an Arrand, which was a bit better but still ran out about 0.002".

I fixed it as follows;

Centre drill the ends of a piece of 10mm bar, slide a 10mm collet onto the bar and mount between centres.

Mount a DTI in teh tool post at exact centre height, bearing on teh cone of the collet.

Adjust the topslide so it tracks along the cone with zero runout i.e. set the topslide angle exacly to the taper angel of the collet.

Mount the ER25 chuck. According to Arrand it should be pulled up tight to the back register then the tommy bar given a tap with a hammer to seat it rock solid on the register.

Uisnig a rigid boring bar, with a fresh carbide insert and wothte cross slide locked, use the topslide to take a cut across the conical surface of the chuck, taking off the absolute minimum of material - I thing I removed less than a couple of "thou".

After this, mine showed zero runout. On removal and replacement it typically manages better than 0.0002.

Worked for me but as others say, this advice is worth exactly what you paid for it!

Garth, very disappointing but unfortunately as was mentioned in answers to your other post not unheard of even with more expensive ones than yours. A couple of options I guess. Ask for an exchange to see if you get a better one although that might appear to be a bit of a lottery, if you are close to the supplier and are not paying postal costs all the time I might try this option once. Correct the problem yourself as Norman has described or ask for a refund and get the backplate mounted type as they can be "tuned" a little more easily.

While it certainly sounds if the chuck register and collet seat are not concentric, I might just confirm that by mounting a suitable bar between centres and mounting the chuck on it with a collet, a DTI on the chuck internal spindle register face will then confirm concentricity or lack of it. I only mention this as you can get chuck mounting problems such as tight threads or oversize internal register that give similar problems and if not corrected will affect repeatability. While your check on the outer chuck surface would seem to indicate that all is well there are some instances where this surface is not machined as part of the manufacture process and I have found examples where the outer surface is not even round. I admit that your conclusion is most likely but being a pedant sometimes I always like to be sure. Also, but again I suspect not in your case, taking a DTI reading off of a tapered face can be a problem. If there is any "end float" in your spindle then the readings will be worthless unless you load the spindle to stop it moving in and out as it rotates.

Once the spindle mounting is confirmed, the correction that Norman details is not as difficult as it might seem and even if the chuck is too hard to machine with a stout carbide boring bar (unlikely); then it can be done with a "toolpost grinder" cobbled together from a dremel type tool and mounted on the topslide. As Norman says the amount of metal to be removed is extremely small. I made two of this type of chuck myself with every success and if I can do it anyone can. The trick here is to take very light cuts. If you need to grind then make sure you cover the lathe ways very well and clean off everything afterwards, grinding grit gets everywhere.

The third collet chuck I have is a backplate mounted version from ArcEuroTrade and I had very little trouble mounting this to a backplate and getting repeatable concentricity of a "couple of tenths". Do choose your backplate carefully though as I have had examples of part machined ones with oversize internal registers (ranging from .0015 - .005" oversize). Once the backplate is mounted on the spindle a quick "lick" across the face and careful turning of the register to fit into the chuck register and drilling (tapping?) the bolt holes is all that is required. On mine this then provided repeatable run out of a couple of tenths. If you mess up the register (yes we all make mistakes - well I do) by making the register slightly undersize the chuck can be attached to the backplate with the bolts just holding and then the chuck tested and tapped until you get the concentricity you require or your patience runs out. The bolts when fully tightened will hold it just fine, obviously there has to be a little clearance in the mounting bolts as well for this to work.

Hope this helps a little and of course apologies if I am "teaching granny".

Keith

Whilst not casting any doubts on other suppliers mentioned here, I can recommend Arc Eurotrade. I understand from people who I trust who are close to Arc Eurotrade that there have been instances where they have refused to continue stock a product as the quality deteriorated when they could have just continued 'box shifting' without caring. I've had excellent service from these guys and would buy again without doubts.

just my 2p

Bob

O.k., so Chronos sent me a replacement collet chuck today, not the sam

as the previous one (apparently this one is top quality British n less).

Strangely I am still getting 0.003" runout on the ground internal cone but less than 0.001" on the machined bit of the chuck body. I'm gettin less than 0.0005" on the lathe mandrel.

Whats going on ? Two different collet chuck, two differen manufacturers but still the same runout? I'm getting about 0.003 runout on by 3 jaw chuck at best, so really using the collets I hav doesnt give me any benefit.

Any thoughts would be greatly appreciated!

Regards,

Garth

-- DR_

----------------------------------------------------------------------- DR_G's Profile:

Garth,

Going back on my experiences of the Chronos product, the best one that I got from them was not a good fit on the register of the nose. I got my register checked by another member of this group and we found it to be spot on for diameter. I measured the bore of the collet holder by turning a plug until it was a tight fit in the bore then measured that and found that it was a thou or two over (I can't remember the exact amount now). You could feel this play when the collet holder was being mounted on the register.

When I measured my runout, I got the same value in the tapered collet socket and along the body of the holder. So I assumed that the holder was being forced slightly eccentric on the register when the holder was tightened on the thread.

I had thought of seeing if I could get the bore plated to add on a thou or so to get rid of the play, or to bore out and re-bush the register bore to a more accurate fit, but I would really want to do that job on another lathe so that I could use the present lathe spindle as a test for fit.

I suspect that the makers are playing safe and the tolerance they are allowing on the register bore is all plus and no minus. As a matter of interest, all the other items I have which fit the lathe register have widely varying fits. My three jaw screwed body Burnerd has as much play on the register as the Chronos collet holder, and it has been like that from new. A backplate I got from Myford a year ago has some play, but not as much as the three jaw and the collet holder. The backplate on my four jaw Burnerd is a wringing fit on the register - all items bought over the years from Myford :-)

I used the backplate I got recently to fit a backplate mounted ER25 holder and I haven't quite got it down to negligible runout, so I might swap it with the Burnerd backplate to make sure that the play on the register is not contributing to the runout.

Jim.

The runout on the chuck body is immaterial so there's no point measuring it. The internal cone is all that matters and if two chucks are showing the same error then it would seem they are not seating properly on the headstock. Check for burrs on the nose of the lathe.

Jim,

Thanks for that. I might try wrapping some ptfe around the register an see if anything changes, but it's strange that I got much less runout o both chuck bodies than on their respective cones.

..or is it? I guess 0.001" on the machined bit of the chuck neares the headstock may well translate to 0.003" further out where the con is.

Cheers,

Garth

-- DR_

----------------------------------------------------------------------- DR_G's Profile:

Garth Hi, I'm not sure how you read these posts but if through thi group I see that several previous answers that were posted were no picked up in this thread. It might be that they said something that yo didn't want to hear if so ignore this one as well.

As Dave (and I in one of the missing posts) said the reading on th outer surface is not relavant at all. It might or might not have bee used as a reference surface during the manufacturing process, mos likely not. There is a way to test the relationship between interna chuck register and the collet seat as I have already mentioned:

"While it certainly sounds if the chuck register and collet seat are not concentric, I might just confirm that by mounting a suitable bar between centres and mounting the chuck on it with a collet, a DTI on the chuck internal spindle register face will then confirm concentricity or lack of it. I only mention this as you can get chuck mounting problems such as tight threads or oversize internal register that give similar problems and if not corrected will affect repeatability. While your check on the outer chuck surface would seem to indicate that all is well there are some instances where this surface is not machined as part of the manufacture process and I have found examples where the outer surface is not even round. I admit tha

your conclusion is most likely but being a pedant sometimes I always like to be sure. Also, but again I suspect not in your case, taking a DTI reading off of a tapered face can be a problem. If there is any "end float" in your spindle then the readings will be worthless unles

you load the spindle to stop it moving in and out as it rotates."

From our earlier discussion you will see that these UK made chucks hav also in the past had problems with concentricity. If you read you question here:

You will see several previous answers that are missing from thi string. Apologies if you have seen them already and don't agree wit the advice posted, I will of course shut up in that case.

Regards

Keit

-- jontom_1u

----------------------------------------------------------------------- jontom_1uk's Profile:

Check particularly the fit between the flat rear end face of the collet holder and the corresponding flat register on the lathe nose. This has to be hard metal to metal contact. Tenths here corresponds to whole thous at the collet business end.

Jim

Perhaps you are aware or someone else has mentioned it but runout of a cylindrical type shape can be of more than one type. It is possible for the runout to be consistent at different points along the rotating axis which indicates that the cylinder is parallel but not concentric with the axis. It is also possible for the cylinder to be "cocked" so that it wobbles around the axis. This causes the runout to vary with distances along the axis.

If you are just measuring the chuck taper I don't think you have enough information to correct the problem. As someone else mentioned, end play or axial runout in the spindle will effect this. I would grip a precision rod in the collet and measure from that. If your collet can be rotated in the chuck you can separate the chuck and collet runouts. Or use a different collet to compare. Keep track of not only the amount of runout but also its direction referenced to the spindle, chuck and collet.

After all, the intent is for the work to run true, not the chuck.

Don Young (USA)

Guys,

I'm reading the thread responses via rcgroups website - Some of th replies mentioned (Keith) I did not see at all. I posted the initia question last week I think, but no responses at all showed up. Also, replied to a post last night, and this morning there is anothe response that appears immediately before mine that was not there when replied! Very confusing. Is there a way of always seeing all posts in thread??

A couple of you said that chuck body runout is irrelevant. I disagre to some extent, since if I'm getting *less* runout along the length o the machined area of the body (ie two places along the machined lengt have similar runouts) than within the cone (0.001" body, 0.003" cone it surely indicates that the cone (at whatever radius is measured) an the body are not concentric? It would be a massive coincidence if a error in circularity in the chuck body were to almost exactly negate a error in circularity on the lathe register? (or in wavyness of th seating face of the register or whatever happens to be the dominan locater).

I have checked the runout with a test bar also, and still get just ove

Whatever, I am sending both chucks back and getting a refund.

Cheers,

Garth

-- DR_

----------------------------------------------------------------------- DR_G's Profile:

Only if you are measuring the runout at the same distance from the headstock in both cases. I believe you indicated that you were measuring the body near the headstock and obviously you'd be measuring the cone at the far end of the chuck.

It would be a massive coincidence if an

I find it hard to believe that two chucks have the exact same problem. I'm sure it's a mounting issue. I obviously don't know exactly how your lathe mounts though. My Student has a bloody great tapered cone with a screw collar that pulls the chucks back onto it. The chucks don't seat against anything flat. I have zero measureable runout on the cone and also zero on my Burnerd collet chuck. Anything less on a collet chuck frankly means it's as much use as a chocolate teapot though.

If you have a cone mount as well I wonder if it's undersize for some reason and the chucks are pulling right back against the flange so they are in fact loose on the cone. That should be easy to check and would explain the problem. It might be as simple as machining a bit off the back of the chuck so it can pull back fully onto the cone.

That seems like a reason to investigate further. You could get this result if it was exactly concentric at the register but was not square. Run out would then be proportional to the distance from the mounting.

I would examine the register very carefully for bright spots (indicating raised areas) on the end face or bits of ingrained swarf. Also worth checking the threads too.

Russell

O.K. I will double check tomorrow that everything is clean and burr-fre

on the register.

Cheers,

Garth

-- DR_

----------------------------------------------------------------------- DR_G's Profile:

Guys,

Regarding the collet runout problems I've been having:

I decided (as suggested) to re-machine the cone on the chuck myself. After carefully cleaning everything, and setting up the topslide angle even more carefully, I took a couple of cuts with a carbide internal boring tool, followed by a polish with some emery cloth and brasso.

After re-checking runout, I'm now getting less than 0.0001" in the cone (which is about the same amount as the lathe nose), and less than 0.001" with a test bar fitted into a collet.

I said previously that I'd be happy with 0.001", so I gues my problem is effectively solved.

If as above, I am getting virtually zero runout within the cone, but a bit more with a collet fitted, I guess this means that the collets and/or test bar are not quite spot on?

Regards,

Garth.

In article , DR_G writes

Garth,

How good is your test bar? When I tested my collet chuck, I used an Arrand 2MT test bar, certified within 0.0001" for roundness and concentricity (putting the 5/8" parallel part in the collet, obviously). If you use something of similar accuracy, then it suggests your collets are, as you say, not spot on. Probably not a lot you can do about this.

If, OTOH, you used a piece of BDMS or something of similar (lack of) provenance, then it could be either. You might learn more if you take a light finishing cut along the MS test bar, then turn it by 180 degrees. If the collet is good, the turning should remove the eccentricity of the test bar, and rotating it will not show any difference. If the collet is out (either on parallelism or concentricity) then rotating the test bar in the collet should show an *increase* in the runout.

David

David,

My test bar is a ground shaft from an old printer, so maybe you are right about its provenance.

I will do as you suggest and turna test piece and turn it throught 180 and repeat.

Thanks for the advice,

Garth.