news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Comprehensive Report

17min read

AI hardware entrepreneurs should read the birth story of the world's first smart bracelet

Jingyu 2024/08/18

summary

How to navigate the pitfalls from idea, technology selection, hardware polishing to production and manufacturing?

Author: TEKLA S.PERRY

原题｜ENGINEERING THE FIRST FITBIT: THE INSIDE STORY

Source: https://spectrum.ieee.org/fitbit

Compiled by Jingyu

Editor's note: From rabbit r1, AI Pin to the AI ​​glasses, necklaces and children's dolls that are now flooding in, AI big models have once again set off a wave of entrepreneurship in smart hardware. Everyone is working hard for the title of "the new iPhone in the AI ​​era". But how difficult is the road to hardware entrepreneurship? We may look back and see how Fitbit, the originator of Apple Watch, created a smart bracelet market and brand from scratch when the mobile Internet emerged? In this IEEE report, the two founders of Fitbit were reviewed in detail, from idea brainstorming, prototype polishing, technology and hardware selection, to the entire process of product birth in the supply chain. And their experience, I believe, can still bring a lot of inspiration and experience to today's hardware entrepreneurs.

It was December 2006. James Parker, a 29-year-old entrepreneur, had just purchased a Wii gaming system. The system included the Wii Nunchuk, a $29 handheld controller with motion sensors that allowed gamers to interact by moving their bodies, such as swinging a bat to hit a baseball or boxing with a virtual partner.

Parker became obsessed with his Wii.

“I’m a tech geek,” he said. “Anyone who held that nunchuck was fascinated by how it worked. It was the first time I’d seen such a compelling consumer use for an accelerometer.”

But after a while, Parker discovered a flaw with the Wii: It got you moving, but it kept you stuck in your living room. What if, he thought, you could take what made the Wii so cool and put it in a gadget that got you out of the house?

“That was an aha moment,” Parker said.His idea was Fitbit, an activity tracker that has sold more than 136 million units since the first version hit the market in late 2009.。

But back to that “aha moment.” Parker quickly called his friend and colleague Eric Friedman. In 2002, both computer scientists, they had started a photo-sharing company called HeyPix, which they sold to CNET in 2005. They were still working at CNET in 2006, but it was a good time to think about doing something different.

Friedman liked Parker's idea.

Fitbit's two founders, James Park (left) and Eric Friedman, took a photo when they first released the product in 2009 | Image source: NYT

“My mother was an active walker,” Friedman said. “She had a walking group and always carried a pedometer with her. My father worked in augmentation engineering (assistive technology) for seniors and people with disabilities. We had played around with accelerometer technology before. So it made sense right away. We just had to refine it.”

The two left CNET and founded the startup in April 2007, with Parker as CEO and Friedman as CTO. Parker and Friedman weren't trying to make the first pedometer -- mechanical pedometers date back to the 1960s.

They weren't the first to create a smart activity tracker, either—medical device maker BodyMedia put accelerometers and other sensors in armbands that measured calorie burn as early as 1999.

The first generation Fitbit was clipped to the side of someone's trouser pocket｜Image source: NYT

And Parker and Friedman aren't the first to bring a smart consumer tracker to market. In 2006, Nike partnered with Apple to launch Nike+, a sports tracking system for runners that required special shoes and a receiver that plugged into an iPod.

Parker hadn’t realized any of this when he had the idea to move fitness out of his living room, but the two quickly did their research and figured out what they wanted to do and what they didn’t want to do.

“We didn’t want to create an expensive product that was for athletes,” he said, “or a product that was stupid and had nothing to do with software.We wanted a product that provided social connection, like photo sharing。」

The device must be comfortable to wear all day, easy to use, seamlessly upload data so it can be tracked and shared with friends, and require little to no charging. Not an easy combination of requirements.

“The simpler something is, the harder it is to design it well,” Parker said.

Tamagotchi and Bra

The first design decision is the most important. Where on the body do they want people to wear this wearable device? They won’t ask people to buy special shoes, like Nike+, or wear a thick band on their upper arm, like BodyMedia’s tracker.

They hired NewDealDesign to work out some of the details.

Gadi Amit, President and Chief Designer at NewDealDesign, said: “During the first two weeks, after many discussions with Eric and James,We decided to target the project towards women. This decision was a driving factor in the design of the form factor.。」

“We wanted to start with something that people were familiar with,” Parker said. “People clip their pedometers to their belts.” So a clip-on device made sense. But women don’t usually wear belts.

To achieve this functionality, the clip-on device had to contain a printed circuit board that was about 2.5 x 2.5 centimeters, Amit recalls. The team had a major breakthrough when they decided to separate the electronics from the battery, which are stacked together in most devices.

“By doing that and lengthening it a little bit, we found that women could put it anywhere,” Amit said.Many people put it in their bras, so we targeted the front of the bra and bought dozens of bras to test.。」

The decision to design for women also affected the overall look, with, as Amit puts it, "a more understated user interface." They hid the low-resolution monochrome OLED display behind a continuous plastic cover, and the display only lights up when you ask it to. This choice helps give the device an impressive battery life.

The earliest Fitbit used a small flower as a pattern that reminded people of Tamagotchi｜Image source: NEWDEALDESIGN

They also came up with the idea of ​​using a flower as a progress indicator — an idea Parker says was inspired by one of the most popular toys of the late 1990s: Tamagotchi. “So we designed a little animated flower that would shrink or grow depending on how active you were,” Parker explains.

After much discussion on control,The team decided to add just one button to the first-generation Fitbit.。

Part engineering, part black magic

Parker and Friedman knew enough about electronics to build a crude prototype, “stuffing the electronics into a box made of chopped balsam wood,” Parker says. But they also knew they needed to hire a real electrical engineer to develop the hardware.

Luckily, they knew who to call.

Friedman's father, Mark, had been working for years to develop a device for use in nursing homes to remotely monitor the location of bedridden patients. His partner was electronics engineer Randy Cacciola, now president of Morewood Design Labs.

Eric called his father to tell him about the gadget he and Parker had envisioned and asked if he and Cassioola could build a prototype.

“Mark and I wanted to do a quick prototype first so they could take the sensor data from it and use it to develop software. Then they would go to Asia and make it in miniature form there,” Casciola recalls. “But one revision led to another.” Casciola ended up working on Fitbits’ circuit design almost full-time until the company was acquired by Google, a deal announced in 2019 and completed in early 2021.

“We were just two guys in a little office in Pittsburgh,” Cacciola said. “We had realized our nursing home product wasn’t likely to be a product before Fitbit came along, so we started doing some consulting work. I had no idea Fitbit would become a household name. I just loved working on anything, whether I thought it was a good idea or a bad idea, or even whether anyone was paying me to do it.”

Cassiolla said,The earliest prototypes were fairly large, about 10 x 15 cm. They were large enough to be easily attached to test equipment., while being small enough to be strapped to a willing test subject.

Since then, Park and Eric Friedman, along with Casciola, two contract software engineers and a mechanical design firm, have worked to transform the bulky prototype into a small, stylish device that counts steps, stores the data until it can be uploaded and transfers it seamlessly, has a simple user interface, and doesn’t require daily charging.

「Finding the right balance between battery life, size, and capacity kept us busy for about a year., "Parker said.

After deciding to include a radio transmitter, they made a major move: They abandoned the Bluetooth standard for wireless communication in favor of the ANT protocol, a technology developed by Garmin that uses much less power. This meant that Fitbit couldn't upload directly to a computer. Instead, the team designed its own base station that plugs into a computer and grabs data whenever a Fitbit wearer passes within range.

Fitbit prototype first powered on in December 2008 | Image credit: James Parker

Casciola had no expertise in RF engineering, so he relied on the supplier of ANT radio chips: Nordic Semiconductor in Trondheim, Norway.

“They would do a design review of the board layout,” he explains. “Then we would send the hardware to Norway. They would do RF measurements on it and tell me how to adjust the values ​​of the capacitors and conductors in the RF chain, and I would update the schematic.”Getting these radio frequency devices to work is part engineering and part black magic。」

Another standard they didn't use was the ubiquitous USB charging connection.

“We couldn’t use USB,” Park said. “It took up too much space. Somebody said to us, ‘Whatever you do, don’t design a custom charging system because it’s going to be cumbersome and very expensive.’ But we went ahead and built one anyway. It’s cumbersome and very expensive, but I think it adds a layer of magic. You just put the device on [the charger]. It looks beautiful, and it works reliably.”

Most of the electronic equipment they use is off the shelf., including a Texas Instruments 16-bit MSP430 microprocessor, 92 KB of flash memory, and 4 KB of RAM to hold the operating system, the rest of the code, all graphics, and at least seven days of collected data.

Fitbits are designed to be sweat-resistant, so they can usually survive showers and quick soaks, Friedman said. “But hot tubs are the bane of our lives.People clip it to their bathing suits and forget to wear it when they jump into the hot tub。」

Do you think counting is easy?

So far, the company has been able to stay afloat with $400,000 in investments from Parker, Friedman, and several people who backed their previous companies. But more money is needed to expand production. So the next key step is a live public demonstration, which they plan to do at the TechCrunch conference in San Francisco in September 2008.

Live demonstrations of new technology are always risky, and this one veered right into disaster. The plan was for the audience to call out a number, and Park would take that many steps while wearing the prototype, which was enclosed in a balsa wood box. The steps would be wirelessly synced to a laptop projected onto a screen on stage. When Friedman refreshed his browser, the number of steps appeared on the screen. Was there anything wrong with that?

Totally wrong.

“You think counting steps is easy, but let’s say you take three steps,” Friedman explained. “One, two, three. When you bring your feet together, is that a step or is that the end? It’s much easier to count 1,000 steps than it is to take 10 steps. If I take 10 steps and miss one, that’s a clear error. If it’s 1,000 steps, that variance becomes noise.”

After a lot of practice, the two thought they could pull it off. Then they began the demonstration. “As I was walking, the laptop crashed,” Parker said. “I didn’t realize it at the time. I was just happily walking along. Eric had to restart everything while I was walking. But the numbers showed up; I don’t think anyone realized what was happening except Eric.”

On that day,About 2,000 pre-orders came in. Fitbit closed a $2 million round of venture capital the following month.。

Although Parker and Friedman had hoped to have Fitbits in people's hands (or clipped to their bras) by Christmas 2008, they missed that deadline by a year.

One of the challenges Fitbit has faced in going from prototype to shippable product is software development. They can’t expect users to walk as precisely as Park did in the demo.Instead, the device's algorithms need to determine what is a footstep and what are different types of motion.—For example, someone scratches their nose.

“Data collection was difficult,” Parker said. “Initially, a lot of us were wearing the prototype device doing a variety of different activities. Our research director, Shelten Yuen, would follow up, videotaping so we could go back and count the exact number of steps. We would wear multiple devices at once to compare the data we collected with each other.”

Fitbit and its packaging when it was first released in 2009 | Image source: NEWDEALDESIGN

Friedman remembers one such outing. “James was chained to his computer, pretending to take his dog for a walk in the Haight [San Francisco], narrating this little play he was putting on: ‘Okay, I’m going to stop. The dog was peeing in this tree. Now he’s going to go over there.’ The great thing about San Francisco is that nobody looks at two people chained together walking around, talking to themselves, in a weird way.”

“Pushing a stroller is a problem” because the wearer’s arms can’t swing, Parker said. “So one of our guys put an alien doll in a stroller and pushed it all over the city.”

Road noise is another big issue。

“Yuen is in charge of the algorithms, and he lives in Cambridge, Massachusetts,” Park said. “They have more potholes than we do. When he rides the bus, the bus hits a pothole, and [the device] bumps along the way, counting steps.” He added that they couldn’t solve the problem by just looking for regular rhythms to count steps, because not everyone has a regular rhythm.

“Older people tend to have irregular rhythms — to the device, they look a lot like a bus going over a pothole.”

The Curse of Hardware

Consumer electronics meant mass production, potentially in huge quantities. Parker recalled that they talked to many contract manufacturing companies. They realized that as a startup with an unclear future market, they would not attract the interest of top manufacturers.But they can't choose the lowest budget option because they need a reasonable level of quality.。

“We saw some manufacturers that were horrible,” Park said. “The equipment was dirty, there were burr marks on the molded plastic (a sign of a bad seal or other error), and the precision was very low.” They eventually found a small manufacturer that was “pretty good but still hungry for business.” The manufacturer was based in Singapore, while their surface mount supplier was in Batam, Indonesia, and they mounted the components directly on the printed circuit boards.

While working with the manufacturer, Park and Friedman made some adjustments to the circuit design and the shape of the case. They struggled with how to keep water and sweat from getting into the device, ultimately deciding to ultrasonically weld the case and add a spray coating to the circuitry after some devices were returned due to corrosion in the electronics.

This required adjusting the layout to ensure the coating could get between the chips. The coating had to be inspected and touched up by hand on each board. When they realized that the coating increased the height of the chips, they had to adjust the layout further.

In December 2009, just a week before their shipping date, Fitbits started to fall off the assembly line.。

Fitbit co-founder James Parker (above, standing in the middle) helps manufacturers debug the device shortly before the product launch in 2009 | Image credit: James Parker

“I tested one of the first fully integrated devices in a hotel room in Singapore,” Parker said. “It wouldn’t sync with my computer. Then I put the device next to the base station and it synced. OK, that’s great, but what’s the maximum distance it can sync? It turns out to be a few inches. In every other test we did, it was fine. It synced from 15 or 20 feet (5 or 6 meters) away.”

Park eventually discovered that the problem was when the two halves of the Fitbit case were ultrasonically welded together. In previous simultaneous tests, the case had not been sealed. The sealing process pushed the two halves closer together, so the display's cable touched or nearly touched the antenna printed on the circuit board, which affected the radio signal. Park tried squeezing the two halves together and reproduced the problem.

“I thought if we could just push the cable away from the antenna, it would be OK,” Parker said. “The only thing I could find in my hotel room was toilet paper. So I rolled up some toilet paper really tightly and stuffed it between the cable and the antenna. It seemed to work, although I wasn’t super confident.”

The next day, Parker came to the factory and discussed the problem and his solution with the manufacturing team.They refined his solution — replacing the toilet paper with a small piece of foam — and that’s how the first Fitbit shipped.。

The company sold about 5,000 of its first-generation, $99 devices in 2009 and more than 10 times that number in 2010. The launch wasn’t smooth. Casciola recalls that Fitbit’s logistics center sent him an alarming number of corroded devices that customers had returned. Casciola was tasked with taking them apart and diagnosing the problem.

「Over time, one of the contacts on the device began to develop green corrosion,” Casciola said. “But the other two contacts didn’t.” It turned out the problem was with Casciola’s design of a system reset trigger that allowed users to reset their devices without a reset button or a removable battery. “Inevitably,” Casciola said, “the firmware would crash. When you can’t remove the battery, you have to have another way to force a reset; you don’t want to have someone wait six days for the battery to die before they can reboot.”

The reset mechanism Casciola designed was “a button on the charging station that you poke with a paper clip. If you do that while the tracker is on the charger, it resets. Of course, we had to have a way for the tracker to see this signal. When I designed the circuit to allow for this, I ended up with the nominal voltage on one pin.” This low voltage led to corrosion.

“If you clip the tracker to sweaty clothing—remember, sweat has a lot of salt in it—a very small current will flow through it,” Casciola said. “It’s only a few microamps, not enough to cause a reset, but enough to cause green corrosion over time.”

A partially assembled first-generation Fitbit. Photo credit: James Parker

With the Fitbit that launched in 2012, the Fitbit One, Casciola added a new type of chip that hadn’t been around when he designed the original product. It triggers a reset if a single button is held down for a few seconds while the device is on the charger. This eliminates the need for an activity pin.

The charging port was another source of early problems。

In the original design, the trim of the Fitbit's plastic case was painted chrome. "We wanted a real metal trim," Friedman said, "but that would interfere with the radio signal."

Chromium wasn’t a good choice either. “It created problems for the charger port,” Park added. “We had to do a lot of work to prevent shorts there.”

After shipping tens of thousands of units, they dropped the chrome, and the new chrome-free look was praised by buyers.

Devices are evolving very quickly, especially when it comes to how data is transmitted.In 2012, when Bluetooth Low Energy (Bluetooth Low Energy) as a new low-power communication standard is widely used, the base station is replaced by a small Bluetooth communication adapter. Eventually, the adapter disappeared altogether.

“We had a big debate about whether to keep shipping the adapter,” Parker says. “It cost a lot, and if you had a new iPhone, you didn’t need it. But we didn’t want someone to buy the device and then return it because their phone wouldn’t connect.” The team closely tracked the adoption of Bluetooth Low Energy in phones; when they felt the number was high enough, they canceled the adapter.

Finally, the bracelet arrived

After several iterations of the original Fitbit design, sometimes referred to as a “clip” because of its shape, the fitness tracker was later ported to the wrist. This wasn’t just a redesign of how the device connected to the body, but a rethinking of the algorithm.

The impetus for this initiative comes from some users’ desire to better track their sleep. Fitbit’s algorithms enable it to recognize sleep patterns, a design choice that Park says is “critical because it transforms the device from a pure activity tracker to a 24/7 health tracker.” But the pajamas have no obvious attachment points.

therefore,Fitbit came with a thin fabric wristband that was only for nighttime use. Users began asking customer support if they could wear the band all day. The answer was no; Fitbit's step counting algorithm at the time didn't support this.。

At the same time, a cultural phenomenon is emerging.

In the mid-2000s, yellow Livestrong bracelets, made of silicone and supporting cancer research, suddenly became ubiquitous, and other causes and movements jumped on the bandwagon with brightly colored wristbands.

By early 2013, Fitbit and competitors Nike and Jawbone had introduced wrist-worn fitness trackers in much the same style as these fashionable bands.

Fitbit's version is called Flex and is also designed by NewDealDesign.

The early versions of the Fitbit Flex wristband were very simple to interact with (Flex 2 in the picture) | Image source: hithacking

The Flex's interface is even simpler than the original Fitbit's single button and OLED screen:It has no buttons or screen, just five LEDs in a row and a vibration motor.To change modes, just tap the watch face.

“We didn’t want to replace people’s watches,” Parker said. The technology wasn’t there yet to “build a compelling device — a device with a large screen and computing power that enabled really amazing interactions on the wrist that matched the screen. It wasn’t until 2014 or 2015 that the technology trends came together to make that possible.”

“It was amazing the amount of information the team was able to convey using just the LEDs,” Friedman recalls. “It was obvious how close you were to reaching your [step] goal. But [there were] light cycles to show it was searching for something, it vibrates when you hit your step goal, and so on.”

Still, Park admitted that the tapping part of the interface was "probably something we didn't do very well." After launch, we spent a lot of time fine-tuning the algorithm to better distinguish which actions weren't taps — like clapping. What's more, some users didn't intuitively know the correct way to tap.

“If 98 percent of users can use this feature, but that number grows to millions, then that 2 percent is really important,” Park said. They brought the button back on the next generation of Fitbit devices.

Google acquires Fitbit for $2.1 billion in 2021 | Image source: Google

The rest is history

In 2010, Fitbit's first year on the market, it sold about 50,000 units. According to Statista, Fitbit sales peaked in 2015, when it sold nearly 23 million units.

Since then, Fitbit's sales have declined as multi-function smartwatches have fallen in price, gained in popularity, and Fitbit knockoffs have entered the market. As of 2021, Fitbit still has more than 31 million active users, according to Market.us.Media.

Fitbit may be riding the wave of a return to simplicity right now, as people find themselves wanting to ditch distractions and return to simpler devices. I’ve seen this happen in my own family: My daughter wore a smartwatch, and earlier this year she traded that wearable for a Fitbit Charge 6.

Fitbit went public in 2015 with a valuation of $4.1 billion.

In 2021, Google acquired the company for $2.1 billion and incorporated it into its hardware divisionParker and Friedman left Google in April. Early retirement? Unlikely. The two, now 47, have started a new company and are currently in stealth mode.

The idea of ​​encouraging people to be physically active by electronically recording their steps has had staying power.

“My dad turned 80 on July 5, and he’s obsessed with his step count,” Friedman said. “From 11 p.m. to midnight, he’s in the parking lot, climbing the stairs. He’s in better shape than I can remember him being in.”

What better reward could there be?

Google Fitbit Smart Bracelet Wearable Device