Adjusting Bearings on a watchmaker's lathe.
The following article is from the British Holological Institute. The link to this fine article follows.
Adjusting The Bearings on a Watchmakers Lathe.
The British Horological Institute has archived and edited the following from e-mails sent to the Clock/Clockers mailing lists on the Internet. The information here does not necessarily indicate a method approved by the BHI, we are only publishing this digest so that others can decide for themselves whether the methods listed below will suit them.
From: John C. Losch, Leon E. Levasseur,
There are several things to know about a WW watch lathe, or any size cone bearing lathe, for that matter. Once you under- stand what the lathe is all about there will be no problem maintaining and adjusting it. Lathes of this type are called cone bearing lathes because the spindle and bearings are tapered. The spindle shapes are actually sections or "frustums" of cones, with bearings to match.
In nearly all such lathes, both ends of the spindle have a thrust angle of 45 degrees, and a lateral bearing angle of 10 degrees. When the lathe was made both the bearing and the spindle were ground to these angles after hardening, or turned in the case of bronze bearings, then lapped to a perfect fit.
If you will look at the front, or collet end of the lathe headstock of most American watch lathes, you will see stamped in the face of the casting either the word "hard" or "soft." This indicates whether the bearings inserted in the cast iron headstock are hardened and ground steel or machined bronze. There is a reason for making the lathes with both kinds of bearings beyond economics, although bronze bearing lathes were cheaper to produce and sell.
According to my friend Leon LeVasseur, who worked at the Rivett Lathe Co., manufacturer of watchmaker's lathes, bronze bearings can last as long as hardened bearings. Bronze bearings will seat in faster, and they are more tolerant of grit than the combination of both hard spindle and hard bearing because an occasional particle of foreign matter will bed-in to a bronze bearing rather than continually rotating and scoring either hardened surface at random. In his opinion, bronze bearings are better, but he agrees that there are arguments on both sides, largely controlled by the kind of service required of the lathe.
As a side note, "Frenchy" LeVasseur revolutionised the method of fitting bronze bearings and spindles when he decided lapping bearing and spindle took longer than necessary. One day, to the astonishment of workmen at the surrounding benches, he picked up a mallet and slammed the lathe spindle into the bronze bearing. Since the bearing was soft, it took the form of the spindle. The job was completed with lapping, but at the cost of considerably less time than the usual method. With soft bearings, the lapping compound was degradable so that if any bedded into the bronze, its effect on the steel spindle was short lived.
Both the life and dependability of cone bearing lathes hangs on adjustment. If one of these lathes were always used with thrust pressure such as occurs from drilling, or working with the job between tight centres, the lathe would always be in adjustment and would virtually never wear out. lateral turning, such as staff turning, polishing pivots, or any operation largely concerned with cylindrical surfaces causes the 45 degree thrust bearing to force the spindle out of the bearing. This means the spindle is actually contacting only a small part of the bearing, usually the spot diametrically opposite the pressure from turning. Obviously, enough of this causes an egg-shaped bearing.
Accuracy and reliability of a cone bearing lathe is not a given under all conditions. Cylindrical bearings when VERY well made, and ball bearing spindle lathes with the best bearings now easily available, are more versatile, but not necessarily more accurate than a well maintained cone bearing lathe. This is why all kinds are still made. When a cone bearing lathe is adjusted for the type service required of it, it is the equal of any kind of lathe.
For frequent normal use on the bench of a watchmaker, the bearings should be adjusted as tight as possible so that the lathe will not seize up in use. Prolonged drilling, extensive work turning between centres, or high speed turning for any length of time requires that the spindle be adjusted very slightly looser so that radial expansion of the spindle from heat does not cause the bearings to seize. Common sense suggests that heating would lengthen the spindle, and that the cast iron head would lengthen as well from heat. Actually, the heating is so localised that there is little likelihood of the entire assembly warming enough to change length dimensions, thus the contradictory rules.
To adjust the lathe. Dismantle and clean it first. Liberally coat both the bearings and the spindle with #10 or lighter oil. A good grade of sewing machine oil is suitable. Re-assemble the spindle, being sure to locate the flat spot on the spindle intended to receive the set-screw which holds the belt pulley. Failure to align this pulley makes it possible for the pulley to slip, and efforts to tighten the set screw excessively will lead to damage to the threads in the pulley.
At the rear of the spindle the lathe will have either a single split "nut" which tightens against the rear cone of the spindle assembly, or there will be a nut followed with a lock nut. Most of these nuts are round, have either a slot resultant from being "split", or there will be holes to fit a spanner wrench. In fact, either of these type nuts can be tightened by inserting a small screw driver in the split, or a piece of rod in the spanner hole. Gradually tighten the nut against the cone until there seems to be a little "drag" or resistance as the spindle is turned by hand. Wait a minute or two so that excess oil can be expressed out of the bearing, then see if the bearing still seems tight. If so, back the nut off just a hair, tap the rear of the spindle gently with a light mallet or block of wood, then re-tighten the nut half a hair.
Regarding split nuts. Keep the nut as tight as possible during adjustment to minimise thread clearance as the final tightening will generally consume a proportionate amount of clearance to seat the thread. In general the thrust is on the thread face away from the spindle and final tightening will drive the nut forward reducing the clearance of spindle thrust by the amount of forward thread clearance.
Regarding double nuts. The same as above, but more severe and with a new twist. For example, the initial clearance attained by adjusting the primary nut causes this nut to bear on the thread angle face away from the spindle thrust clearance. Considerable clearance is now available on the opposite thread face, commensurate with the degree of diametrical clearance of the nut and spindle. Now, in the process of using the secondary check nut, lo and behold the clearance goes to pot. Why? When two nuts are tightened on a common thread (spindle) each nut is driven to each outside thread face, and the thread backlash inherent in both nuts is equally divided between them at the centreline face of the two nuts. This explains, at least in my experience, why final adjustment with two nuts is not as simple as it appears. One must always under adjust the first nut to compensate for the above conditions.
To facilitate the understanding of this, perhaps another example of a commonly used adjustment that performs in the same fashion. A set screw used as a pusher for a gib slide adjustment. No problem. They can normally be adjusted equally by feel. However, some machine tools use a double set screw, and or and check nut on the screw when vibration or otherwise can cause loss of adjustment. The same inaccuracies as with the two nuts but of a lesser degree. Any additional force exerted on the screw by tightening with a check nut will cause the screw to be retracted from its original position dependent upon the torque used. Which means the screw must be tightened on its own slightly more than is required to compensate for the check nut torque. Using a secondary set screw as a locking mechanism will give identical results as using two nuts. Two nuts or two screws are identical in their performance, the only difference being in gender. I think most competent mechanics instinctively get around this problem without really giving it much thought as to the cause by maintaining snug adjustments and leaving a tweak for finally locking.
Check the ease of turning again, and repeat the adjusting process if necessary. On lathes equipped with a lock nut in addition to the adjusting nut, tighten the lock nut against the adjusting nut, but be sure to hold only the two nuts with either screw drivers, rods, or spanners, as appropriate. The reason for this is to assure the lock nut is turned against the adjustment nut without changing the setting of the adjustment nut.
At this point, whenever possible, it is reasonable to check end-play of the spindle with a dial indicator. If there is no end play and the spindle is free, the lathe is properly adjusted for most repair procedures especially. There should be no run out of the throat of the spindle, where the collet is inserted. If there is, the lathe needs special repairs, another subject.
The final test, optional, is to let the lathe run at moderate speed for five minutes or so. Unless there is perceptible heating of the bearing areas of the headstock, the lathe is fine for ordinary use. If it heats, the spindle is too tight. If the lathe is to be used for extensive single use, as in manufacturing, or if it is to be used for long periods at high speed, a looser setting is best. Under these conditions, it may be necessary to back the adjustment off slightly so that the bearings will be free when they warm up. This requires vigilant checking to get the right combination of tightness relative to use.
Don't forget about the lathe after it is adjusted. Keep it clean, and dismantle, clean and re-assemble the headstock as often the lathe has been used for dry grinding. This rule is a pain in the neck, but the time required to honour it is less than the time required to earn the price of a replacement headstock. The seals of these lathes are not really more than light dust shields. Oil the lathe frequently. Try to get the bearing to take oil as it is revolving so that you seem to be over oiling. You are really flushing the bearings, and you will have to wipe up the excess oil after oiling.
Finally, the rule of rules. Always loosen the belt on a lathe as soon as you have finished with it for the moment. If a belt is left tight on a lathe it pulls the spindle towards the countershaft or motor, and this squeezes the oil out of the bearing at that spot. When you start the lathe later, it takes several revolutions of the spindle to bring oil to that spot, so that for a few turns there is a portion of the bearing and the spindle running against each other without oil. Like the Chinese water torture, these events have a cumulative bad effect.
In the repair shop, a watchmaker's lathe should outlast at least one watchmaker. I use an American Watch Tool lathe, later Derbyshire, which I bought in 1949 from the son of a watchmaker who had used it all his productive life. The lathe is still good as new, but that is because I have given it the care it deserves.
The following article is from the British Holological Institute. The link to this fine article follows.
Adjusting The Bearings on a Watchmakers Lathe.
The British Horological Institute has archived and edited the following from e-mails sent to the Clock/Clockers mailing lists on the Internet. The information here does not necessarily indicate a method approved by the BHI, we are only publishing this digest so that others can decide for themselves whether the methods listed below will suit them.
From: John C. Losch, Leon E. Levasseur,
There are several things to know about a WW watch lathe, or any size cone bearing lathe, for that matter. Once you under- stand what the lathe is all about there will be no problem maintaining and adjusting it. Lathes of this type are called cone bearing lathes because the spindle and bearings are tapered. The spindle shapes are actually sections or "frustums" of cones, with bearings to match.
In nearly all such lathes, both ends of the spindle have a thrust angle of 45 degrees, and a lateral bearing angle of 10 degrees. When the lathe was made both the bearing and the spindle were ground to these angles after hardening, or turned in the case of bronze bearings, then lapped to a perfect fit.
If you will look at the front, or collet end of the lathe headstock of most American watch lathes, you will see stamped in the face of the casting either the word "hard" or "soft." This indicates whether the bearings inserted in the cast iron headstock are hardened and ground steel or machined bronze. There is a reason for making the lathes with both kinds of bearings beyond economics, although bronze bearing lathes were cheaper to produce and sell.
According to my friend Leon LeVasseur, who worked at the Rivett Lathe Co., manufacturer of watchmaker's lathes, bronze bearings can last as long as hardened bearings. Bronze bearings will seat in faster, and they are more tolerant of grit than the combination of both hard spindle and hard bearing because an occasional particle of foreign matter will bed-in to a bronze bearing rather than continually rotating and scoring either hardened surface at random. In his opinion, bronze bearings are better, but he agrees that there are arguments on both sides, largely controlled by the kind of service required of the lathe.
As a side note, "Frenchy" LeVasseur revolutionised the method of fitting bronze bearings and spindles when he decided lapping bearing and spindle took longer than necessary. One day, to the astonishment of workmen at the surrounding benches, he picked up a mallet and slammed the lathe spindle into the bronze bearing. Since the bearing was soft, it took the form of the spindle. The job was completed with lapping, but at the cost of considerably less time than the usual method. With soft bearings, the lapping compound was degradable so that if any bedded into the bronze, its effect on the steel spindle was short lived.
Both the life and dependability of cone bearing lathes hangs on adjustment. If one of these lathes were always used with thrust pressure such as occurs from drilling, or working with the job between tight centres, the lathe would always be in adjustment and would virtually never wear out. lateral turning, such as staff turning, polishing pivots, or any operation largely concerned with cylindrical surfaces causes the 45 degree thrust bearing to force the spindle out of the bearing. This means the spindle is actually contacting only a small part of the bearing, usually the spot diametrically opposite the pressure from turning. Obviously, enough of this causes an egg-shaped bearing.
Accuracy and reliability of a cone bearing lathe is not a given under all conditions. Cylindrical bearings when VERY well made, and ball bearing spindle lathes with the best bearings now easily available, are more versatile, but not necessarily more accurate than a well maintained cone bearing lathe. This is why all kinds are still made. When a cone bearing lathe is adjusted for the type service required of it, it is the equal of any kind of lathe.
For frequent normal use on the bench of a watchmaker, the bearings should be adjusted as tight as possible so that the lathe will not seize up in use. Prolonged drilling, extensive work turning between centres, or high speed turning for any length of time requires that the spindle be adjusted very slightly looser so that radial expansion of the spindle from heat does not cause the bearings to seize. Common sense suggests that heating would lengthen the spindle, and that the cast iron head would lengthen as well from heat. Actually, the heating is so localised that there is little likelihood of the entire assembly warming enough to change length dimensions, thus the contradictory rules.
To adjust the lathe. Dismantle and clean it first. Liberally coat both the bearings and the spindle with #10 or lighter oil. A good grade of sewing machine oil is suitable. Re-assemble the spindle, being sure to locate the flat spot on the spindle intended to receive the set-screw which holds the belt pulley. Failure to align this pulley makes it possible for the pulley to slip, and efforts to tighten the set screw excessively will lead to damage to the threads in the pulley.
At the rear of the spindle the lathe will have either a single split "nut" which tightens against the rear cone of the spindle assembly, or there will be a nut followed with a lock nut. Most of these nuts are round, have either a slot resultant from being "split", or there will be holes to fit a spanner wrench. In fact, either of these type nuts can be tightened by inserting a small screw driver in the split, or a piece of rod in the spanner hole. Gradually tighten the nut against the cone until there seems to be a little "drag" or resistance as the spindle is turned by hand. Wait a minute or two so that excess oil can be expressed out of the bearing, then see if the bearing still seems tight. If so, back the nut off just a hair, tap the rear of the spindle gently with a light mallet or block of wood, then re-tighten the nut half a hair.
Regarding split nuts. Keep the nut as tight as possible during adjustment to minimise thread clearance as the final tightening will generally consume a proportionate amount of clearance to seat the thread. In general the thrust is on the thread face away from the spindle and final tightening will drive the nut forward reducing the clearance of spindle thrust by the amount of forward thread clearance.
Regarding double nuts. The same as above, but more severe and with a new twist. For example, the initial clearance attained by adjusting the primary nut causes this nut to bear on the thread angle face away from the spindle thrust clearance. Considerable clearance is now available on the opposite thread face, commensurate with the degree of diametrical clearance of the nut and spindle. Now, in the process of using the secondary check nut, lo and behold the clearance goes to pot. Why? When two nuts are tightened on a common thread (spindle) each nut is driven to each outside thread face, and the thread backlash inherent in both nuts is equally divided between them at the centreline face of the two nuts. This explains, at least in my experience, why final adjustment with two nuts is not as simple as it appears. One must always under adjust the first nut to compensate for the above conditions.
To facilitate the understanding of this, perhaps another example of a commonly used adjustment that performs in the same fashion. A set screw used as a pusher for a gib slide adjustment. No problem. They can normally be adjusted equally by feel. However, some machine tools use a double set screw, and or and check nut on the screw when vibration or otherwise can cause loss of adjustment. The same inaccuracies as with the two nuts but of a lesser degree. Any additional force exerted on the screw by tightening with a check nut will cause the screw to be retracted from its original position dependent upon the torque used. Which means the screw must be tightened on its own slightly more than is required to compensate for the check nut torque. Using a secondary set screw as a locking mechanism will give identical results as using two nuts. Two nuts or two screws are identical in their performance, the only difference being in gender. I think most competent mechanics instinctively get around this problem without really giving it much thought as to the cause by maintaining snug adjustments and leaving a tweak for finally locking.
Check the ease of turning again, and repeat the adjusting process if necessary. On lathes equipped with a lock nut in addition to the adjusting nut, tighten the lock nut against the adjusting nut, but be sure to hold only the two nuts with either screw drivers, rods, or spanners, as appropriate. The reason for this is to assure the lock nut is turned against the adjustment nut without changing the setting of the adjustment nut.
At this point, whenever possible, it is reasonable to check end-play of the spindle with a dial indicator. If there is no end play and the spindle is free, the lathe is properly adjusted for most repair procedures especially. There should be no run out of the throat of the spindle, where the collet is inserted. If there is, the lathe needs special repairs, another subject.
The final test, optional, is to let the lathe run at moderate speed for five minutes or so. Unless there is perceptible heating of the bearing areas of the headstock, the lathe is fine for ordinary use. If it heats, the spindle is too tight. If the lathe is to be used for extensive single use, as in manufacturing, or if it is to be used for long periods at high speed, a looser setting is best. Under these conditions, it may be necessary to back the adjustment off slightly so that the bearings will be free when they warm up. This requires vigilant checking to get the right combination of tightness relative to use.
Don't forget about the lathe after it is adjusted. Keep it clean, and dismantle, clean and re-assemble the headstock as often the lathe has been used for dry grinding. This rule is a pain in the neck, but the time required to honour it is less than the time required to earn the price of a replacement headstock. The seals of these lathes are not really more than light dust shields. Oil the lathe frequently. Try to get the bearing to take oil as it is revolving so that you seem to be over oiling. You are really flushing the bearings, and you will have to wipe up the excess oil after oiling.
Finally, the rule of rules. Always loosen the belt on a lathe as soon as you have finished with it for the moment. If a belt is left tight on a lathe it pulls the spindle towards the countershaft or motor, and this squeezes the oil out of the bearing at that spot. When you start the lathe later, it takes several revolutions of the spindle to bring oil to that spot, so that for a few turns there is a portion of the bearing and the spindle running against each other without oil. Like the Chinese water torture, these events have a cumulative bad effect.
In the repair shop, a watchmaker's lathe should outlast at least one watchmaker. I use an American Watch Tool lathe, later Derbyshire, which I bought in 1949 from the son of a watchmaker who had used it all his productive life. The lathe is still good as new, but that is because I have given it the care it deserves.
