One of the important design considerations in order to get the most out of the battery life for an IoT device is how to connect the battery (or batteries) to the load. The best option is often to connect the two through a DC/DC converter, enabling a power regulation that secures relevant power supply for a specific application, see our comparison to LDO or direct connection in our tutorials below. This way many application runs with optimal drainage of the battery, and with minimal unnecessary energy consumption. To get this optimal case, the DC/DC converter needs to exhibit high efficiency for the voltage and current required for the IoT device over time. So how do you know if a certain DC/DC converter is the right one for your system, matching the battery and the device or sub-system in the device?
The DC/DC converter data sheet gives an obvious indication however the best way to find out is to characterize DC/DC efficiency as part of the whole system under design. In this tech post we showcase how this can be done.
We investigate the DC/DC efficiency for a system that consists of a 1.5V Alkaline battery, and a DC/DC converter that boosts the voltage to 3.3V to an Arduino Pro Mini, see figure 1. The timer on the Arduino board starts the system at certain intervals. During these power-on intervals we want the system to run as efficiently as possible. We want the system to utilize the high efficiency of the DC/DC converter in the range of 1 – 10 mA (active mode) for the entire battery life cycle, with the decreasing battery voltage over time in mind. For systems where the sleep mode power consumption dominate over the active mode power consumption the efficiency of the DC/DC is not as important.
When connecting a load, RL, to a battery with an unloaded voltage, E, the voltage over the load will drop due to the internal resistance of the battery, Ri. The battery voltage, U, will decrease as the current into the load increases due to the voltage drop over the internal resistance. Coin cell batteries and AAA batteries have a high internal resistance so it is not possible to load them with high current as the output voltage would be too low.
If the load is resistive, like with a simple resistor, the current drained from the battery will decrease as the battery discharges and voltage drops. When a DC/DC converter is connected to the battery, presenting a fixed output voltage to the same load, the current drawn from the battery will increase when the battery voltage drops. This will accelerate the discharge, make the voltage drop further which leads to further current increase and so forth.
In the datasheet (see figure 3) for the DC/DC converter we use here, efficiency is presented as a function of load current at a specific voltage, 1.2V in this case. More information of our system is needed, for example the efficiency for the range of ~0.8V – 1.5V. It would be great to have this curve as a function of the current drawn from the battery with the DC/DC self-consumption and the losses in its components (mainly the inductor) included.
When the voltage starts to drop, the current drawn from the DC/DC converter starts to increase which leads to decreased voltage from the battery (note that we are only checking behavior at room temperature in this study). The efficiency of a step-up DC/DC converter typically decreases with lower input voltage.
For the DC/DC characterization, Otii is used as the supply for the system, acting as the battery, but also as a measurement unit. The fact that Otii can supply the system while measuring both input voltage and current at the same time as output voltage and current from the DC/DC converter, comes in handy. By dividing the output power with the input power, the efficiency can then be calculated.
To be able to measure the output current from the DC/DC converter a sense resistor is needed, see Figure 4. The value of the resistor must be high enough to allow sufficient resolution for the measurement but low enough to not introduce a too big voltage drop and to not exceed shunt voltage limits.
In this example, a 4.7 ohm sense resistor is used. This will result in a voltage drop of 37mV at 8mA current, which will not affect the performance of the Arduino. Otii’s ADC expansion port voltage range is -81.9175mV to 81.2mV, so 37mV is well within this range.
A simple application is run on the Arduino Mini, flashing an LED, while measuring the DC/DC efficiency for that specific use case.
A Lua measurement script is run in the Otii desktop application to set input voltages from 1.5V (fully charged battery) to 0.9V (almost empty) and calculate the DC/DC converter efficiency over time to reflect a real use case. One .csv file per voltage is exported, containing the measurements and calculations.
Between each measurement, the power supply is disabled to reset the system. By doing this, it is possible to see the entire startup phase and the active phase of the system. The script file and saved .csv files can be found in the download section below.
Figure 5 shows that the output current from the DC/DC converter (ARC ADC Current in the figure) is relatively constant, regardless the input voltage. This can be expected as long as the DC/DC converter works as designed. However, the input current (ARC Main Current, see Figure 5) increases when the input voltage decreases.
The figure 6 shows the DC/DC efficiency for different battery voltages. The average DC/DC converter efficiency varies a lot with battery voltage. At 1.5V the efficiency is more than 87% while at 0.9V it has dropped below 79%. This is measured with an Arduino as load, consuming roughly 5.1mA when the LED is on and 3.6mA when it is off. As expected the data sheet values provide a good indication of this for higher battery voltages, however for lower voltages this data is not provided, hence importance of characterization is apparent.
Most battery powered IoT devices spend most of their time in sleep mode, consuming very little power, and then periodically wake up, performing tasks for a short time that consume more power. The activity profile is usually unique for each type of product, as well as the design.
There is no single design for a power supply that fit all these different products. In order to find an optimal design you need to characterize the battery and the load, design the power management, and then validate the entire system design. In this tech post we showed how to characterize a power management that is a DC/DC converter, for a specific IoT design example.
The DC/DC converter data sheet gives an obvious indication of the performance but it is general and often limited, hence the best way to find out is to characterize DC/DC efficiency as part of the whole system under design. A task that is straight forward when using the Otii tool.