Blog

How To Optimize IoT Device Battery Life

IoT devices permeate our world at a rapid pace with a big share of them running on batteries. For both primary and rechargeable batteries there is a need to optimize the available runtime. Sensor nodes are often deployed with a battery that needs to cover the device’s lifetime of 5 or even 10+ years, some IoT devices are installed in places difficult to access.

Replacing or recharging batteries more frequently than absolutely necessary always adds maintenance expenses to the Total Cost of Ownership calculation. There are many aspects to optimizing the power consumption of an IoT device and many of them deserve an in-depth discussion which would go beyond the scope of this article.

It is a known fact that design decisions in the concept phase of the project have a major impact on the cost of the final product and can cause or avoid technical difficulties in the subsequent development phase. A simplified generic IoT device would look something like this:

DECISIONS

A non-exhaustive list of decisions to be taken:

  • Which Microcontroller gives us the perfect balance of cost, performance, power-consumption, development tools, available libraries, long-term availability, …
  • Which wireless Interface suits our requirements in terms of range, power-consumption, required throughput, infrastructure needed / available, …
    Choosing the perfect battery type, chemistry and capacity is a non-trivial task that will profoundly impact the overall cost and performance of our product.
  • Do we have a choice of different sensors / actuators to choose from that differ in price, power-consumption, size and other relevant aspects?
    The power management can be as simple as a timer to wake up the system in predefined intervals or as involved as a customized power management chip.

After we have carefully considered all relevant aspects of the hardware we will ultimately end up with a functional prototype or the first version of the PCB on our desk.

The part not visible on the block diagram but with a great impact on performance and power-consumption is the software. This usually offers a significant potential for optimizations.

In the remainder of this article we will focus on this important optimization process which will define the final specifications of our product.

SOFTWARE OPTIMIZATION

At this point we should know what the best current profile should be to achieve the best battery life. Sometimes we can use the current spikes caused by the wireless transmission to mitigate passivation effects of the battery. Things start to come together now. We work through the design verification step by step, always with an eye on the coulombs we burn.

A few bullet-points (non-exhaustive) on our to-do list:

  • Implement production test scenario
  • Test Connection scenarios including failure or difficulty to connect to network
  • Security
  • Interoperability
  • Reliability and environmental tests
  • Battery-life confirmation measurements
  • EMI/EMC compliance
  • Certification

In order to verify our device’s power-consumption through all of the above steps we can measure the power-consumption using different approaches.

USING A DEBUG PROBE WITH BUILT-IN POWER MEASUREMENT

In a lot of battery powered IoT-designs the microcontroller turns out to be an ARM based chip.

To go with this example the ULINKplus can be utilized to perform software synchronized current profile measurements on Cortex-M based microcontrollers.

Source: ARM/Keil ULINKplus -User’s Guide, https://www.keil.com/support/man/docs/ulinkplus/ulinkplus_measuring_power.htm

The debug adapter features an isolated current measurement input that can be used with different shunt-resistors to scale the current measurement range.

With this solution we can use the full software debug features and see the impact of software changes in real-time in the System analyzer window.

Source: ARM/Keil ULINKplus -User’s Guide, https://www.keil.com/support/man/docs/ulinkplus/ulinkplus_measuring_power.htm

Power measurement features of ULINKplus:

Isolation for independent supply measurement and optimal signal/noise ratio

  • Wide current range with supplied shunt boards (2.5 – 250 mA)
  • High resolution with linear behaviour (detect misconfigured I/Os)
  • Time-synchronized with Event Recorder and SWO trace streams
  • Various filter frequencies for better power visualization

Since IoT devices often feature standby currents in the µA range and 10’s to 100’s of mA during wireless transmit we may need a higher dynamic range without a shunt resistor (since we cannot afford the voltage drop over the 100Ω shunt needed to accurately measure the standby current) to get good measurement results.

USING A DC POWER ANALYZER

A DC Power Analyzer sources the supply and measures the current simultaneously, removing the shunt resistor from the equation.

A good solution for this approach would be a Keysight system consisting of an N6705C DC Power Analyzer, N6781A Two-quadrant SMU for battery drain analysis and BV9200B control and analysis software.

With this setup it is possible to measure µA – 100’s of mA with seamless range changes for gap-free recording and high accuracy at 200kSa/s. The device provides logging up to days. The output impedance is programmable to emulate the behavior of the battery like voltage drops that result from current pulses.

This SMU has a “current measurement only” mode that sets the power supply to behave like a zero-ohm shunt. When the output is connected in series with the battery and the battery-powered device, as in below diagram, the SMU emulates a zero-burden ammeter. The voltage is regulated to zero volts where the remote sense lines are connected.

Source: Keysight Application Note “Evaluating Battery Run-Down with the N6781A or N6785A 2-Quadrant Source/Measure Unit and the BV9200B Control and Analysis Software”

A detailed Application Note on how to use an SMU for Battery Run-Down Evaluation can be found here: https://www.keysight.com/ch/de/assets/7018-02880/application-notes/5990-7370.pdf

In case the application requires an even wider dynamic range from pA to A or a significantly higher bandwidth up to 200MHz you may also consider to use a Device Current Waveform Analyzer like the CX332xA Instruments from Keysight.

COMPARISON

When we compare the Debug Probe with built-in Power Measurement and the DC Power Analyzer approach we conclude that both have their pro’s and con’s and often the optimum solution is to use a combination of both in the design-, optimization- and verification-process.

GET IN TOUCH WITH COMPUTER CONTROLS AG

We have the expertise and a carefully selected product portfolio available to assist you throughout your concept, design, optimization, test, certification and deployment cycle with tailored solutions.

We are only one click away.

Share

Related Posts

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *