In the process of integrated circuit chip operation, power loss is inevitable, and most of these power losses will be converted into heat energy and dissipated. In abnormal conditions such as high ambient temperature and short circuit, the heat inside the chip cannot be dissipated in time, which inevitably causes the chip operating temperature to rise. Excessive operating temperature has a great impact on the chip's operating performance, reliability and safety. Studies have shown that for every 1°C increase in chip temperature, the driving ability of the MOS tube will decrease by about 4%, the connection delay will increase by 5%, and the integrated circuit failure rate will double. Therefore, there must be an over-temperature protection circuit inside the chip to ensure chip safety.

This article will introduce an over-temperature protection circuit that can be implemented using standard CMOS technology. In the circuit design, a current source proportional to temperature (PTAT current) and a PNP tube (parasitic in CMOS technology) junction voltage with a negative temperature coefficient are used as two differential temperature sensing units. This differential sensing method can improve the circuit's sensitivity to temperature changes. At the same time, its hysteresis function can effectively avoid damage to the chip caused by thermal oscillation.

1 Analysis of the Architecture Principle

1.1 Analysis of the Working Principle

 Figure 1 shows the principle architecture of this design. Q1 is a parasitic PNP transistor in the NWELL CMOS process. Its collector must be connected to the ground. In order to utilize the negative temperature characteristics of the parasitic PN junction forward conduction voltage, Q1 is connected as a diode (the base and collector are also connected to the ground). In this way, the voltage VA between point A and the ground has the property of the PN junction forward voltage being inversely proportional to the temperature. Since the bias voltage output by the reference circuit is added to the gates of M0, M1, and M5, PTAT current (Proportional to Abso-lute Temperature) will be generated on the branches where they are located; while providing the bias, a voltage VB proportional to the temperature is also generated on the resistor R0, that is, the voltage at point B increases accordingly. When a certain temperature TH (the set shutdown temperature) is reached, VH ≥ VA, the comparator Comp outputs a high level, and after passing through the inverter INV, the output TSD is a low level; after this signal acts on other modules of the circuit, the entire chip stops working and realizes the thermal protection function. At the same time, the TSD signal positively feedbacks the gate of M2, turns on M2, increases the current on resistor R0, and makes VB higher.


After the chip is thermally protected and stops working, the temperature on the chip will drop from TH, causing the voltage VA at point A to rise slowly and the voltage VB at point B to drop slowly. Since the positive feedback of TSD has already increased VB, in this state, in order for VA≥VB to occur and the comparator output to flip, point A needs to have a voltage greater than the previous voltage VA, and accordingly, the recovery temperature TL corresponding to the flip point when it drops will also be lower than TH. When the temperature is lower than TL, VB≥VA, and after the comparator acts, TSD will output a high level and the chip will resume working. At the same time, the TSD signal will still act on the M2 gate through positive feedback again, turning off M2, further reducing the current on the resistor R0 and making VB lower.

During the entire working process, the positive feedback of TSD plays a hysteresis role. When working normally, TSD outputs a high level to act on other modules of the circuit. When the temperature is too high, TSD outputs a low level to act on other modules of the circuit, causing the chip to stop working and protect the chip.

1.2 Calculation of temperature flip point

The voltage VA at point A is the voltage on both sides of the PN junction, and the expression of the PN junction voltage can be written as

　
Where VG is the bandgap voltage, γ and α are device parameters, and A represents the constant that is independent of temperature in the process of deducing those equations. Because Vt=kt/q, it can be obtained that VBE decreases as the temperature rises. The equation for temperature change is


If we ignore the effect of the lnT term change caused by temperature change on equation (2), dVBE/dT can be approximately equal to the constant C0. At room temperature, C0 is approximately -2 mV/K. To simplify the calculation, the equation for the change of the PN junction with respect to temperature can be approximately linear as follows:


The voltage at point B is the voltage on the resistor, which can be calculated by Ohm's theorem. Calculate the temperature rise flip point TH. When the temperature rises, from the previous analysis, we know that TSD is high and M2 is cut off. Therefore, the current flowing into R0 is only I1. The expression of VB voltage in this state is

The corresponding equation at the flip point (VA=VB) is

　
Calculate the temperature drop turning point TL When the temperature drops from higher than TH, it can be seen from the previous analysis that TSD is low and M2 is turned on. Therefore, the current flowing into the resistor R has two paths, I1 and I2. The expression of VB voltage in this state is

The corresponding equation at the flip point (VA=VB) is:


Formula (10) is the hysteresis when the over-temperature protection is working.

2 Actual Circuit Design

The designed over-temperature protection circuit is generally divided into three parts, as shown in Figure 2.


Startup circuit: The startup circuit will only work when the power is just turned on. When the power supply voltage slowly increases from 0 V and the input signal Shut is at a low level, the switch tube M37 will be turned on and the MOS tube M38 will also be turned on, which will cause the voltage on the M38 branch to slowly increase. The gate end of M40 connected in the form of a diode will also increase accordingly, and also provide a gate voltage for M45, M43, and M42 on the right end, thereby destroying the balance of the reference circuit and enabling it to start. When the gate voltage of M40 rises to the threshold voltage of M44, it will be turned on, so that a path from power to ground is generated, thus completing the function of the startup circuit.

Bias circuit: The working principle of this part of the circuit is basically the same as the principle of generating the PTAT current circuit, and the voltage at the output end provides bias for other circuits. The branch composed of M36, M45, and Q4 will take out the current from the right branch, filter, amplify, and then mirror it back, and output a bias voltage signal on the branch of M34, M42, and Q1. Due to the working principle of the current mirror, M34~M36 and M42~M45 are required to be the same pair of tubes. In addition, in this circuit, M33 is used as a shutdown tube. When the Shut signal is high, M33 will be in the on state, so that M34 and M35 will be turned off, so that the entire circuit will be turned off.
Thermal shutdown protection circuit: Since the bias voltage output by the reference circuit is added to the gates of M39, M51, and M52, PTAT current will be generated in these two branches. A two-stage comparator composed of M41, M47, M48, M49, and M50 is used to implement the function of Comp in the principle equivalent diagram. The first stage of this comparator is a PMOS differential input.

The role of using PMOS as a differential input is: (1) Reduced input noise. Under normal circumstances, the temperature cannot change drastically, so the frequency of temperature fluctuations cannot be very high, so the flicker noise 1/f will become the main component of the noise. Since the PMOS input can reduce the impact of noise on the circuit; (2) The PMOS input reduces the lower limit of the common-mode input range. The comparator of this circuit compares the voltage close to the PN junction VBE. The input terminal composed of PMOS can better meet the requirements of this low common-mode input voltage. The second stage of the comparator is a common source amplifier. As the second stage of the comparator, its main function is to increase the output swing, improve the gain and input resolution, and speed up the conversion rate of high and low levels. The role of capacitor C0 in the circuit: capacitor C0 can suppress the voltage fluctuations generated by some interference on the resistor wind leading to the input terminal of the comparator to prevent the system from being disturbed.

3 Simulation test

According to the circuit designed above. It is simulated by Cadence Specture, and the model parameters of the device adopt 0.35μm CMOS process. Figure 3 is the output control signal TSD of the over-temperature protection circuit, and the curve of the temperature rise and fall. The power supply voltage is 3.3 V. It can be seen from the simulation results that the circuit has achieved good "temperature hysteresis" characteristics. The hysteresis function effectively avoids the problem of thermal oscillation of the chip. The shutdown temperature TH160° and the recovery temperature TL140°.


4 Conclusion The overheat protection circuit designed in

the language uses PTAT current to detect temperature changes and converts it into a voltage signal to input into a two-stage comparator for comparison, thereby generating an overheat protection signal. The hysteresis effect of the comparator can effectively prevent the occurrence of thermal oscillation. The circuit has high sensitivity to temperature sensing and is very suitable for integration into integrated circuit chips.