View Full Version : Collet Questions

Herb Helbig
12-06-2003, 01:39 PM
I want to drill and ream a series of 1/4" indexing holes in the cast iron back plate for a 6" Bison 4-jaw independent chuck.

1) Can I safely hold a C drill (0.008" under a 1/4") in a 1/4" R-8 collet so I can center drill, drill and ream with the same tool holder?

2) I couldn't find the answer in either a 1976 edition of Machinery's Handbook or the new Handbook for the Metalworking Industries by Hanser Gardner. The only acknowledgement of the existence of R-8 collets I could find was on page 767 of the latter book. Are the limits of R-8 collets in there somewhere?


12-06-2003, 01:55 PM
It should work. Give it a try.
Why not just use a drill chuck in an R-8 holder?
Has anyone noticed that genuine Bridgeport collets have a step machined in the relief slot to prevent over stress by closing too much? I have never seen them on other collets.

[This message has been edited by JCHannum (edited 12-06-2003).]

12-06-2003, 04:24 PM
Another Collet question..

I am wanting a collet set to use the Morse #3 drills in the chuck, add a vise to the table on my lathe.

What kind of collet do I need? It has to be chucked.. Is there a chuck specially for this?


12-06-2003, 04:48 PM
There are MT sleeves with straight OD's. Use them all the time in my turret lathes.

12-07-2003, 03:04 AM
Collis makes the sleeves you are looking for. They are a little expensive and are available in diameters from 1 1/4" to 2-1/2" for MT#1 through MT#6

12-07-2003, 04:47 AM

My Japanese MT3 set of collets have a 5/32 hole machined at the narrow end of the three slots to allow for closing to the limit of the slots without exceeding the elastic limit of the metal. Is that what you mean? I imagine the hole is there to prevent cracking as a round hole in a straight line cut will distribute the stress over a larger area and reduce the chance of cracks.

A properly designed collet should be able to close to the design limit without deformation.

12-07-2003, 11:29 AM
It looks like this:


It is interesting, I have only seen it on Bridgeport, I have Royal, Lyndex, Hardinge and others and they don't have it. I assume that it is to keep the collet from closing too far.
Once a certain point of over closing has been passed, the gripping power and accuracy of the collet will be lost, especially with three segmented collets. This most likely will happen before the elastic limit has been reached. The more segments, the wider the gripping range TG, ER and DA collets have 8, 16 or more as size increases.
Just curious, since Bridgeport developed the R-8 Collet, it might be a patent thing.
Incidentally, the size of available R-8 collets seems to be by 1/32"s, 5-C is 1/64"s so that is probably limit to which they can be closed and retain acceptable accuracy.

12-07-2003, 12:48 PM
JC. I don't think thats the reason. Since their spring tempered you'd really have to pull it up into the spindle to over stress it. I think that what your seing is the web that is left before the collet is heat treated. It is then is removed after it is HT and then ground to size. If the collets were slit completely through , after HT they'd look like a big three petaled flower. Depending on the hardness the web can be sawn or slit with an abrasive saw after grinding.

12-07-2003, 01:00 PM
I got the bid on some MTaper sleeves extensions that are "Flat" on the female end. 1 1/8th through hole in my spindle.

Perhaps, Thou I have lost about everything I have bid on lately. Seems like a wildfire the flagrant bidding on "own" items by sellers.


12-07-2003, 02:45 PM
"1) Can I safely hold a C drill (0.008" under a 1/4") in a 1/4" R-8 collet so I can center drill, drill and ream with the same tool holder? "


Yes you probably can "safely" hold the drill. Whether it will run concentrictly is something else.

.008" undersize plus the shank of most drills are slightly under the drill's size so you could be .009" or more undersize.

The R8's are tool holding collets, not so tolerant of holding under/oversize. On the other hand 5C collets are generally workholding and seem to a little better with undersize, but .008" under on a small diameter 5C and you lose a major portion of your grip.

In my shop we almost exclusively use 5C's up through 2" with power closers. I know in 1" and under bar work if the stock was .008" undersize we would be limited to very light cuts and the concentricity would suffer.

We recently ran an undersize job. It was nominally 1/4" stock, but the customer purchased the raw material (1030? steel) sheared to length form a wire company (5000 pieces). Wire gages are not fractional, IIRC the material was about .243". We had to die cut a 1/4-20 thread on one end and found with collet closer turned up as high as possible a 1/4" collet wouldn't hold the material well enough. We had to order a wire size collet from Hardinge ($$$$) to complete the job.

12-07-2003, 04:47 PM
Rustybolt, that is a possibility, but I doubt it. It would be easier to slit completely through rather than re-index and stop at the same point each time. The step appears to be uniform in size and location for many years worth of B'port collets which are of different manufacture, but I have never seen on one by another manufacturer.
Another alternative would be to use a metric R-8 collet. 6mm=0.23622".
Emergency collets are also available, they are soft, can be machined to size and hardened if desired.

Paul Alciatore
12-07-2003, 05:31 PM
Three thoughts:

1. If it's a 4 jaw independent, why is such great attention to accuracy needed in it's mount. All work will be indicated to center anyway.

2. Use a drill chuck.

3. On the Bridgeport collets, the slit looks like at least 0.02" at it's narrowest. AT LEAST that much. That's 3 x 0.02" = 0.06" total distance off the circumference if you tighten it up until they actually close. Since the diameter is the circunference / pie, that's about a 0.019" difference in the diameter. I would never tighten any collet of less than 2 or 3 inch size by that much.

Those slits must have some other reason for being. Perhaps a rough cut to 75% of the depth to save wear on the final saw or wheel.

Paul A.

12-07-2003, 06:44 PM
A collet should hold work +/- a couple thou of its intended diameter. I think a "C" drill in a 1/4" collet is probably pushing your luck.

You might try making a simple split bushing, "C" size i.d., 5/16" o.d., and holding the drill in a 5/16" collet. It won't be the same collet for all your steps, but changing only collets may be slightly less of a pain than putting in a drill chuck. Of course, if the machinist gods are smilling on you, you could try making a 0.004" wall split bushing to hold the "C" drill in a 1/4" collet.

12-07-2003, 07:11 PM
The MT sleeves that have a flat in the socket are "usem-up" sleeves. When the taper on the drill has become galled from spinning you gring a flat along the bad taper and put it into the socket with the flat. The flat keeps it from turning. These adapters are usually #2 to #2, #3 to #3, etc. They really extend the unusable length of the drill.

12-07-2003, 07:55 PM

If the elastic limit of the material has not been exceeded then there is no change in the shape or properties of the collet (NB). Under the elastic limit the metal acts as a spring and returns completely to the "as made" shape. There is stress that acts upon the metal's crystalline structure and the elastic limit will change with time as the metal is repeatedly stressed. The usual failure mode in that case is cracking when the elastic limit has not been exceeded, not bending, bending being elastic deformation until the yield strength is reached.

12-07-2003, 09:38 PM
C drills a .242 drill. Might be close. I use thre .010 rule of OS and US for colletts, but a tooling collett is closer as stated. The major problem is with U.Size tools, you can "spring" the collett to the smaller size.

What about a split bushing? Drill a piece of 3/8 aluminum with the desired drill, split with a hacksaw, use the split bushing in a 3/8 collett. I use this method very commonly on both R-8 and my TG 100 colletts. Just be sure to keep the "split" line(s) off the collett splits for the best holding.

12-07-2003, 09:45 PM
I realize the elastic limit situation. That is not what I am referring to at all.
The accuracy of a collet like the R-8 or 5-C will suffer when closed too far as will the holding power.
As it is closed, the bore will go from straight to tapered, and the hole will begin to resemble a three leaf clover, reducing the gripping power and damaging the collet if the tool slips.
The TG and ER collets are slotted from either end, and have a larger number of slots to negate this. They are more accurate, have a wider size range and have greater gripping power for this reason.
BTW, I, and I am sure anyone else who has kicked around machine shops for any amount of time, have seen more than one collet that has been destroyed by overtightening or jamming in a tool large part. It is actually pretty easy to do. They seem to lurk in the bottom of tool chests or the back of the tooling cabinet for some reason.

12-08-2003, 01:05 AM

Correct on all counts.

Herb Helbig
12-08-2003, 12:15 PM
Thanks to all for your help. Very instructive as usual. I started with a drill chuck, but found that the reamer showed a lot of run out. That's when I thought of using a collet, and laziness suggested that I might even be able to use it for the C drill.

A knowledgeable friend suggested the index holes in the back plate. Instead of radial holes in the rim, I decided to use the x-y dials of the mill to lay out the holes near the rim and parallel to the rotation axis. That way one can interpolate ANY angle between index holes by using the index pin and spindle as the rolls of a sine bar. Anyone tried that?


12-08-2003, 03:14 PM
I saw a very interesting design for an indexing attachment for a SB9 change gear lathe. In this case the fellow was using it to scribe dials. It works on the sine bar principle and attaches to the left end of the spindle. It has two arms, one stationary and one fixed on the same axis as the spindle. The device clamps into the spindle tube with an expanding tapered split tube with drawbar and tapered plug. The stationary arm is held by a fixture that clamps on the change gear mounting arm. The two arms of the device project to the front and are a little over five inches long. A couple of toggle action clamps allow the moveable arm to be locked and unlocked to a clutch plate that is locked to the spindle via the plug assembly. The stationary arm has a hardened insert inset and the moveable arm has a hardened pin exactly five inches from the center of rotation. The amount of index desired is set by a block of the desired thickness between the stationary arm hardened insert and the pin on the movable arm. Any division may be obtained and the thickness may be calculated on a pocket calculator. The thickness of the block is equal to the sine of the desired angle times 5.

The article with pictures is here:


12-08-2003, 07:19 PM
I haven't seen a problem that I think is from closing too far. But I HAVE seen a problem with used collets that get spread too far out.

Then they won't go into the closer right, you need a hammer.

I don't know what causes that, since I don't abuse collets enough myself to find out. But they seem to still work OK holding-power-wise, once (or if) they actually go into the closer.

Herb Helbig
12-08-2003, 08:51 PM
Evan - Thanks for the link to homemetalshopclub and Ed Gladkowski's design for a sine bar indexing system. It reminds me of the way a transit works. The only nit I can pick with it is the potential accumulation of error as one leap frogs repeatedly from one setting to the next. The holes of an index plate will never be perfectly positioned, but pretty much have to average to 360 degrees!


12-09-2003, 03:25 AM

I think you could reduce if not eliminate that indexing error by using several different blocks. If you needed, for sake of argument, holes every 15 degrees, you could use a block that gave 45 degrees and make those cuts/holes/whatever. Then start over with a smaller block and make the intermediary marks/cuts..etc. You can reference to the previous operations to keep the accumulated error very low.

12-09-2003, 11:09 AM
The sine bar attachment is an interesting, elegant method. It can introduce stackup as noted.
But, attaching a cheap spindex to the spindle by an expanding mandrel will permit indexing any whole angle to 360*. Much simpler to install and less prone to error as well.
If you want finer accuracy, install a dividing head in the same way, or expand your energies to make a dividing accessory to attach to the spindle. Now counts and degrees are available, and the indexing plates do cancel out minor errors.

12-09-2003, 01:29 PM
Now that I have finished the Christmas present my next project will be an expanding mandrel for the end of the spindle that will allow me to mount any one of the change gears directly to the spindle for dividing purposes. I intend to use a mechanism somewhat like a spring loaded ratchet or even similar to an escapement to sequence the positions.

12-10-2003, 05:10 AM
To do the sort of work that you are trying to do accurately consider getting an ER Collet system for your mill. and ER-25 set would allow you 1mm-15mm range (1/16"-5/8") and ER-40 would give you (1/4"- 1-1/2") ech collet close on a 1mm range and holds the tool accurately and very tight. so tight that you can infact damage hss shanks if you overtighten them (great for carbide though). The collets self extract when loosened so you don't have to tap on the spindle to loosen the collet (great for te spindle bearings). They are far more accurate than chucks. You can also get floating reamer holders for them if you do lots of reaming.

Herb Helbig
12-10-2003, 12:28 PM
Thank you, Thrud - You may just have defined my Christmas present!

I'd like to have a collet system that I can use in both mill and lathe.

I'm always surprised that the ads for collets are almost never accompanied by ads for collet holders. Does that suggest that everyone HAS their collet holders (like Boston ladies HAVE their hats), but wear out a few collets every week?


12-10-2003, 05:40 PM
Colletts closed too far will cause grip at the "tip most" end of the collett, or the mouth, thus leaving the rest of the bored opening tapered away from the piece being held. The result in drilling is a wandering drill, the wander and wobble being exponential to the distance out fom the collett face. his is shown in indicating the tip, but upon drilling 1/4 inch in with a 15/64" (about double the .008 proposed) drill, I measured the center area of the drill, and found a jumprope shape occuring to .008. Broken drills will result in deeper holes without a doubt, as I have been there, done that. Even on spotted or center drilled holes. The tip straightens, but the center section of the drill now takes stress, and eventually the "jumprope" shape has to release stress, thus the dill breaks, usually at about the 1/2 distance between collett and drill tip, and more often or not, above the part surface, not even with.

The same goes with lathe parts. I will post a series of pictures where I RAN SOME 19/32 STOCK IN A 5/8 COLLETT. The piece with 1" overhang from collett face, three inches held. The parts come out triangular, but the triangular shape spirals almost 90 degrees on each corner of the triangle from end to the spiral ending at the collett splits. The stock is being held by the outer most "end of the collett", or the tips of the jaws, oe even more precise, only a small fraction of that area, usually the cneter most area between the splits, or the high points of the colletts, I have determined that each of the three collett "jaws" was holding with only about .030 of the collett surface, and only at the face, and this with some goodly pressure.

I did this first by accident, then decided to repeat this for the adult and secondary students many tiomes, I have ten pieces of the same type. the proof was dykkeming the part OD, and looking at the marks where the collett held. Also used prussian blue, and magic marker. Each time consistant .030 spotting - this determined by optical comparator. The part inside the collett "whips" to the sides, but is held somewhat steady by the bore to keep consistancy with the whip so it is not out of control. This is a .031 out of control situation, but is true after .010 on any collett in any holding situation. You will get "triangulation" on measurement, especially further out on the part.

meanhile, my 5/8 collett is now sprung too small for 5/8 material, and scrathes the stock. These are hardinge colletts and royal Colletts I use, no cheap stuff, but I had several donated 5/8 colletts, so i could prove the point with one.
Drills will show this on a "non spotted" situation using a .484 drill in a .500 collett.. Did this this a.m. to prove it to myself, the thread on the computer. My hole measured .487 to .499 triangular, and the triangulation was visable.

Smaller drills break. Even center drilled.

All proofs with steel, 1117 and 1018. Aluminum would surely show this even worse?

Just FYI from controlled experimentation with this as a purpose, and fom my resulting aditional experimentation due to this thread.

[This message has been edited by spope14 (edited 12-10-2003).]

Herb Helbig
12-10-2003, 08:51 PM
Many thanks, Spope, for the information and your hard work!


12-11-2003, 02:38 AM

An addendum to Scotts notes: insert a piece that is too large and it tends to grip at the rear of the collet and not at the front - opposite to the effect he is noting. This is common with 5C, R-8, Mt#x, 16C, etc. type collets.

This does not happen with ER type collets as they close from the front and rear and evenly. The only problem with these types of collets is frittering from high speed operations and build up of foreign material on the collet internal and external surfaces and in the collet holder itself. they should be kept oiled in storage and cleaned and dried before use - ceanliness is important to extend the life of the collets. Under extreme high speed use they can be expected the be replaced once or twice per year under heavy use.

Make sure that the nut you get is a ball bearing type nut and not a plain nut - Lyndex makes the best one (they are all standarized sizes). I have the Lyndex Precision Grade ER-25 collets and I am very pleased with them. I use the Lyndex BB nut on a Schnaublin E-25 shank.

An ER-40 set covers 1/8"-1" and is avalable in R-8 (Bison from KBC)

[This message has been edited by Thrud (edited 12-11-2003).]

Herb Helbig
12-13-2003, 12:39 AM
Thrud -

Thanks for your extra input. Oddly, I was given a copy of the latest KBC flyer a day before you mentioned them. So I went to their web site hoping to find stuff about ER collet chucks. All I could find on their site was pictures of the flyer and of their machinery catalog. Neither had any links to further information. Maybe I missed something, but at first glance it doesn't look like a very helpful web site. If they sell any tools besides what appears in the flyer, I couldn't find it.

Even in the MSC catalog I could find only one reference to ball bearings in connection with ER collets, and that related to CAT40, 50, and BT40 tooling. Not much help from a Lyndex search, either.


12-13-2003, 03:00 AM

[This message has been edited by Thrud (edited 12-13-2003).]

12-13-2003, 03:05 AM

The full KBC catalog is about 1.5 inches thick. A bit much to put on the web site. Ask them to send you one.

12-13-2003, 03:16 AM
<font face="Verdana, Arial" size="2">Originally posted by Thrud:
KBC sku# 7-171-4108 Bison R-8 ER-40 Collet Chuck and kit 572.19 (canadian dollars) page 413 2003/2004 Catalog includes 15 collets 1/8"-1", wrench, nut, and fitted wood case</font>

If you decide you want a Lyndex Ball Bearing ER-40 nut you can order an ER-40 nut from them or through KBC as a special order.

Mine was $75 for the ER-25 but was flown up for me along with a collet tray.

Herb Helbig
12-13-2003, 09:42 AM
Thanks one more everyone - I better stop whining and get back to work!