Did you even wonder what’s the deal with mechanical watches and their timekeeping variations? Some days they run a little fast - sometimes slow. Some guys say they can lay them in a certain position at night to gain or loose a little time. I will attempt to explain positional error.

Before we start, let me say that I will be discussing a "perfectly" running watch with no other problems that would cause positional error. I am talking about a balance wheel that is not perfectly balanced. And after all, nothing is perfect. I will talk about some other causes of errors at the end.

A balance wheel should be just that - balanced. That is how it got it's name. The reason that such great effort is made to make sure it is balanced is to avoid positional error. If there is a heavier place on the balance, it will have error.

Watchmakers have worked since the invention of the balance wheel to minimize the effects of this problem. Along with the efforts to compensate for temperature, this is the reason for most of the material and design changes for the hairspring and balance wheel. They use specialized tools to analyze and minimize these problems.

There is no positional error when the balance is flat in the dial up and dial down positions. Positional error only shows itself in the sideways positions. In the flat positions, the pull of gravity has no special effect on the rate of the oscillations.

In the sideways positions, any heavier spot on the balance causes it to move faster and slower as it moves up and down. Picture a balance scale with more weight on one side. The same physics that causes the heavy side of the scale to go lower is what causes the balance wheel to run at a different rate.

Now it gets even more complicated. Most balance wheels move back and forth about 285 degrees from the center point. As the mainspring unwinds, the degrees of oscillation become less and the error starts to diminish until it gets to the 180 degree mark. At this point, there is no positional error. Since the balance wheel is turning exactly half way around, it cancels out. As the motion continues to lessen, the error once again starts to increase, but this time the opposite way. A watch that was running slightly fast will run slow and by the same amount.

So why don’t they just balance that wheel?

It’s easier said than done. Every effort is made, but once the hairspring is attached, other problems arise. The outside end of the hairspring is fixed (having no effect on the balance), but the rest of the hairspring winds and unwinds as the balance oscillates. This introduces a constantly changing distribution of the weight. So even if the balance wheel itself was perfect, the hairspring throws it off. Through the years, many different designs of overcoil hairsprings have been used to help minimize this problem.

What’s the answer?

Dynamic poising helps. A watchmaker can check the rate of the movement in several different positions using his timing equipment. If he understands the theory, he can then tell where the heavy spot is. He removes a little weight or adjusts screws on the balance wheel perimeter and then tries it again (yes, trial and error). It sounds difficult, but with practice it is a very efficient way to reduce positional error.

In practice, dynamic poising is not done during the manufacturing process on any but the highest grade watches. Most watches, direct from the factory and properly adjusted, will meet the standards for which they are designed without dynamic poising.

Incidentally, this problem with positional error is why the tourbillon was invented. In a tourbillon movement, the entire balance wheel assembly rotates around about once a minute. If there is a positional error, it will cancel out when the tourbillon gets 180 degrees around. So rather than try to solve a problem which can’t be completely solved, they figured out a way to cancel it out!

So what’s the deal with my watch?

Most watches are going to have some degree of positional error. But there is no set standard for which positions might be fast or slow. This is determined when the watch is made and there is no guarantee that the same models will have the same error in the same positions. So just because Joe’s watch gains a couple of seconds when left overnight in a certain position, it doesn’t mean yours will too. Each watch is different and you will just have to experiment on your own.

Is that all there is to it?

There are other reasons for positional error that don’t have to do with balancing. Most have to do with use (or abuse), wear and lubrication.

There are pivots on the end of the axel that the balance wheel turns on. They are less than a tenth of a millimeter and rest in spring-loaded jewels to absorb any shocks. The condition of these pivots is extremely important to good timekeeping. They must be perfectly straight, properly shaped, polished, and lubricated. Any changes in this condition will effect the timekeeping. If the oil runs dry or if they get dirty, they will start to wear. If the watch receives a hard enough shock, they can be bent and this will throw the balance out of poise. Shock damage to the hairspring can bend it slightly and cause similar problems. Most of these issues are not seen until a watch is a few years old, but you never know when they can appear.

Magnetism is another problem that should be mentioned. I recently had a watch that showed more that a minute fast in all positions. I demagnetized it and it immediately returned to COSC specifications - an instant cure. This was a dramatic example. Usually there is just a slight amount of magnetism that causes odd and unpredictable readings. Watches have been antimagnetic for years, but every watchmaker has a demagnetizer near his bench and uses it before he even tries to make any adjustments.

Remember I said that positional error is not seen in the dial up or down positions? Theoretically it is impossible on a “perfectly” running watch. This type of error is caused by a difference in the top and bottom pivots - depending on which is carrying the weight. If one is dirty, worn or improperly oiled, there will be a difference. How much of a difference depends on the condition of the pivots.

My advice

First, don’t go crazy. There is no perfect watch. It is completely normal for a watch to vary in different positions and from day to day. The better the movement, the less the variation. As you wear it, the rate on a watch changes throughout the day. Second, remember that there are many factors influence the rate. This is why the COSC standard is a range and an average. Get to know your watch’s “personality”. It is the changes in it's behavior that will let you know if something is going wrong. I am continually amazed that these little pieces of machinery are as accurate and durable as they are.

Thanks again to Steve / N2FHL for writing this installment!

 

Positional error can occur face up and face down, even if the wheel is properly poised. One significant cause can be pivots on the balance staff that are too small for the hole in the jewels. As the pivot “wobbles” in the jewel hole, the effect is to increase the diameter of the balance wheel which can result in a loss of rate.

 

Indeed - what the squirrel actually means is that positional error due to poise error is only in vertical (what he calls "sideways") positions, but positional error exists in all positions. The most common cause of positional variation in horizontal positions is on watches with regulating pins, where the pins are not parallel.

But certainly worn pivots (usually the ends leading to amplitude differences as well) can cause errors in horizontal positions...

 

I know that this discussion is pretty old but I'll try and ask. Is it normal for my cal 265 to have a maximum positional variation of 55 seconds? (Dial up -55 seconds per day crown down +-0)

 

That series of watches, the 30mm calibers, were pretty good timepieces. While they are not adjusted to positions like a chronometer or railway watch a 55 second difference between two positions is quite a bit I would imagine that a service may be due.

perhaps @Canuck or @Archer will chime in and lend their expertise.

 

Thanks for your reply! Could the mainspring have a role in this? AFAIK it's still the original one from 1952 and the power reserve is lower because of that (around 24 hours). I've talked to my watchmaker and he'll switch it for free as soon as he has one for my caliber.

 

While the power available to the machine drops as the spring unwinds I wouldn’t expect that to have much effect on positional accuracy.

i think your issue is due to irregularities in the balance or its pivots.

 

It makes sense. I've bought this cal 265 from my local watchmaker and he said he had serviced it, I'm wondering what he meant by "service". Still, the watch is under warranty so I'll just swing by and ask him to solve the issue. How precise do you think a cal 265 from 1952 can be? What is a reasonable expectation?

 

@MaxJeix asks about a 55 second per day variation between two extreme positions with his calibre 265 Omega. I would start out by asking what the service history of the watch might be? The rule is, repair (and that includes stripping the watch and examining every component for wear, conditioning thoroughly, including installation of new parts where necessary, re-assembling, check thoroughly every phase of its operation, lubricating, etc.). Once the watch has been brought up to standard in every respect, rating the watch. Rating means recording the rate in all positions, and adjusting as is necessary to even out the rate in all positions. Only after all this is done, regulate the watch. Because a watch “runs”, it does not necessarily mean it will meet the timekeeping expectation of the owner. Repair, rate, then regulate.

Some time ago, I had a Zenith wrist watch brought to me for “demagnetizing“, as its rate was poor. It was minutes fast in a day. Close examination of the hairspring (not a Breguet hairspring in this case) showed that the hairspring had too much freedom between the curb pins of the regulator, and quite frequently, the second coil of the hairspring was “tapping” the inner curb pin. Closing the curb pins didn’t stop the second coil from touching the curb pin. A microscopic adjustment of the outer coil (outward) solved that problem. Then checking the rate on the timing machine. I didn’t have to move the regulator, NOR demagnetize it. The owner reported that the rate went from minutes per day gain, to an 8 seconds per day gain, on the wrist. Moving the regulator would not have solved the irregular rate, not the tendency for this watch to gain. Moving a regulator is not intended to bring a watch to time if adjustments to the operation of the watch aren’t made.

 

Thank you for your answer. Unfortunately I don't know the service history but my watchmaker told me he had serviced it before selling it to me. I suspect he just checked that there were no broken parts. I think that the truth is that this watch hasn't been properly serviced in decades and that would justify the issue I've encountered. On the wrist it loses consistently 35-40 seconds, but if left in specific positions on a table this changes dramatically (+10 even).

 

It appears that the watch needs a proper service, as this watch should run longer than 24 hours (32 is the official power reserve, but generally they rung longer than that when properly serviced). The positional tolerances of this movement are not very tight, so measuring the rates on a timing machine, at full wind, this check is required over 3 positions by Omega. These are dial up, crown down, and crown left. Over those 3 positions the variation is allowed to be as much as 40 seconds at full wind, and 50 seconds 24 hours after full wind.

Of course these can be made to run much better, but that depends on the condition of the movement, and the skill of the watchmaker.

Cheers, Al

 

Thanks for your answer!

 

After changing the main spring and demagnetizing the watch my watchmaker called me and told me I could pick it up. Now it loses around 3 seconds per day which I find a very satisfying result.