What's new
What's new

Watchmakers lathe accuracy

henrylr

Plastic
Joined
Jan 13, 2018
This may be hard to believe but the Boley Leinen 8mm lathe, I've owned for 15 years has amazing, to me, spindle runout accuracy.
About 5 five years ago I measured the spindle runout, starting with a last word indicator with .001 divisions and the indicator never moved. I switched to a .0005 division dial indicator and again no movement. So I switched to a .0001 division indicator and saw very slight hard to see movement.
Since then I’ve acquired a Mahr Klein Millimess 9129 and last night I went through the .001 to .0001 indicators again and got the same results. So I set up the Mahr indicator and the pointer moved about 1um which is equal to .00004 (1um actual vale is .00003937…..). This is amazing to me or is it normal for watchmakers lathes? All my indicators are calibrated using my Mitutoyo height gage.
 

Attachments

  • DSC00552lathe10.JPG
    DSC00552lathe10.JPG
    96.8 KB · Views: 231
  • DSC00561dialindicators2.jpg
    DSC00561dialindicators2.jpg
    75.6 KB · Views: 188
  • DSC00563dialindicators5.jpg
    DSC00563dialindicators5.jpg
    53.8 KB · Views: 194
Seems like normal results for a good WW-style lathe like a Leinen with hardened steel spindle running in diamond-lapped hardened steel conical bearings. The spindle bearing play that allows oil in the bearing area is adjustable, so the thickness of the oil film and the side load on the bearings is the main source of runout when testing while turning the spindle by hand. Testing while running at normal speed will probably give the best result unless there is motor vibration that shakes the indicator.

Many newer (post-1940) watch lathes have precision ball bearings and they will usually lose their preload after years of use, making measurable runout and endplay. The ball bearing lathes were favored by industrial mass production users with operators who did not own the lathe, so those lathes often show abuse and obvious wear.

Ball bearings seem like a good thing, but a good Webster Whitcomb lathe can be expected to have great accuracy after 100 years or more of the light use seen by actual watch repair shops where the operator owns the lathe.

Larry
 
While not a watchmakers machine its not much bigger my uw1 gets about a division on a 0.5um indicator. 1946 built but has taper rollers with a coil spring to preload them i was blown away by it.

Sent from my SM-G973F using Tapatalk
 
I agree with Larry that seems normal to me. A well made plain bearing spindle can approach perfection.

I have a Leinen 8mm lathe, WW82, with ball bearings. I've had it about 22 years, put new bearing in it about 18 years ago, Barden P4 anglular contacts. The old ones were fine, just seemed slightly noisy. I don't have a Mahr, but it won't even try to move the needle on a 0.002mm Interapid.
 
all these instruments have stiction/hysteresis and deflection. microkator on a stand has least. dont neccesserily believe what you see on the needle. a push-on cable is a good idea with these.
 
I have a late 1920’s Derbyshire WW pattern lathe and a full set of carefully maintained collets. I have always been astounded at the accuracy in doing staff and pivot work. This lathe was used a significant amount by the original owner and then sent back to Derby for replating and refitting. He had another lathe on his bench and never used the reworked one after he got it back. When I got it from him maybe thirty years ago, it had a Derbyshire hang tag still on it. The collets are a wonder also, with half sizes up through 10 1/2.
 
all these instruments have stiction/hysteresis and deflection. microkator on a stand has least. dont neccesserily believe what you see on the needle. a push-on cable is a good idea with these.

Yes, this is why I do not trust mechanical measurements at such a small level. I think .0001" is a good limit for mechanical means. I lack the money for it, but I would want to measure it electronically with an LVDT indicator and amplifier.
 
This may be hard to believe but the Boley Leinen 8mm lathe, I've owned for 15 years has amazing, to me, spindle runout accuracy.
About 5 five years ago I measured the spindle runout, starting with a last word indicator with .001 divisions and the indicator never moved. I switched to a .0005 division dial indicator and again no movement. So I switched to a .0001 division indicator and saw very slight hard to see movement.
Since then I’ve acquired a Mahr Klein Millimess 9129 and last night I went through the .001 to .0001 indicators again and got the same results. So I set up the Mahr indicator and the pointer moved about 1um which is equal to .00004 (1um actual vale is .00003937…..). This is amazing to me or is it normal for watchmakers lathes? All my indicators are calibrated using my Mitutoyo height gage.

It looks like good enough for the purpose. Watches are not "accurate", they're only small. Magnified 10 times a watch is horribly sloppy.
The expectation, in a former USSR factory was max 0.4um roundness. It is vague and I do not know what to make of it.
 
Yes, this is why I do not trust mechanical measurements at such a small level. I think .0001" is a good limit for mechanical means. I lack the money for it, but I would want to measure it electronically with an LVDT indicator and amplifier.

LVDTs show up pretty often on ebay for peanuts. If they're in serviceable condition they have almost unlimited resolution. I bought amplifiers a few times for under $100. Same time, there is an Analog Devices IC which does that, does it well and for little money. An alternative is to hunt ebay for a Mahr uMaxum II ( red buttons !) . That's basically an LVDT and will go down to almost whatever.
 
Levin claimed to produce spindles in the few millionths accuracy, and Crystal Lake, had special order workheads of the same accuracy, evidently targeted to the military and space work, all plain bearing.
Ball bearing spindles seem to be in the 12 millionths range, perhaps better these days.
 
Back when I had Levin stuff as I recall they claimed "under 40 millionths" for ball bearing machines and didn't see anything for plain but it could be and certainly was better. They were using New Departure bearings in their early bb lathes, and would give better spec on demand, I imagine from careful bearing selection. I think an average P4 pair from Barden is way under 40 millionths, mine certainly seem to be. Of course as Levin and Moore and anyone else making this class of machine knows (and some weirdos in the wild like Donie) the bearing class is just one aspect of the whole, and there are multiple points along the way where you can totally negate several hundred bucks(or much more) of precision.

I'm pretty sure Levin lost thier mojo some time back and am surprised they're even still kicking. I sent in a headstock in the late 90s for new bearings, old were a little loud but still good; came back well _over_ 40 millionths with a weird rubbing noise to boot. They did get them right in the end but I sold that one on and switched to Leinen. Still use a Levin slide, nobody ever made on close for WW machines.
 








 
Back
Top