A Day in Watch School Part 5

Aug 07, 2013,16:24 PM

A Day in Watch School Part 5

Getting Serious

by ei8htohms
© 11-10-2002
oiling an ETA 2892

There's nothing like going back to school after a long summer vacation.  At least, that's the way I remember it.  Of course, a long summer vacation when I was growing up was three months or so.  Going back to school after five weeks off is only OK but it was fun to play the role of the wizened second year students (hazing is not encouraged in watch school).

 1st year students everywhere!

The incoming class was full to the gills with twelve students at twelve benches and they even had a new instructor, the inimitable Henry Hatem - okay, actually we do sometimes imitate his sinister laugh, but only with the fondest of intentions.  Henry brings a wealth of manufacturing experience and know-how to the classroom from his years of proto-typing and precision machining (on top of his watchmaking experience that is).  Elaine Rolf would be our (the second year) instructor for another year, where her extensive experience with repair and adjustment could be properly exploited.  While we certainly felt well schooled by Elaine in the manufacturing portion of the first year class, I often find myself bending an ear to catch the technical explanations, tips and tidbits that Henry heaps upon the unsuspecting heads of his pupils.

By the end of the fourth quarter, we had completed the checks and adjustments of the barrel and power train, dabbled with the lubrication of shock jewel settings, and toyed with hairsprings for a moment.  As the fifth quarter began, we turned our attention to lubrication issues in earnest.

I need a shave

Properly lubricating the power train of a watch is something that really must be learned from experience.  While any amount of lubrication, from nearly non-existent to swimming with oil, will probably perform adequately in the short term, the long term performance of a timepiece relies on precisely allocated amounts of lubrication, applied with great precision and scrupulous cleanliness.  One of the best ways to see if a pivot has the right amount of oil is to remove the bridge and inspect the underside of the jewel bearing.  A clean looking "doughnut" of the appropriate size in the middle of the jewel is a good indicator.  The problem being, once the bridge has been removed, the jewel and pivot must be cleaned of any remaining oil before the bridge can be reinstalled and then the pivot can be lubricated again.  What this means is, there's no good way to tell if a pivot has the right amount of oil without disassembling it and starting over, so that's what we did; over and over again.

After doing this umpteen times, one begins to understand how much oil on the oiler will translate to a doughnut of the proper size on the jewel and one can confidently leave a bridge in place with the knowledge that the amount of oil around the pivot will protect it for many years without running down the pinion leaves or slopping onto other nearby surfaces.  Of course, confidence on the part of the student does not necessarily translate to confidence on the part of the instructor, so oiling was the focus of our first test of the quarter.

little doughnuts of oil roughly half the size of the jewel

From there we went on to learning how to re-staff a watch.  Staffing involves removing a broken balance staff and replacing it with a new one.  While modern shock protection has greatly lessened the frequency of broken balance staff pivots, it is still often necessary to replace a balance staff because the pivots are worn, bent or otherwise imperfect.  Elaine relayed that when she worked in a service center, if they had trouble gettting a watch to perform well and all other possibilities had been exhausted, they would frequently replace the balance staff and as often as not that did the trick.  It is one of the most common repairs of a mechanical watch and should ideally be a quick and routine operation even though it involves highly invasive operations on one of the most sensitive and critical components in the movement.

my favorite kind of tool: cheap, readily available and effective

The first part of staffing a watch involves removing the old staff.  Before you can do this, you must remove the hairspring (after marking the position of the hairspring stud on the balance rim so you know where to line it up when you put it back on) and the roller (after likewise marking where the roller jewel lines up on the underside of the balance).  The collet is lifted off with a pair of levers while being careful not to mangle the hairspring and the roller is removed using any of a number of roller removers.  Conveniently, my favorite variety for most situations is the cheapest and most readily available.  While there are adjustable jaws that fit into a staking set, exotic scaffolds designed for removing rollers of various sizes and other elaborately conceived tools, this simple piece of nickel with a slot with blades on either side of increasing thickness seems to do the best job.  You simply slide the balance into the slot with the blades under the roller until it's snug, then squeeze the open end of the tool together until the blades pop the roller off.  Don't get me wrong, it's still possible to crush or otherwise damage the roller or snap the staff in the middle, but I have much better luck with this one than with the more exotic and expensive tools for the job.

cutting out the old balance staff

Once the hairspring and roller have been removed, the staff can be removed from the balance arms.  The most traditional and often safest way to do this is it to cut the staff out on a lathe and this is the first method we learned.  Although some texts recommend just cutting the small rivet out from the top of the staff, our instructor recommended cutting through the entire hub from the bottom.  The advantage being that you have more material to cut through and a larger diameter to keep you from slipping and gouging the arms.  I tried cutting the rivet out of a  staff from the top a few times before taking the class and had some luck with it but I must admit that cutting it out from the bottom is safer and only takes a few minutes longer.  The only real disadvantage I can see to cutting out the entire hub is if one needed the staff more or less intact to use as a reference for making a new one.  In that case, carefully cutting out the rivet would probably be advised.

Many modern balances employ a collet that cannot easily be removed and in these situations, it's best to knock the staff out on a staking set or using a specialized tool.  The crimped Greiner collets used by ETA and most other manufacturers as well as the laser welded Nivatronic and Rolex collets are all of this variety and it's much safer to punch out the staff without trying to remove the collet using conventional methods.  The punch out method can also be used successfully on most modern balances (with the hairspring removed) and is much quicker and easier than cutting the staff out on a lathe.  Many staking sets come with a tool for supporting the arms of the balance while the staff is punched out but it is not advisable to try the punch-out method on a split balance as you will likely distort the arms and rim considerably.

Once the old staff has been removed from the balance, it's a good idea to clean up the hole in the balance arms a little.  Any burrs around the hole can keep the balance from sitting flush on the hub of the new staff and result in trueness issues after riveting.  With the hole cleaned and double-checked, the balance and new staff are placed on the staking set for riveting.  A round-nosed punch with a hole just large enough to fit over the staff (down to the rivet) is selected and, after carefully checking that everything is properly aligned and the balance is sitting squarely and flush on the hub of the staff, it is used to spread the rivet out with a few taps of a watchmaker's hammer.  Some watchmakers prefer the single, good-strong-whack approach but, at least as a newbie, I'm more comfortable giving it a series of taps while rotating the balance and punch in between taps.  This way I'll know that any irregularities in the punch or staking table can be accounted for and I'm less likely to destroy the staff and/or the balance with an overzealous blow.  Once the rivet has been pushed out enough to securely hold the balance, a flat nosed punch is used to flatten the rivet in place.

cutting into the underside of the hub on the staking tool whacking it with a hammer riveting

With the balance riveted onto the new staff, the balance must be checked for trueness, just as train wheels are.  Subtle corrections are made to insure that it is absolutely flat at the rim and then the roller is replaced so that it can be checked for poise.

my second favorite kind of tool: effective and readily available but very expensive checking the balance for truth in the flat isn't that little level cute?  it's just for show

Static poising is accomplished by placing the balance on a perfectly level poising table with the cylindrical portions of the pivots resting on the ruby jaws.  Don't be fooled by the cute little level/bubble on the poising tool shown.  All the levels in the latest batches of poising tools the school has received have been completely useless.  We picked up some cheap plastic levels at the hardware store that work much better (which is to say that they work).

careful, careful

If the pivots are perfectly smooth and clean and the jaws are perfectly smooth, clean, level and at the same height, the balance will come to rest with the heavy spot at the bottom.  If the balance has screws on the rim, some weight can be removed from the screw nearest the heavy spot using a variety of different methods (undercutting, beveling, countersinking, beveling the slot, etc.) or if the balance is smooth, a tiny amount of material is removed from the underside of the rim with a small drill.  The balance is then checked again for poise and the process repeated until the balance will stop freely in any orientation, indicating that it is in poise.


While the poising process can sometimes result in a number of small holes in the underside of the rim, with some care and forethought, a balance can be poised with a maximum of three holes.  By removing the largest amount of weight from the hole closest to the heavy spot and a smaller amount from the holes next closest (the amount of weight removed being proportional to the proximity to the heavy spot), the balance can eventually be brought to poise without scarring the rim as noticeably.  Once the balance has been brought to poise, the hairspring can be replaced and the assembly can be cleaned and placed in the movement for further checks, adjustments and dynamic poising.

After practicing staffing for awhile, we moved on to what WOSTEP calls the "divisions".  This means checking and adjusting the endshake and vertical alignment of the escape wheel, pallet fork and balance.  This can be a complicated process as several components have to line up with each other with one or two 100th's of a millimeter of endshake on each one.  The escape wheel must strike the center of the pallet stones, the fork horns must positively contact the impulse jewel without touching the roller and the guard pin (safety finger) must contact the center of the safety roller in order for the escapement to function properly.  These checks can be performed in just about any order, but the best method I found was to start with the component with the least possibility of adjustment.

Viet sabotages poor Evan (1st year student) almost daily with dubious information and questionable techniques

For instance, often times the upper jewels for the balance pivots can not be adjusted (in older watches neither the upper or lower balance jewels can be adjusted without bending or otherwise altering the balance cock) so it's a good idea to start with the balance.  After setting its endshake properly, the correct height for the fork horns and guard pin will have been established by the height of the roller.  After setting the pallet lever to the correct height for the horns to positively contact the roller jewel (and sometimes bending the guard pin slightly up or down to line it up with the safety roller), the escape wheel can then be set to the right height to contact the pallet jewels squarely.

While this sounds pretty straight forward, it's not always so easy.  Sometimes the pallet fork cannot be made to line up with the roller without rubbing the underside of the pallet bridge or the bottom of the plate.  In this situation it might be necessary to bend the pallet lever slightly to correct it (this is pretty rare unless some other watchmaker [or watchmaking instructor] has sabotaged the watch beforehand).  The pallet lever can be moved up or down on its arbor or its jewels can be raised or lowered (slightly), so deciding which way is best might require a little judgement.  Occasionally by the time you get the height of the escape wheel set properly, the escape-wheel pinion is no longer lining up well with the fourth wheel so the entire power train (or some portion thereof) must be adjusted to make everything work.

The real trick though is seeing what the heck is going on in the first place.  It's not so hard to see all of the interactions in a Unitas 6497 (because of its large size and good visibility) but when you're trying to see inside a 5 or 6 ligne lady's wristwatch movement, it can be quite a challenge.  Sometimes the only way to see what's going on is to put a little grease on one of the components, let it interact with the others and then remove it to see where the grease left a trail.  This works especially well on escape-wheel teeth and pallet stones but can also be used on the fork and guard pin if necessary (of course, they must be cleaned thoroughly afterward).

Once the divisions are all dialled in, the real fun with the escapement begins.  From here we began to study the multitude of checks and adjustments that might be necessary when observing the escapement from above (as opposed to from the side).  Horn shake and symmetry, guard-pin shake and symmetry, proper drop and locking depth are all critical for a properly functioning escapement.

Testing the horn shake begins with just making sure the roller jewel fits in the horns of the pallet lever.  The only time this will really be an issue is if the lever, the roller jewel, the roller or the entire balance complete has been changed so it's not really necessary to perform this check on its own in most situations.  Basically though, you want to make sure the roller jewel can fit easily inside the fork.  As with most of the escapement checks, there is a certain amount of judgement that must be exercised with regards to what is a good fit.

horn shake

Once it has been confirmed that the roller jewel can easily enter the horns without binding, the balance is held so that the roller jewel is just entering the fork, and the freedom of the pallet lever - between the lever hitting the banking and the inside of the fork horn hitting the roller jewel - is examined.  This distance must be small enough so that the pallet stone cannot come unlocked but slightly larger than the guard pin shake (which we'll get to in a moment).  If the shake is not right, it might be necessary to adjust the banking.  In a watch with banking pins, this is accomplished by turning their eccentric mounting posts (in older American pocket watches especially) or (lacking those) bending the pins slightly while making sure that they remain parallel.  It is also critical that the horn shake is the same on both sides, so the banking pins might have to be adjusted for this reason as well.  If the watch has fixed banking there is really nothing that can be done except in extraordinary situations where filing down the banking might be necessary (assuming changing out the fork and all other less invasive options have been exhausted).  Although some "watchmakers" like to fiddle with the banking pins to try to correct all manner of ailments, it is really only in relation to the horn shake that the banking should be adjusted at all.

guard pin shake

Once the proper horn shake has been established, we can check the guard pin shake.  This is tested by rotating the balance until the roller jewel is completely clear of the pallet lever and checking the freedom of the pallet lever between the banking and the safety roller.  The guard pin shake must be slightly smaller than the horn shake so that the roller jewel will not strike the tip of the fork horns when it tries to enter the fork, and must be equal on both sides.  If the guard pin shake is too large, the guard pin must be pulled out a little (assuming that's an option), peened (tapped with a rounded punch to stretch it a little) or replaced.  If the guard pin shake is too small, the guard pin must be pushed in a little (assuming that's an option) or filed to the proper length.  If the guard pin shake is not equal on both sides, the guard pin should be bent until it is perfectly centered in the pallet fork.  The guard pin shake should be tested at several orientations of the balance so that any inconsistencies in the safety roller might be observed.

Now that the proper functioning of the escapement on the roller and fork side of the equation has been established, the much more subtle interactions between the escape wheel and pallet stones can be observed and adjusted if necessary.

Most of the proper functioning of the escape wheel and pallet stones has already been ensured (or not) by the time a watchmaker sees it.  The geometry of the escape wheel and pallet lever are critical and subtle changes can have dramatic effects.  If the angle of either of the stones is off or the spacing between them is not correct, the chances for a fatal failure are great and there's not a whole lot that your common watchmaker could do about it.  In a rare and extreme circumstance it could conceivably be necessary to bend one or the other of the lever's arms slightly (and it would take a very experienced and knowledgeable watchmaker to be able to make this determination), but other than that diagnosing the problem and replacing the parts is the best one could hope for.

The primary adjustment that a watchmaker can readily make to the pallet stones is to adjust the depth of their engagement with the escape wheel.  This alone can make the difference between a watch that is barely running (or not at all), a watch that is tripping and knocking (with excessive amplitude) and a watch that is running optimally.

locking depth

The line on pallet jewels from our textbook is that they should lock the escape wheel to a depth of between 1/4th and 1/5th of the impulse plane of the jewel.  That is, the distance between the leading corner of the escape wheel tooth and the locking beak of the pallet jewel at full lock should be equal to approximately 1/4th to 1/5th of the entire length of the adjacent, impulse face of the jewel.  In practice, some very fine watches might have a locking as light as 1/6th or so and some, uh, not so fine watches might lock up close to 1/3rd.


The important thing is that they lock as shallow as possible while also locking consistently and securely.  The drop (the act of the escape wheel tooth "dropping" onto the locking plane of the pallet) and locking (including the draw action of the escape wheel tooth on the pallet stone) are checked with the balance removed from the movement and approximately one turn of power on the barrel (it's important not to check the drop and lock with too much power on the mainspring or the escapement will seem to skip ALOT).


The pallet lever is gently manipulated back and forth (I use a thin brass pin to keep from scratching the steel lever but an oiler works fine if you're very careful) and the depth of locking is observed on the entry and exit stone.  The lever should return to the banking when disturbed slightly and snap over to the other banking when pushed a little farther.  This action, the security and consistency of the drop and the depth of locking are observed for at least a full revolution of the escape wheel so that any eccentricities in the escape wheel teeth are accounted for as well.

checking the drop and lock

If the pallet lever flicks from one side of the banking and back again (once or several times), this is called skipping or tripping and means  one or both of the stones need to be pulled out a little.  The drop on either stone is dependent on the depthing of both stones so when skipping is observed, if one of the stones has a shallower total lock than the other, it is pulled out a little.  If both stones are equal in that regard, they must both be pulled out slightly.  An invaluable tool for this procedure is this little pallet stone adjusting tool and the accompanying electric heater.

Traditionally a brass clamp and/or plate of various shapes was heated over a flame with the pallet lever in place until the shellac got soft at which point a very steady hand and some stout tweezers were used to move the stone in or out as the well-trained eye deemed necessary.  As the difference of only a hundredth of a millimeter or so can be crucial when dealing with small watches, this could be a very frustrating activity and almost always involves a little trial and error.  There is also the danger of heating the pallet fork a little too much and discoloring it or, god forbid, annealing it.

This little escapement tool allows for very precise adjustment of the stones by reading the amount of change on the little dial while the whole tool is heated just enough to melt the shellac with no risk of adversely affecting the steel.  The increments on the dial don't relate to any actual measurement as far as I can tell (and couldn't really with much accuracy as the exact angle the pallet stone engages with the measuring arm varies depending on the size of the pallet lever), but you can use them as a rough guide and then make any further adjustments relative to the number of increments the previous adjustments included.  With a little practice and a little trial and error, very precise adjustments can be made.  For instance, by adjusting the depth of the pallet stones on an ETA 2892, it was relatively easy to give it a running amplitude of 320°, 300° or 280° without changing any other variables.

pallet stone adjusting tool and heater adjusting the pallet stone depthing

You'll notice that the pallet lever shown in close-up above has no visible shellac on the underside (facing up in the picture).  Although in some finer watches of yesteryear it was common to scrape any excess shellac from the underside of the pallet lever (leaving just the small amount in the slot around the jewel), it is common practice at all levels of manufacturing today to leave a small dot of shellac on the jewel and lever.  Although arguably this convention has changed as a symptom of a general decline in craft, a visible spot of shellac will undoubtedly hold the stone a little more securely.  The lever shown has had all the shellac removed (to start from scratch) and is being adjusted prior to putting the shellac in place.

After practicing all of the steps of re-staffing a balance and checking and adjusting the escapement on our kit watches, we took a test on these skills (the fourth intermediate exam).  The test consisted of this:  Cut out the balance staff and rivet the new staff (including truing and poising of course); check the endshakes and divisions of the escapement and balance, and make corrections; and finally check the various escapement interactions and make corrections including horn shake, guard pin shake and pallet stone depthing (inserting pallet stones from scratch).  All of this in under 4 hours.  It was a bit of a crunch and frankly I didn't do as well as I might've hoped on this one.  I did pass however, and that's what's really important, isn't it?

After the intermediate exam, we began to work in earnest on hairsprings.  That though, is a story for another day.

vibrating a hairspring


For more information about career opportunities in watchmaking, check out the Watch Technology Institute at North Seattle Community College.
Visit Now


Copyright March 2002 - Mr. John Davis and ThePuristS - all rights reserved


  login to reply

Comments: view entire thread


Outstanding lessons

 By: nilomis : August 8th, 2013-07:07
John, Thank you very much. For sure, I go to share this and the other lessons with my watchmaker team. Cheers, Nilo

Went through the same things at my school

 By: Ophiuchus : August 9th, 2013-09:06
We did the exact same things, with minor differences. I noticed a very exotic balance truing tool- we did not have one of those. Any advantage over simple bergeon lyre poising calipers, which we used for wheels, and balances? Good article, great coverage ... 

The exotic truing tool belonged to a student

 By: ei8htohms : August 9th, 2013-11:08
It's a great tool that makes truing easier because of the stable base and adjustability, but obviously not necessary to do the job. A luxury tool. :-) _john