While the performance of wearable devices continues to improve engineers continue to struggle when it comes to delivering improvements in power management.
The power/performance profile for wearables isn’t helped by the fact that they can only use small batteries. Bigger batteries mean bigger devices, and that's not what the consumer wants so there is intense interest in reducing battery size, increasing battery efficiency and researching innovative ways to power these devices.
The power/performance issues is accentuated by the growing use of sensors and the fact that they need to be ‘always-on’ in order to sense, track, classify and store data.
According to Managing director, Dr Jacob Skinner of Thrive Wearables, the rapid improvements being made in silicon could provide a solution. This, he explains, sees processor designs using lower and auto-adjusting silicon voltages, deploying more efficient process architectures and optimising instruction sets.
“The backdrop to processor developments is a continuing drive to integrate processing 'nodes' right at the front end of wearables sensors, so that the data transfer (and power drain) 'up the chain' is minimised,” he explains.
A good example of this trend is Bosch’s BMA400 low power accelerometer series. This accelerometer draws less than 1µA in full operation, but can independently process sensor data. For example, it can convert the long data stream of three-axis motion into step counting.
It does this by allowing the main (host) microcontroller to remain off for most of the time that’s required for tracking a user’s activity and is then woken up by the accelerometer itself every 100 steps, for example. The sensor becomes the component that manages the overall duty cycle of the microcontroller, helping to reduce system power while increasing the efficiency.
Richard Edgar, director of comms tech ensigma at Imagination Technologies, believes that the isolation of device functionality is one possible way forward. He describes this approach as creating ‘power islands’.
“Although you may acquire data, you don’t necessary have to do anything with it and for the data you do need to transfer, you don’t need high data throughput,” Edgar contends. “By designing a system that reacts only under certain scenarios or perimeters, we can preserve power.
| "Usually the market wants you to deliver improved specifications and even more sensitivity, but by focusing on performance it will be a hindrance if you're trying to save power." Richard Edgar, director of comms tech ensigma at Imagination Technologies |
“What we need to weigh up is how much we can reduce the time the cell is awake, against the possibility of missing important information and processing data accurately or in real-time.”
Edgar suggests that power amplifiers are one of the main problems when it comes to power consumption, with around 80% of today’s communication solutions having most of their power spent by one.
He says this is ‘avoidable’ and can be easily rectified by replacing existing power amplifiers with a low range 0dbm one. “If you do that, and the device is only communicating a few 100bytes on a daily basis, you should be able to get the battery life to last up to a year.”
Engineers are looking to use smaller nodes as a possible route to better manage power consumption and are also considering desensitising their devices.
“Usually the market wants you to deliver improved specifications and ever more sensitivity,” explains Edgar, “but by focusing on performance it will be a hindrance if you’re trying to save power.
“You have to establish what the best compromise is,” Edgar continues. “For example, do you need a wearable that transmits 100m or 1m? If it’s the latter, you don’t need as high performance and therefore, you can create a device with a lower spec.”
The problem lies in whether consumers will be happy to accept lower performing devices in exchange for better battery life.
The cost of manufacturing smaller using smaller nodes is also problematic, for example, according to Edgar, moving from 40nm to 20nm can result in a 30-40% cost increase.
New research
Research in to alternative power technologies is throwing up some interesting developments.
One alternative to traditional battery technology to emerge is the development of a stretchable and twistable power device, a biobattery, which could be integrated into wearable electronics.
According to Assistant Professor, Seokheun Choi of Binghamton University, this technology can be directly printed onto a single textile substrate and establishes a standardised platform for flexible biobatteries.
Assistant Prof. Choi believes that textile-based wearables hold promise, but says the challenge lies in creating a ‘truly self-reliant and stand-alone wearable sensing system that does not rely on an external power source’.
“The microbial fuel cells (MFCs) used in this work are arguably the most underdeveloped for wearable electronic applications because microbial cytotoxicity may pose health concerns,” he says. “Work into wearable MFCs has been quite limited but, if we consider that humans possess more than 3.8×1013 bacterial cells compared with 3.0×1013 human cells in their bodies, the direct use of bacterial cells as a power resource is conceivable for wearable electronics.
“Most microorganisms use respiration to convert biochemical energy stored in organic matter into biological energy, adenosine triphosphate, in which a cascade of reactions through a system of electron-carrier biomolecules sees electrons transferred to the terminal electron acceptor,” Assistant Prof. Choi continues. “Most forms of respiration use a soluble compound as an electron acceptor, such as oxygen. However, some microorganisms respire solid electron acceptors to obtain biological energy.
“These microorganisms can transfer electrons produced via metabolism across the cell membrane to an external electrode. MFCs typically comprise of anodic and cathodic chambers separated by a proton exchange membrane so that only H+ or other cations can pass from the anode to the cathode. A conductive load connects the two electrodes to complete the external circuit.”
Unlike traditional batteries and other enzymatic fuel cells, MFCs have whole microbial cells that can be used as a biocatalyst to provide stable enzymatic reactions and longer lifetimes.
Assistant Prof. Choi believes that it’s possible for organic fuels, such as wastewater, sweat and urine, to be used as fuel to support bacterial viability, providing the long-term operation of the MFCs.
Another possible energy source is energy harvesting which Edgar believes could pave the way for longer lasting wearables, with movement as the most plausible way to generate power for wearables. He does raise some concerns over the cost of placing an energy harvesting technology into silicon and reasons that complete reliance on energy harvesting is not feasible, rather it should act as a complementary way to power devices.
Overall though, Edgar believes that at least in the short terms the solution to extending battery life is to ‘design chips that consume less power and to make compromises when it comes to the design taking into account its market application’.
“There’s only so much we can do on the engineering side. It does rely upon other changes and techniques, like developing new battery technologies,” he says. “That is being held back as the most promising research seems to use rare materials and they are, by their very nature, hard to come by.”
Alternative power sources and techniques are emerging but whether or not they will provide long term solutions is another matter. It appears the answer for now is to assess whether performance can be lowered without affecting the overall purpose of the device too drastically.
That trade-off will be dictated by consumer reaction but they will need to be better informed. Having their technology cake and eating it is, as yet, not possible when it comes to balancing power and performance issues.
