I admit this thread/task might not be as basic as some of the other watchmaking tips I’ve posted in these other threads:

https://omegaforums.net/threads/basic-watchmaking-tips-cleaning.56365/#post-696021

https://omegaforums.net/threads/basic-watchmaking-tips-oiling-part-1.62310/

https://omegaforums.net/threads/basic-watchmaking-tips-oiling-part-2-the-mainspring-barrel.71246/

https://omegaforums.net/threads/basic-watchmaking-tips-oiling-part-3-the-wheel-train.84482/

https://omegaforums.net/threads/basic-watchmaking-tips-spotting-wear.81025

/https://omegaforums.net/threads/basic-watchmaking-tips-oiling-part-4-the-escapement.87072/

However I wanted to address one step in this process that is often cited by amateur watchmakers as the reason they can’t change a balance staff. To do that I might as well go through the entire process.

So first off why do we need to change a staff? Well the most obvious is a broken staff, or a bent staff, but those are actually not common on watches with shock protection on the balance, so that leaves us with a worn or rusty staff. In the example I’ll show you, it’s a worn staff in a Cal. 321 movement that I have in the shop for servicing that is in rough shape in several ways. Here is the “good” end of the balance staff:



And here is the end that has a clear groove worn in it:



This warrants a closer look, so I used my higher power microscope to take this photo again of the good end:



And the end with the groove in it:



Note that in addition to that groove, the very end of the staff at this end has been worn flat from running without lubrication, so this staff certainly needs to be replaced.

So first off there are different methods in how balance staffs are actually attached to the balance. For example some vintage American pocket watch makers used what are known as "friction fit" staffs, and as the description implies they are fitted by friction:



In this thread I will address the most common method of staff attachment - the riveted staff.

So the first step in replacing a balance staff is to mark the balance so that the balance spring and roller are placed back on in the same orientation. The next step is to remove the balance spring, so here I have set the balance on a piece of plastic that has a hole in it so the wheel can sit flush on the plastic, and I’ll use a pair of balance spring levers to get under the balance spring collet and lift the spring off the balance:



Now removed:



The next step is to remove the roller, and there are a few different tools that can be used for this, but a simple roller remover works for me:



The remover has an inclined surface that acts as a wedge as you slide the balance into the tool, and when you get it in as far as it will go, a very light squeeze of the handles together will pop the roller off:



Here is the roller:



Now we have a bare balance – this is the rivet side:



And this is the hub side:



Let’s take a closer look at both sides under the microscope – the rivet side:



A bit closer:



The hub side:



So one area of controversy in watchmaking is how to remove a staff like this one that is riveted in place. One school of thought is that the staff can be punched out by force, so striking the staff from the rivet side, which will break off the rivet and free the staff. Personally I’m not a big fan of this method no matter what tool you might use to accomplish it (Platax tool, Hoira tool, staking set, etc.) because even if the rivet breaks off cleanly, there will be an enlarged portion of the staff (bulge under the rivet) that is forced through the hole in the balance, and this will potentially enlarge the hole and make riveting a new staff on more difficult. Now there are plenty of watchmakers who swear by the “punching it out” method, but it’s not how I do it.

I was taught to cut the staff out using a lathe, and looking back I have some photos from a previous staff replacement already in my Photobucket account, so this shows the process of cutting the staff out. So here is the balance with the spring and roller removed, and in my lathe ready to cut the hub end of the staff:



Here the cutting is in progress, and for this work I use a hand held carbide graver:



By cutting from the hub side, when you remove the hub it allows you to pull the balance straight off the staff, not having the riveted side go through the hole in the balance at all. Here I have cut the hub off and you see a small ring of material that has been cut free (I recall in watchmaking school having to hand in that little ring, so it was something all the students had to be able to do):



And the staff removed from the wheel:



Now this method requires proficiency with the lathe, and it doesn’t come without risks. One slip of the graver and you could throw the balance out of true, gouge it up, or otherwise damage it. I don’t have any difficulties using the lathe to do this sort of work, but I often come across amateur watchmakers who don’t have a lathe, so how would they remove the staff other than punching it out? For the 321 balance I thought I would use a tried and true method of removing stuck steel parts – soaking it in alum:



NOTE – since alum is very good at dissolving carbon steels, you obviously can’t use this method on a balance that is made of steel or has steel as a part of the actual wheel portion. This method is only for balances made of materials like Glucydur (as in this case) or other non-ferrous materials that are not affected by alum.

So I mixed a heaping teaspoon of alum with some warm water:



Here is the small glass with the alum dissolved in the water on my bench – you want a saturated solution of alum, so dissolve as much in the water as it the water will take:



And I place the balance in the water, and wait for the alum to start working:



It doesn’t take long before I see a good stream of bubbles coming from the staff:



Now I’m not in a rush doing these jobs since I always have several watches in progress at the same time, so on this one I left it for several hours. It didn’t appear that much had happened, but when I removed the balance and checked how tight the staff was, it came right out:



Now one small drawback of this method is that the balance was a little discoloured at first:



So I did a little manual cleaning of it, then placed it in a jar of cleaning solution:



And dropped that in my ultrasonic tank for a while:



Once that had gone through a few cycles, I dried it off and then ran it through my regular cleaning machine in a small basket:



This is what the staff looked like by the way:



So here is a photo of the hole in the balance after it was cleaned, and you can see the hole is pristine and has no damage:



Here is the new staff:



Now it’s always a good idea to check the fit of the new staff before you rivet it to the balance, so here I have installed it in the movement, and I check things like end shake and side shake to make sure this is really the right staff:



I then test the fit of the staff in the balance – often with a balance that has had the staff punched out, the balance will essentially fall right on the staff, and in severe cases the hole will be significantly larger than the staff. This can lead to issue riveting the staff on, centering of the balance on the staff, etc. But in this case, I had to use a tiny bit of force to seat the staff into the balance, so the fit is very good:



This is what the rivet side of the staff looks like before it is riveted to the balance:



So next we want to rivet the staff to the balance, and to do that I use my staking set. I have a K&D “Inverto” staking set, which means all the punches (that come down from the top) can also be used as stumps on the bottom, and to me this is a must have in terms of staking sets as it provides you will endless options. This is important when staffing because you want to select punches that have very little clearance around the items you are using them on, so having more options will give you the best fit possible.

Here I have selected a punch to use as a base for riveting the staff, and I have placed the staff (rivet side up) in the punch:



Next I place the balance on the staff:



I start with a round nose punch, as I want to spread the rivet on the staff out first:



Lining the punch up exactly is critical, so you don’t damage the new staff:



You bring the punch down on the staff, and use your watchmaker’s hammer to tap the punch. It’s good practice to rotate the punch as you tap, as that will even out any errors in the event that your punch face it not perfectly flat to the base. Since these parts are all very small, you don’t have to hit hard, so really just letting the weight of the hammer fall on the punch:



I the switch to a flat faced punch, to fatten out the rivet a bit:



Same procedure...tap and rotate:



A look at the rivet, and it looks good:



But how do we know it is tight enough? There is a simple test – take a pin vise and secure the rivet side of the staff firmly in the vise, and place a very small rolled up “log” of Rodico putty on the opposite pivot. You then hold the balance with your fingers, and try to turn the staff with the pin vise – if the staff turns you will see it via the log of Rodico turning, and you need to go back and rivet the staff more. This one was fine – it didn’t budge:



The next step is to check and true the balance as needed, so I use the truing calipers for that. They have jewels on either end that the pivots of the staff fit into, and you spin the balance and use the gauge at the side to look for a wobble. Since we didn’t distort the wheel by punching the staff out, this one spins perfectly true so no adjustments required:



Note that one common side effect of having a hole in a balance that is too large is that due to the excess amount of riveting required to spread the rivet out to fill the hole, this can lead to the balance getting distorted in the flat or in the round. I’ve had to fight balances where the hole is oversized, and it almost always ends up in more work truing the balance after.

The next step is to install the roller, so again switching punches around to find a close fit, and then place the roller on the staff:



You bring the punch down on the roller, and you just press it on – no hitting with a hammer required!



So this is the stage where many books would tell you that you need to statically poise the balance. I’m going to give a brief review of that before I tell you why you probably shouldn’t do it. So you get out your poising tool, set it up directly on the solid wood surface of your bench, and place the bubble level on the ruby jaws of the tool:



You level the tool, and when done I use a pencil to mark where the feet of the tool were when level in case is moves. I also level the tool with my arms on the bench as if I was working, because if I level it with no load on the bench and then lean on the bench, it will go out of level. Once the tool is level, you place the balance on the tool, and I use a hair from a dial brush on a stick to move the balance and get it rolling:



The poising process is pretty simple – add or remove weight as needed so that the balance will not stop in any particular position, or swing like a pendulum. It should roll and stop in random positions, and then you have perfect static poise – great!

The problem is you likely just messed up the factory poise.

When a balance is poised statically, there are some obvious parts still missing from the assembly – collet, collet pin, balance spring, stud, and stud pin. So all this trouble perfectly poising the balance statically essentially goes out the window once you install the balance. When you mount it in the watch, and fix one end of the balance spring using the stud, now that adds other effects on the balance that make the poise act differently in use. There is a method of poising that accounts for all this called “dynamic poising” and it’s done using a timing machine and taking specific tests to determine where material must be added or removed from the balance. So for this watch I’m going to make an assumption that it was poised well as a system already, and I’m going to check the timing before I erase the factory poising.

One reason I can do this is that again I didn’t have any distortion in the balance wheel when riveting, because the staff fit the hole perfectly. If the hole was oversized, or I had to perform a lot of truing of the balance, then I would be more inclined to static poise the balance since I know the staffing operation has likely thrown everything out of whack anyway. So I would typically only static poise if I feel there is some sort of gross error that needs to be corrected.

So next I need to install the balance spring, and this requires a different thought process in the selection of the punch at the bottom – here is the balance showing the roller installed:



If I draw an imaginary circle want to select a punch as the base for this operation that has a hole in it that is just larger than the roller jewel, like so:



This will support the balance in a way that won’t cause any pressure or deflection on the arms or rim of the balance. So here I have that punch selected and installed in the staking set:



I place the balance spring on the balance:



One again no hammering, just pressing it on using my fingers on the punch:



The staff has now been changed:



From here I carry on with my normal servicing procedures, checking balance amplitudes, beat error, and of course positional variation:




So although this can be a daunting task for people the first time through, it’s actually a very straightforward task in many ways. I hope that with the use of alum it’s something that more people who are already working on watches may try, and of course if anyone has questions, please let me know.

Cheers, Al

 

Comprehensive description, Al. I use a lathe to do this as well but, I like the alum method you show. I might give that a go on a spare balance as you can be doing something else while it's happening. Thanks!

Cheers, Chris

 

I'm amazed!

Amazed you give us work this good when surely you could get paid well for such information

Thank you.

 

I love these basic watchmaking threads!
I happen to have a watch that needs this work done to it, and I appreciate knowing what it will take to get the job done right

 

I really appreciate you educating us this way.
Thanks Archer.

 

So interesting, I wish I had become a watchmaker. Thank you.

 

Always fascinating.. now I understand the consequences of dropping my 13.33z Longines upon unpacking a few years back. Broke the staff ...

 

I’m in awe, just out of interest does anyone have an opinion on how many young people are going into the watchmaking profession? I see so much in the media about how coding and app development has fantastic prospects, I can’t imagine it gives the same sense of satisfaction as the work Archer does.

 

Glad I was a butcher by trade after school after reading that @Archer

 

Superb post Al.
I had a recent balance staff change. I turned the rivet side down on the lathe and even though I was down to he spoke, it still looked too big .
I punched the remainder out and when I tested the new staff it was loose to the point I realised I would need a new wheel.

Even if I had been aware of turning off the hub, I don't think it could have saved it.

But now I know another "trick" or two.

Thanks mate.

 

Very informative, thank you!
If i understand the title correctly, these are basic watchmaking tips, not to confuse with basic watchmaking techniques
Otherwise i would wonder what are the advanced techniques are..

 

If I had to guess, less than 100/year if you look at the school options in the U.S. I’ve started taking classes the AWCI when I can, with the goal of 2 weeks of hands on training per year. I’m doing this from a hobby stand point, hoping to build some decent skill over the next 5 or so years. Full time watch school isn’t an easily accessible option for most people I’m guessing, and skilled trades aren’t something that people want to go into. It’s something we struggle in my current profession as well

 

Archer,
Once again thanks for the great post! I’m going to try the alum next time I replace a staff in a balance that would allow it. I’ve be buying cheap broken stuff off eBay to practice with. The last staff I replaced was from an elgin pocket watch. That staff had been affixed with shellac from a previous repair. I was able to tighten the hole in the balance a bit with the staking kit and rivet a new staff in place. The balance was way out of true, it took quite a while to true it. It still have to replace the roller jewel which I broke in the process and poise it.

Joe

 

So I wanted to follow-up on this post and show how it all turned out. So for a start the watch that this balance came from was in rough shape when I received it. The watch arrived with the upper balance jewels missing, and the shock spring laying in the movement:



In addition the regulator on this watch had been worked on by someone who...well I'm not really sure what they were thinking. I'm guessing one of the pins for the regulator had broken off, and this was their attempt at correcting that problem, using copious amounts of glue:



So in addition to needing a new staff, it needed some other work certainly. A new regulator was installed, the staff was changed, and here is the first set of timing checks right after the service:



Closer look shows a watch running very fast with the regulator set in the middle of the range, a high beat error, but the Delta isn't bad considering the initial condition. The Delta is the difference between the fastest and slowest position. To me this result of 24.2 seconds over 6 positions at full wind confirms that the balance was poised pretty well to start with:



Now one way of correcting the rate is to simply move the regulator, but it would have had to be way off to the slow side to get the rates in a proper range, so the right way of doing this is to add weight to the balance, and for that I use timing washers:



The balance is removed and hung from the balance tack:



I remove one of the balance screws, and here I've placed it between a regular ball point pen, and the screw that holds the balance cock in place just to give you an idea of the scale:



After adding washers to two screws that were opposite each other, the rates came down to a more normal range. I then corrected the beat error, but at this point the Delta was still in the mid-20's range:





Omega requires that this movement have a Delta of 15 seconds at full wind, measured over just 3 positions. I always measure over 6 positions, so when you do that the Delta is more likely to be a higher number. Looking at the last result there is a clear poise error between position 4 and 6, so to correct that I use the dynamic poising process. I am measuring just vertical positions (8 readings) and these are taken with the balance amplitude reduced, which makes the errors more visible in the timing results:



The machine takes 6 readings on the screen, so these are the last 2 on a second screen:



I use this information to know where to add weight on the perimeter of the balance in order to correct the errors. This process is similar to balancing a car tire, but of course on a much smaller scale. It is an iterative process that requires very small changes, and in the end this is the final result:



As you can see the Delta is now just 10.5 seconds over 6 positions, and this is actually within Omega's standards for their COSC watches, as those (and many of the METAS calibers) allow up to 12 seconds over 5 or 6 positions:



Not bad for a watch from the mid-60's. But this confirms the point that often static poising simply isn't required after a staff change, if the staff has been changed in a way that doesn't distort the balance.

Cheers, Al

 

Thanks Al for posting all this. I must say I do a lot of DIY stuff just for fun, but watchmaking I will stand in the sidelines and just enjoy your work!

Keep up the great content.

 

Thank you so much @Archer ! Your knowledge and willingness to take the time to share with us is huge. Although, I will never do this sort of work myself it is good to understand what is involved. It makes me appreciate watchmakers and understand what they need to charge.

 

@Archer Thank you so much for your post. Very informative. Would you mind explaining how the position of the roller and hairspring are marked? Fine point sharpie? If so, how does this interact with the alum?

 

Yes, that would work, and since you are marking the rim of the balance it has no affect on removal with alum.

So since the debate rages on in some circles about the best way to remove a staff, and the "punchers" still say it's fine to do, someone on another forum took a brand new balance that had never been staffed, installed a riveted staff, and then ground the hub off and carefully removed the staff.



The red arrow points to the area of the staff that has been riveted, so you can see how much expansion there is when this is done.

Here is another photo the same person posted of a staff that has been punched out, and you can see how "cleanly" the rivet breaks off (or not in reality):



You can see how ragged it is, and that there is a portion of staff that is enlarged that had to have gone through the hole in the balance.

To me, punching out staffs (no matter what organization says it's okay to do) is hack work...

Cheers, Al

 

I'm also curious about marking the balance to preserve the phase of the balance vs collet and balance vs roller table. I presume both can be installed in any position but both should be restored to their original positions with regard to some fixed point on the balance.

Do you put a mark on the balance for the collet and another for the table, and what features do you chose? I would suspect the impulse jewel and the slot in the collet.

You confirm a sharpie but also imply that it's removed by the alum. Doesn't the mark have to survive the alum bath? Perhaps I'm misunderstanding (imagine that) but I would think that whatever you use as a mark should survive the alum and cleaning but then be removed another process once the balance is completely reassembled.

I hope I can get through life without ever having to restaff a balance, but I do plan on removing the hairspring from one watch to straighten it out at some point in my future. I will have to remove and reinstall the collet and will face the indexing problem.

Looks like I need to start thinking about a staking set.

 

I use a marker - it stays on there fine. I mark on the rim of the balance - actually on the OD. I use the roller jewel for one side, and the balance spring stud for the other.

The reason this is necessary, is that the balance has been poised with those parts on the balance in the specific locations they are in. Since these parts are not uniform weight, the balance has been poised to correct the errors they create (as well as any that might be in the balance wheel itself). If you change the location of where those parts are, it throws off the poise that was already done on the balance at the factory, and possibly through the work of other watchmakers who have fine tuned things.