How to prevent system errors during power outages?
Applications that require heavy computing and data processing, such as embedded systems, often use devices such as microcontrollers, microprocessors, and field programmable gate arrays (FPGAs) to perform complex computing routines because of their versatility, high speed, and flexibility. However, these recommended devices also have limitations and different power requirements, which can affect the performance and reliability of the system if not considered in the early stages of system development. One of the limitations is the potential for system failure during power-off conditions. When the power supply voltage drops below the minimum operating voltage, the microcontroller may malfunction and cause system errors. Fortunately, voltage supervisors are designed specifically to address this problem.
This article discusses high performance voltage supervisors, including some from the Analog Devices, Inc., portfolio. It describes voltage supervisor functionality, input and output basics, and other fundamentals of high performance power monitoring products.
A voltage supervisor is a type of device that monitors a voltage supply rail and provides an output that can be used to perform some action as long as the monitoring condition is met. It detects whether the monitored voltage supply rail falls below or exceeds a predefined voltage level (called a threshold). The output signal it provides is usually called a reset signal and is used to put another device into another operating mode, such as reset mode or active mode. Voltage supervisors are also very suitable for applications where operating outside a specific voltage range will cause errors and failures. Sometimes, the reset output is also used to enable and disable another device, such as in any application that requires a certain input voltage range for normal operation. A typical application example is using a voltage supervisor to allow the regulator to operate normally, as shown in Figure 1a. To ensure normal operation during startup, the LDO regulator requires sufficient energy in the input, or a high enough input voltage level.
As we all know, voltage monitors are a close partner of microcontrollers or MCUs. When the power supply voltage drops below the minimum operating range while commands are being executed, the MCU is at risk of malfunctioning and causing system errors. In this case, the MCU's power supply voltage is the monitored voltage, and the MCU's minimum operating voltage should be the threshold voltage. We will discuss how to define the threshold level further in the article. A simple example of a voltage monitor for monitoring the microcontroller power supply is the ADM809 shown in Figure 1b. The monitor detects the monitored voltage level and feeds it into the VCC pin. Once the monitored voltage falls below the threshold, the active low reset output puts the microprocessor in reset mode until the voltage supply returns to normal levels.
Figure 1. The ADM809 is a simple example of a voltage supervisor that monitors the input voltage to (a) enable the LDO regulator when the input voltage level is within the correct range and (b) place the microprocessor system in reset mode during a brownout condition.
There are four important input specifications to understand about voltage supervisors. This will help system designers implement voltage supervisors to improve system reliability in their applications. These specifications include reset threshold, threshold accuracy, reset threshold hysteresis, and power-on reset.
Reset Threshold
The reset threshold is a voltage level; when the monitored voltage drops below this value, it issues a reset signal. In voltage supervisor products, the reset threshold is usually marked as V TH . When the monitored voltage V CC drops below the reset threshold voltage V TH , it generates a low reset output, as shown in the timing diagram in Figure 2. In the application, the threshold voltage is set to the minimum voltage that allows the system to operate normally.
Figure 2. Timing diagram of the voltage supervisor's monitored voltage, V CC, and reset output signal.
One way to set the reset threshold is through an external resistor divider. A small portion of the monitored voltage is compared to a reference voltage to see if the monitored voltage is above or below the reset threshold, as shown in Figure 3a. The ADM8612 is an example of this configuration. The reset threshold of some voltage supervisors is set by an internal resistor divider through laser trimming at the factory, such as the MAX16140. This brings some advantages, such as fewer external components, which can free up extra space for the solution and meet the needs of compact applications, as shown in Figure 3b. It also achieves higher accuracy because it does not rely on external factors (such as using standard value resistors with tolerances). However, the external resistor scheme allows flexibility in adjusting the reset threshold level.
Figure 3. How to set the reset threshold: (a) the ADM8612 reset threshold is set using an external resistor divider, and (b) the MAX16140 reset threshold is set using a factory-trimmed internal resistor divider.
Threshold accuracy
Threshold accuracy refers to how close the actual threshold is to the calculated reset threshold or target reset threshold. Several factors affect the accuracy of the threshold, including the resistor divider and the reference voltage. Both the resistor divider and the reference voltage are analog circuits that are affected by environmental factors such as temperature. This results in a certain tolerance on the reset threshold. The more robust the reference voltage and resistors, the tighter the tolerance and the higher the threshold accuracy. Threshold accuracy is usually expressed as a percentage. Assuming the threshold accuracy of the voltage monitor is ±1% and the threshold is set to 3.3 V, the actual threshold may be around 3.267 V to 3.333 V.
It is important to understand the threshold accuracy as this is critical to setting the reset threshold. If the accuracy is not considered when setting the reset threshold, the system may fall into an undesirable failure region.
Reset Threshold Hysteresis
Reset threshold hysteresis is the additional voltage required to cancel the reset signal after the monitored voltage returns to the normal area. In voltage supervisors that monitor for undervoltage, the reset threshold hysteresis is usually expressed as V HYST or V TH+HYS . Hysteresis has several benefits. First, it ensures that the monitored voltage returns to the normal level and has a certain safety margin relative to the threshold. Second, it allows the power supply to stabilize before the reset is canceled, which helps solve power supply noise and instability problems. Without hysteresis, the voltage supervisor will repeatedly issue or cancel the reset signal when the monitored voltage exceeds the threshold. This can happen in applications with power supply noise or in battery-powered systems because the voltage drops with load current due to internal resistance. An example is shown in the purple shaded area in Figure 4. At the same time, due to the presence of hysteresis, the reset output will keep the system in reset mode until the power supply stabilizes, eliminating unstable and oscillatory behavior of the system, as shown in the blue shaded area in Figure 4.4.
Figure 4. Comparison of reset output behavior with and without hysteresis.
Power-On Reset
During startup, when the supply voltage begins to rise, the internal circuitry of the voltage supervisor is not adequately biased. Therefore, the reset output is in an undefined state. As the supply voltage continues to rise, it will reach a certain voltage supply level that causes the voltage supervisor to leave the undefined state and issue a valid reset signal. The minimum supply voltage that allows the supervisor to be in a specified state and provide a valid reset output is called the power-on reset voltage or V POR . Consider the simplified schematic of the voltage supervisor in Figure 3b. Assuming the open-drain reset output is pulled up to V CC , in the undefined state, the reset output will reflect the supply voltage V CC . This will produce a glitch in the reset output, called the power-on glitch. When the supply voltage reaches V POR , the supervisor will issue a valid reset output signal, as shown in Figure 5 .
Figure 5. Power-on glitch and power-on reset voltage VPOR during startup.
In some applications, the power-on glitch is negligible and insignificant, such as in high-voltage systems. However, for some applications, such as devices with lower logic-high voltage thresholds, this is undesirable.
One factor to consider when designing a voltage supervisor is the reset output polarity and timing. You can choose the polarity—active low or active high—depending on the application.
Low level is effective
An active low output means that the reset output goes low as long as the monitored voltage is below the threshold voltage. The timing diagram in Figure 2 shows the response of a voltage monitor with an active low output. For easy identification, the active low reset output is labeled RESET (read as RESET bar). When the monitored voltage rises above the threshold voltage, the RESET output remains active for a specified time before going high. This time delay is called the reset timeout period (t RP ), which can be a fixed time or adjustable with an external capacitor.
High level is effective
Depending on the output requirements, the system may require an active high output. In contrast to an active low output, in an active high output, the reset output goes high when the monitored voltage is below the threshold and goes low when the monitored voltage rises above the threshold voltage after the reset timeout period tRP. See Figure 6 for an illustration .
Figure 6. Timing diagram of VCC and reset signals for an active-high reset output.
Another factor to consider, depending on the application, is the output topology. There are two main output topologies used – open-drain and push-pull.
Push-Pull Output Topology
The push-pull output topology consists of a pair of complementary MOSFETs, as shown in Figure 7. When the bottom FET is off and the top FET is on, the reset output goes high; when the bottom FET is on and the top FET is off, the reset output goes low. The push-pull output provides a high-speed response that is almost rail-to-rail from low to high and from high to low.
Figure 7. Push-pull output topology.
An active-low push-pull reset output is suitable for most applications, but other output types are available. As shown in Figure 8, a push-pull output in a single-voltage system is straightforward, but a push-pull output in a multivoltage system requires more care, especially when the microcontroller has only one reset input.
Figure 8. Single voltage system.
Open-drain output topology
For the open-drain topology, the reset output of the supervisory circuit is the drain of the internal MOSFET. To produce a logic signal output similar to that shown in Figure 3b, an external pull-up resistor is connected from reset to the supply voltage. When the MOSFET is turned on, the reset signal goes low; when the MOSFET is turned off, the reset signal goes high. The pull-up resistor can be connected to a voltage rail other than the supervisory circuit supply. This is very beneficial for systems that require a reset level that is different from the supervisor supply voltage.
Another advantage of open-drain outputs is the "wired-OR" function. Connecting the open-drain outputs of two or more monitoring circuits to the same bus allows you to implement a "negative logic OR" circuit. This means that when the reset output of any one of the monitoring circuits goes low, the bus is low. The bus is high only when all reset outputs are high. This topology is convenient if you want to monitor multiple power supplies and trigger a reset when any one of them drops.
Application Cases
Figures 9, 10, and 11 show some typical application cases of different output topologies and polarities of voltage supervisors. Figure 9 shows an example of a multi-voltage system using an open-drain topology. In a multi-voltage rail system, daisy-chained active-low outputs can be used to perform sequencing, as shown in Figures 10a and 10b. In some applications, proper power supply sequencing may be one of the first considerations. Multi-rail systems, such as FPGA-based solutions, often require and specify proper power supply sequencing to prevent system failures and instability. Figures 11a and 11b show examples of applying active-high outputs. For these cases, the active-high output is used to enable or disable the high-side MOSFET to implement an on/off control scheme. Such configurations can be used in circuits such as overvoltage protection, low-voltage sequencing, etc. The high-side MOSFET can also be driven using the active-low output of the voltage supervisor. For more information, refer to the article "Using Active-Low Output to Drive High-Side MOSFET Input Switches for System Power Cycling."
Figure 9. Multivoltage systems share a single microprocessor reset input.
Figure 10. Multi-rail sequencing using active-low outputs (a) push-pull topology and (b) open-drain topology.
Figure 11. Applications of active high output polarity. (a) N-channel MOSFET low-voltage sequencing using push-pull topology. (b) P-channel MOSFET overvoltage protection circuit using open-drain topology.
Voltage supervisors are used to enable, disable, or reset another device. A common application for supervisors is to reset a microcontroller. Supervisors protect the system from errors and failures, thereby improving the overall reliability of the application. The input, output, and timing specifications of the voltage supervisor need to be considered during design. Supervisors have different output topologies and polarities, which can play different advantages in different application scenarios, thereby achieving the intended function and improving system reliability.
????Click here to explore the ADI "chip" world
京公网安备 11010802033920号