How much run out with a 5C collet?

[jeastwood]

I'm building a 5C collet chuck for my little Sherline lathe. For you lucky folks with big iron that already use 5C collets, this question: if you meaure run out on a 1/2" piece of drill rod held in the collet, what do you expect to read 3" away from the collet? I'm tryng to figure out when I'll be "done" tweaking my chuck. If when I've machined it properly I should get 1/2 mil at the collet and 1/2 mil 3" away, then I have some error I still need to find and correct.
I have cheap-o Enco collets; perhaps I can't expect great accuracy out of them?

Thanks guys,

[Jim Caudill]

For total indicated runout, you're going to see the combination of the spindle error + collet chuck + collet + dowel or gage pin (whatever you're using) + any error as a result of clamping force. You are looking for 1/2 millionth of an inch error????? I can't even measure that error let alone have a lathe that can produce it. My spindle runout is a couple of tenths of a thousandth. When I maeasure my collet adapter, now we are close to TIR of 1 to 1-1/2 thousandths. My Hardinge collets are pretty true, so that is what my gage pin will run. TIR of about 1 to 1-1/2 thou.

[jeastwood]

Thanks Jim. Sorry my terminology was misleading; by "mil" I meant "thousandth of an inch". So with a decent collet, I should read a variation of 0.0015 or so 3" away from
the collet, is what I get from your response.

Thanks again,

[SGW]

Nothing like going down and measuring it.....

I have Hardinge 6K collets for my lathe, but that probably doesn't matter in terms of what accuracy you can expect. Right at the chuck, I see virtually no runout at all on a 5/16" dowel pin. Out about 1 1/4", I see about 3/4 thou runout.

[nheng]

Jeastwood:

Must be an electronics guy with PC board copper, dielectrics and everything else in mils

I've got an older Sherline and was going to suggest ER32 or ER40 collets but it sounds like you're almost done. One of the other collet families could cut your overhang length to about half of what the 5C will require.

There's an outfit that has some lower cost collet chucks for threaded spindles in HSM and on the web but the name escapes me.

Den

[nheng]

It's www.bealltool.com and they have chucks for fairly small spindles. They don't mention what type of collet is used but it looks like the ER series and goes up to 3/4".

Den

[docsteve66]

jeastwood: you may have a problem seperating the ACTUAL TIR from the APPARENT TIR.

If your spindle axis of rotation is truely pointing down the lathe bed parallel to the carriage travel axis at every point, you can begin to find the apparent TIR and the Actual TIR.

If the colllet center hole is Parallel to the spindle axis and you measure TIR you are seeing actual TIR. actual TIR should be a constant (again assuming the spindle axis and collet axis are PARALLEL, but off set some what) where ever you measure.

If the colllet center hole is NOT Parallel to the spindle axis and you measure TIR you are seeing actual and Apparent TIR. This figure will vary (most like increase as you move from the head) with the carriage. But the CENTER of the cone of TIR will point and tell you how much deviation from the spindle axis you have.

So: first: align your head stock (spindle)axis of rotation to be parallel to the travel of the carriage. Thats parallel side to side and up and down.

Second: be sure your collet hole is centered around that axis of rotation. The deviation is "actual" TIR. It should be constant where ever measured. This in only true if the spindle axis is parallel to the carrage travel.

Third; Be sure your collet hole center axis is parallel to the spindles axis of rotation.
The deviation caused by the non parallel collet hole will appear to increase as yo move away from the head stock. but the cone's center will co-incide with the spindle axis and the ACTUAL TIR axis. but the Cone will have a "thick" center line if the TIR is not zero.

Finally, lets assume you have the Spindle axis pointing in the right direction, your collet made with zero Actual and apparent TIR. when you remove the collet and remount it, the TIRs may change because your taper is not really centered and pointing in the right direction. So mark your spindle and collets so they can be installed thesame way each time.

WHEEEEEEWWWWWWWWW!!!!! I bet I made some gross errors! Comments please. Talk about a fellow asking the time and being told how to make clock!!!! but i see no simpler way of explaing what is being seen.

Steve

[jeastwood]

Thanks VERY much for the long and thoughtful reply; details and insight are just what I need.

I did figure out ahead of time that I need the spindle axis to be parallel to the ways
of the lathe, in both axes. I had an interesting time measuring this and correcting it on my Sherline. The up and down paralleism was pretty mush dead on; side to side took a little adjustment. So, when I bored the chuck, and turned the 10 degree taper to fit the collet, I had the spindle axis lined up as you described.

The run out I see is consistent with the axis of the collet diverging at an angle from the axis of the spindel, since I see small run out at the collet and it increases as I move rightwards. I'm starting to think that much of the error is in the collet; I hope so. I'll do some more measuerments inside the collet, looking for variations along the spindle axis. I note that MSC has a spec. of .0005 TIR at 1" from the end of the collet for their higher cost collets; they don't show a spec. for the lower cost variety. For a cheap collet set like I have, perhaps I shouldn't expect any better accuracy than what I'm seeing.

Thanks again for the info,

[Treven Baker]

How about mesuring just the collet bore and forget having a collet and rod? It would simplify things.

[Thrud]

jeastwood:
The very best Hardinge Precision Grade (Select) 5C collets are .ooo2". For maximum accuracy you have to bore blank collets to the work size. The Hardinge are the very best collets that hard earned money can buy. The Enco are likely closer to .oo1".

I would consider a 3" 6 Jaw precision grade Chuck by Buck, Bison, or Pratt-Bernard instead. Your spindle cannot pass any larger diameter stock anyway, so the chuck will be far more useful. The six jaws have a more even clamping pressure so there is less distortion with tubing and light parts.

Take a look at the Beal chucks.

The suggestion of ER collets make more sense for a lathe this size as well. A set of 15 ER-25 collets over the entire range of 1mm-16mm with no gaps. 5c collets only hold one diameter and that has to be very close for proper clamping. You can buy the ER-25 collets, a ball bearing collet nut and a wrench, and make a collet holder for it. It would require a Metric thread to be cut, however for the standarized ER nut. It should also be hardened after machining as the collets exert considerable force on the walls of the holder.

[docsteve66]

jeastwood: you are getting some good advice, but don't give up yet. At least you are learning! And so am I!.

Yu say the increasing deviation from the center axis leads you think the collet is bored at an angle. Not neccessarily so. check wherethe center of deviation goes. thas wherethe axis the collet is turning about is.

Try marking the installed positions, then removing and re-installing. The errors should follow the defective parts. they may (probably are) "defective" (what ever that means). Nothing is perfect (ask the wife), but it can all be used if you know where and what the defct is and hhow it affects what you are trying to do. You may never get it even "good enough", but you will (already have so far as I can tell) learn a lot about those imaginary lines and how to find them. some day, some job, you will find the ability to "see " the invisible is invaluable.

[Thrud]

jeastwood:
I was reading your problem of error. The problem you may be having might be if an undersized (say .2485" instead of .2500" in a .25oo" collet) rod is used the front edge of the collet grips the piece while the back does not. The opposite is true with an oversized workpiece. When over or under sized bar is used the bar can "wobble" as you are experiencing. This is why precision jobs always use blank collets and are bored to the workpiece size.

This is a functional defect of the single angle collets like the 5C. The Precison collets such as the ER series are double angle - having force cones on the front and rear of the body of the collet. The ER collets always close completely in a parallel fashion compared to the limited geometry of the 5C. This is why I recommended them over the 5C collets - greater versitility and precision. 5C only close properly on the specific size they are cut for.

Sorry I did not make that a little clearer before.

[This message has been edited by Thrud (edited 12-12-2002).]

[jeastwood]

Thanks again for all the good info, guys.

I have measured the taper in the bore; it's less than 1 thou over 3 inches or so.

I've also found that at least some collets do have a "sweet spot" where the run out 3" from the collet drops to 1 thou or less; I find it by closing the collet, measuring run out, loosen collet, twist it 30 dgrees or so, re-clamp, and measure again. A few cycles of that usually locates a sweet spot where the run out is at a minimum, sometimes down to a thou or less. My next test is to see if this angle is constant for a give collet, or if it varies.

I like the idea of making another chuck for the ER-40 metric collets; smaller, lighter, more accurate? Cool!

Regards,