How To Repair A Double Roller Balance Staff

I bought a 'working' Waltham 1899 model Riverside 19J from ebay and it arrived with a broken balance staff. My 'limited' experience so far has been simple stuff - stripping/cleaning pocket watches + fitting new mainsprings - with some success!

As the staff is a push-fit, and after watching some Youtube videos, I thought I'd have a go trying to change it.

So, my 'new' old Boley and Leinan staking set has arrived and I have a spare broken Waltham balance to play with. I've got the roller off and have a diagram as to what needs doing :

But now I'm sat looking at the stakes and stumps wondering which to use.
Any suggestions - apart from taking it to someone who knows what they're doing?
Cheers
Mike

Dave Coatsworth

One thing to keep in mind with the model '99 is that they actually used two different friction staffs during production - the #4860 'straight shoulder' (earlier) and the #4861 'taper shoulder' (later). So check your staff when you have it removed. If it looks like the staff in your diagram, it's the taper shoulder. If the hub is more square, it is the straight shoulder.

It's the tapered #4861. I'm wondering if there are special tools just for this job, or whether standard staking sets will have the necessary bits.

Dave Coatsworth

Yes, K&D did make specific stakes for Waltham friction staffs. These are shown in your diagram. Unfortunately, the diagram does not give the stake numbers and I don't remember them off the top of my head. I'll try to look them up later.

If the watch was listed as working, I'd send it back. It happened to me a few times, and the seller always took it back, even sellers that say they don't accept returns.
Meanwhile, you're right. Practice with the broken staff, paying attention to the positions of the roller jewel and hairspring. Practice pays off big time.

Last edited: Feb 13, 2015

Dave Coatsworth

The K&D punch number for removing Waltham friction staffs is 39. The punches for installing the staffs are 40, 41, and 42. Marshall sets also had punches for this purpose, numbered 1K, 2K and 3K. (Thanks to Archie Perkins' Antique Watch Restoration book for this info.)

The hairspring came off easily, as did the roller, once I'd filed down the roller removel tool so that it would fit under the roller.

I found a good fitting stake and stump and it was easy. So I refitted it and had another go!
Feeling confident I moved on the the Waltham Riverside staff with double roller - just as easy.
I managed to measure the pivot size (bought some calipers from ebay that measure to 0.001mm) and now waiting for the new part. I just hope it fits
Thanks for the advice

Agree with everything stated and practice and study is the answer. I forgot to mark the location of the roller jewel and hairspring on my first one so rest assured I remembered on my next one. Broke off a roller jewel on the second one because I forgot to use the slotted punch when re-installing the roller table so now I remember every time. LOL
I guess the most interesting thing I have learned in all of this is just how fragile these pivots are. Then I have to ask did they really need to make the pivots so thin to reduce friction? Would it really have killed them to make a pivot a few MM larger instead of having them so thin they stick you like a needle? I actually plan on doing some research on this particular subject. I am guessing that there is a fine line between pivot size and friction and also pivot size verses jewel strength. If the pivot is to thick it then becomes very strong but then would easily crack the jewel if the watch was dropped.

So- you either get a broken pivot or a cracked jewel. What should be the weakest link? In my opinion I would rather replace that cracked jewel! Think about it in most cases you take out 2 screws and push the setting cup out and then remove the broken jewel. Reassembly is just as easy. Much easier than a balance staff replacement and much less to get wrong like damaging the Hairspring or roller jewel plus you have to worry about the correct timing and beat.. Getting everything on the BS to be perfect is not always an easy task and you must be very careful to not break the new pivots when you are reassembling the balance cock.
So I threw it out there as a newbie anyone wanna enlighten me? I'm sure it's buried in one of my unread watchmaker books.

Then I have to ask did they really need to make the pivots so thin to reduce friction? Would it really have killed them to make a pivot a few MM larger instead of having them so thin they stick you like a needle? I actually plan on doing some research on this particular subject. I am guessing that there is a fine line between pivot size and friction and also pivot size verses jewel strength. If the pivot is to thick it then becomes very strong but then would easily crack the jewel if the watch was dropped.

So- you either get a broken pivot or a cracked jewel. What should be the weakest link? In my opinion I would rather replace that cracked jewel! Think about it in most cases you take out 2 screws and push the setting cup out and then remove the broken jewel. Reassembly is just as easy. Much easier than a balance staff replacement and much less to get wrong like damaging the Hairspring or roller jewel plus you have to worry about the correct timing and beat.. Getting everything on the BS to be perfect is not always an easy task and you must be very careful to not break the new pivots when you are reassembling the balance cock.
So I threw it out there as a newbie anyone wanna enlighten me? I'm sure it's buried in one of my unread watchmaker books.

There is a correlation between smaller pivot sizes and better time keeping, all things being equal of course. USA companies used the smallest pivots on their RR watches, largest pivots on their 7 jewel watches of the same model.

One problem with larger pivots and the jewel breaking instead of the staff: The watch will still run until the broken jewel scores and damages the pivot. The owner won't realize right away the jewel is damaged. So you end up replacing the jewel and the staff anyway.

I find that on pivots .12mm or larger, 90% or so of the time, if it needs a staff, it also needs at least one jewel. Staffs with pivots .09mm or less rarely need a jewel when replacing a staff.

Working on a 12s Rockford, 21j, with pivots .08mm. Believe they are original pivot sizes since my staffs are also pivot .08mm. Really way to small for the weight of the balance assembly. I guess .09mm is very small for a an American 12s. Swiss balance assemblies generally are quite lighter than American ones of the same size. So the Swiss can get away with smaller staff pivot sizes without increasing the likelihood of breaking pivots.

"One problem with larger pivots and the jewel breaking instead of the staff: The watch will still run until the broken jewel scores and damages the pivot. The owner won't realize right away the jewel is damaged. So you end up replacing the jewel and the staff anyway.

I find that on pivots .12mm or larger, 90% or so of the time, if it needs a staff, it also needs at least one jewel. Staffs with pivots .09mm or less rarely need a jewel when replacing a staff.

Both points and observations above are great and I would not have thought about the broken jewel then damaging the pivot problem at all but you bare completely correct. Well- at least I thought through it far enough to halfway figure out why they might have made them so small you just took it the rest of the way for me. Thanks a million for the insight. This is one of the things I find fascinating about watches is learning the history and the procedures they used back then. I am still amazed at the quality and precision they were able to achieve in relation to what was available in machinery and energy sources.

Well the #4860 'straight shoulder' pivot arrived - and it didn't fit. I bent the balance trying to knock it in.
I'm confused. It's definitely not the tapered pivot #4861 as I have one of those in another broken balance, which I practised on. The shoulder looks sort of half way between the two?
So now I've ordered a #4861 to use in my 'practise' balance. This came from a PS Bartlett which only has a single roller. Will everything line up if I use it on my Riverside with double rollers?

Dave Coatsworth

What's the movement serial number?

Dave Coatsworth

Per Jerry's comment on seeing some poorly formed taper shoulders out there... Take note of the staff in the middle in the photo below. Not a very good taper and perhaps what you have? The staff on the left is the correct #4860. The photo also demonstrates the height difference that Jerry mentions.

Dave you are right - it's just like the staff in the middle of your pic

The top is the tapered shoulder balance I took out of my PS Bartlett #4861.
The middle is the one I took out of the Riverside that I'm trying to replace.
The bottom is the #4860 I bought to replace it.

As the shoulder didn't look tapered like the top one, I assumed it was a #4860

So I screwed up. At least I have the other balance to use, and I know that's a #4861

The 'new' staff turned up in the post today. New old stock - a bit of rust which I rubbed off with a fibreglass pen.
The bit of the shaft that pushes in to the balance measure 0.95 mm diameter, whereas the old one is 0.92 mm (the correct spec).

Can someone please tell me, will that push in OK, or do I somehow have to take it down a bit - bearing in mind I don't have a lathe.

As it's such a small amount, and I don't know anyone with a lathe ....
Just a crazy idea, but :

If I used a small strip of VERY fine wet and dry, and wrapped it round. Wouldn't that do it ?

You need a lathe if you are replacing pocket watch staffs. Exceptions would be Hamilton 10s and 16s friction staffs, and Waltham friction staffs, but only if you had genuine staffs, and the balance never had a non genuine staff. Or if you get very very lucky. Sorry, but that is the way it is. Work arounds are a disservice to the watch, the owner, and future owners, and the watch repair industry.

I have quite a few genuine staffs. But often on older pocket watches I can't use them due to changes made to the balance assembly components. Also I think roller tables can get oversize if removed and installed many times.

The chances of buying a non genuine staff and all the diameters and lengths are correct for your older pocket watch are slim. And even if you get a genuine staff, most come in different pivot sizes. (But not Hamilton friction staffs. They are great). And some non genuine staffs are junk. Poor metal, varying diameters even in the same lot. Seen some where the upper and lower pivot diameters are not the same.

Your .03mm oversize is too much for the balance wheel. .01mm you can probably get away with. And cleaning the staff with a fiberglass brush?

Thanks for the advice Kevin. I'm a bit disheartened. I bought a staking set, and got the old (Waltham friction) staff out easily enough, and assumed that fitting a replacement staff would be straightforward. I have read that replacement staffs can be of variable quality and tolerances.

Thanks Graham, I've replied.
Out of interest, how do you measure to these tolerances. Most micrometers I've seen are only good to +/-0.01 mm. I needed to measure the pivots, and without a pivot gauge you need better accuracy. I managed to find a digital micrometer with an accuracy of 0.002 mm. How did they do it 100 years ago - making pivots to 0.12 or 0.13 ?

Ok I got another staff and this has the correct dimensions.
It went in easily, as did the rollers and hairspring. Feeling pround I dropped it in place, laid the balance cock on and all looked good, moving nicely, until the last turn of the screw holding the balance cock - zero end float. Locks up.

I've checked the length is correct, 5.41 mm. And I removed both end jewels to make sure they fit over the pivots.

What is wrong ? Bent balance cock ?

Unlikely to be a bent bridge. But check to see if it has been played with.

In your post #14, you said you bent the balance wheel. Did you true it? Possibly the wheel is hitting the center wheel or pallet bridge, causing the lock up.

Sometimes when trueing a balance wheel, you don't know whether to go up, or go down. Often I install the balance and staff without the roller table and hairspring. Then check endshake, sideshake, and the wheel running true, without hitting anything in all positions. Also check for freedom of friction. Should come to a very slow stop, and easily turns again when changing position. If it does hit something due to not being perfectly true, I then know whether to bend up or bend down when trueing.

Are you sure the pivots are the correct size? Also lengths of staffs can be a little tricky to measure accurately. Harder than diameters. Staff must be exactly perpendicular and level to the jaws of the micrometer.

Kevin, I'm using another balance wheel - it's not perfectly flat, but it's not rubbing anywhere (I'm trying to get some truing calipers from ebay).

I've just checked and am pretty sure nothing else is rubbing. I haven't got the pallet bridge or lever in yet.

What is the typical end play i.e. how much tolerance can there be on the lenght of the staff ? - I didn't want to do the micrometer up too tight.

I'll try your suggestion tomorrow Graham - take the top jewel off.

What is the typical end play i.e. how much tolerance can there be on the lenght of the staff ? - I didn't want to do the micrometer up too tight.

I'll try your suggestion tomorrow Graham - take the top jewel off.

The endplay or end shake is hard to measure. I would guess on your watch about .04 or .05mm. Tolerance on the length of the staff is different. The okay range would be long enough but there is enough endshake, and short enough that there is not too much endshake. Too much endshake and you run into parts hitting - rubbing. And poorer oil retention. And for it to work with excess endshake, you would need longer pivots, which are weaker than a properly proportioned pivot.

On a 16s USA watch like yours, I would guess the tolerance range to be about plus or minus .01mm. So a staff measuring 5.40 to 5.42mm would work, usually. Because of the variations of older staffs, that is why I like to test without the hairspring or roller. If the staff is slightly long, I can check if I should shorten the lower side or the upper side. Or take a slight amount off of both.

And do Graham's test.

Do you have Fried's book? The Watch Repairers Manual? He covers this very well. I think this is the best book for a beginner.

Well I solved the problem. I got the wrong pivot size and it was too tight in the jewels. I had to measure it with a micrometer and was pretty sure it was 0.12 mm. I guess the jewels are 0.11 mm. The bottom jewel was broken which I hadn't noticed before, so it was probably caused by me fitting the new staff.
I have a scrap Walham PSB so fitted the jewels from that, and the balance now swings freeely !

I wish it was easier to find Pivot gauges !

Well I solved the problem. I got the wrong pivot size and it was too tight in the jewels. I had to measure it with a micrometer and was pretty sure it was 0.12 mm. I guess the jewels are 0.11 mm. The bottom jewel was broken which I hadn't noticed before, so it was probably caused by me fitting the new staff.
I have a scrap Walham PSB so fitted the jewels from that, and the balance now swings freeely !

I wish it was easier to find Pivot gauges !

After buying a lathe, making your own pivot gauges is great practice work, and you end up with useful tools.

In addition to needing to have a lathe to do balance staffs on old watches, a good measuring tool is very important. I think the JKA Fienstater dial indicator in MM is really a must. Only it is not so good with staff lengths.

I only started collecting a few months ago. I was just planning to clean the watches and fit new mainsprings (and any other simple parts I could change) to get them running to reasonable accuracy. I bought this Waltham Riverside as a 'runner' from the US and it turned up with a broken balance. As it was a friction fit, I thought it was a relatively simple task. So I bought an old staking set from ebay, and the fun started. I've enjoyed the challenge but have found that balances are best left to the experts! I've no plans to buy a lathe .... not at present anyway.
Thanks

I got it all together today, and running on the old mainspring (weak, I'm waiting for a replacement) it's showing 18012 bph (on the WildSpectra App on my phone). Good enough for me, and should improve with the new spring(?)

The balance does need truing (in the flat) when I get some calipers. I did have a go by using the staking anvil as a flat reference. I read the chapter from the Chicago School of Watchmaking, but couldn't figure out how I should be bending it with my fingers. Does anyone have a better explanation/ video etc ?

Rob, thanks for the info on Elgin. I chose Waltham as my first make as they look so damned good, and as the first to mass produce, I guessed that getting parts, and interchanging parts would be easiest. I may try Elgin next.

A seitz jewel plate is great for figuring the pivot size of any staff or wheel.