Up to 10 years of battery life for LoRa sensors? This is what it looks like in practice!

Sensors are necessary to collect data, for example from machines or buildings. In many cases, this means laying a lot of cables, both for power and for transmitting the data. However, depending on the location of the sensor, this can be time-consuming or simply impossible. If you want to avoid breakthroughs in walls, cable clutter and distribution stations, you will use wireless sensors if possible. LoRa-based sensors in particular have great advantages due to their long battery life, long range and penetration in buildings. They are simply mounted at the required location and the desired values are transmitted – at least on the basis of the LineMetrics platform.

High-quality and tested sensors are extremely low-maintenance. Only the battery needs to be replaced at the end of its service life or, in the case of models with a permanently installed battery, the entire sensor.

Sensor manufacturers make grandiose promises on their websites and in data sheets: battery life of up to 5 years, up to 7 years or even up to 10 years. The question is whether this is realistic and, if so, under what conditions. These are exactly the questions we will answer below.

These factors affect battery life:

Technological advances are enabling ever longer battery life. But change only once every 10 years? Is this oft-communicated promise of many sensor manufacturers realistic?

How long a sensor actually lasts without changing the battery depends on three basic factors:

• Battery capacity
• Power consumption for the acquisition of the measured values
• Power consumption for the transmission of the measured values

In order to achieve the longest possible battery life, all three parameters should be optimized. If this is not the case, dramatically lower battery runtimes can be expected.

Self-discharge and battery life

If the aim is to achieve a runtime of several years without changing the battery of a sensor, the battery used must meet a number of requirements.

A commercially available alkaline manganese battery has a self-discharge rate of around 6% annually. This means that after 5 years of storage (without electricity consumption) it contains only 73% of the original energy, after 10 years only about 50%. These types of batteries are de facto not suitable for long-term operation.

For professional purposes, therefore, special lithium thionyl chloride batteries are usually used. Here, self-discharge amounts to only 1% per year. So after 5 years they still have 95% of the original energy, after 10 years they still have a whopping 90% of the original energy. This battery is often used in the well-known form of an AA battery – with a total capacity of around 2600mAh.

Power consumption for the acquisition of the measured values

Depending on the measured values a sensor collects, such as temperature, motion or air quality, different measurement methods are used with different power consumption. Especially in the latter case, i.e. CO2 sensors for monitoring air quality, the energy consumption is generally higher than with other sensors, for example.

Regardless of the sensor type, in order to obtain energy-saving sensors, this must already be taken into account in the circuit design and firmware. Circuit components should have a high degree of efficiency, active components should have a low quiescent current or a dedicated “sleep mode”. If this is not possible for a sensor element, it should be possible to disconnect it from the supply as an alternative.

The firmware should also be designed in such a way that each component is only active for the minimum necessary time and then deactivated again. It is crucial to keep the time between the sensor waking up and sending the data as short as possible. In this way, the energy-saving potential of a sensor can be optimized.

Power consumption for data transmission

A large part of a sensor’s power requirement is not needed to collect the readings, but to send “LoRa messages”. Depending on the defined parameters, different amounts of transmission power or more or less transmission time are required. Both, added together over several years, have a significant effect on battery life.

Transmitting power: How much power is used to transmit?

The first parameter, the transmitting power, has a direct proportional effect on energy consumption. The higher the transmitting power, i.e. the more “power” the transmission takes place, the higher the range of the signal but also the consumption. In order to consume less power and thus save battery capacity, it should be kept as low as possible.

How far the transmitting power can actually be reduced depends on the distance of the sensor to the gateway and obstacles in the radio link.

Broadcast time: How long does the transmission take?

The second crucial parameter is the transmission time of a message, i.e. how long the sensor has to send in order to transmit the data in full. Depending on the configuration used (see blue box), this may take longer or much shorter.

Background: Transmission speed and configuration

The transmission time is made up of the spreading factor (SF) and the bandwidth (BW). Depending on the SF and BW, there are different data rates and thus the duration of the transmission process.

As can be seen in this figure, a transmission with SF7 and a bandwidth of 125kHz is approximately 22 times faster than a configuration with SF12/125kHz.

Source: official LoRaWAN Specification V1.0

Why don’t they always send at the fastest transfer rate?

The choice of transmission rate depends on the distance and environmental conditions. In addition, the sensitivity of the gateway chip for receiving the messages varies depending on the LoRa parameters. The slower the transmission rate, the weaker the signal can be so that it can still be detected and captured by the endpoint.

For distant or shadowed sensors, this means that transmission must be slower and thus more energy is consumed for the transmitting process.

Ideally, the LoRa parameters should be adjusted for each sending process. In the LoRaWAN specification, a feature called Adaptive Datarate (ADR) is implemented for this purpose. This means that the LoRaWAN parameters are continuously adjusted automatically. An algorithm calculates the ideal parameters from the last 20 transmissions, thus enabling the most energy-efficient transmission process possible.

The transmission interval

The best state of a sensor, in terms of its battery life, is the one in which the sensor sleeps, calculates nothing, and transmits nothing. That’s why LoRa (and low-power applications in general) should pay close attention to when and how often data needs to be transmitted.

The number of transmissions is additionally limited by the ISM regulations. Since LoRa broadcasts in an unlicensed frequency range, it is subject to certain rules. The “duty cycle” indicates the active transmission time, based on one hour. For LoRa in the 868MHz range, this may be max. 1%. This means that one sensor per hour for max. 36 seconds.

There are various computers on the Internet that can be used to determine the sending time of a LoRa message.

Example: How many times per hour can I transmit data?

As an example, we choose the following LoRa parameters for transmission:

Bandwidth (BW): 125kHz
Spreading factor (SF): 12 (lowest data rate)
Sensor data to be transmitted: 100Byte

The transmission of such a message takes (see here) 3.94 seconds. According to ISM regulations, a message may be sent every 6 minutes and 34 seconds.

If the sensor is close to the gateway and can transmit at SF7 (maximum data rate), the transmission process takes only 174 ms and a value should be transmitted every 17 seconds.

As shown in the example above, a massive amount of energy can be saved by improving the quality of the connection. The battery of a sensor that can transmit with SF7 instead of SF12 would last 20x longer. So, the distance and environment of a sensor has a big impact on battery life.

Result:

It is theoretically quite possible to achieve a battery life of 10 years with a LoRaWAN sensor. However, this is very much dependent on the positioning of the sensor, as the greater the distance or more obstacles in the signal path, the more power it transmits and the more it takes.

If the battery life of LoRaWAN sensors is to be maximized, more gateways can be operated than would normally be necessary. This shortens the distance to the respective sensors and, in some cases, allows for up to 20-fold improvement.

In reality, sensors will not always be able to transmit at the maximum transmission rate, receive firmware updates via LoRaWAN, or low temperatures will put more strain on the battery.

The values specified by the sensor manufacturers are achievable under optimal conditions. However, the reality is usually different. The actual battery life cannot be predicted in general due to the numerous influencing factors. Shorter runtimes by a factor of 2 or 3 are not uncommon – depending on the environmental conditions.

You can find more information about the advantages of LoRa technology and areas of application in the field of IoT at LoRa Radio.