Equalizing balance amplitudes and pivot polishing
I received a very nice Omega 28.10RA SC PC (Omega also call this a 350) bumper movement watch from about 1945.
It was not running as these bumpers do tend to loosen screws if they are not tight (as will any other automatic but, these are particularly efficient at that) and one of the dial screws had worked out because of under tightening and slipped into the train. On strip down, it was also clear that there was a lot of wear on a balance cap jewel.
The picture below shows the balance only fitted to the plate with its Incabloc shock protection. This consists of a jewel with a hole for the pivot and a cap jewel so that the balance pivot is resting on that jewel when in either dial up or dial down positions. Someone has tried to regulate this with the existing issues and scratched the balance cock badly but the rest of the movement is very nice.
Usually, if a watch is not maintained well, one location that can be damaged is the upper cap jewel due to the pivot of the balance using the oil and dirt to grind into the jewel when it is left dial up overnight. Oddly, this was was the jewel on the dial side which implies it was mainly run dial down. Here’s a picture of the jewel when removed showing the indentation that the pivot has ground into the jewel, which should be flat on this side.
Replacing these cap jewels is not an issue as Incabloc still support most of their designs. However, the effect on the balance pivot due to this grinding leads to flattening of the end of that pivot. As an exercise, I built the watch with new cap jewels but without addressing the pivot. I know many people only show dial up (CH) positions on test and here is that position directly on start up after base movement build.
So, this is looking pretty good and with a small adjustment to the beat error would look better. Good amplitude which will lift a little as the oils distribute. But, turning over into dial down (CB) position shows a very different performance.
This is a far lower amplitude and it’s oscillating causing the waviness of the curve. The first point in adjusting the watch performance is to equalize the dial up and down positions for amplitude and rate. In order to resolve this major difference, the balance is removed from the movement and the damage to the pivot is clear. Here is a picture of the lower (dial side) balance pivot at this point and the balance is resting here when in the dial down position. This is under a lot of magnification as the pivot diameter is only about 0.1 mm. What’s clear is that the end of the pivot where it sits on the cap jewel is flat.
The balance pivot needs to be rounded again to reduce the friction with the cap jewel and this can be adjusted with a Jacot tool or, as in this case, I use a Bergeon 5482 which is specifically for reshaping the pivot. After a little work, the pivot looks like this. It’s difficult to see but the top of the pivot is now domed as it should be.
The balance is cleaned/fitted back into the base movement and the end shake checked. The tester now shows the following. Dial up is virtually the same although the rate and beat error have changed due to the strip down of the balance. The amplitude is still settling but is again between 260 and 270 degrees.
And dial down has now improved dramatically showing a smooth curve and an amplitude gain of over 40 degrees to bring it closer to the dial up position.
This is an iterative process as too much polishing can bring the dial down position to be higher than the dial up so, the process is repeated. Usually, it’s a maximum of two tries and for a very old movement like this, a 10 degree variation is acceptable. Once the dial up and dial down amplitudes are close, the issue of the overall rate and the rate difference between the two can be resolved with regulator pins and hairspring adjustments. The lower pivot gets another small polish and then the beat error is addressed.
These older calibres don’t have a moveable stud carrier for the hairspring so, correcting beat error means removal of the balance and hairspring and rotating the spring via the collet that holds the spring to the balance staff (the centre part with the cut-out in below) then re-pressing with a staking set. Again, it’s iterative so, obtaining a beat error of zero is not always possible. Less than 0.8 ms is a good value. Typical balance shown below.
Beat error adjusted and the two positions are also much closer in amplitude. Just the variation in rates needs to be addressed although this is fairly good for this vintage movement. I usually aim for under 5 seconds/day maximum variation (that’s less than 0.006% or 6 in 100000 variation) but, there is a limit to what can be achieved in a reasonable time with a movement that is over 70 years old.
There are some other notes on improving positional performance between vertical and horizontal positions here.
The full service document for this particular watch including this write up is here.