Inside the testing facility where Dyson breaking its own devices
Have you ever dropped your vacuum cleaner? How about dipped your hair dryer in a bucket of water? Or slammed a drawer shut on your desktop lamp? It might not be something most people do to with their prized possessions, but in Dyson’s weird and wonderful Malaysian testing facility this precisely what happens 24 hours a day.
Although Dyson claims to enjoy designing new technology, when you visit their testing facility you’d be forgiven for thinking they secretly prefer breaking it.
The first thing that strikes you when you enter their Malaysian test centre (and if you aren’t paying close enough attention it could very well be a vacuum cleaner being waved around by a robotic arm) is the noise.
While the rest of Dyson may be full of aesthetically pleasing, clean white surfaces, the facility they use to test machines could be diplomatically described as: ‘practical’. The rigs used to test new technologies are reminiscent of creations employed by Willy Wonka in his chocolate factory.
Forget about brushed aluminium and robotic arms they use to make their motors, in the testing spaces you’ll find more concrete, wood, and bits of old rope. But don’t mistake hardy materials for a hard up facility. As testing facilities go this one’s about as tough as they get.
A common misconception about testing is that the purpose is to see if an individual product has a fault, whereas it is actually a game of statistics. Like polling, testing products involves taking a sample batch and extrapolating the results across the production line as a whole.
When you make over 25 million high quality products, the last thing you want to do is smash all of them to bits before putting them on the shelf to be bought. Instead, representative sample groups are tested regularly for weaknesses.
“Materials aren’t static, they change over time and under different conditions,” explains Dyson’s head of testing, Marco Li.
“When we select the little rubber seals we want to use in vacuums, for example, we have to account for the fact that when we ship it in a container it may get cold and contract on a cold night in the mid-Atlantic.
“Or when it is stored before it reaches a shop in Dubai it may get really hot and expand. The tolerance we test for takes these eventualities into account. It means that we can predict that the product that ends up in a consumer’s hand will be the same wherever they are.”
“Each vacuum cleaner alone is dropped from a meter up exactly 100 times and will vacuum over 750 miles of surfaces.”
This begs the question, do products ever fail tests?
“Yes,” says Marco with a wry smile, “everything we test will fail, in one respect, because the way we test our products we want to make sure it fails. That’s the only way we can be sure we know what the limits of our technology are and make sure that what we say on the box isn’t the best, but the worst-case scenario.”
This is a subtle but interesting point. At Dyson the “worst possible outcome is their testing baseline”. According to Marco, “there are always variations in the manufacturing process and because we make such highly engineered machines, slight differences in the manufacturing process that can potentially change the performance of the machine. That’s why the claims we make are based on the worst possible version of our product — and we try to set that bar higher than most companies’ best estimate.”
This doesn’t always make Marco and the testing team particularly popular with the product engineers and designers.
“The tests we are doing are very aggressive. Often the engineers ask to see how we are actually testing and they’ll say: ‘This is too aggressive.’ At the beginning of the process we have to send a lot of technology back to development. Sometimes, I have to say: ‘It is failing. Give me a better one’.”
“One of the tests is just dropping the motor on the floor over and over again, it looks needlessly aggressive, but it ensures the final motor is really robust. Another test we do is called ‘motor life testing’. We want to see how long it can keep running and running until it fails.
“Even when it does fail, we want to make sure it fails safely. And that’s not all; we don’t test it in conditions that will help the motor perform better. Instead we actually keep adding different types of dust and different types of contaminants to see whether the machine can take it.
“For the new V10 motor this is maybe 400 hours with this type of non-stop loading. By benchmarking against the last motor we can tell the engineers who designed it that this is clearly a superior machine to the last generation.”
This aggressive approach isn’t limited to the motor either. Each vacuum cleaner is dropped from a meter up exactly 100 times and will vacuum over 750 miles of surfaces. Banks of testing rigs automatically smash the devices and products Dyson engineers have spent years crafting into concrete blocks. There are even ‘obstacle courses’ which even the most brutal drill sergeant would consider excessive where vacuum cleaners are scraped over breeze blocks and dented on step corners.
The testing professionals even tell stories about Sir James visiting the facility and rebuking them for “not slamming the drawer on a hairdryer hard enough” before reportedly demonstrating the required forcefulness himself.
“Most of our testing starts with two areas,” explains Marco. “The first asks: how robust is the product? The second is: does the performance of the product meet the claims and standards we need to deliver at the (usually global) launch of a new product?”
During our visit in February, it is the new cord free V10 vacuum cleaners that are taking the brunt of the bashing. Everywhere you look hundreds of distinctive copper vacuum cleaner wands are being dropped, slammed, bashed, and bent. To one side of the room a person in a lab coat is cataloguing the inflicted damage like a pathologist.
Ahead of the launch of V10, over 900 vacuum cleaners were tested in February alone. Considering each testing cycle takes 600 hours to complete, the combined time Dyson spent testing their latest machine added up to well over 700,000 hours. All, just to ensure that Dyson can say confidently that they understand their technology’s limits.
“The investment we have in testing is huge. I have never seen a company that invests as much into testing products outside of heavily regulated industries like automotive or aerospace manufacturing,” says Marco.
“When I arrived [at Dyson] I came in and explained that, in the wider industry, the standard is three drops from one metre high. ‘Three drops. It survived? It’s good.’ In fact, the first thing I learned when I came to Dyson was that three drops isn’t enough. There isn’t any regulation that says you have to drop a vacuum 100 times.”
Nevertheless, this is exactly what Dyson chooses to do to its machines. In fact, each prototype’s building process includes over 150 individual tests.
And the world of testing in Dyson is still growing. In the past five years the company has invested heavily in new technologies including AI, robotics, and software. This has opened up a new frontier in digital testing.
As Marco explains, “in 2017, we placed a strong focus on software at a product level. As we started to deliver connected products, we realized that we needed to enhance our testing strategy to include software so that we could confidently cover the whole customer experience, even on their devices. For example, we take the product out of the box and check, how well and easily it connects to the cloud. To do that we have a variety of different types of phones and routers. They may all perform differently. So, we try to make sure that the roll out of any software will work at a set benchmark level across the all the different scenarios — even if someone is using the old versions of that system. Imagine that, all these different household experiences, the number of possibilities is vast. There are so many different configurations that need to be tested. This is the next frontier in testing.”
Words: Henry Tobias Jones | Photography: Matthew Beedle
This article was originally published in on:Revolutions in May 2018.