Second of two parts: Energy harvesting and improved battery technology are no substitute for best practices in electrical and software engineering.
The earliest recorded mention of a wristwatch dates back to the late 1500s, but it really began gaining adherents in Great Britain’s Boer War campaign as a way of synchronizing military actions beyond the line of sight. Strapping a pocket watch to a horse or a camel simply didn’t work, and pulling it out of a jacket pocket was not only inconvenient, it was dangerous.
Advertised as a “campaign watch,” for the next couple decades, the wristwatch got its real boost after World War I, when soldiers walked off the battlefield with their watches still strapped to their wrists. And so began a trend to wear time pieces rather than carry them, which has evolved into a very lucrative market. As of last count, an estimated 1.2 billion wristwatches are being sold each year. There were eight watches sold in 2014 with list prices of more than $1 million.
Whether the kind of wearable electronic devices strapped to the wrist are considered watches or something else is debatable, but they certainly are an outgrowth of the wearable timepiece industry coupled with some “gee whiz” technology thrown in (see part one of this two-part series). How much technology, and which technology will ultimately win out, is a big question, which explains why there were an estimated 1,800 companies pitching wearable electronics at this year’s CES show. Not all of them were for the wrist, of course. There are smart glasses and belts, and a slew of health- and fitness-related devices in that number. Still, it does show just how nascent and fluid the market is at this point. Sometime within the next few years, most analysts expect the market to sharply consolidate and rally around a handful of key ideas.
It’s clear that none of these devices will warrant a $1 million price tag, and many of them will never make it out of the showroom. In fact, at this point there are lots of questions about what a successful wearable device will look like, what the market will bear in terms of cost, who might subsidize the costs, and whether technical issues can be solved such as battery life and security. But the market is heating up with expectations that at least something—or some things—will catch on.
“We need to go from technologically possible to technologically meaningful,” said Shawn DuBravac, CES’ chief economist, in a speech at SEMI’s Industry Strategy Symposium earlier this month. He noted that the Internet of Things will change perceptions of everything we touch, from wearables to cars. “Right now we go to our car. In a driverless world, our cars will come to us.”
From a semiconductor perspective, this is almost a blank slate. That’s both good and bad. The criteria for a successful device in this space will be extremely high energy efficiency, with enough performance to satisfy users and enough security to safeguard critical data. Beyond that, however, features are likely to be added, subtracted and rethought for years to come, and probably at a pace that will make design teams look back fondly on the current state of engineering change orders.
“The question is how much you really need versus what’s nice to have versus who cares,” said Jamil Kawa, group director of the Solutions Group at Synopsys. “If it’s life saving—if it can identify certain enzymes associated with blockages in your arteries, that’s critical. But if all it does is tell you, ‘You’re still alive and congratulations,’ that’s not important.”
Yet within this uncertainty, there are some things that are very clear. Best practices in electrical engineering still apply.
“If you don’t get the power right, you’re hosed,” said Kawa. “All wireless communications, by definition, are power hungry when they’re on. There are two aspects we need to worry about. One is peak current versus average current versus sleep. The second is how to best use capacitors on a chip. Capacitors are good because you can recycle them hundreds of thousands of times, but retention is low. Still, you cannot afford spikes. You need to optimize the RMS of the current.”
That’s not so simple when you’re swapping IP blocks in and out to try to get products right, however. And it’s especially difficult when power is the gating factor for a successful wearable design.
“There are three main problems you need to solve with wearables,” said Aveek Sarkar, vice president of product engineering and support at ANSYS. “One is the digital side.That’s a compact problem and it can be solved. The second is analog. For that you need more on-chip regulation of power. And the third involves communication with the antenna design. If you don’t place the antenna properly you absorb energy. You may do everything right and get the signal propagated properly and send it to the antenna, and then someone on the marketing side decides to put some chrome on the outside and it acts like a shield for the antenna. So now you have to figure out how to solve that and you’re doing double work. Wearables require a joint discussion between mechanical and electrical engineering teams.”
Power on the go
One of the big unknowns in all of this is just how much energy generation will contribute to designs in the way of wireless charging and energy harvesting.
“There’s been a lot of attention focused on inductive charging, which is the latest trend for easy charging, but inductive charging is wasteful,” said Bernard Murphy, chief technology officer at Atrenta. “About 30% to 40% of that is dissipated as heat. That creates a challenge for wearable devices because you potentially can get burned. Starbucks is installing charging tables http://news.starbucks.com/news/national-rollout-of-wireless-charging-by-duracell-powermat-begins-in-starbu but that won’t work for wearables. There also is the notion of ‘power snacking,’ where you don’t charge it up fully every time but you do it more often.”
Murphy noted that energy harvesting is an interesting idea, but it won’t eliminate the need for extreme energy efficiency because there is no free lunch. Energy from activity is a fixed number, and if you insert springs into the bottoms of shoes to harvest some of that energy, each step will be require more effort. How much more depends on how much you remove. The same goes for heat harvested from the human body. But in small enough amounts, that shouldn’t be noticeable. That’s the idea behind a Chinese device called Skinny Player, which uses heat from the human body for power and straps on like a Band-Aid.
A second approach is to improve batteries, and there has been much work on this front with relatively little to show for it. The lithium ion battery remains the workhorse, and efforts to improve that technology win headlines but little else. There even has been work on nuclear isotopes, notably at the University of Missouri, which has developed the prototype of a battery based on Strontium 90. The isotope has a half life of 29.1 years and emits beta radiation, according to the U.S. Environmental Protection Agency.
But while the power budget may be architected at the system level, a third approach is to have many of the blocks within a design energy self-sufficient. This builds on the current approach by IP vendors to focus at the block level, where the most improvements can be made in efficiency and compatibility.
“The key is to provide different functional blocks, like LEGOs, and provide chipmakers the capability to tune what they need and add custom instructions,” said Larry Przywara, product line group director for audio/voice at Cadence. “No one size will fit all architectures.”
In conjunction with that approach, Synopsys’ Kawa said modules can be created with sensors and energy scavenging circuits that are self-sustaining. “Most sensors should be autonomous,” he said.
Business models to pay for all of this technology
One of the more interesting questions behind this market isn’t about technology, though. It’s about business models. A full-fledged strap-on electronic device with sophisticated SoC engineering, possibly at the latest technology nodes, with built-in communications technology and custom software, isn’t going to be cheap. So is there a way to subsidize that cost for consumers? The answer is yes, but it likely will come in multiple ways.
First, there is the software marketplace and the ability to download apps through whatever connection ultimately is used, whether it’s by smartphone as the controller or directly into the device.
“These devices can be revenue generators,” said Andrew Caples, senior product manager for the Nucleus product line at Mentor Graphics. “Devices can be connected to an app store to upload or download. It’s also a great way to provision and update a wearable device. But that also requires you to do things like partition memory so you can start, stop and reload, and to update firmware and software. That’s where the value proposition is from a software perspective.”
A different possibility is in the health-care realm, where insurance companies might offer devices that can track habits of their customers. Numerous sources talked about this concept, but none wanted to be quoted for this story. As with connected cars that reward drivers in real time for staying within the speed limit, imagine a device that can reward you for your restaurant choices or the number of steps you take within a given day—and that penalize you for being a couch potato.
As one executive described it, “This is the crossover point for someone watching everything you do.”