To minimize power consumption, a simple MOSFET is often used to gate power to unused circuits. However, a better option is to use a load switch as it has additional functions to handle the many subtleties and vagaries of power rail management.
Load switches are offered with an array of performance parameters and ratings from multiple vendors, which enables a good fit between application priorities and available parts.
This article will briefly discuss IC and circuit power-down concepts, and then introduce suitable load switch options and how they should be used.
Minimizing power consumption by temporarily shutting down unneeded functions is a standard system tactic. For this reason, many ICs have a user-directed, ultra-low power quiescent mode.
However, putting an IC in a quiescent mode only shuts down the IC and not the associated circuitry, which includes other power dissipating passive devices (primarily resistors), as well as active discrete devices such as transistors. Therefore, designers often turn to the simple MOSFET to gate the power to the entire subsection to be shut down.
This MOSFET may be needed even if the power supply (either LDO or switching) can be turned off via an enable control line to reduce the idle mode power consumption of its load subcircuit. The reason is that while the savings can be significant, the leakage current of many power supplies is relatively large even in their shutdown mode, so the power savings may not be sufficient.
Although using a properly sized MOSFET as a power-rail on/off switch does work, the capabilities and functions which a MOSFET alone can provide are marginal, and often unable to support other on/off switching requirements. Also, the circuit designer must provide a suitable gate driver for the MOSFET, which becomes another item on the “to do” list, and so adds to design complexity, time, space, and cost.
The load switch provides an “all-in-one” solution
A better approach is to use a “load switch” IC, which is a pass element MOSFET plus additional power management functions in a tiny package. Most load switches have just four pins, with one each for input voltage, output voltage, logic level enable, and ground (Figure 1).
Figure 1: The basic load is a four-terminal device which combines a MOSFET and MOSFET driver in a single, easy-to-use package. (Image source: Texas Instruments)
The operation is simple: when the load switch is enabled via its ON pin, the pass FET turns on and allows current to flow from the input (source) voltage pin to the output (load) voltage pin. As with a basic MOSFET, the DC resistance through the “switch” is just a few milliohms (mΩ), so the voltage drop is low, and so is the associated power dissipation.
A load switch is more than just a MOSFET and a driver which allows it to be turned on/off by a simple logic level signal. While that capability alone might make the load switch a better solution than a MOSFET plus separate driver, a load switch does much more (Figure 2).
Figure 2: Load switches are often enhanced with other functions, including discharge control, slew rate control, various forms of protection, and fault monitoring. (Image source: Vishay Siliconix)
Why use the load switch, besides the logic level control feature? There are several reasons:
- The integral driver manages charging and discharging the gate, thereby providing slew-rate control of the rise/fall times of the MOSFET’s turn-on/turn-off cycle. This optimizes the MOSFET performance, avoids overshoot and ringing, and minimizes undesirable EMI/RFI.
- Further, control of the turn-on time of the MOSFET in the switch prevents the input rail from sagging due to the sudden increase in load from inrush current resulting from quickly trying to charge the load’s capacitor. This sag is a problem if the same input rail also supplies power to other sub-systems which must remain fully powered.
- Some load switches offer a quick output discharge (QOD) feature via an on-chip resistor between the output and ground; this mode is activated when the device is disabled via the ON pin. This will discharge the output node and prevent the output from floating, which can cause unwanted activity when the load circuitry is not powered down to a defined state.
Note that this feature is sometimes undesirable: if the output of the load switch connects to a battery, such quick output discharge would cause the battery to drain when the load switch is disabled via the ON pin – not a good thing! Therefore, some vendors offer it as a selectable feature in a single device, while others offer two variations of a load switch, with one having it and the other not. The former option allows multiples of the same parts to be used in a single product but in different scenarios.
- Load switches can incorporate other features which are desirable whenever there is a power source and rail, such as a thermal shutdown, undervoltage lockout, current limiting, and reverse-current protection. These protection features contribute to system-level integrity.
Compared to starting with a basic MOSFET to switch the power rail and adding these features and capabilities, the overall BOM, design time, and real estate cost can be greatly reduced.
Going further, load switch use is not limited to simple shutdowns to save power. By using an array of load switches, a single larger supply can power multiple circuit subsections, with up/down power to these subsections implemented via a prescribed sequence and timing under the control of multiple digital outputs (Figure 3). In this way, the load switches act as the gating elements of a broader and effective power management control scheme.
Figure 3: Load switches allow a single supply to drive multiple loads, each having independent turn-on/off and relative timing. (Image source: Texas Instruments)
Keep in mind that load switches need a capacitor (typically 1 microFarad (μF)) on their input side to limit the voltage drop on the input supply caused by the transient inrush currents into the discharged load capacitors. They also need to “see” a load capacitance which is about one-tenth the value of the input capacitance; if the load is less than that, a small output capacitor should be added.
Load switch parameters
The performance attributes of a load switch begin with those of a standard FET used as an on/off switch. These include:
- ON-state resistance (Ron) determines the voltage drop across the load switch and also the power dissipation of the switch. Typical values are in the tens of milliohms range, but will vary with individual vendor products and load switch current capacity. The designer must do some basic calculations to determine the maximum allowable value in the application.
- Maximum voltage (Vin) and current (Imax) ratings specify how high a voltage the switch can tolerate, and with how much maximum current flow. The designer should check both the steady-state values as well as the transient and peak values of these factors.
- Other parameters are the quiescent current and the shutdown current. The quiescent current is the current that the load switch consumes when the load switch is ON, and thus becomes wasted power. This is negligible compared to the power that the load itself is dissipating. The shutdown current is the current that “leaks” from the load switch to the load when the switch is in OFF mode.
Loads switches span simple to complex
A good example of a load switch with extra features is the NCP330 from ON Semiconductor. This is a basic N-channel MOSFET load switch, but it includes a 2 msec soft-start mode for circumstances where the application of sudden loads may be detrimental. This is often required in mobile applications, where there is a limited capacity battery (Figure 4).
Figure 4: ON Semiconductor’s NCP330 load switch includes a 2 msec slew mode so that the load is not suddenly attached to the source. This prevents various operational and performance issues for both supply and load. (Image source: ON Semiconductor)
The NCP330’s very low on-resistance of just 30 mΩ makes it a good fit in a system battery charging up to 3 amperes (A) (5 A peak). The 1.8 volt to 5.5 volt device is automatically enabled if a supply is connected to the Vin pin (active High). If there is no input voltage, it remains off via an internal pull-down resistor. Reverse-voltage protection is built in as well.
Vishay Siliconix offers the SiP32408 and SiP32409 slew-rate controlled load switches (2.5 msec at 3.6 V) designed for 1.1 V to 5.5 V operation. The SIP32409 is identical to the SiP32408, but with a fast turn-off circuit for output discharge. A key feature is that their on resistance, typically 42 mΩ, is flat over most of the supply range, from 1.5 to 5 volts. Another attribute is that the controlling enable voltage is also low, so it can be used in low-voltage circuits without the need for a level shifter (Figure 5).
Figure 5: Relationship between controlling enable signal low and high logic-levels thresholds versus the input voltage for the SiP32408 and similar SiP32409 load switches from Vishay Siliconix. (Image source: Vishay Siliconix)
Although load switches are relatively simple devices in terms of number and function of package pins, layout can still be an issue when there is current flow and likely parasitics. For this reason, it is best to use the company’s suggested pc board layout (Figure 6), as well as top and bottom side layouts for a 1 × 1 inch (2.5 × 2.5 cm) evaluation board (Figure 7).
Figure 6: It takes a carefully planned pc board layout and component placement to realize the full performance of load switches such as the SiP32408 and SiP32409, so that ground noise, parasitics, and current flows do not impede maximum performance. (Image source: Vishay Siliconix)
Figure 7: In addition to showing a preferred pc board layout for the SiP32408 and SiP32409, Vishay Siliconix also provides the layout for a small evaluation board for these devices. (Image source: Vishay Siliconix)
A load switch for use at the lower voltages which are increasingly common is the TPS22970 from Texas Instruments, which can operate from an input as low as 0.65 V up to 3.6 V (Figure 8). On resistance is also low, from a typical 4.7 mΩ at an input of 1.8 V, rising slightly to 6.4 mΩ at 0.65 V. The switch handles a continuous current of 4 A, with an on-state quiescent current of 30 μA (typical) at an input of 1.2 V, and an off-state current of 1 μA at inputs above 1.8 V.
Figure 8: The basic application of the TPS22970 shows the critical input (source) capacitor and the sometimes unnecessary output (load) capacitor; it also makes clear that a load switch is a simple four terminal device. (Image source: Texas Instruments)
The TPS22970 has a 150 Ω on-chip resistor for quick discharge of the output when the switch is disabled. This avoids any unknown state caused by a floating supply to be seen by the load. The slew-rate controlled turn-on times are 1.5 milliseconds (ms) and 0.8 ms, at input voltages of 3.6 volts and 0.65 volts, respectively. The comprehensive data sheet (25 pages long, for a four terminal device) contains numerous detailed tables and graphs which fully characterize its performance across a wide array of perspectives. For example, it shows the rise and fall times versus temperature for each of four input voltages (Figure 9).
Figure 9: Rise time (left) and fall times (right) versus temperature with load resistance of 10 Ω and load capacitance of 0.1 μF, for the TPS22970. (Image source: Texas Instruments)
MOSFETs by themselves can provide a simple solution to switch DC on and off to minimize power usage, implement sequencing of multiple loads, and control power timing. However, a load switch with an integrated MOSFET, driver, slew-rate control and various forms of fault protection is often a better choice, as it can provide all these extra features in a single, small footprint device.
Load switches are offered with an array of performance parameters and ratings from multiple vendors, enabling a good fit between application priorities and available parts.