Last week I was at Logitech’s headquarters in Lausanne, Switzerland, ostensibly to take a look at the G Pro X2 Superstrike but also to get a peek behind the curtain and see how Logitech designs and tests its gaming mice. We actually got a peek behind this curtain over a decade ago, but things have changed a lot since then.
The non-PC gamers in my life were a little bewildered by the fact that I was flying to another country to look at ‘just a mouse’—and even more bewildered when they heard about the $180 price tag—but you and I both know: there’s a lot that goes into making a good gaming mouse.
That being said, even I was surprised at just how much really goes into the design and testing of its mice once I saw it all in person at Logitech HQ. From hand-carving to mechanical flinging, it seems there’s little the company doesn’t account for.
It’s not all giant Faraday-esque rooms and mechanical arms, however—though there is a fair amount of that, too. A lot of what impressed me about my visit was the very human element. I’ve already spoken about how impressed I was by designers using (fake) wood to model mice by hand, but from design and ergonomics through to new tech innovations, I like how grounded it’s all kept.
Anyway, enough rambling. Here’s how Logitech makes its gaming mice.
Design
(Image credit: Future)
It all starts at a coffee machine. Logitech’s chief engineer, Regis Croissonnier, said as much:
“The coffee machine is an interesting area where [we discuss] ideas. We don’t set meetings to [do that], we go to the [coffee machine]. Everyone here in this building, in engineering, has some kind of innovation mindset. So you can exchange by talking.”
I suppose the ideation phase can take place wherever. Though ideas can’t just stay inside heads forever. In the design studio, on a table surrounded by plywood display walls featuring all kinds of peripherals, I was shown a bunch of mouse ideas amidst row after row of prototypes and finished products.
Nick Jinkinson, head of industrial design at Logitech, explains how different mouse ideas are tried out:
“We do cool experiments of, what if a mouse had nothing inside it, you know, how light could it be? What if it’s just made of, you know, strands of plastic? We’ve got designers thinking about that stuff—more far out, you know, wacky ideas—but you need to do that to keep fresh, basically. What if there’s literally no internals, or the internals are a tiny little sliver of something there?”
(Image credit: Future)
These “wacky” experiments are brought to life by 3D printers, three of which were whirring along one side of the room. But it’s not all CAD and 3D printing. I was particularly struck by the fact that there was a chemical wood block with pencil drawings on it, and Jinkinson told me more about this:
“Our designers actually start quite a lot by hand… It’s kind of old school, it’s analogue, which we like. The reason for that is because, as a designer, you need to both see the shape you’re making, but also touch the shape you’re making, right? Because they’re handheld products at the end of the day.
“Any new shape, anytime we need to come up with something new… You sketch over on a block, and then use a band saw, and you start to cut out the profiles, and then you fasten it to a vice, and then you shave the shape.”
Presumably, this helps a bit with figuring out the ergonomics. On that front, and in addition to learning from the academic literature, Logitech also does its own testing. I got to witness how some of this is done, and it involves wrist pressure pads and electrode pads. With these fitted, Logi can see how much different muscles are used when holding different mice, and heat signatures can also show where on the mouse there’s the most contact and pressure.
One little tidbit that I took from seeing this, which I found interesting, is that one reason vertical mice are so comfortable and ergonomic compared to regular mice is because, unlike holding a regular one, holding a vertical mouse doesn’t make your two forearm bones cross over each other. I knew vertical mice put your arm and wrist in a more natural position, but I hadn’t put two and two together to consider the bone structure behind it.
So, with ergonomics in mind when necessary—gaming mice often forego ergonomics for performance, of course—and hand-carved designs, Logitech can land on a design to 3D print and test out.
Sensor testing
(Image credit: Future)
There’s more to a mouse than its physical design, of course. Just as important is what goes inside it: its sensor and its wireless receiver. Logitech does a ton of testing and validation of these, too.
It all goes on in the basement, either in machine-laden labs or spooky pointy-wall rooms. Here, engineers put mouse internals, and the mice themselves, through their paces.
For the sensor, there are four main machines. Two of them test the sensor accuracy, one tests its speed and acceleration capabilities, and the final one tests lift-off. It’s important to have machines test these things, rather than humans, because they can be programmed to repeat the same tests exactly. Plus, they can spin under the sensor at speeds that I think even the best plate-spinning magician would struggle to match, which is useful to test the sensor’s limits.
(Image credit: Future)
Accuracy is measured by two machines, one that tests just the sensor across X, Y, and Z axes, and another that tests just the X axis and Y axis, but with the sensor inside the mouse. François Morier, principal engineer of optical sensing at Logitech, explains:
“The motor is moving a given sequence that has been chosen [for a] parameter that we want to watch, and the motor has the ground floor movement. We record the data from the sensor, and we compare with the ground floor movement, so that we can see where we have good performance, what parameter is good, what is wrong, where we have a margin, where we can improve, etc.”
The “ground floor movement” Morier mentions refers to the movements the machine is calibrated to make and what perfect sensor data for those movements would look like. In other words, it’s a perfect baseline that the sensor’s own reports can be compared to. The sensors are plugged into “specific, dedicated acquisition boards” via serial peripheral interface (SPI). This should be more accurate and less prone to system influences than if you just plugged directly into the computer.
With these machines, Morier says, “It’s not quick… But here the goal is really to validate that the base is really good.”
There’s another machine that tests how well the sensor deals with speed and acceleration. It’s basically a big plate that spins around very quickly. In fact, it’s so quick that Morier had to pull up a wire mesh around it before starting the test, for safety, and mouse skates have to be removed so they don’t melt from the friction.
“Here”, he says, “we are kind of looking at the extreme capability of the sensor… It has a 4.5 kilowatt motor to drive this disk and spin at different speeds or accelerations… In terms of speed, we can go up to 1,000 IPS with this setup [about 25 metres per second]. And in terms of acceleration, we can go up to 43 but we can evaluate up to 100 as well. We have parameters so we can validate this.”
Another machine—which I like to call the fling-a-tron but which probably has a much more serious name—flings the mouse around in exactly the same fashion, over and over. Swipe, and lift, and swipe, and lift. This one is used to test sensor lift-off.
Wireless latency can also be tested by comparing when a sensor reports the first data after starting one of the machine’s motors, but for this, a bunch of tests have to be done, and their results averaged:
“The reason why there are multiple results for multiple starts is because when you stop, you never know how far you are from the next count [ie, the next little bit of sensor data]… So when you move, you stop somewhere, are you very close to the next integer, or are you far from this integer? You don’t know… So it is normal to have a distribution of results.”
Wireless testing
(Image credit: Future)
Logitech has a whole separate area to thoroughly test a mouse’s wireless, i.e. RF (radio frequency) connectivity, and it’s by far the freakiest lab of the lot.
Some rooms are anechoic chambers, which are essentially faraday cages on steroids, blocking out all kinds of interference, from sound to electromagnetic waves. And yes, the air felt peculiarly quiet in there, in case you were wondering—having no echo to your voice at all is a weird experience.
Before putting an actual mouse to the test in these rooms, though, first there’s simulation work to figure out where it’s best to put the RF chip and antenna in the mouse.
Frederic Fortin, senior manager of RF at Logitech, explains:
“The game here is to find the right position for the RF chip and the antenna which is linked to the RF chip… Once we are happy with our design, we run the simulation, we see the results. What we see here is what we call the radiation pattern, in 3D. The goal is to have a nice red apple.”
(Image credit: Future)
The “red apple” he’s referring to is a blob that shows a visualisation of how the signal will project from the antenna. Red means “maximum power”, which means “maximum performance.”
After simulating and figuring out where to put the RF chip and antenna, this can be put to the test in the two anechoic chambers. The first one has a giant antenna—a few feet tall, pictured below—next to a smaller one:
“These two antennas are covering the full frequency band. We have to measure in this frequency band to make sure we are complying with the standards: FCC, CE, and other standards, to make sure we are not emitting too much power—but still, we can have good performance.”
(Image credit: Future)
At the other end of the room, the device sits on a table which mechanically rotates around, so the signal from the RF and antenna is measured in 2D.
The second anechoic chamber looks a little like an alien torture chamber, but I’m assured it’s just for testing the wireless signal again, though this time in 3D rather than 2D. Here, tests are only running across 2.4 GHz, as opposed to the previous room, which runs across the whole spectrum.
There’s also an ominous-looking plastic hand, but again, I’m assured it’s nothing nefarious and is used simply to cover up the mouse during testing, to see how it blocks the signal and how much of a problem this will cause.
In another section, there’s also a wooden bench with a mouse atop a spinning rod. The wood looks a bit slapdash in such a carefully crafted facility, but Fortin says it’s used “not because it’s nice, but because it’s transparent at 2.4 GHz, and we don’t want equipment to interfere with the device.”
The rod spins underneath the sensor to get the mouse running up at 8 kHz, and the goal is to measure how good the wireless connection is to its dongle about a foot away. Specifically, Logitech is testing how strong the connection is even when there’s outside interference:
“Thanks to this antenna and this noisy and very hot equipment, we’re able to emit some noise. We are able to mimic the noise from a LAN party… We are playing these different scenarios, and we are measuring the quality of the link, and we are ranking our device and the competitor’s device.”
New innovation
(Image credit: Future)
All of the above is what Logitech does to design and test any mouse, but there are sometimes new and innovative technologies that require different testing. I am, of course, primarily thinking of the Superstrike, the first mouse to use inductive switches to allow for ultra-low actuation and rapid trigger.
The focus on ultra-low and adjustable actuation here requires a focus on click latency, and during my tour of the labs, I got to see some of what goes into testing that.
I’ve written more extensively about the Superstrike’s technology and how Logitech tests it, but the long and short of it is that there’s a good combination of machine testing and human feedback. Logitech also says it gets a lot from its collaboration with EPFL university, upon whose campus the company’s HQ is based, because it gets a ton of interns and academic collaborators, and these collaborators get to use Logitech’s equipment and research capabilities—a win-win scenario.
In addition to getting lots of pro and non-pro gamers to test out the click feel and adjust the haptics accordingly, these gamers also had their clicks measured to test for reductions in click latency. Even the different ways that gamers hold their mouse were observed.
Machines were also used for testing and improvement, of course. In particular, myself and the others being shown around got a look at a big actuator machine, which pushes down the mouse button at a very precise speed. The time taken for a click to actuate is then accurately and precisely measured, and this is where Logitech’s claims about exact click latency reductions for the Superstrike come from.
I think that’s part of what I liked most about what I saw of Logitech’s labs in Lausanne: the mix of machine accuracy and human-first design and testing. Overall, I got the impression that the engineers and designers genuinely do keep end-user usability at the forefront, not allowing that focus to be overshadowed by the expensive machines and data crunching.
