The article mentions, but does not describe, tourbillons, which are ingenious mechanisms to continuously rotate the escapement around one or more axes, so that the effects of orientation are averaged out.
My uncle always said you have to put your watch onto the rest/stand that comes with it as those are designed to hold the watch at the optimal position.
I never understood what he meant since I've always worn a Casio F-91 for as long as I can remember.
That's a little interesting, if only because a lot of watches don't come with stands/cases - I remember when I was a kid I got a couple of watches that had little plastic holders, that they were packaged with.
These days I've got watches ranging from the low-end Casio Terrorist Watch to mid-range Rolexes, and a lot in between, and they typically come in boxes or cases without a stand of any kind.
Mechanical clock accuracy used to be pretty important for navigation to determine longitude. A large prize was offered by the British government to whoever could invent a way to accurately determine longitude.
It's an interesting look at the high-tech world of yesteryear.
In an alternate universe, people might've discovered simple radio first, and skipped all the fancy clock developments!
1. Set up a big transmitter in some major city that fires off a big noisy pulse at every day at solar-noon.
2. Equip your long-distance ships with mediocre clock and a radio-receiver that is just good enough to detect the pulse if a crewmember is waiting and listening for it.
3. Using the low-accuracy clock, measure the difference between (A) solar noon where the ship is and (B) when the radio-pulse is received. Every 4 minutes is 1 degree of longitude.
P.S.: The pulse-listener will have to shift their listening-window as the ship travels, but it shouldn't be too tough since they already have a so-so clock and know their latitude. Sailing 200km/day along the equator means the pulse-listener would have to adjust their listening-window by ~8 minutes/day to keep "centered" on the pulse moment.
> Mechanical clock accuracy used to be pretty important for navigation to determine longitude.
Specifically:
> One degree of latitude equals approximately 364,000 feet (69 miles), one minute equals 6,068 feet (1.15 miles), and one-second equals 101 feet. One-degree of longitude equals 288,200 feet (54.6 miles), one minute equals 4,800 feet (0.91 mile), and one second equals 80 feet.
> Sextants measure the angle between the sea horizon and a celestial body. These angles are measured in degrees and minutes of arc (1/60th of a degee). Measuring this angle to an accuracy of 1 minute of arc (1') will result in a positional accuracy of 1 nautical mile. Accurate sextants can measure this angle to an accuracy of 0.2'. This means that theoretically one can determine their position to 1/5 of a mile [~320m]. Additionally, a good clock is required to accurately compute the GP of the celestial body. An error of 1 second in the clock will create a positional error of up to 1/4 of a mile [~400m].
Airplanes also used to use sextants for trans-oceanic flights (when beyond radio beacon range). Part 63 of FAA regulations has a section on Flight Navigators:
My dad was navigator on a B-17 in WW2. The B-17 crews had to fly their own birds from Newfoundland to England. Many crews and their airplanes were lost without a trace on this trip. (There was no hope of rescue if you went down in the North Atlantic. You just prayed nothing went wrong with the engines or the navigation.) My dad was proud he crossed the coast of Ireland within a mile of where he was aiming for.
His tool was a sextant and math.
He was made lead navigator for his bomb group.
He wanted to be a pilot, but because he was a math nerd, the Air Corps told him he was gonna be a navigator.
It's actually surprisingly unknown that throughout the war there was a lot of cat and mouse games over electronic navigation aides for bombers (the so-called "Battle of the beams").
It wasn't until those beams and later air to ground radar (H2S, allegedly named as such because it "stunk" that no one else had done it already) that the truly destructive strategic bombing came about.
Indeed, airliners used to have a domed window on top, just behind the cockpit, where the navigator could stick their head and take star fixes. The Apollo astronauts also did this when they needed to realign their gyros.
Technically yes but the user of that system doesn't have to have an accurate clock, the GPS system itself provides the time signal. I think the first stage of GPS positioning is clock signal acquisition and training the local oscillator. Once synced you can proceed with almanac download or whatever comes next.
(The accurate (atomic) clocks are on the satellites and base stations.)
GPS receivers are required to have excellent stability (<<1ppm), but not accuracy. The problem of marine chronometry was also about stability and not accuracy. After all, if you have a stable but inaccurate clock and can't tune it for some reason, it's straightforward to compensate for the offset when measuring to calculate your position.
Seiko also builds watches that use GPS signals to set the time zone and the time.[1] Seems like a cool idea if you cross time zones often and have an extra $2,000 on your hands.
I think most phones with a GPS receiver don't use it for time, they use the native network time signal instead. This is annoying as the (formerly) GSM time signal in many cases used to be _way_ off, like, 5-10 seconds off. This situation has improved somewhat in the last 10 years or so.
The more I learn about old-timey maritime navigation, the more I realize that (like a lot of skilled trades) it must have been nearly indistinguishable from literal magic to laypeople.
I wonder if that fact inspired the mysterious and bizarre Guild Navigators in Dune.
Funny to see a fairly technical article on mechanical watches rank so high on HN.
Yes, quartz watches are simpler and more accurate yadayada, but there is an interesting engineering challenge in trying to improve accuracy (1 second/day = 10 ppm!) through purely mechanical means.
Long ago my dad (an experimental physicist) had a self-winding mechanical watch.
At the end of the day, he would check if the watch had gained or lost time during the day (compared to the broadcast bell chimes radio stations used to broadcast on the hour).
Based on the observation he would set the watch on its right side or left for the night, causing the spring in the self winding mechanism to be slightly more or less tensioned, leading to a slight gain or loss of time by morning to tune the watch to the precise broadcast signal.
Yes! On average, my Orient watch runs about 15-20s fast per week when I lay it flat on its back at night. But when I position it crown down at night, it runs about 4s slow per week. I usually check it every Sunday morning, and if it's running a bit slow, then I'll put it on it's back for a couple of night to let it catch up. Haven't had to pull the crown out and adjust it for about 2 months now with this system.
Thanks for that story. While reading the article I always wondered what the value of the inaccuracy they where talking about really is. I guess it varies for different manufacturers. But seeing this shows me that even in the past it was only a matter of a few seconds.
An interesting race will be improvements to the local oscillator versus continued size and power reductions for grabbing a GPS signal every hour to correct your clock. No predictions here, though, of course, an increasing number of watches do have GPS for other reasons. (And the GPS solution doesn't help you if you live on a submarine)
RFID-style sync to some device with accurate time (probably a phone) would be the best way to keep an electronic watch mechanism accurate on a day-by-day basis.
Something that really blew my mind is how far mechanical accuracy can be taken when money is no object.
It's said that George Daniels could make mechanical watches that were more accurate than the quartz watches of his day[1], 0.4 seconds per year.
Of course that doesn't really matter if the thing is a priceless work of art that no one can afford to wear day to day but it's still really interesting to know how he did it.
Quartz watches' real advantage is just their accuracy but that they are accurate and cheap to make. There's something rather wonderful about a practically disposable Casio like a Bic Biro.
Well, both things are true. The technological advancement required to make them was immense. Now that we can stand on the shoulders of those giants, they are ~1 order of magnitude simpler to make.
Do you think that people buying mechanical watches are under the illusion that they're more accurate?
Some people surely buy them for the status symbol aspect, and those are definitely the ones we notice because their watches are big and tacky.
There are far more mechanical watches who don't wear brazen Rolex-esque "look at me" watches. You may not notice us because we're wearing Seikos, Citizens, Timexes, etc.
Ironically the accuracy and ubiquity of mobile phone based timekeeping has made mechanical watches more practical. You don't need your wristwatch to be hyperaccurate if you're carrying a phone.
when a $10 digital watch does the job of timekeeping just as well
Surely not everything in your life is chosen strictly on terms of absolute performance?
I'm sure you own some articles of clothing that were chosen despite the existence of cheaper alternatives that are functionally equivalent for your use cases.
And yet, arguably, that's exactly what most app store subscriptions are, priced at several multiples of what that lightweight utility would have cost to purchase before the subscription model.
There's some really expensive software for CAD etc.
And there's bespoke software, that can be almost arbitrarily ludicrous in terms of cost. I'm talking about eg corporate in-house software that has whole teams developing it, but only a few users.
But I guess these examples are mostly just very expensive, but you don't use this software to brag to your mates.
> Typical crystal acceleration sensitivities will range from about 1 x 10-10 per g for specially constructed precision SC-cut and AT-cut resonators to the order of 10-7 per g for tuning fork resonators.
https://youtu.be/O-6KhwwM66k?si=W0-jQ033m_kkv7-R&t=200