A2824 Running Fast Phenomenon

[johnchpark]

Hi, I have an A2824 in a ZF Black Bay 58 that's about 2 months old.

I regulated it perfectly about a month ago, but haven't worn it much since. It's on a watch winder at the lowest setting (spin 2 minutes/rest 18 minutes).

Every time I walked past the shelf, I noticed the A2824 was running faster than my other watches.

So today, I graphed it again and saw it is running 40~60 sec/day fast. This was at a low amplitude (~200), because it came straight out of the watch winder.
I suspected magnetism first, so I demagnetized it. But it didn't change the measurements.

INTERESTINGLY, when I wound up the watch more and got the amplitude to +300, the measurements went back down to 0 sec/day.

I've never experienced something like this before on my ETA 2824's or SW200, and I'm very curious as to why a lower hairspring tension would result in faster beat rate.

Has anyone experienced this or know why this might happen?
thanks

 

[dogwood]

Yes, this is actually expected. Although usually the effect isn't as pronounced. The theory of how a balance wheel works is very similar to the theory of how a pendulum works. (and by "works" I mean "works to keep time)... a pendulum of a given length and mass will swing back and forth at a fixed frequency as long as the distance the pendulum swings through is "small" relative to the length of the pendulum. And by "small" usually mean a factor of at least x100... so for a 1m long pendulum an oscillation of 1cm would be good. This is useful because it allows us to solve the 2nd order differential equation that governs the pendulum's motion in a way were the frequency comes out as a constant. But this only works if the oscillations are small relative to the length of the pendulum since the solution uses a Taylor series expansion of the Sine function where we ignore higher order terms and assume the Sin(Theta) = Theta for Theta close to zero... basically we assume that the sine function is linear. It's not. But it kinda acts linear as long as the angle is small (i.e. the oscillation of the pendulum is small relative to its length).

The same basic differential equation applies to a balance wheel -- as long as the oscillation is within a certain range relative to a geometric parameter of the balance wheel, then the period of the balance wheel will be a constant. However, reps being reps don't have the best manufacturing standards... It's possible (likely) that the hairspring force isn't behaving linearly (magnetism is the common cause for this, but other things like kinks in the spring could also make it behave non-linearly). If the restoring force (aka the hairspring) doesn't act linearly, then the assumption that Sin(x) = x no longer applies, so the solution to the differential equation of motion will no longer product a result where the period of the balance wheel is a constant regardless of amplitude.

TLDR; you probably need to replace the hairspring.

 

[Raddave]

or mainspring

 

[johnchpark]

Wow. I've never read up on the physics behind a balance wheel, but very interesting and very impressive.

My A2824 issue doesn't seem catastrophic and I'm not too bothered by keeping it wound or resetting the time every few days. So I guess I'll get it serviced at a later point.

I've read that the A2824 are basically identical and among the most reliable in the rep market, but I guess even these come with some flaws.

 

[derjenigewelcher]

You may also test if its magnetized