Using power switches can be complex or even confusing for most electronic designers, especially for those who are not power management experts. In a broad range of applications such as portable electronics, consumer electronics, industrial or telecommunication systems, designers are increasingly working with power switches that can be used in a variety of ways including control, sequencing, protection, power distribution or even system supply turn-on management. Of course, each of these need power switching solutions with different characteristics.

This article summarizes important specifications and concepts that designers need to consider when using power switches in different applications, and reviews possible solutions to help designers select an optimized solution.

The first question you should ask yourself before selecting a power switch obviously is: "What do I want to achieve with this switch?" This is an easy question, but the answer will help to define the perfect product to use. There are several ways to use power switches. The most popular are to:

• control, distribute & sequence (i.e., turn on/off a power rail to enable a subsystem or distribute power to multiple loads)
• protect against short circuits or any kind or any kind of over-current or over-voltage (USB current limiting, sensors protection, power rail short circuit protection)
• manage the turn-on inrush current (i.e., when charging a capacitor)
• select power supplies (i.e., muxing or ORing) or load sharing.

Table 1 summarizes which feature is important to consider for each specific use of a power switch.

Table 1: Application-specific Requirements
(Click on Table to enlarge)

ON Resistance, Maximum Currents and Input Voltage Range
ON Resistance (rON), maximum continuous current, and input voltage range are always key characteristics to consider. These are the basic characteristics you need to study before looking at any device. Depending of the application, the designer can easily determine the current that needs to be switched, and at what voltage. Based on this information, a first selection can be made. Indeed, if you need a switch to pass 1.2 V or 36 V, two distinct product ranges can be identified.

ON resistance impacts the dropout you will see across the switch. Designers must be cautious to understand what the maximum acceptable dropout is with regard to their particular application set-up (voltage, current). This can be calculated easily using Equation 1:

where the dropout is V_DROP, the pass FET ON resistance is r_ON, and the current through the switch is I.

If the application needs to switch a lot of current, or switch a low-voltage rail (like 1.0-V), then dropout needs to be minimized. Therefore, the ON resistance needs to be as low as possible, for example, as with the TPS2292x series featuring a 14-mΩ r_ON at 3.6-V.

However, if the current to be switched is small, the ON resistance won't be a key concern, and you can select a higher ON resistance device like one of the TPS2294x series, which is about 1 Ω. The ON resistance is a key contributor to the die size of a power switch device and, hence, to the cost of the device. You will want to look closely at this to select the most cost-effective solution possible.

In addition to the maximum continuous current the designer targets to switch, another important characteristic is the maximum pulsed-current the switch can accept. In certain applications, the load requested most of the time consists of moderate, continuous currents. However, spikes are evident when a subsystem requests additional power. A good example is the GSM/GPRS transmit burst, which sinks up to 1.7 A during 576 μsec with a duty cycle of 12.5 percent. Make sure the selected part can support such pulsed current.

Power Dissipation and Protection Features
Power dissipation is also an important characteristic to consider. During normal operation as a pass switch, power dissipation can be calculated considering the switch's ON resistance as well as the current being switched. You can easily calculate the maximum power dissipated through the device by using Equation 2:

If the part's ON resistance is selected low enough, the power dissipation is small and has little effect on the part's operating temperature. However, be careful if you plan to use the switch to protect the rail against an over-current, or a short circuit as with USB ports, or fingerprint sensor protection. In this case, you must select a current-limited switch like the TPS22944.

If you are not using a current-limited switch, power dissipation can be a major issue for system reliability. For instance, a 0.9-Ω short applied to a non-current-limited-load-switch with a 3.3-V input voltage (switch ON resistance being ∼100-mΩ like for the TPS22902) translates into a dissipated power as shown in Equation 3:

(Click on equation to enlarge)

Usually, this power dissipation is too high for most packages in the market, resulting in failure and reliability issues.

In the same manner, the designer using a current-limited switch needs to make sure the package can support a short-circuit condition. If the part goes into current limit, maximum power dissipation occurs when the output is shorted-to-ground. For devices like the TPS22945 featuring auto-restart time, t_RESTART, and the overcurrent blanking time, t_BLANK, the maximum average power dissipated is shown in Equation 4:

(Click on equation to enlarge)

For devices that do not feature auto-restart loops like the TPS22944, a short on the output causes the part to operate in a constant-current state, dissipating a worst-case power level until the thermal shutdown activates. It then cycles in and out of thermal shutdown so long as the ON pin is active and the short is present.

Several current-limited switches exist in the market and the two main characteristics to look at are the current-limit minimum value (fixed current limit or programmable using an external resistor), and the current-limit accuracy and response time. In most applications the current-limit accuracy isn't a key concern because the device is used as a circuit breaker (i.e., the switch is turned off in case of a short circuit). However, accuracy in some application like USB current limiting can be important, as the switch is used as a constant current source.

For applications where it is expected to switch large current or to face over currents, it is recommended that you select a device that features some kind of thermal protection. When the device temperature is identified as too high, most devices will activate the thermal shutdown, which will turn off the FET in order to protect the device itself against any potential thermal damage.

Aside from the current limit (or over-current protection, OCP), which is mandatory to protect against short circuits, other protection features like reverse current blocking can be interesting to consider.

Reverse current blocking (also known as reverse voltage protection) is mandatory when designers are trying to design a power selector (ORing), or to make some load sharing.

Figure 1 shows an example of power switches configured to supply a load from two potential power sources (i.e., a DC input and a battery):

Figure 1: Dual-Source Power Selector
(Click on image to enlarge)

For a device that doesn't have reverse voltage protection, it is important that the input voltage of the pass FET stays higher than its output voltage. Otherwise the input will be clamped via the body diode of the FET, which will cause significant current to flow from the output to the input.

In the example in Figure 1, if the battery is a lithium-ion (Li-ion) battery at 4.2 V (max), and if the DC input is enabled and at 5.0 V, then a potential large current will flow from the load to the battery which, of course, is undesirable!

A solution to make this work is to use a device featuring reverse-voltage protection. Reverse-current protection usually can be implemented using back-to-back FETs or by switching the back gate of a PMOS FET when a reverse voltage condition is detected. You will want to look at the reverse-voltage comparator trip point (V_OUT " V_IN value threshold above which the reverse-current feature is activated), as well as the time from reverse-voltage condition to MOSFET turn off.

Another protection that can be useful for some applications is the overvoltage protection (OVP). This feature protects the switch and the system, if an overvoltage is applied to the switch. It can be useful for example in some USB applications or in some battery applications.

Inrush Current Management
Another common use of a power switch is to manage the inrush current when the system is being turned on. If the switch turns on without being controlled, a large inrush current is created that could result in a supply rail drop at the input of the switch. This could ultimately impact the system's entire functionality.

When charging large output capacitances, inrushes will be large and need to be controlled and/or limited. The inrush current can be calculated using Equation 5:

For example, with a C_LOAD = 1 μF, V = 3 V, and a rise time of 1 μsec, the inrush current could be as high as 3 A.

An easy way to avoid this inrush current is to slow-down the switch's rise time. This will slowly charge the output capacitor and reduce the current peak. In the example in Equation 5, a rise time of 200 μsec would result in a 15 mA inrush, which is acceptable.

In some cases, you may want to charge extremely large capacitors (several hundred of μF). A very slow rise time is usually recommended, but you can also select a switch with a high current limit. The device will go in current limiting at power on and the capacitor will be charged at the current limit value, which will be the maximum power dissipation capability of the power switch.

System interoperability
Whenever selecting a power switch, system interoperability needs to be cautiously considered. For instance, when using a power switch to enable and disable loads to optimize power consumption in a portable application, the switch's control inputs must be compatible with general-purpose, low-voltage (1.8 V), GPIO is critical.

In addition, when turning off the switch, make sure that the floating output of the switch does not impact system performance. Therefore, some users may tie the power-switch output to ground when disabled with an additional transistor, or use an integrated device, which integrates this pull down-to-ground like in the TPS22902.

Another important point to check is the input and output capacitance used to design a stable system. Although an input capacitor is usually not required to stabilize power switches available in the market, it is considered good analog design practice to connect a 0.1 μF to 1 μF, low equivalent series resistance (ESR) capacitor across the input supply. This capacitor counteracts reactive input sources and improves transient response, noise and ripple rejection.

Depending on the load of the switch, you may want to consider some additional storage capacitors at the switch's output. If the switch doesn't have reverse-current blocking, then an input capacitor bigger than the output capacitor is strongly advised, otherwise the input will be clamped via the FET's body diode, which will cause significant current to flow from the output to the input.

References
For additional information on the technologies and products discussed in this article, visit: www.ti.com/loadswitches-ca.

Related link
See the online, on-demand course, "Fundamentals of MOSFETs for Switching"

About the author
Philippe Pichot manages the strategic marketing for the load switches product line at Texas Instruments. Philippe received his MSEE from the Institut Superieur D'Electronique du Nord (ISEN) in Lille, France. Philippe can be reached at ti_ppichot@list.ti.com.