Mechanical Watch (2022)

In the world of modern portable devices, it may be hard to believe that merely a few decades ago the most convenient way to keep track of time was a mechanical watch. Unlike their quartz and smart siblings, mechanical watches can run without using any batteries or other electronic components.

Over the course of this article I’ll explain the workings of the mechanism seen in the demonstration below. You can drag the device around to change your viewing angle, and you can use the slider to peek at what’s going on inside:

What you see here is known as the movement – the inner part of a mechanical watch that’s usually enclosed in a metal case. In this article I’m focusing on a watch movement itself, since beautiful watch cases merely hide the intricate mechanisms which are the real stars of the show.

The entire watch movement has a lot of parts, and in this blog post I’ll explain the purpose of each one. The world of watchmaking is jargon-heavy, so many of the components may have unfamiliar names, but you shouldn’t feel pressured to remember them – the names and parts will be color-coded for easy reference.

In a functioning watch many parts are in constant motion. By default all animations in this article are enabled, but if you find them distracting, or if you want to save power, you can globally pause all the following demonstrations.disabled, but if you prefer to have things moving as you read you can globally unpause them and have animations running.

While the entire watch movement has many parts, the timekeeping system, which forms the core function of any watch, consists of just seven major elements which we can lay out in a straight line:

It may not look like much, but these parts still have a lot of interesting details about them that contribute to the second hand rotating at a correct pace. We’ll start exploring these details by focusing on the source of power for this entire contraption.

Purely mechanical devices have a few different ways to power themselves, but one of the simplest methods to store energy is to use a spring. Most springs we see in daily life are coil springs. In the demonstration below, you can move the mass attached to this type of spring to see it bounce:

When a spring like this is compressed, it stores some energy that is then released when the compressing tension is removed. Mechanical watches typically use a different kind of spring – a spiral torsion spring. This type of spring is loaded when it’s twisted. When let go, the spring unwinds in the opposite direction to eventually settle in its natural state:

In a mechanical watch, we ultimately want to show rotating hands, so a spinning motion that a torsion spring provides is particularly useful. A spring in a typical mechanical watch has a slightly more complicated shape – you can see it below in its relaxed state. By dragging the slider you can try to wind it midair, but as soon as you let go, it will snap back to its original shape:

As you can see, this spring is quite strong and it wants to expand very rapidly. To contain the spring we have to put it in a casing known as a barrel:

Once in the barrel, the spring still wants to expand to its original state, but the barrel’s wall keep it in place. This spring is the storage of energy for the watch and its name, the mainspring, reflects its importance.

Unfortunately, we can’t really get any useful work from the mainspring in this state – it has already expanded to the largest possible size. To store more energy in it we need to wind it tightly using the arbor that we’ll first attach on the inner side of the mainspring:

If you look closely, the mainspring has a little hole near its end – you can see it in the center of the demonstration. The arbor has a little hook that grabs onto that hole:

When the arbor is turned, it pulls the mainspring with it, causing it to wind. In the demonstration below, we’re holding the barrel tight, and you can turn the arbor by dragging the slider:

Notice that as soon as you let go of the arbor by releasing the slider, the mainspring will turn the arbor right back. This is less than desired – we want the barrel to turn instead, so that it can power the other parts of the watch. To get some useful work from the mainspring, we’ll have to keep holding on to the arbor and instead let the barrel go when we want to use the stored energy:

We’ll soon see how this is accomplished in practice, but for now we’ll assume that the arbor is held tight and the mainspring ends up rotating the barrel, just like in the demonstration above. Before we finish up with the mainspring and the barrel, let’s discuss two other details that make this mechanism more reliable. Let me bring up the relaxed spring one more time:

The metal strip attached to the mainspring provides additional tension to its outer part. That metal strip really wants to snap back to its straight shape, so it pushes against the wall of the barrel, creating a lot of friction that keeps the mainspring in place:

This locks the outer end of the mainspring when the arbor moves the inner. If we were to keep winding the spring past its maximum capacity, we’d overpower that friction letting the mainspring slip inside – this acts as a safety mechanism to prevent the parts from breaking.

As we’ve seen, in its relaxed state, the mainspring forms an S-shape with varied curvature throughout. This helps to balance the tension in mainspring’s different sections when it is inside the barrel. Notice that the inner sections of the wound spring have a much smaller radius than the outer parts. If the relaxed spring was just a straight piece of metal, then after winding, the inner parts would be bent much more than the outer parts. With the S-shaped spring the outer sections of the spring are also under a similar tension because they want to get back to their curve that is bent in the opposite direction.

To secure the mainspring and prevent dust from getting in we close the barrel with a lid that snaps into its place:

We’ve managed to make some parts rotate and one could naively think that we could just attach a watch hand to the barrel to make it track time. Unfortunately, that won’t really work – you can witness this in the demonstration below. You can see how this “watch” behaves after you wind the mainspring with the slider and let it go:

We clearly have some work to do – the hand spins way too fast and it only does a few rotations before the mainspring inside the barrel runs out of the stored energy. Clearly, this contraption won’t let us track time in any reliable way.

If we wanted our watch to run continuously for around 40 hours on a single wind, we’d need the minute hand to complete 40 rotations in that time. Moreover, the second hand should cover around 40 x 60 = 2400 complete rotations in that time. We need to find a way to convert a small number of revolutions of the barrel into a large number of revolutions of the hands. This is where gears come in.

I’ve talked about gears on this blog before, so let me just recap things very briefly. Gears can be used to change the speed of rotation between two different axes. In the demonstration below, you can witness that by observing little dots I put on each gear – the yellow gear, which is powered by the bigger red gear, takes much less time to finish a single revolution:

An important aspect of two matching gears is their number of teeth. Each tooth in one gear meets with a space between teeth in the other gear, so within a unit of time both gears rotate by the same number of teeth. If the number of teeth in two gears is different, those gears can take a different amount of time to complete a single rotation. In the demonstration below, you can change