How a manual transmission works: sliding mesh, constant mesh and the reverse idler

Originally published on MotoSutra in January 2021, and re-verified against current reference material during this recovery. A gear slid into mesh, a shaft that turns inside a set of gears spinning loose around it, a brass ring whose only job is friction, and a third idler that turns the direction of the world around: the explanation the 2021 article built, in the order it built it.

On the missing figures. The original kept pointing at diagrams — a basic mechanism is shown, as shown here — and those drawings did not survive the archive. Every sentence that leaned on one has been rewritten to carry the idea in words instead. No substitute artwork has been added, and nothing here should be read as a figure that once existed. Two overstatements in the original, one about which gear is direct and one about reverse, are corrected where they appear.

The basic question: why does an automobile need a transmission?

Power generated by the engine flows through the transmission before it reaches the drive wheels. The transmission’s basic job is to control the speed and the torque delivered to those wheels, for different driving conditions.

The need comes from a property of the engine. A given amount of power can appear as a lot of force moving slowly or a little force moving fast, and the engine itself cannot choose between those — it makes its power within a limited band of engine speed. So if you want to climb a hill you need more torque, and you get it by reducing the speed at the transmission: less output speed, more output force, from the same power input. Conversely, when the torque demand is low, the transmission ratio can be raised so the engine turns more slowly at the same road speed.

Now let’s look at its inner workings

Manual transmissions work on the simple principle of the gear ratio. The basic layout is an input shaft and an output shaft, connected through a countershaft — a layshaft carrying its own fixed gears, so that the two shafts are always in line with each other but never touch.

The oldest way to change ratio on that arrangement is to slide a gear along the output shaft until its teeth mesh with a matching gear on the countershaft. Different gears give different ratios, so a five-speed box has five of them to choose between, and this design has a name: sliding-mesh. It controls speed well enough, but it has an inherent disadvantage, and it is a mechanical one rather than a driving one.

To slide a rotating gear into mesh with another rotating gear means forcing the teeth to meet while they are moving at different surface speeds. That is difficult, jarring and destructive, which is why sliding-mesh boxes need double-declutching and why the arrangement disappeared from cars long ago.

Constant mesh solves it permanently, and the fix is to stop moving the gears into mesh at all. In a constant-mesh box every gear pair is always in mesh. What changed is where the gears sit: the gears on the output shaft are cut with plain bores and ride loose on it, spinning freely whenever they are not connected. Because each loose gear turns at its own speed, whichever one you happen to be driving through, the box can stay permanently geared and still offer a choice of ratios.

Now the problem is smaller and sharper: connect one loose gear to the shaft, and the shaft takes on that gear’s speed. Different ratios are then just different choices of which gear to lock down. It is worth being clear about what “lock down” means, because the whole difficulty has moved here — the gear you want is already spinning at some speed and the shaft is spinning at another, and locking them together means making them agree.

The art of joining a freely turning gear to a rotating shaft, smoothly and in a fraction of a second, lies at the heart of the manual transmission. Everything below is that one problem.

One ratio in the set is special: the gear that connects input to output directly, with no countershaft in the path, gives a 1:1 ratio called direct drive. Gears taller than direct are overdrive, and give an output shaft turning faster than the input.

Corrected. The original described a five-speed layout but then pinned specific behaviour to a “sixth gear”, and stated flatly that “fifth gear is used to turn the output shaft at a higher speed than the input shaft.” Those two statements contradict each other, because which gear is direct and which are overdrive depends entirely on how many ratios the box has: in a five-speed, direct is typically fourth and fifth is the single overdrive, while a six-speed puts direct further up and has two overdrives. The underlying principle — one direct ratio, taller ones above it — is the reliable part, so that is how it is stated here rather than by naming a gear number the drawing would have settled.

Let’s see how this is done in actual practice

The mechanism that answers the locking problem is the synchroniser, and its parts are worth naming because each has one job.

On the main shaft there is a hub fixed in place, splined so it turns with the shaft and cannot move along it. A sleeve rides on that hub, splined the same way, but free to slide axially. The sleeve is the part your shift fork actually pushes. The loose gear, meanwhile, carries two features: a set of internal dog teeth for the sleeve to grab, and a coned face next to them.

If the sleeve were simply driven straight at the dog teeth, the gear and the shaft would have to be at exactly the same speed already, which is the thing you are trying to achieve. So the sleeve first meets a synchroniser ring — a brass-coloured ring with a coned outer surface that fits over the gear’s cone, and with four pins or slots so it can rotate along with the hub while still being free to slide.

The sequence during a shift is then:

1. The clutch pedal is pressed, so the flow of power from the engine to the gearbox input is discontinued and the shafts are unloaded.
2. The shift fork moves the sleeve toward the gear. The sleeve’s internal teeth contact the synchroniser ring and push it against the coned face of the gear.
3. Friction does the work. The cone pair is the whole point: rubbing together, the gear’s cone and the ring’s cone drag their two speeds toward each other. Because the ring is wedged against the gear by the sleeve’s pressure, the gear is pulled toward the hub’s speed rather than the other way around.
4. Once the two are essentially matched, the ring can no longer resist being pushed further, the sleeve slides past it, its teeth engage the gear’s dog teeth, and the gear is now locked to the shaft.

The gear and shaft turn together, and the drive path is complete. The same mechanism serves every gear in the box. And the reason the cone has to do its work first is that a dog tooth is a square block: forcing two square blocks together while they spin at different speeds is exactly the grinding you hear when a shift is made badly, or when a synchroniser is worn out.

Now let’s see how the reverse gear works

Everything above is forward motion, and reverse needs a different answer, because a gearbox that reverses its output shaft’s direction of turn would have to do so somewhere.

Reverse uses a three-gear arrangement, and the extra gear in the middle is called an idler. Two gears in mesh turn in opposite directions; putting a third between them flips the direction once more, so the output shaft turns the way it would not have. When the idler is slid into mesh with the other two, the output shaft reverses.

It is a simple device, and its simplicity shows in how it must be used.

Corrected. The original stated as fact that “the reverse gear does not have a synchronizer ring mechanism”, and concluded that the gearbox rotation must stop completely before reverse can be selected. That is a sound description of the traditional layout — reverse is commonly unsynchronised, which is precisely why it sometimes grinds when you slot it in while the shaft is still turning, and why the clutch-down-and-wait habit works. But “does not have” is a universal claim, and plenty of modern boxes do synchronise reverse. So the correct reading is a practical one: reverse may or may not be synchronised, you cannot tell from the gear lever, and the habit of stopping the rotation before engaging it is good advice whatever your car has in there.

The full taxonomy of gearbox designs, including how a manual relates to automated and dual-clutch versions of itself, is covered in types of automobile transmission.