lathe woes

[arron]

Hi Guys.  

I bought an unbranded used lathe this past summer and only recently started to fool with it (lesson learned about not immediately trying out used watch repair tools).  Come to find out the collets are very hard to seat all the way and then they get very stuck; I had to use a dowel and hammer to pound it out.  Looking inside the spindle i can see a very minute remnant of the key (it's not big enough to show in a pic).  It looks like it was smashed; and looking at one of the collets (pic 2) it appears possibly that someone got a collet stuck then damaged the key trying to get it out.  I'm guessing that the key should look similar to the key on the collar that goes on the opposite end of the spindle (pic 1, top circle).

i think it is still useable although it's going to be an on-going hassle to get the collets in and out.  I'm wondering if anyone has suggestions for grinding down the old key and installing a new key in a way that doesn't do more damage.  Picture 1 bottom circle shows the location of the key.  Perhaps that opening is how they put in the key.  Or could i just grind the old key and not install a new one.  How important is that key to prevent the collett from spinning independent of the spindle?  Any ideas on how to fix this issue would be greatly appreciated.  Arron.

[Michael1962]

Looking at the second picture and blowing it up as much as I can on my phone, I would suggest that the ‘key’ is a grub screw with the end of it machined cylindrical in shape to engage the keyway in the collett. If this is how it is done, then from a fitter and machinist point of view, there should be a grub screw on top of the ’key’ grub screw to stop it moving. I am not sure of the thickness of metal in the spindle as to whether it is thick enough to have two grub screws. So Loctite would be the answer in that case.

I agree with you though, that I would have expected there would be enough friction between the collett taper and the spindle taper when it is tightened to not slip. It is not as if you are taking hugely deep cuts.

I assume they are the correct taper colletts for the spindle taper?

[JohnR725]

4 hours ago, Michael1962 said:

I assume they are the correct taper colletts for the spindle taper?

The watchmakers lathe an interesting device that they all look basically the same but? I have a link to an interesting book that you can download from the link below. The book is titled The watchmakers' lathe, its use and abuse.  To understand the collet problem you want to look at page 65 in the book. There were unfortunately quite a few different variations and that means you can end up with a lot of collets that may or may not actually fit the Lathe you have.

https://archive.org/details/watchmakerslathe00good/page/n1/mode/2up

[nickelsilver]

The key is usually soft (soft compared to the spindle itself), and can be drilled. The spot you see on the tapered bearing surface is it, you can drill out the old one. Some folks stop there, as the key isn't strictly necessary and if you have a mix of collet brands might hang up on some and be loose on others. I like having a key. To make a new key you will need to turn a plug the correct diameter to fit the hole, with a (tapered) shoulder. It should be a tight fit. Get it in, file flush- a file shouldn't even mark the spindle if the spindle is properly hardened. Then give it a few punches with an automatic center punch. It's very fiddly work. Finally, you can go up inside the spindle with a fine file and fine tune the size.

 

Collets that bind on the key will not turn true, and collets that have burrs raised by the key must have those burrs removed with a stone or perhaps a file and then a stone. Be as careful as you can to not remove more material than the burr itself.

 

On a proper* 8mm lathe, the bore in the headstock is exactly 8mm, -0 +0.005mm; the collets are specced at 7.99 +0 -0.005. With a total window of 0.02mm but most often 0.01mm of freedom in the spindle it doesn't take much of a burr to bind up.

 

*Some cheaper makes of lathe will have the bore loose on size to the large size, and collets loose on the small size, for one because it's easier to manufacture to loose tolerances, and also to make it easier to ensure different makes fit. There is a loss in precision though.

[arron]

Thank you guys.  my friend from switzerland gave me an idea.  could i just file down the sides of the key that extend beyond the key way?  what kind of file would a person use for that?  any old flat metal file?

and if i go with the option of drilling it out, what kind of file would i use to remove the burrs and left over key material?   some kind of round file?  a dremel with some sort of sanding cylinder.  i'm a bit worried about make unnecessary grooves inside the spindle, or making it out of round.

by the way i looked at it under a microscope and it doesn't look like a screw.  there is a center punch mark on top like a plug might have.

Edited February 3, 2022 by arron
another idea

[Michael1962]

There is nothing to stop you drilling the current key out. Then drill and tap in the same place to take a grubscrew. File the end of the grubscrew to fit your colletts keyway. You can then screw the grubscrew and set it to a depth which will engage the keyway without being too deep to snag on the bottom of the keyway. A bit of loctite and Bob’s your uncle.

[LittleWatchShop]

I have 5 WW lathes and many!! collets. One thing I have noticed is when there is just a small amount of rust on a collet, it will not fit in one or the other spindle.  So be sure your collets are clean. As @nickelsilvernoted, tolerance is tight.

 

[arron]

Michael, i like the idea of putting in a grub screw.  will an ordinary tap cut through whatever metal the spindle is made of?  someone mentioned its harder metal than the key metal itself.  

[JohnR725]

6 minutes ago, arron said:

will an ordinary tap cut through whatever metal the spindle is made of?  someone mentioned its harder metal than the key metal itself. 

It probably depends upon who manufactured it but they can be really really hard and brittle.

When I first had my very first watchmaker's lathe I had issues it didn't seem to be right? Fortunately my father was around you put a dial indicator on and it wasn't right. It was disassembled and found that it was bent the spindle. No problem we had access to a hydraulic press and that I had a two-piece spindle? When we look at how it broke you can see that when the original pin Was driven in versus your screw it cracked so it was partially cracked for a very long time. But as a warning the spindle was very very hard.

Fortunately the story has a happy ending I was able to purchase another lathe with a complete set a call it's cheaper than the first saw it worked out just fine. But definitely a caution you'll find out when you try to tap As to whether even going to people who do this.

Then the label work fine without the whatever being there to keep the collet from turning. You just have to hold onto the collet when you're turning the other end not as nice but it does work.

[arron]

Thanks John.  I hope you are doing well up there in seattle.  seems like our winter is already over here in eastern oregon.  of course, now i've jinxed it.  take care.

[rodabod]

On 2/2/2022 at 12:07 PM, nickelsilver said:

Collets that bind on the key will not turn true, and collets that have burrs raised by the key must have those burrs removed with a stone or perhaps a file and then a stone. Be as careful as you can to not remove more material than the burr itself.

It has occurred to me now that this is most likely why some of my existing chucks would not turn true in my new lathe. The fussiness when fitting is the giveaway. I'll need to dress the key possibly. Cheers. 

[Michael1962]

12 hours ago, arron said:

Michael, i like the idea of putting in a grub screw.  will an ordinary tap cut through whatever metal the spindle is made of?  someone mentioned its harder metal than the key metal itself.  

The hardness of the spindle is not something that I know. You could test it with a file. On the non collett end of the spindle. If the file 'skates' over the metal  without really marking the metal, tapping a thread will be fraught with danger. Especially as it will not be a very large tapping size.

I have actually just thought of another idea which would be a fair bit easier and would allow you to have different setups depending on which collect/keyway size you have to accomodate.

Presumably, you can access the hole when the spindle is mounted in the headstock bearings? Let me know and I will try and draw my idea and then scan it in to show you. It would allow you to have different 'pins' for different size keyways. Wouldn't take 30 seconds to change over 'pins' and easy to replace as they get worn.

Edited February 4, 2022 by Michael1962

[nickelsilver]

There's no access to the key with the spindle assembled. I haven't seen one that had the key threaded, always staked in; there really isn't enough wall thickness to thread in my opinion - plus the real danger of breaking a tap in there!

[Michael1962]

And that is all of my ideas washed away with one stroke. Sigh.

I shall now go and sit in the corner.

[jdm]

4 hours ago, Michael1962 said:

And that is all of my ideas washed away with one stroke.

Watchmakers lathes are different from machine shop or even model engineer ones, where in many cases the use of collets is unknown. An error of 0.01mm really matters, but would be negligible in many other applications. Precise and beautiful items, worth to be collected as well.

[LittleWatchShop]

36 minutes ago, jdm said:

An error of 0.01mm really matters

I have one of those dials to measure "trueness" or "runout" (not sure the proper term).  I have measure this for all of my lathes, but the question in the back of my mind..."where to measure it?"  I measure it in multiple places,
-spindle outside end
-spindle inside end where the collet is squeezed
-on the collet head itself (sometimes, depending on the collet, this cannot be done)
-on a piece of work mounted in the collet
-etc.

So at the end of the day, what matters (I think) is the work piece, but the work piece cannot be better than the collet, and the collet cannot be better than the spindle inside edge, and it cannot be better than the top outside of the spindle itself (or maybe it can??).  All of the errors add up as you go from the outside edge of the spindle to the work piece via the collet.  A scratch, a bit of rust, or collet malformation and perhaps other things can accumulate to the error in the final work piece.  I was obsessing about this for awhile--hoping to find (among my five lathes) a perfect one...like 1 micron runout!!  In the past I have posted about this and received sage guidance from @nickelsilver and @jdm.

...and then someone on the forum told me that the measurements should be made at such and such RPM.

Now I am going to start obsessing about this again...ugh!

[nickelsilver]

Make the measurements turning the spindle by hand, whatever rpm that is. If you see that you have 0.05mm runout, turn the lathe on with the indicator still in contact- it will look way better, haha.

 

Generally you would check the collet taper inside the spindle. The outside front diameter should be as good, but not necessarily (better makes will be great, lesser, less good). Unless a lathe has seen a lot of abuse the collet taper should indicate darn close to zero. My lathe doesn't move an Interapid 0.002mm indicator, so it's quite good.

 

Yes, lots of errors can stack up- but sometimes they cancel out, and a dinged up spindle, with a manky collet, and a workpiece that isn't straight, can end up looking great- until you pull the part out and try to put it back like it was.

[LittleWatchShop]

I am going to fork this discussion over to "collets" which seems reasonable for this thread.

I have been amassing a collection of collets via ebay purchases.  The goal has been to get a complete set from 1-50.  I have done that already which also includes all but one of the half sizes between 1-10.  I have spent a lot of money!

Then some time back I decided I needed LEVIN collets, so I started that journey ($$$).  I have been under the impression that they are the best (vs. starrett, boley, peerless, victor,hrp, whitcomb, unbranded).  Frankly (as would be obvious from my posts) I am clueless about whether and how much Levin is better than the others.

I am sure someone has an opinion.

[jdm]

29 minutes ago, LittleWatchShop said:

.The goal has been to get a complete set from 1-50.  I have done that already which also includes all but one of the half sizes between 1-10. 

That would be grabbing flush outside grabbing only. Then you have funnel inside and outside, and other types too. Never try to play complete set, it's the same a martingale strategy at the casino.

[LittleWatchShop]

Here is a video and photo showing the testing I just now did on what I think is my best lathe.

[LittleWatchShop]

5 minutes ago, jdm said:

That would be grabbing flush outside grabbing only. Then you have funnel inside and outside, and other types too. Never try to play complete set, it's the same a martingale strategy at the casino.

I have read and re-read this several times.  Not sure what it means.  I cannot tell if it is good news or bad.  I am a native Texan...that may explain my confusion.

[jdm]

4 minutes ago, LittleWatchShop said:

I have read and re-read this several times.  Not sure what it means.

Just a way to say.. one can't win a this game

https://en.wikipedia.org/wiki/Martingale_(probability_theory)

[LittleWatchShop]

8 minutes ago, jdm said:

Just a way to say.. one can't win a this game

https://en.wikipedia.org/wiki/Martingale_(probability_theory)