There are several ways to use a p-channel MOSFET (PFET), the most common way of usage being as a load switch. When using it together with batteries there are a few things one should keep in mind.
The battery as a source is not a constant voltage source. The battery voltage is influenced by a lot of factors, including state of charge (SoC), the amount of current discharging the battery, and temperature. This means that the PFET as a load switch needs to handle different battery voltages during the duty cycle and lifetime of the product. By handling we mean it needs to perform without wasting unnecessary energy.
Hence, if you want to use a PFET as a load switch for your battery-powered device, you want to know how the battery behaves for your device. Let’s go through some practices for matching a PFET to a battery. Keep in mind: The lower the power dissipation of the PFET the better.
The power dissipation for the PFET depends on Rds(on) – the resistance for the FET when it’s conducting. Looking at datasheets for PFETs, you will see that Rds(on) is usually displayed as a table.
Fig. 1 PFET data sheet, in this case SSM3J338R from Toshiba (link to the datasheet)
The resistance depends on the voltage between the gate pin and the source pin. That means to avoid a high-power dissipation at the load switch, you want to make sure that your PFET works well with your battery voltage.
Assuming you are using a single alkaline cell, you can expect the battery voltage to be between 1.6V and 0.8V. If you go back to the datasheet displayed in Fig.1, you most likely find that there is nothing specified for that low voltage level (Vgs is the same as your battery voltage, but negative, when the logic signal is low). This would lead you to the conclusion that this PFET is not a good match for this battery.
Fig 2 Simple schematics example of PFET as a load switch
Let’s look beyond the datasheet. The PFET needs to have a decently low power dissipation for the whole battery lifetime. In the datasheet the Vgs voltage of -1.8V states that below a battery voltage of 1.8V the set-up would not work, but it doesn’t state the firm cut off. Perhaps, it covers the single alkaline cell down to 0.8V.
So, we measure the Rds(on) for different Vgs voltages.
Fig 3 Otii Arc setup to investigate PFET resistance
Setting up an Otii Arc as shown in figure 3, you can observe the Rds(on) for different Vgs voltages by changing the voltage of the GPO1 pin (Vgpo1). Vgs in this case is (Vgpo1) – (V+) with V+ being the voltage on the main terminal of Otii Arc. For this investigation we have set V+ to 4V, and you can change Vgpo1 to anything between 1.2V and 5.0V. This allows us to investigate PFET resistances for Vgs between -2.8V and 0V.
The reason for not just changing V+ is that you want to have the same load situation. If you lower V+, then you will also decrease the load current. By changing Vgpo1 instead, you keep the load current the same. The PFET does not see the difference.
You can do these tests manually; we chose to speed it up by writing a simple lua script that adjusts Vgpo1 to measure the current flow for different Vgs voltages. Based on that we can calculate the PFET resistance. If you’re recreating this automatic calculation on your own desk keep in mind that for this setup you will need the Otii Automation toolbox as a software add-on to your free Otii Arc standard software.
Further, we have used several different resistors as switched load, ranging from 3ohm to 220ohm. The recording below shows a load of 22ohm.
Fig 4 Otii Arc recording with a switched load of 22ohm
Very short on the topic of temperature in this investigation of PFET resistance. Do try to avoid heating effects! We have paused 60s between each measurement for cooling, so, that the starting temperature should be roughly the same for the benchmark purposes.
Here are the results from the measurements:
Based on the datasheet the PFET doesn’t match the battery and can be dismissed. However the measurements above show that, for the battery voltage between 1.6V and 0.8V the load switch will have a resistance of roughly 35mohm while using the battery when full. Hence, the load switch could be used with the battery despite the datasheet numbers for that particular case. It will, however, have several ohms before the battery is empty, as seen at Vgs=-0.9V in the table. Consequently, you will end up with an unnecessary power dissipation and most probably get a shorter battery life.
When using PFET as a load switch, make sure that you use a PFET that works well, without unnecessary high power dissipation, throughout the whole discharge life of the battery. Check the datasheet but measure it yourself to get insight into the real behavior throughout the lifetime of your product.
Of course, this is just one of many parameters you should take into account when picking a PFET, nevertheless, it is also the one that many miss.
So, #measureeverydamnday and good luck in your design endeavour!
Products used in this article:
Power Analyzer, Log Sync & Power Supply in one Product. Otii Arc comes with the featured-packed standard Otii software (perpetual license).