Elgin 12s 15j 315 gaining time like mad.

[ANDY YALE]

I got this watch as a complete, non running movement. On examination, the balance was good, and the train ran well on the high speed test. But the pallets would not function- a busted pallet arbor and cracked pallet bridge jewel.
I replaced both the pallet fork and bridge with parts from the same model Elgin. I got good action on the pallets with the watch wound, and with the balance in, fair amplitude, in all positions, but a little short on the arc.
I completely disassembled it, ran it through the ultrasonic, cleaned the hairspring in benzine, trued the balance in the flat, tightened some loose balance screws, reassembled and oiled the watch and demagnetized it.
Amplitude is better at about 220, but the watch picks up time like crazy. About an hour in every 5 hours.
The hairspring is between the pins on the regulator, but they look a little wide, so the spring is sitting on the outside pin, not in the center of the two pins.
The only other thing I can observe is that the overcoil looks a little bit unusual, the curve is a little tighter than usual to my eye.
Besides the pallets, all the parts of the watch carry the same serial number.
What else could I look for to cause this crazy rate of gain?

 

[topspy]

Sometimes demagnetizing goes wrong, depending on method used. Check no coils are sticking from that or oil, or are rubbing anywhere. Are the "loose balance screws" you tightened meantime screws? If so that would increase the rate, but not by that much I think. Another thing to check is if an improper gear (train or motion work) has been substituted. To check this you need to count all the wheel teeth and pinion leaves and work out the ratios. Check that the cannon pinion is not slipping, if loose the hour hand can jump around during handling.

 

[karlmansson]

"Tightened loose balance screws". Could it be that someone in the past used the balance screws as substitute mean time screws and that seating them would cause the increase in rate? Maybe balance screws have been substituted for lighter ones?

 

[Smudgy]

Take the balance out and compare it's vibration to a known balance. That will at least let you know if the problem is in the balance/hairspring or in the movement,

 

[ANDY YALE]

The watch was gaining before I tightened the balance screws. I didnt tighten the meantime screws. There is one balance screw much smaller than the others, but it is not paired with another like it.
After I demagnetized it I checked it with a compass and it seemed OK.
Smudgy, what would a "known balance" be? One from a properly running watch with the same train?

 

[Electro]

Hairsprings are not the only cause for a watch to run too fast. You might also have a weak mainspring.

 

[Smudgy]

Andy, yes another 18000 train that is running to time.

 

[karlmansson]

I didn't think you

 

[karlmansson]

I didn't think you moved the meantime screws but that someone else had used the balance screws as such.

If you get the amplitude up to above 270 degrees a weak mainspring should effect the rate. It would be something to fix later but it wouldn't explain your problem.

The watch could also be skipping teeth in the escape wheel. The lock would be insufficient in that case.

 

[DeweyC]

Hour every five? Pallet busted upon receipt? Low but timeable amplitude (Vertical amp of 220 is the magic number for positional timing). Balance spring looks proportional to balance wheel? Cannot assume anything is right.

It could be the balance spring ("overcoil looks funny") which can be checked by replacing it with one from a known good watch.

The gain sounds an awful lot like an incorrect train wheel/pinion; especially the motion work.

Why not try the balance assembly from the other Elgin movt that provided the pallets? Or from a good known example?

The other thing is to see how the balance is loaded compared to another of the same grade. Maybe screws were removed from the balance.

Discussed problems caused by "making a spring look good" in another thread. Timing is shot the instant a balance spring is altered. This is a real problem with railroad watches that are so often used as learning material; and have been for 50 years.

And then comes all the hairspring services offered years ago. Given what I now know about how each caliber has its own spring pattern in order to produce precision timing results (collet pinning point, number of coils, shape of terminal coils are unique to each caliber); I am very skeptical that these service accomplished much more than making a watch tick unless they used their stock of replacement balance springs from the manufacturers (which in the 50s was very likely for many popular but obsolete watches like Illinois and Southbend. People bought up stock then just as do they now.) But to make from the raw to get proper results on just one caliber takes months of practice in the factory even today. And they know what the final spring should look like.

So don't blame age or yourself if you can't bring that Hamilton 940 or 992 to factory timing specs across positions. It may well be something that looks good, keeps an acceptable rate in use; but which was altered. A watch at the bench brought to 30 seconds difference across the 6 (or 10 depending up your method) positions, WILL deliver that performance when worn. However you can often regulate a watch to keep time as worn by an individual with the regulator even though it cannot be adjusted to position.

The former is positional adjustment to reduce rate differences across positions; the latter is taking advantage of the various positional rate errors that occur in the way a person wears the watch to get it to show time in a reliable manner.

Does it matter? To some yes, to others no.

FWIW, I sold my vibrating tool and balance spring stock after I got back from school. Now if I encounter a rusty or badly distorted spring, I find one from the same grade.