Applying power management solutions in IoT designs -

Applying power management solutions in IoT designs


This article explores IoT battery technology. It describes some of the problems that designers face with power sourcing and provides some solutions.

This article explores internet-of-things battery technology. It describes some of the problems that designers face with power sourcing and provides solutions from Analog Devices. These solutions are highly efficient and can help curb other problems in your IoT devices, including size, weight, and temperature.

With the growing use of IoT devices in industrial equipment, home automation, and medical applications, there is increasing pressure to optimize the power management portion of these devices — either through decreased size, better efficiency, improved current consumption, or faster charging times. All of this must be achieved in a small-form factor that does not negatively impact thermals nor interfere with the wireless communication implemented by these devices.

What is IoT?

This IoT application area comes under many different guises. It generally refers to a smart, network-connected electronic device that is likely battery-powered and sends precomputed data to a cloud-based infrastructure. It utilizes a mixture of embedded systems such as processors, communication ICs, and sensors to collect, respond to, and send data back to a central point or other node in a network. This can be anything from a simple temperature sensor reporting room temperature back to a central monitoring area all the way up to a machine health monitor that tracks the long-term performance data of very expensive factory equipment.

Ultimately, these devices are developed to solve a particular challenge, whether that be to automate tasks that would typically require human intervention, like home or building automation, or perhaps to improve the usability and longevity of equipment in the case of industrial IoT applications, or even to improve safety if you consider condition-based monitoring applications implemented in structure-based applications such as bridges.

Example applications

The application areas for IoT devices are almost endless. Smart-transmitter–based applications gather data about the environment in which they sit to make decisions about controlling heat, setting off alarms, or automating tasks. In addition, portable instruments like gas meters and air-quality measurement systems provide an accurate measurement through the cloud to a control center. GPS tracking systems are another application. They allow the tracking of shipping containers as well as livestock through smart ear tags. These comprise just a small area of cloud-connected devices. Other areas include wearable health-care and infrastructure sensing applications.

A significant growth area is industrial IoT applications, which are part of the Fourth Industrial Revolution. There is a broad range of IoT applications that are ultimately trying to automate as much of the factory as possible, whether that be through the use of automated guided vehicles, smart sensors such as RF tags or pressure meters, or other environmental sensors positioned around the factory.

From an ADI perspective, the high-level IoT focus has been in five main areas:

  1. Smart health: supporting vital-signs–monitoring applications both at a clinical level as well as consumer applications
  2. Smart factories: focusing on building Industry 4.0 by making factories more responsive, flexible, and leaner
  3. Smart buildings/smart cities: using intelligent sensing for building security, parking space occupancy detection, and thermal and electrical control
  4. Smart agriculture: using the technology available to enable automated farming and resource usage efficiencies
  5. Smart infrastructure: building on our condition-based monitoring technology to monitor movement and structural health

More information on these focus areas and the technologies available to support them can be found at

IoT design challenges

What are the key challenges facing a designer in the ever-growing IoT application space? Most of these devices, or nodes, are being installed after the fact or in hard-to-reach areas, so running power to them is not a possibility. This, of course, means that they are totally reliant on batteries and/or energy harvesting as a power source.

Moving power around large facilities can be quite expensive. For example, consider powering a remote IoT node in a factory. The idea of running a new power cable to power this device is costly as well as time-consuming, which essentially leaves battery power or energy harvesting as the remaining options to power these remote nodes.

The reliance on battery power introduces a need to follow a stringent power budget to ensure that the lifetime of the battery is maximized, which of course has an impact on the total cost of ownership of the device. Another downside to battery usage is the need to replace the battery after its life has expired. This includes not only the cost of the battery itself but also potential high costs of human labor to replace and possibly dispose of the old battery.

An additional consideration on the battery cost and size: It is very easy to just overdesign the battery to ensure that there is sufficient capacity to achieve the lifetime requirement, which is very often greater than 10 years. However, overdesign results in additional cost and size, so it is extremely important to not only optimize the power budget but to minimize the energy usage where possible to install the smallest possible battery that will still meet your design requirements.

Power in IoT

For the purposes of this power discussion, the power sources for IoT applications can be seen as three scenarios:

  1. Devices that rely on non-rechargeable battery power (primary battery)
  2. Devices that require rechargeable batteries
  3. Devices that utilize energy harvesting to provide system power

These sources can be used individually or combined if the application requires it.

Primary battery applications

Non-rechargeable battery applications are geared toward applications in which only occasional power is used — that is, the device is powered up occasionally before going back into a deep-sleep mode in which it draws minimal power. The main advantage of using this as a power source is that it provides a high energy density and a simpler design, as you don’t need to accommodate battery charging/management circuitry, as well as a lower cost, being that batteries are cheaper and fewer electronics are required. They fit well into low-cost, low-power–drain applications. However, as batteries have a finite lifetime, they are not well suited to applications in which power consumption is a little higher.

Consider a large IoT installation with many nodes. As you have a technician on-site replacing the battery for one device, very often all the batteries will end up being replaced at one time to save labor costs. Of course, this is wasteful and just adds to the global waste problem. On top of that, non-rechargeable batteries provide only about 2% of the power used to manufacture them in the first place. The ~98% of wasted energy makes them a very uneconomical power source.

Obviously, these do have a place in IoT-based applications. Their relatively low initial cost makes them ideal for lower-power applications. There are many different types and sizes available, and as they don’t need much by way of additional electronics for charging or management, they are a simple solution.

From a design perspective, the key challenge is making the most use out of the energy available from these power sources. To that end, it is important to create a power budget plan to ensure that the lifetime of the battery is maximized, with 10 years being a common lifetime target.

For primary battery applications, two parts from ADI’s nanopower family of products are worth considering: the LTC3337 nanopower coulomb counter and the LTC3336 nanopower buck regulator, shown in Figure 1.

The LTC3336 is a low-power DC-to-DC converter running from up to a 15-V input with programmable peak output current level. The input can go as low as 2.5 V, making it ideal for battery-powered applications.

The quiescent current is exceptionally low at 65 nA while regulating with no load. As DC-to-DC converters go, this is easy to set up and use in a new design. The output voltage is programmed based on how the OUT0 to OUT3 pins are strapped.

The companion device to the LTC3336 is the LTC3337, a nanopower primary battery state of health monitor and coulomb counter. This is another easy device to use in a new design — simply strap the IPK pins according to the peak current required, which is in the 5-mA to 100-mA region. Run a few calculations based on your selected battery, then populate the recommended output cap based on the selected peak current, which is noted in the datasheet.

Ultimately, this is a fantastic pairing of devices for IoT applications with a limited power budget. These parts can both accurately monitor the energy usage from the primary battery and efficiently convert the output to a usable system voltage.

Figure 1: LTC3337 and LTC3336 application circuit

Rechargeable battery applications

Let’s move on to rechargeable applications. These are a nice choice for higher-power or higher-drain IoT applications in which primary battery replacement frequency is not an option. A rechargeable battery application is a higher-cost implementation because of the initial cost of the batteries and the charging circuitry, but in higher-drain applications in which the batteries are drained and charged frequently, the cost is justified and soon paid back.

Depending on the chemistry used, a rechargeable battery application can have a lower initial energy than a primary cell, but in the longer term, it is the more efficient option and is overall less wasteful. Depending on the power needs, another option is capacitor or supercapacitor storage, but these are more for shorter-term backup storage.

Battery charging involves several different modes and specialist profiles depending on the chemistry used. For example, a lithium-ion battery charge profile is shown in Figure 2. Across the bottom is the battery voltage, and charge current is on the vertical axis.

Figure 2: Charge current vs. battery voltage

Figure 3: Charge voltage/current vs. time

When the battery is severely discharged, as on the left of Figure 2, the charger needs to be clever enough to put it in precharge mode to slowly increase the battery voltage to a safe level before entering constant-current mode.

In constant-current mode, the charger pushes the programmed current into the battery until the battery voltage rises to the programmed float voltage.

Both the programmed current and voltage are defined by the battery type used the charge current is limited by the C-rate and the required charge time, and the float voltage is based on what is safe for the battery. System designers can reduce the float voltage a little to help with lifetime of the battery if required by the system.

When the float voltage is reached, the charge current drops to zero, and this voltage is maintained for a time based on the termination algorithm.

Figure 3 provides a different graph for a three-cell application showing the behavior over time. The battery voltage is shown in red and charge current is in blue. It starts off in constant-current mode, topping out at 2 A until the battery voltage reaches the 12.6-V constant-voltage threshold. The charger maintains this voltage for the length of time defined by the termination timer — in this case, a four-hour window. This time is programmable on many charger parts.

For more information on battery charging, as well as some interesting products, read the Analog Dialoguearticle “Simple Battery Charger ICs for Any Chemistry.”

Figure 4 shows an example of a versatile buck battery charger, the LTC4162, which can provide a charge current up to 3.2 A and is suitable for a range of applications, including portable instruments and applications requiring larger batteries or multicell batteries. It can also be used to charge from solar sources.

Figure 4: The LTC4162, a 3.2-A buck battery charger

Energy-harvesting applications

When working with IoT applications and their power sources, another option to consider is energy harvesting. There are several considerations for the system designer, but the appeal of free energy cannot be understated, especially for applications in which the power requirements aren’t too critical and in which the installation needs to be hands-off — that is, no service technician can get to it.

There are many different energy sources to choose from, and they don’t need to be an outdoor application to take advantage of them. Solar as well as piezoelectric or vibrational energy, thermoelectric energy, and even RF energy can be harvested.

Figure 5 provides an approximate energy level when using different harvesting methods.

Figure 5: Energy sources and approximate levels available for various applications

As for disadvantages, the initial cost is higher compared with the other power sources discussed before, as you need a harvesting element such as a solar panel, piezoelectric receiver, or a Peltier element, as well as the energy conversion IC and associated enabling components.

Another disadvantage is the overall solution size, particularly when compared with a power source like a coin-cell battery. It’s difficult to achieve a small solution size with an energy harvester and conversion IC.

Efficiency-wise, this can be a tricky one to manage low energy levels. This is because many of the power sources are AC, so they need rectification. Diodes are used to do this. The designer must deal with the energy loss resulting from their inherent properties. The impact of this is lessened as you increase the input voltage, but that’s not always a possibility.

The devices that pop up in most energy-harvesting discussions are from the ADP509x family of products, and the LTC3108, which can accommodate a wide range of energy-harvesting sources with multiple power paths and programmable charge management options that offer the highest design flexibility. A multitude of energy sources can be used to power the ADP509x but also to extract energy from that power source to charge a battery or power a system load. Anything from solar (both indoor and outdoor) to thermoelectric generators to extract thermal energy from body heat in wearable applications or engine heat can be used to power the IoT node. Another option is to harvest energy from a piezoelectric source, which adds another layer of flexibility — this is a nice option to extract power from an operational motor.

Figure 6: Block diagram of the ADP5090 in a harvesting application

Another device that is capable of being powered from a piezoelectric source is the ADP5304, which operates with a very low quiescent current (260 nA typical with no load), making it ideal for low-power energy-harvesting applications. The datasheet shares a typical energy-harvesting application circuit (see Figure 7), powered from a piezoelectric source and being used to provide power to an ADC or an RF IC.

Energy management

Another area that should be part of any discussion relating to applications with a limited power budget is energy management. This starts with developing a power budget calculation for the application prior to looking at different power management solutions. This essential step helps system designers understand the key components used in the system and how much energy they require. This impacts their decision to select a primary battery, rechargeable battery, energy harvesting, or a combination of these as the power supply methodology.

The frequency of the IoT device gathering a signal and sending it back to the central system or cloud is another important detail when looking at energy management, which has a large impact on overall power consumption. A common technique is to duty-cycle the power usage or stretch the time between waking the device up to gather and/or send data.

Making use of standby modes on each of the electronic devices (if available) is also a useful tool when trying to manage the system energy usage.

Figure 7: ADP5304 piezoelectric source application circuit


As with all electronic applications, it is important to consider the power management portion of the circuit as early as possible. This is even more important in power-constrained applications such as IoT. Developing a power budget early in the process can help system designers identify the most efficient path and suitable devices that meet the challenges posed by these applications while still achieving high energy efficiency in a small solution size.


1 F. Dostal. (Sept. 2015). “New Advances in Energy Harvesting Power Conversion.” Analog Dialogue, Vol. 49, No. 3.

2 S. Knoth. (Jan. 2019). “Simple Battery Charger ICs for Any Chemistry.” Analog Dialogue, Vol. 53, No. 1.

3 G. Murphy. (Jan. 2018). “Internet of Things (IoT): What’s Next.” Analog Devices, Inc.

4 Z. Pantely. (Sept. 2018). “One-Size-Fits-All Battery Charger.” Analog Devices, Inc.

>> This article was originally published on our sister site, Power Electronics News.

Diarmuid Carey is an applications engineer with the European Centralized Applications Center based in Limerick, Ireland. He has worked as an applications engineer since 2008 and joined Analog Devices in 2017, providing design support for the Power by Linear portfolio for European broad market customers. He holds a Bachelor of Engineering in computer engineering from University of Limerick.

Related Contents:

For more Embedded, subscribe to Embedded’s weekly email newsletter.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.