Producing a ER20 collet nose spindle, what is least worse, too steep or too shallow

Hi,

This is both a theoretical and practical question in terms of educating on best practice. I am about to cut an 8° taper for an ER20 collet nose spindle. On the assumption that I can't get it perfect what is recommended:

1. Assuming I use commercial collets, for collet closing concentricity. What is best practice too big >8° or too small an <8° angle?

2. In terms of the practicality of doing a nice job my inclination is to make it too shallow so with a bit of polishing it would tend to open out.

Thanks,

Steve

Edited By SteveI on 01/11/2017 12:39:52

I would play about with a bit of scrap to get it right then not alter the topslide until you have made the actual part. With a bit of blueing you should be able to get contact just about right.

Any polishing near the open end would risk making the taper bellmouthed.

For most other collets that have slits only on the outer end you want the closed down position to still be gripping the outer end. Since ER have two sets of slits they should close up concentricly but any error would mean the central bore is not parallel. So you still want the outer end to be tightest. I think that relates to <8.

Tangent of 8° is 0.1405, sine is 0.1392. Set up a decent dial gauge in the tool post running against a really straight bar and it should be no problem to get sufficiently close that the direction of error is moot with an inch or, better, 2 inches, of travel to generate the appropriate movement on the gauge. Tangent for moving along the bed as with a proper taper turning attachment. Sine if swivelling the topside.

Metric mavens could use 25 mm travel. Look for 3.51 mm gauge deflection with a taper turning attachment and 3.48 mm for swivelling the topside.

As I have proper taper turning attachment and a bed stop I use a gauge block from an old, no longer wrings together, set. Not too hard to make a suitable block of your own. Using the dials to measure travel works but its easy to make mistake and have to start over.

If swivelling the topside make use some sort of opposing screws pusher to give controlled movement. Use nice fine threads. 40 tip ME is good if you have it. Pushing and tapping tends to deteriorate into a long sequence of "nearly there, "one more... rats too far..... back a bit..... rats too far" very naughty words.

Clive.

Hi Steve,

Harold Hall provides this diagram of a set-up similar to Clive's to achieve as close to 8 degrees as makes no difference...

**LINK**

It's part of a description of an ER collet chuck here... **LINK**

HTH

Jon

With patience you should be able to get a result close enough that you won't know if its over or under. (If I can anyone can...)

As suggested use a collet with some blue/felt pen as a gauge. Don't push it in as you don't want to close it up at all.

Neil

Hi,

Thanks to all for the responses. I am not worried about my method for setting the top slide, in that it is limited to the accuracy of my DRO, dial indicator and rigidity of my top slide which I can't do much about. My issue is being able to accurate move it in to place when those final tiny adjustments are needed. On reflection I think I am simply asking what is the best compromise for an ER collet.

As an exercise the best I can get it to is 7.9984° which I expect makes no difference, assuming perfect DRO and dial indicator. I've wasted 3 hours getting it that close. I had it at 8.0101° within 15 minutes owing to a bit of luck with the raw hide hammer on the cross slide. As per Bazyle's post my "guess" is that being under is better than being over. And in the act of getting just under I found my luck was used up for the day and it took 2.45 to get it as good again. And thanks to Jason for the bell mouth warning.

So for ER collets is being a tad under better than being over the 8° nominal?

Thanks,

Steve

Trial and error on a scrap piece worked for me. If you use a collet as a gauge it is a good idea to insert a drill bit of the appropriate size to prevent the splits closing and distorting.

I would only use those numbers as a guide to initial setup, bit of spring in the boring bar or cutting forces down through the gibs etc and you won't cut what you have managed to set up for. That's why it is better to compare the cut female taper to a male collet by blueing and adjusting.

Here is my GHT pusher:

Also in the photo is a Martin Cleeve-inspired bushing type steady in use with an ML7 cross slide co-opted as a top slide for its extra travel.

ER collets rely on the holder for their shape and are less assured to be at a stress free dead true taper when empty. So for your trial fit insert a straight parallel bar of nominal or larger size into the collet.

Edited By Bazyle on 01/11/2017 15:31:52

I set my top slide to the graduations and unbelievably it was spot on, wouldn't like to bet on getting it right first time again though.

Mike

According to my calculations, Sin(8) - Sin(7.9984) over a length of say 1" results in a diametric error of 5/100ths of a thou and ER20 collets are probably less than 1" long.

For 1 thou diametric error over 1" length I think you need an angle within about 0.028 degrees of 8 or 1.7 minutes.

HTH

Jon

One way to set the taper exactly: Chuck up a piece of bar. Turn the OD so a collet will just slip on (a little force required). Slip the collet over the bar. Set an indicator in the toolpost. Adjust the angle of the topslide so the indicator stays on zero as you sweep the length of the taper. The indicator point needs to be at the same height as the spindle center or error will result.

Hi,

Ok job done on a test piece it blues up nicely. I just cut the taper with the angle I had set and it looks good. I've used the test taper work piece and made it in to an ER20 tool holder for my QCTP. Although run out is not important for that it should come in handy to hold small boring bars. Thanks for the replies. I think the best 2 things to come out of the thread for me personally was the advice that <8 is better than >8 which makes sense, but more interesting was Clive Foster's advice:

I'll be thinking about a design for a fixture (?) for my lathe to allow screw adjustment of the top slide angle. Clive do you have any pictures of your solution?

I also recall that GHT made a worm wheel adjustment modification for his taper turning attachment and the pusher for his top slide so that is tonight's reading material sorted.

Thanks to all,

Steve

Edited By SteveI on 01/11/2017 19:08:32

So far, I've been lucky, and got it pretty accurate. But I finish off by wrapping a strip of emery round a collet, oiling it, and then GENTLY pushing it into the slowly rotating workpiece.

Push hard and the emery and collet will be snatched out of your hand!

You can determine the depth to which you bore the taper by measuring an unloaded collet at the mouth of a known good taper bore. If you are going to lap using a collet, boring should stop a thou (0.025mm) or so undersize, and let the emery do the rest.

DON'T forget to clean everything afterwards, to remove the abrasive dust.

Howard

That's sounds a completely pointless exercise I'm afraid.

Over 4" of cross slide the difference between 7.9984° and 8.0000° is +0.00013", little more than a tenth of a thou. Over the length of an ER collet you are talking a wavelength of light.

Forget any temperature variations in your workshop, how confident are you 'that your really straight bar' is aligned to a significantly greater accuracy than a tenth of a thou over its length? Or that the toolpost won't flex by at least that much?

Sorry if that seems harsh but with typical workshop measuring equipment and conditions it is simply impossible to work to such a level of accuracy by measurement.

If you want a good fit, work to a known good collet and you will achieve a better and more reliable fit than you are ever likely to get by measurement alone.

A good way to use a collet as a gauge is to find a bar of silver steel, or a cutter, which is a snug fit in a collet. Not tight, just firm enough not to slide out when you tip the axis 20 degrees. Offer up the collet with the steel/tool in place and you can be sure it won't compress to give you a false reading.

Hope this helps.

Tim