How Much of an Effect Does Altitude and Humidity Have on the Accuracy of a Watch?

[grtnev]

I have been using the same gentleman to service my watches for about 20 years now and, anecdotally, I have observed that when I receive a watch back from him, it is not unusual to find, after running on my desk for at least 24 hours in the face up position, that the watch will be gaining 6-12 seconds/24 hr as opposed to being in agreement with his electronic timing machine at his location.

At my location:

• Typical uncorrected barometric pressure: 25.10 in Hg @ 4,722 ft (1,439 m)
• Typical humidity varies from: 30% to single digits depending on the time of the year
• Typical temperature varies from: 23°F to 95°F depending on the time of the year.​

At his location:

• Typical uncorrected barometric pressure: 28.5 in Hg @1518.7 ft. (462.9 m)
• Typical humidity varies from: 20-70% depending on the time of year
• Typical temperature varies from: 28°F to 86°F depending on the time of the year.​

Temperature variations are similar at both sites, being a little warmer (and much drier) at my location during the summer.

However, there are significant differences in the uncorrected barometric pressure and humidity variations:

• Difference in barometric pressure from his location to mine, typically: -3.4 in Hg (-1.67 psi)
• Difference in humidity can be as much as 60-65% higher at his location during the summer and 5-15% higher at his location during the winter.

Would these typical variations in humidity and barometric pressure from his location to mine explain why a watch would run a few seconds faster at my location as opposed to at his location; i.e. lower humidity at my location (much lower during the summer) with a lower barometric pressure?

Thanks in advance for your thoughts.

Richard

 

[musicguy]

I would assume hypothetically that altitude, pressure, humidity and other factors
could effect how a watch runs. But 12 seconds a day is 1.4 minutes a week
or 5.6 minutes a month that sounds like a big error that might not
be explained by location after a complete service.


Rob

 

[John Matthews]

I assume the comparison is face-up in both locations. Have you asked the watch maker to run the same 24 hour test for comparison?

John

 

[karlmansson]

Does your watch guy time your watches after 24h after fully wound? As well as on full wind?

 

[svenedin]

You don't say what these watches are or how old they are. I would be rather pleased with an accuracy of 6-12 seconds in 24 hours.

 

[grtnev]

Any watchmakers have any input? DeweyC?

1) Differential barometric pressure between where the watch resides and where it was serviced: -4.28 inHg

2) Average humidity:
15% where the watch resides
65% where it was serviced

What change in the watch's performance would you expect? None? Watch will run slightly faster where it resides?

Richard

 

[Skutt50]

Here is my take on your questions:

I would not expect any practical change based on air pressure or humidity.

Theoreticly air pressure would have the balance run in a "more dence" environment and it would slow down the balance but from a practical point of view No.

Humidity should not make any difference as long as it does not build up any drops on the parts. If condensation occurs you have much bigger problem than a small change in accuracy......

 

[praezis]

Hi,
different from clocks, barometric pressure and humidity have no direct effect on performance of a watch, imho.
Its rate is ruled by moment of inertia (balance wheel) and deflecting force (hairspring).
I heard that even on the moon a mechanical watch was running well...

At best temperature differences will affect the rate. But most probably your rate differences come from different test condition, as already mentioned.

Frank

 

[DeweyC]

Any watchmakers have any input? DeweyC?

1) Differential barometric pressure between where the watch resides and where it was serviced: -4.28 inHg

2) Average humidity:
15% where the watch resides
65% where it was serviced

What change in the watch's performance would you expect? None? Watch will run slightly faster where it resides?

Richard

Richard,

As you may know, my main bread and butter are RRG watches returned to factory function, aircraft clocks used in working aircraft, and chrononometers. For 35 years I was as close sea level as you can get without getting your feet wet.

First, this efffect was not even disucssed in class at Neuchatel. Secondly, it takes a very precisely timed instrument to detect barometric changes (in another I related a physicist at Los Alamos who sent me a plot of M21 rate vs barometric pressure). Third, no one from the mountain states or pilot ever complained (except when I screw up).

So my experience and education suggest something else is going on.

According to your observation, the difference is systematic; not random. Knowing nothing about the two test procedures, my hypothesis is: Isochronal rate vs instantaneous rate; Are you comparing his (at whatever wind) instantaneous reading to your 24 hour timing test? Big difference.

The other thing I would be interested in is who came up with temp/barometric pressure? As you know, precision timepieces are adjusted to temp by matching the balance spring's elasticity to the balance wheel's moement of inertia (via the mass screws on the rim). Plus, no matter how badly adjusted, the difference would be very minor even if one room temp was 72F and another was 68F.

In other words, this explanation is (yes I am being definitive) moving into the real of personal magnetism.

You should find out how your watchmaker regulates the watches before delivering them. Does he do it using the machine or does he do a 24 hour hour test? AS you know, many (including watchmakers) conflate the terms regulate and adjustment.

 

[Jerry Treiman]

Watchmakers and collectors familiar with the watches that Waltham made for Ball may have noticed that the Ball models have a different balance wheel than the standard Waltham models. The Ball balance wheel has a larger diameter and shorter (lower profile) timing screws. I was told once that the shorter screws were to reduce the effect (air friction or drag) of the thinner atmosphere over the Rockies relative to the denser atmosphere at lower altitudes. It makes theoretical sense to me.

 

[DeweyC]

Watchmakers and collectors familiar with the watches that Waltham made for Ball may have noticed that the Ball models have a different balance wheel than the standard Waltham models. The Ball balance wheel has a larger diameter and shorter (lower profile) timing screws. I was told once that the shorter screws were to reduce the effect (air friction or drag) of the thinner atmosphere over the Rockies relative to the denser atmosphere at lower altitudes. It makes theoretical sense to me.
View attachment 680655

Jerry,

Prior to 1948 watch designers could not calculate the needed mainspring for a new caliber (Ben Hummel; private com). There was a deficit of knowledge regarding applying physics to watch manufacture. Even in 1948, watch engineering was empirical endeavor. As a result, many explanations (like personal magnetism) were merely speculations.

Even if the goal was to reduce drag (which a person competent in calculating fluid dyanamics could calculate), its effect would be consistent. It is possible to speculate that it impacts the work (as defined by a physicist) done by the balance. But Ben told me that the reason reason he had to publish his nomograph in 1948 on estimating required mainsprings was because there was not yet a method to make that calculation. So I am not sure how the air drag hypothesis would have been tested by Ball or Waltham.

I would posit that the difference in balances has to do with the moment of inertia. The total OD across the screw appears almost identical. They have to fit the within the constraints imposed by a 16s watch.

The mass in the Ball design (rim plus screws) is more evenly distributed around the lever arm for the moment of inertia. This may provide an advantage for adjusting to temp and position. Like with the balances in modern wrist watches. Considering motion of the arms in response to temp change (and amplitude; the arms deflect more at higher amplitude), this more even distribution may offer an advatage to temp and isochronal adjustment.

Almost every explanation for watch behavior given by watch repairers with no education in physics or engineering should be viewed with caution. I even offer my notes only as a working hypothesis that should be tested by my betters who can do the calculations and physical testing.

Air drag did not seem to be a concern in the writings of Kullberg, Frodsham, Potter or even Jendritski of today. And most watchmakers of the past were not interested in reading these works. Many watch repairers could not even adjust the locks of the escapement in a consistent and repeatable manner.

 

[John Runciman]

I have been using the same gentleman to service my watches for about 20 years now and, anecdotally, I have observed that when I receive a watch back from him, it is not unusual to find, after running on my desk for at least 24 hours in the face up position, that the watch will be gaining 6-12 seconds/24 hr as opposed to being in agreement with his electronic timing machine at his location.

The problem is failure to grasp what the timing machine is telling you. The timing machine is telling you at the exact instant the watchmaker had the watch on the machine what the watch is doing. The timing machine is not making a prediction of what the watch is going to do over the next 24 hours.

You don't say what these watches are or how old they are. I would be rather pleased with an accuracy of 6-12 seconds in 24 hours.

This comment brings up a problem? The problem is the watch itself basically how good is the watch at running on time for 24 hours. Depending upon the watch it could be extremely bad at running on time for 24 hours. Even though the timing machine will show it looks wonderful at the exact instant it was on the machine.

This is why companies have specifications for timing their watches. Rolex has nice rigid specifications. Like putting the watch on the timing machine in timing and more than one position makes it really easy to spot problems if you do that. Then they also time again 24 hours and make a comparison. Because of course a Rolex watch should keep decent time over 24 hours. Then they even have a timing procedure over several days then of course they cheat the average the numbers together but still they watch the watch over several days to see what the watch does they do not trust the timing machine to predict that their watch will keep perfect time they still verify that it really will over several days. That even includes letting the watch run down in power so it's not left on the auto winder where it's at an even power.

Your timing machine is only making basically a prediction of what the watch might do perhaps.

 

[karlmansson]

To my, albeit limited, knowledge of watch testing I gather it's standard procedure to check a watch both at full wind after a service but also after 24h from full wind to see how it behaves at both ends of the functional power reserve. This would be even more important with watches with an older type mainspring that has a much less linear power curve than modern alloy versions. I always aim for amplitudes of 180 or above after 24h.

Rate and beat errors are exaggerated at these lower amplitudes (which is also why I strive to make dynamic poising adjustments at below 220 deg amplitude, easier to see what effect your adjustments have) and if your watch is losing power and therefore amplitude right after the initial, full wind power bonanza that usually lasts for a couple of hours, you will start to see problems that were not apparent at full wind. Could very well account for the rate issues you are having.

Regards
Karl

 

[John Runciman]

To my, albeit limited, knowledge of watch testing I gather it's standard procedure to check a watch both at full wind after a service but also after 24h from full wind to see how it behaves at both ends of the functional power reserve. This would be even more important with watches with an older type mainspring that has a much less linear power curve than modern alloy versions. I always aim for amplitudes of 180 or above after 24h

In modern service centers yes they do this. The watchmaker is required To follow a standard testing procedure as outlined by the technical specifications. Then the watch goes to a quality control person. The quality control person visually examines the watch usually with really bright lights high magnification they want their watches perfect. Then they follow a several day procedure of making sure the watch actually keeps time.

Problem becomes what does the average watchmaker do? Some of them still have their paper tape machine conceivably they don't even have a machine though measure amplitude. How many watchmakers look at the watch beyond one position? How many even look at the watch 24 hours later to see what it's doing? The problem with a lot of people as they learned watch repair in school they went out into the field and that's it. They don't have access to modern technical literature because it's hidden behind firewalls with passwords you have to have an account. They don't know what a proper timing procedure is. They know that if they regulate their watch and maybe run it for a couple of days and it keeps time that's good enough for them. Oh and how do I know this I've actually met a watchmaker that that's basically what he did times in one position watched it for a couple of days good enough any in charge very much either. Of course his customers are happy and their watches probably keep somewhat time it's only when you get really obsessed does this procedure become a problem.

 

[karlmansson]

In modern service centers yes they do this. The watchmaker is required To follow a standard testing procedure as outlined by the technical specifications. Then the watch goes to a quality control person. The quality control person visually examines the watch usually with really bright lights high magnification they want their watches perfect. Then they follow a several day procedure of making sure the watch actually keeps time.

Problem becomes what does the average watchmaker do? Some of them still have their paper tape machine conceivably they don't even have a machine though measure amplitude. How many watchmakers look at the watch beyond one position? How many even look at the watch 24 hours later to see what it's doing? The problem with a lot of people as they learned watch repair in school they went out into the field and that's it. They don't have access to modern technical literature because it's hidden behind firewalls with passwords you have to have an account. They don't know what a proper timing procedure is. They know that if they regulate their watch and maybe run it for a couple of days and it keeps time that's good enough for them. Oh and how do I know this I've actually met a watchmaker that that's basically what he did times in one position watched it for a couple of days good enough any in charge very much either. Of course his customers are happy and their watches probably keep somewhat time it's only when you get really obsessed does this procedure become a problem.

That's a pretty bleak picture you paint of the average watchmaker.

I would then suggest that we strive to be above average and actually check our work. No equipment is necessary to check a watch again after 24h.