Timing tolerances - a guide to understand how they work

I often see threads where people are questioning if their watch is running outside the manufacturer’s specs. I thought I would post some information to explain how these specs work and why your watch may actually be within specs, but not appear to be.

So when a watch is timed at the factory, it’s done under very controlled conditions on a timing machine. The duration of these tests is usually very short, so only over a matter of a few minutes depending on the specific watch in question. The watch is typically timed with two different states of wind on the mainspring - the first is fully wound and the second is 24 hours after being fully wound.

For now I’ll just talk about the fully wound state to keep things simple. If the watch is not a chronometer, it will typically be tested over just 3 positions, and those are dial up (CH), crown down (9H), and crown left (6H). If the watch is COSC rated, it will be tested over 5 positions, so the three above plus dial down (FH) and crown up (3H). If the watch is a METAS caliber, then it will have all those 5 plus crown right (12H) tested.

Each individual test lasts maybe a minute, and between each test the watch is left to stabilize for a few seconds. Each position will have a rate in seconds fast or slow per day, balance amplitude, and beat error. So taking a watch like the Speedmaster Pro with a Cal. 1861, let’s put some rate numbers for each of the 3 readings:

Dial up = +7.5 s/d
Crown down = –7.5 s/d
Crown left = 0 s/d

So to calculate the average daily rate we add up the 3 numbers and divide by 3, and that gives us a result of 0 seconds gain or loss - great! 👍

But there is another number we must calculate from these three readings, and that is the Delta, which is the difference between the fastest and slowest position. So here that is the difference between –7.5 and +7.5, so 15 seconds. Omega’s tolerances for this watch are that the Delta number must be no greater than 15 seconds, so on that measure this watch passes. The tolerances for the average daily rate are from –1 to +11 seconds per day, and since the average here is 0, this also meets the requirements Omega set out.

So if you buy this imaginary Speedmaster and get it home, and it runs at –5 seconds per day, then is it faulty?

Keep in mind that the average daily rate calculation above assumes that each position is given equal weight in the calculation, so if you were to wear this watch so that it spends 8 hours in the dial up position, 8 hours in the crown down position, and 8 hours is the crown left position, then it theoretically would average out to 0 seconds gain or loss over 24 hours. But no one wears watches like this, so in real life wearing conditions the average rate number can be skewed towards the position that the watch spends the most time in. For example, if you stand a good part of the day with your left wrist down at your side, then crown down will be the dominant position in the weighted average, and this could lead to the watch running at –5 seconds over the 24 hours, even though it’s running completely within Omegas tolerances.

Also keep in mind that any of these tests that use 3, 5, or even 6 positions don’t reflect the positions the watch will actually be in, as there are a virtually infinite number of variations with how your wrist can be positioned during the day that don’t match these exact test positions.

So there’s another thing to touch on, and that is isochronism - this is what affect the state of wind has on timekeeping. So as the mainspring winds down, the torque from the spring that is delivered to the balance is reduced and this can cause the timekeeping to drift. Omega measures this on the Speedmaster at 24 hours after the watch is fully wound, and they allow the Delta number to increase from 15 seconds to 20 seconds. So now our imaginary Speedmaster could have the following rates 24 hours after being wound:

Dial up = +10 s/d
Crown down = –10 s/d
Crown left = 0 s/d

So again the average is zero, but the two positions that deviate from zero are spread even further apart. Since Omega doesn’t have a tolerance for the average rate at 24 hours after full wind, but they clearly acknowledge that the timekeeping will likely get worse, this is another reason why this imaginary Speedmaster could be running outside of the range for the average rate, but still be within Omega’s tolerances.

Note that although I’m using a Speedmaster in this example to keep things simple, this same process applies to all Omega’s watches. Of course the tolerances are different, but the key take away here is that if the specs for average rate are from –1 to +11, or –1 to +6, or from 0 to +5, that doesn’t mean that every position the watch is in will run between those numbers. If your watch happens to spend time in a position that is outside the range for the average, then it could very well run outside the range on your wrist, and yet still be within Omega’s tolerances.

If you have any questions on this, please let me know.

Cheers, Al

It depends. Some movements do use different parts for different quality grades, where some do it through adjustment.

Because there is clearly something wrong with the Omega

Is there a different tolerance stated by Omega for the Speedmaster if the chrono is engaged?

In your experience how big of a difference would you accept with and without chrono activated? (Is total reserve time reduced?)

No, the rate tolerances are all with the chronograph stopped. Rates are not checked with the chronograph running but the balance amplitudes are.

The reserve is a function of mainspring length and gearing, so it is designed into the movement and doesn’t change. However the increase in load from the addition of the chronograph mechanism will mean that a greater torque is required to keep the watch running, so it will stop sooner than it would than if the chronograph was off. But when the watch stops, if you turn the chronograph off the watch will start running again.

Have you ever tested with and without chronograph engaged to see the difference?