A Novel Approach to Implementing Precise, Low-Power, and Compact Temperature Monitoring

Heat can present challenges for designers of almost any electronic system such as wearables, white goods, medical equipment, and industrial equipment. Unnoticed heat buildup can be especially troublesome. To avoid such a problem, there are several options available to detect heat including temperature-sensing ICs and positive temperature coefficient (PTC) thermistors. However, these have their limitations. Each sensing option uses multiple components, requires a dedicated connection to the host microcontroller unit (MCU), occupies valuable board space, takes time to design, and has limited precision.

That said, designers have a new option. ICs have been developed for use with multiple PTC thermistors that enable a single IC to perform precise overtemperature detection with one connection to the host MCU. To provide high levels of design flexibility, these ICs select output currents to support various PTC thermistors. They are available with a choice of MCU interfaces and may include a latching function. They come in a tiny 1.6 x 1.6 x 0.55 millimeter (mm) SOT-553 package and have a current consumption of 11.3 microamperes (μA), thereby enabling compact and low-power solutions.

This article reviews the heat sources in an electronic system and examines some temperature-monitoring solutions using PTC thermistors combined with sensing ICs or discrete transistors. It also compares those solutions with temperature measurement ICs. The article introduces and explains how to apply ICs from Toshiba that exemplify low-power, cost-effective thermal protection.

Heat sources

The heat generated by electronic components negatively impacts user safety and device/system operation. Large ICs like central processing units (CPUs), graphical processing units (GPUs), application-specific ICs (ASICs), field programmable gate arrays (FPGAs), and digital signal processors (DSPs) can produce substantial amounts of heat. They need protection, but they are not the only devices that must be monitored for excessive heat.

Current flowing through a resistance causes heat, and in the case of large ICs, there are thousands or millions of micro heat sources that can add up to a large thermal management challenge. Those same ICs often need precise voltage regulation directly adjacent to their power pins. This can require multiphase point of load (POL) DC-DC converters or low dropout (LDO) linear regulators. The on-resistances of the power MOSFETs in POLs and the pass transistors in LDOs can cause the devices to overheat, reducing voltage regulation accuracy and compromising system performance.

It is not just POLs and LDOs that generate heat. Heat needs to be monitored and managed across a range of systems, including AC-DC power supplies, motor drives, uninterruptible power systems, solar inverters, electric vehicle (EV) drive trains, radio frequency (RF) amplifiers, and light detection and ranging (LiDAR) systems. Those systems can include electrolytic capacitors for bulk energy storage, electromagnetic transformers for voltage transformation and isolation, optoisolators for electrical isolation, and laser diodes.

Ripple currents in electrolytic capacitors, eddy currents in transformers, current flow in the LED in optoisolators, and laser diodes in LiDAR are among the potential heat sources in these devices. Temperature monitoring can help in all these instances to improve safety, performance, and reliability.

Conventional PTC thermistor approaches

Monitoring temperature is the critical first step in thermal protection. Once an overtemperature condition has been identified, remedial actions can be taken. PTC thermistors are often used to monitor temperatures on a pc board. A PTC thermistor experiences an increase in electrical resistivity as its temperature rises. PTC thermistor designs are optimized for specific functions like overcurrent, short-circuit protection, and temperature monitoring. Temperature monitoring PTC thermistors are made using semiconductor ceramics with a high temperature coefficient. They have relatively low resistance values at room temperature, but their resistance rises rapidly when they are heated above their Curie temperature.

PTC thermistors can be used individually to monitor a specific device, like a GPU, or several can be used in series to monitor a wider group of devices, like the MOSFETs in a POL. There are several ways to implement temperature monitoring using PTC thermistors. Two common methods are the use of a sensor IC or discrete transistors to monitor the resistance of the PTC thermistors (Figure 1).

Figure 1: Two common temperature monitoring schemes with PTC thermistors involve sensor interface ICs (left) and discrete transistor solutions (right). (Image source: Toshiba)

In both cases, there is a single connection to the host MCU for a chain of PTC thermistors. There are several tradeoffs between these approaches:

Component count: The IC solution uses three components, compared with the six devices needed with the transistor approach
Mounting area: Since it uses fewer components, the IC solution requires less pc board area
Precision: Both approaches are susceptible to changes in the supply voltage, but the transistor approach is also susceptible to changes in transistor characteristics as their temperature rises. Overall, the IC approach can provide better precision
Cost: The transistor approach uses inexpensive devices, which can provide a cost advantage compared to the IC approach

Sensor ICs and Thermoflagger

Multiple temperature sensing ICs can be used instead of PTC thermistors. Temperature sensing ICs measure their die temperature to estimate the temperature of the pc board. The lower the thermal resistance between the pc board and the IC, the better the temperature estimate. When correctly mounted on the pc board, temperature sensing ICs can provide highly accurate measurements. Two limiting factors of using temperature sensing ICs are that it is necessary to place an IC at every point where the temperature needs to be measured, and each IC needs a dedicated connection to the host MCU.

Thermoflagger from Toshiba provides a fourth option. Using Thermoflagger, temperature measuring circuits can be implemented with only one additional component, compared to the use of temperature measuring ICs. Instead of having multiple connections to the host MCU, the Thermoflagger solution requires only a single MCU connection, enabling the use of inexpensive PTC thermistors for simultaneous monitoring of multiple locations (Figure 2).

Figure 2: Temperature sensor IC monitoring typically requires an IC at each potential heat source and a connection to the MCU for each sensor IC (left); a Thermoflagger plus multiple PTC thermistors solution has a single MCU connection (right). (Image source: Toshiba)

More reasons to consider Thermoflagger include:

It occupies less pc board area compared with other solutions
It is unaffected by power supply voltage variations.
It can be used to implement simple redundant temperature monitoring

What does a Thermoflagger solution look like?

Thermoflagger supplies a small constant current to the connected PTC thermistors and monitors their resistance. It can monitor an individual PTC thermistor or a chain of PTC thermistors. At an elevated temperature, depending on the specific PTC thermistor being monitored, the resistance of a PTC thermistor rises rapidly and Thermoflagger detects the increase in resistance. Thermoflaggers with different constant currents, like 1 or 10 microamperes (µA), accommodate a variety of PTC thermistors. With a current consumption of 11.3 μA, Thermoflagger is designed to enable low-power monitoring.

The detection trigger temperature is determined by the specific PTC thermistor used and can be changed by substituting a different one. If an overtemperature occurs, the Thermoflagger detects the increased resistance in the PTC Thermistor and triggers a change in the PTCGOOD output to alert the MCU (Figure 3).

Figure 3: Thermoflagger senses the rise in resistance of a heated PTC thermistor (bottom), compared with the low resistances associated with normal operating temperatures (top). (Image source: Toshiba)

How Thermoflagger works

Thermoflagger is a precision analog IC with an output optimized for connection to a host MCU. The following description of its operation refers to the numbers in Figure 4 below:

Constant current is supplied from the PTCO terminal and converted to voltage using the resistance of one or more connected PTC thermistors. It is the internal constant current source that makes a Thermoflagger solution insensitive to supply-voltage variations, a significant differentiator compared with other temperature monitoring techniques. If a PTC thermistor gets heated and has a substantial increase in resistance, the PTCO voltage increases to the supply voltage (V_DD). The PTCO voltage also rises to V_DD if the PTCO terminal is open.
If the PTCO voltage exceeds the detection voltage, the output of the comparator inverts and sends a ‘Low’ output. The PTCO output accuracy is ±8%.
Thermoflagger ICs are available with two output formats: open-drain and push-pull. Open-drain outputs require a pull-up resistor. No resistor is needed for push-pull outputs.
After the comparator output is inverted, it is latched (assuming the Thermoflagger includes the optional latching function) to prevent the output from changing due to a drop in temperature of the PTC thermistor.
The latch is released by applying a signal to the RESET pin.

Figure 4: A block diagram showing the key functions of Thermoflagger, a precision analog IC with an output optimized for connection to a host MCU. (Image source: Toshiba)

Application considerations

Thermoflagger solutions can be especially useful for monitoring MOSFETs or LDOs in power supply circuits for large ICs like systems-on-chip (SoCs) and for motor drive circuits in industrial and consumer systems. Typical applications include notebook computers (Figure 5), robot vacuum cleaners, white goods, printers, battery-powered hand tools, wearables, and similar devices. Examples of Thermoflagger ICs include:

TCTH021BE with a 10 µA PTCO output current and a non-latching open-drain output
TCTH022BE with a 10 µA PTCO output current and a latching open-drain output
TCTH021AE with a 10 µA PTCO output current and a latching push-pull output

Figure 5: Shown is a typical Thermoflagger implementation in a notebook computer. (Image source: Toshiba)

Like all precision ICs, Thermoflagger has specific system integration considerations, including:

The voltage applied to the PTCO pin should not exceed 1 V
Thermoflagger should be protected from system noise to ensure reliable operation of the internal comparator
The Thermoflagger IC and the PTC thermistors should be spaced far enough apart to prevent heat from being transmitted through the pc board to the Thermoflagger IC
A decoupling capacitor placed between V_DD and GND will help ensure stable operation
All GND pins must be connected to the system ground

Simple redundancy

Some systems can benefit from redundant temperature monitoring. This can be especially true if an expensive IC is being monitored or if a critical function is involved. The simplicity and small solution size of Thermoflagger make it easy to integrate an additional layer of temperature monitoring, resulting in a robust and reliable temperature monitoring system (Figure 6).

Figure 6: Thermoflagger can add a layer or redundancy (right) to a basic temperature monitoring solution based on temperature monitoring ICs (left). (Image source: Toshiba)

Conclusion

To ensure reliable system performance, designers need to monitor excess heat. Several heat monitoring options are available, including temperature sensing ICs and PTC thermistors. A newer option is Toshiba’s Thermoflagger, which provides many advantages including the use of multiple low-cost PTC thermistors, a smaller footprint, lower component count, a single connection to the MCU, immunity to power supply fluctuations, and the option to implement simple redundant temperature monitoring.