How watches work
Discussion
I hope starting a new thread is ok, but cyberface gave me a really great reply to a question I asked in another thread and I thought if I posted his reply here I could avoid cluttering up the wrist check thread with off topic discussion, and maybe some others can benefit from it.
I'm trying to remember how I started. I think it may have been a couple of fakes, a starter box of Chinese tools, a break between consultancy gigs and too much free time, and lots of beer 
I would have got all I know off the internet, which is sometimes dangerous... firstly I'd look at nailing the jargon, knowing what the different parts of a normal mechanical watch are called. A mechanical watch is a pretty understandable piece of kit when the torque flow 'clicks' in your head. Basically a big mainspring tries to spin the hands round freely, but the release of energy is controlled by the pallet fork oscillating on the escape wheel, allowing one tooth to pass the fork for each oscillation of the fork. The fork is then connected to a specific oscillating mass that has a known rate of oscillation (old clocks used pendulums - same concept) - this is the 'balance wheel' and has a hairspring attached to it so the balance wheel can swing round one way until the hairspring tightens, then swings back the other way (switching the pallet fork's position and allowing the escape wheel to advance one tooth) until the spring is over-extended, at which point the balance wheel swings back again, etc. etc.
Timekeeping is all about keeping the balance wheel oscillation steady. Accuracy is then mathematically determined by the rate of oscillation, with faster oscillations allowing greater accuracy (statistically it makes sense to me, but in the real world I can imagine that a very fast oscillating system could theoretically have *more* deviation from the mean if lots of random quantum effects are present, so I'm interested in how the atomic clocks are so accurate...). So pendulum clocks are *theoretically* less accurate than old 1 Hz pocket watches, then a low-beat wristwatch movement at 21,600 bph (6 half-oscillations per hour i.e. 3 ticks per second), then the average 28,800 (4 beats per second, a general standard for wristwatches), then the 36,000 bph specials (5 ticks per second - notably, the Zenith El Primero movement was this fast - and it's a challenge getting balance wheels and springs handling this type of acceleration). Think of the engineering challenges as exactly equivalent to increasing the maximum rev limit in a piston engined car - the higher the engine speed, the faster the oscillation of masses, and the greater accelerations the masses (pistons in the car, the balance wheel in a watch) have to endure at each reversal of motion.
Balance wheels don't tend to oscillate faster than this. To get greater accuracy, first a tuning fork was used (Bulova Accutron, a transistor and coil-pumped tuning fork that oscillated at (initially) 300 times a second (hertz, or Hz), with each oscillation of the tuning fork pushing an exquisitely fine toothed wheel (the wheel had 320 teeth and was 2.4 mm in diameter!) to drive the entire movement. The tuning fork technology didn't last long, as quartz came on the scene, which is effectively a solid-state tuning fork, with the quartz crystal resonating at 32,000-odd Hz. The super-accurate atomic clocks used in astrophysics, GPS, etc. use the oscillation of a big group 1 metal atom (rubidium and caesium, usually) between two electronic states (as the atom switches from an excited electronic state to the less excited state, it spits out the energy in the form of a photon which can be captured and measured) which goes up to 9 billion Hz. I've just checked my facts and apparently someone is using strontium now for a super-accurate clock... that's group 2 on the periodic table but the concept is the same.
Veering *hugely* off topic, sorry. Back to where I started with mechanical watches? I've checked my bookmarks, and The Watch Guy Beginners Guides look like where I started. Get a basic hand-wound really cheap movement for this. The Chinese low-grade fakes are good for this. I've already mentioned counterfeit tools elsewhere IIRC - the big two 'cartel' guys who sell most of the high-end 'replica' fakes also sell nice counterfeit copies of Bergeon watchmaking tool sets. As a beginner it's crazy to buy Bergeon tools as they're really expensive. The Chinese copies of the starter set I bought are decent, usable, average quality tools. The only area where they're shoddy is screwdrivers - it's best to get German or Swiss screwdrivers as a bad screwdriver can wreck the slot of a screw and make reassembly of your movement very hard. Apart from buying expensive screwdrivers though, the Chinese tool sets are perfectly adequate for a beginner. If you're messing about with fake Rolexes (probably the most popular counterfeited brand, so lots available at low prices) then you'll need a Rolex caseback opener. Again the fake cartel guys sell a nicely made Chinese aluminium Rolex caseback opener set, very similar to the Bergeon set in appearance and even the same colour and design box!! The aluminium chucks are great since they won't turn the fine 'teeth' on a steel Rolex (or fake Rolex) case back. It's worth having a Rolex caseback opener anyway - fakes have got so good these days that if I was buying a used Rolex, I'd only part with my cash after having opened the watch and seen the movement...
Once you've done a couple of basics, then there's the TimeZone Watch School which looks pretty comprehensive for a beginner. I was all up for starting this (it's only $50 a course, but you've got to buy tools and the blank movement to practice on) but I found myself a consultancy job and was back working flat out. Now I've got the money to buy watches but not the time to tinker with them!
One thing that is essential really is a controlled working environment - ideally a clean, flat, empty desk with good lighting, and the floor below being tiles or lino... carpet eats tiny movement parts that inevitably ping from your tweezers. A 'magnetic brush' can be bought to sweep over the floor and pick up parts, but will obviously only work on ferrous metals. Parts pinging away into space when you've desperately tried for the 83rd time to get the screw in the hole without dislodging the keyless works... part of the learning process I haven't followed my own advice and try to tinker at my regular office desk, which has thick pile carpet underneath, loads of stationery and computer equipment everywhere, and terrible lighting. I've wrecked a couple of watches simply by an essential part pinging from my tweezers into the twist in space where time becomes a loop. In other words, don't decide to completely strip, clean, lubricate and reassemble your JLC gyrotourbillon or Patek perpetual calendar as a first project 
Apologies, this has veered *wildly* off topic. Mods, feel free to shift this and Debaser's question to a new thread if you want - I'm sure there'll be plenty of people who know MUCH more than me who can correct my little mistakes and suggest other resources or approaches to the job.
My next post will be of my latest new acquisition so this post can be moved if desired.... hope this helps, Debaser.Great post cyberface, many thanks! I'm going to read up on the parts of the watch you mentioned (and no doubt be back with some questions!)
cyberface said:
Debaser said:
cyberface said:
The fine adjuster is dead on centre and the Seiko watchmakers don't mess about so I'd expect it to be as good as it's going to get. But any consistent slow or fast running can be adjusted by the fine adjuster on the balance. Variability from day to day can't really be adjusted but I'm not trying to time moon landings, bomb drops or missile launches - I just need to make sure I don't miss my train Everything else is gravy. So a minute a day would have me tinkering, anything else I ignore.
Usually it's just a question of slow running being due to age and lubricants drying up - to a point you can correct the increase in friction by adjusting the balance adjuster... but it's better to strip the movement and re-lubricate really. I'm not a skilled enough watchmaker (yet) - it's a hobby but I'm not confident enough to strip the Seiko automatic yet.
Thanks for your reply cyberface. I've been thinking for a while I'd like to learn more about how mechanical watches work, and how to perform some simple tasks on them. Is there anywhere you would suggest a beginner to mechanical watches could look to learn more about them? Everything else is gravy. So a minute a day would have me tinkering, anything else I ignore.Usually it's just a question of slow running being due to age and lubricants drying up - to a point you can correct the increase in friction by adjusting the balance adjuster... but it's better to strip the movement and re-lubricate really. I'm not a skilled enough watchmaker (yet) - it's a hobby but I'm not confident enough to strip the Seiko automatic yet.
I would have got all I know off the internet, which is sometimes dangerous... firstly I'd look at nailing the jargon, knowing what the different parts of a normal mechanical watch are called. A mechanical watch is a pretty understandable piece of kit when the torque flow 'clicks' in your head. Basically a big mainspring tries to spin the hands round freely, but the release of energy is controlled by the pallet fork oscillating on the escape wheel, allowing one tooth to pass the fork for each oscillation of the fork. The fork is then connected to a specific oscillating mass that has a known rate of oscillation (old clocks used pendulums - same concept) - this is the 'balance wheel' and has a hairspring attached to it so the balance wheel can swing round one way until the hairspring tightens, then swings back the other way (switching the pallet fork's position and allowing the escape wheel to advance one tooth) until the spring is over-extended, at which point the balance wheel swings back again, etc. etc.
Timekeeping is all about keeping the balance wheel oscillation steady. Accuracy is then mathematically determined by the rate of oscillation, with faster oscillations allowing greater accuracy (statistically it makes sense to me, but in the real world I can imagine that a very fast oscillating system could theoretically have *more* deviation from the mean if lots of random quantum effects are present, so I'm interested in how the atomic clocks are so accurate...). So pendulum clocks are *theoretically* less accurate than old 1 Hz pocket watches, then a low-beat wristwatch movement at 21,600 bph (6 half-oscillations per hour i.e. 3 ticks per second), then the average 28,800 (4 beats per second, a general standard for wristwatches), then the 36,000 bph specials (5 ticks per second - notably, the Zenith El Primero movement was this fast - and it's a challenge getting balance wheels and springs handling this type of acceleration). Think of the engineering challenges as exactly equivalent to increasing the maximum rev limit in a piston engined car - the higher the engine speed, the faster the oscillation of masses, and the greater accelerations the masses (pistons in the car, the balance wheel in a watch) have to endure at each reversal of motion.
Balance wheels don't tend to oscillate faster than this. To get greater accuracy, first a tuning fork was used (Bulova Accutron, a transistor and coil-pumped tuning fork that oscillated at (initially) 300 times a second (hertz, or Hz), with each oscillation of the tuning fork pushing an exquisitely fine toothed wheel (the wheel had 320 teeth and was 2.4 mm in diameter!) to drive the entire movement. The tuning fork technology didn't last long, as quartz came on the scene, which is effectively a solid-state tuning fork, with the quartz crystal resonating at 32,000-odd Hz. The super-accurate atomic clocks used in astrophysics, GPS, etc. use the oscillation of a big group 1 metal atom (rubidium and caesium, usually) between two electronic states (as the atom switches from an excited electronic state to the less excited state, it spits out the energy in the form of a photon which can be captured and measured) which goes up to 9 billion Hz. I've just checked my facts and apparently someone is using strontium now for a super-accurate clock... that's group 2 on the periodic table but the concept is the same.
Veering *hugely* off topic, sorry. Back to where I started with mechanical watches? I've checked my bookmarks, and The Watch Guy Beginners Guides look like where I started. Get a basic hand-wound really cheap movement for this. The Chinese low-grade fakes are good for this. I've already mentioned counterfeit tools elsewhere IIRC - the big two 'cartel' guys who sell most of the high-end 'replica' fakes also sell nice counterfeit copies of Bergeon watchmaking tool sets. As a beginner it's crazy to buy Bergeon tools as they're really expensive. The Chinese copies of the starter set I bought are decent, usable, average quality tools. The only area where they're shoddy is screwdrivers - it's best to get German or Swiss screwdrivers as a bad screwdriver can wreck the slot of a screw and make reassembly of your movement very hard. Apart from buying expensive screwdrivers though, the Chinese tool sets are perfectly adequate for a beginner. If you're messing about with fake Rolexes (probably the most popular counterfeited brand, so lots available at low prices) then you'll need a Rolex caseback opener. Again the fake cartel guys sell a nicely made Chinese aluminium Rolex caseback opener set, very similar to the Bergeon set in appearance and even the same colour and design box!! The aluminium chucks are great since they won't turn the fine 'teeth' on a steel Rolex (or fake Rolex) case back. It's worth having a Rolex caseback opener anyway - fakes have got so good these days that if I was buying a used Rolex, I'd only part with my cash after having opened the watch and seen the movement...
Once you've done a couple of basics, then there's the TimeZone Watch School which looks pretty comprehensive for a beginner. I was all up for starting this (it's only $50 a course, but you've got to buy tools and the blank movement to practice on) but I found myself a consultancy job and was back working flat out. Now I've got the money to buy watches but not the time to tinker with them!
One thing that is essential really is a controlled working environment - ideally a clean, flat, empty desk with good lighting, and the floor below being tiles or lino... carpet eats tiny movement parts that inevitably ping from your tweezers. A 'magnetic brush' can be bought to sweep over the floor and pick up parts, but will obviously only work on ferrous metals. Parts pinging away into space when you've desperately tried for the 83rd time to get the screw in the hole without dislodging the keyless works... part of the learning process
I haven't followed my own advice and try to tinker at my regular office desk, which has thick pile carpet underneath, loads of stationery and computer equipment everywhere, and terrible lighting. I've wrecked a couple of watches simply by an essential part pinging from my tweezers into the twist in space where time becomes a loop. In other words, don't decide to completely strip, clean, lubricate and reassemble your JLC gyrotourbillon or Patek perpetual calendar as a first project Apologies, this has veered *wildly* off topic. Mods, feel free to shift this and Debaser's question to a new thread if you want - I'm sure there'll be plenty of people who know MUCH more than me who can correct my little mistakes and suggest other resources or approaches to the job.
My next post will be of my latest new acquisition so this post can be moved if desired.... hope this helps, Debaser.
andy_s said:
Only Cyberface can get the phrase 'periodic table' into a thread about watches! Interesting stuff.
Sorry chaps. Anyone care to explain why the scientist chappies all went with group 1 metal ions for their atomic clocks... and then suddenly the 'most accurate' clock according to the Internet is a strontium (group 2 metal) ion? Is it that the energy difference between two well defined strontium ion excitations is more accurate / more distinct from other excitations / etc. than the group 1 metal clocks (e.g. caesium / rubidium)?
Or am I talking drunken shyte based on university level chemistry that is *well* below the knowledge required to understand how this works (hell, it's not just chemistry (for the ions and electronic energy states) but also physics (for energy state conversion to photons (microwave frequency for detection) and resonance / oscillation of said states)....
One vote for the TimeZone watch school here. Very absorbing and astonishingly all the bits went back in and the watch keeps good time! Taking a watch to bits is without doubt the best way to understand it (although the keyless works are still a bit of a mystery to me - there is a clutch in there somewhere...). The school tool kit is also recommended.
Gassing Station | Watches | Top of Page | What's New | My Stuff