LED overheat detection method without sensor

Through semiconductor theory, LED temperature model is established and LED junction temperature is calculated using voltage and current information, which solves the problem of low monitoring accuracy of traditional sensors, realizes high-precision and low-complexity LED temperature detection, and promotes the miniaturization of the car lights.

CN112731094BActive Publication Date: 2024-11-15MAGNETI MARELLI AUTOMOTIVE COMPONENTS WUHU
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011481230.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-12-16
Publication Date
2024-11-15
Estimated Expiration
2040-12-16

AI Technical Summary

Technical Problem

Traditional LED temperature monitoring methods use sensors to have accuracy errors, increase system complexity and cost, and are not conducive to miniaturization and platformization.

Method used

The LED temperature model is established through semiconductor theoretical formulas and laboratory measurement data, and the LED junction temperature is calculated using voltage and current information, which avoids the use of sensors and additional harnesses, and uses moving average processing to remove interference to improve accuracy.

Benefits of technology

It greatly improves the accuracy of LED temperature measurement, reduces error risk and system complexity, reduces cost and EMC interference, and promotes the miniaturization of car lights.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112731094B_ABST
    Figure CN112731094B_ABST
Patent Text Reader

Abstract

The present invention provides a method for realizing LED overheat detection without using a sensor. First, the formula is used to obtain: #imgabs0#. Next, a quadratic term is introduced to further optimize the parameter model: #imgabs1#. The following model is used to fit the LED, #imgabs2#. The above-mentioned multiple LEDs are connected into a system, and the corresponding values ​​of voltage and current at different temperatures are measured. The parameters d, e, and f are obtained through regression analysis. When the model formula is completed, the current I and the voltage drop information used by the current system are substituted into the model formula to obtain the PN junction temperature T corresponding to the LED. The value of the junction temperature T obtained is the actual application value of the LED. Compared with the junction temperature value obtained by the sensor in the prior art and the LED, there must be errors. The accuracy of the value of the LED junction temperature T is greatly improved by the present invention, and the error risk is reduced without additional work hours. It is simple and accurate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of automobile lighting, and in particular to a method for detecting LED overheating without using a sensor. Background Art

[0002] With the popularization of LED (light-emitting diode) technology, the application of LED in the field of automotive lighting is becoming more and more popular. Since LED is essentially a heat-sensitive semiconductor device, it has poor tolerance to high temperatures. Considering the space and volume limitations in the car. Generally speaking, the heat dissipation of automotive LED lighting faces many challenges, which puts forward many requirements for the temperature control of LED. When the LED temperature is too high due to long-term use of the light, especially when parking, the general driver will adjust the LED current according to the LED temperature or turn on the fan to enhance the heat dissipation to ensure that the semiconductor junction temperature of the LED does not exceed the safe range. The most important point is LED temperature monitoring.

[0003] Traditional LED temperature monitoring is achieved by placing a temperature sensor near the LED. Generally, an NTC (negative temperature coefficient) resistor is placed near the LED, and the possible temperature of the LED is inferred by detecting the NTC resistance value. Of course, sometimes a cheaper PTC (positive temperature coefficient) or more accurate dedicated temperature detection sensor may be used.

[0004] Dedicated external sensors will bring some problems. First, the sensor needs to be placed as close to the LED as possible, which will affect the optical layout, which is not conducive to the miniaturization of lamps. Secondly, the sensor needs additional wiring harness support, which not only increases the cost, but also increases the complexity and weight of the lamp. The additional wiring harness also increases the interference coupling path of EMC, making EMC testing more complicated. In addition, the sensor is not an LED. No matter how the layout is arranged, it can only be placed infinitely close to the LED. However, there is always an error value between the LED junction temperature and the temperature sensor. The LED junction temperature cannot be accurately measured, that is, the accuracy of LED temperature detection using the sensor will inevitably be affected. In order to reduce this problem in the prior art, it is often necessary to design a special design and calibrate the position of the sensor multiple times to try until a small error and a better result is obtained, but this still has a certain error risk and will also require additional work hours. In addition, the accuracy of temperature detection using sensors depends not only on the accuracy of the sensor placement position, but also on the accuracy of the sensor itself. Different designs will lead to different performances in different projects, which is not conducive to platformization and saving of construction.

[0005] In view of the above problems, a solution is needed to detect the temperature of the LED by utilizing the temperature characteristics of the LED's own semiconductor barrier. Therefore, it is very necessary to solve this problem. Summary of the invention

[0006] To solve the above problems, the present invention provides a method for detecting LED overheating without using a sensor. The value of the junction temperature T obtained is the actual application value of the LED. Compared with the junction temperature value obtained by the sensor in the prior art and the inevitable error between the junction temperature value and the LED, the present invention greatly improves the accuracy of the value of the LED junction temperature T, reduces the risk of error and does not require additional work hours. It is simple and accurate.

[0007] The object of the present invention is to provide a method for detecting LED overheating without using a sensor, comprising the following steps:

[0008] First, we can get from the original semiconductor theory formula:

[0009]

[0010] Where: U: PN junction voltage, I: current, T: junction temperature, k: Boltzmann constant C: parameters related to chip manufacturing, Eg: bandgap width, q: electron charge U: PN junction voltage, I: current, T: junction temperature, k: Boltzmann constant C: parameters related to chip manufacturing, Eg: bandgap width, q: electron charge,

[0011] Let a=k / q, b=1 / C, c=kEg / q,

[0012] but ;

[0013] Next, considering that various metal and semiconductor resistors in actual applications will contribute to a part of the voltage drop, that is, the PN node voltage U, especially LED welding and wiring harness connection will produce additional voltage differences under different currents, combined with actual test experience data to consider the relevant impact and introduce quadratic terms to further optimize the parameter model:

[0014] ;

[0015] The total voltage drop, that is, the PN node voltage U model is , where n is the number of LED particles,

[0016] The final model is used to fit the LED.

[0017]

[0018] Among them, the parameters a, b, and c are obtained by the laboratory through regression analysis by measuring the corresponding values ​​of voltage and current of each single LED at different temperatures. In practical applications, for each specific LED, the coefficients a, b, and c are fixed values.

[0019] Connect the above multiple LEDs into a system, measure the corresponding values ​​of voltage and current at different temperatures, and obtain the parameters d, e, f through regression analysis.

[0020] When the model formula is completed, the current I and the voltage drop, i.e., the PN junction voltage U, used by the current system are substituted into the model formula to obtain the PN junction temperature T corresponding to the LED.

[0021] A further improvement is that for LEDs that are directly mounted on the heat sink or used directly and not mounted on the PCB, this method does not require a sensor, and can solve the problem of having no space to place a sensor when the LED is directly soldered to the heat sink.

[0022] A further improvement is that for a light distribution system where there is no space to place a temperature sensor in the surrounding space, this method does not require a sensor, thereby ensuring heat dissipation and reducing the size of the optical system, which is conducive to the trend of miniaturization of vehicle lights.

[0023] Further improvement is: after obtaining the model, the output voltage of the LED constant current power supply is detected by using the detection circuit or the output voltage detection of the headlight driver, the current I value and the voltage U value of the actual electronic system are obtained, and then substituted into the above model , and obtain the value of the LED junction temperature T.

[0024] Further improvements are: firstly, the LED voltage signal, LED current signal and driver ambient temperature data are obtained;

[0025] The LED voltage signal, LED current signal and driver ambient temperature data are processed by moving average to remove interference and impurities and optimize data values;

[0026] Then, the LED voltage signal data after moving average processing and the LED current signal data after moving average processing are substituted into the function of the model formula to obtain the value of the junction temperature T;

[0027] Combined with the driver ambient temperature data after moving average processing, a further corrected T value is obtained.

[0028] Beneficial effects of the present invention: The value of the junction temperature T obtained by the present invention is the actual application value of the LED. Compared with the junction temperature value obtained by the sensor in the prior art and the inevitable error between the junction temperature value and the LED, the accuracy of the value of the LED junction temperature T is greatly improved by the present invention, which reduces the risk of error and does not require additional work hours. It is simple and accurate.

[0029] No need to set up additional temperature sensors and additional wiring harnesses, effectively reducing system complexity and costs;

[0030] Saving extra wiring harness effectively reduces EMC risk;

[0031] No need for sensors means that the position setting and optical layout of the LED itself no longer need to consider the influence of sensors, and the spatial flexibility is improved;

[0032] For special situations where the LED needs to be directly installed on the heat sink or there is no space for the light distribution system to place the temperature sensor, the solution of the present invention can ensure heat dissipation and reduce the volume of the optical system, which is beneficial to the trend of miniaturization of vehicle lights. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flow chart of a specific control detection method according to a specific implementation mode of the present invention. DETAILED DESCRIPTION

[0034] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0035] This embodiment provides a method for detecting LED overheating without using a sensor, including the following steps:

[0036] First, we can get from the original semiconductor theory formula:

[0037]

[0038] Where: U: PN junction voltage, I: current, T: junction temperature, k: Boltzmann constant C: parameters related to chip manufacturing, Eg: bandgap width, q: electron charge U: PN junction voltage, I: current, T: junction temperature, k: Boltzmann constant C: parameters related to chip manufacturing, Eg: bandgap width, q: electron charge,

[0039] Let a=k / q, b=1 / C, c=kEg / q,

[0040] but ;

[0041] Next, considering that various metal and semiconductor resistors in actual applications will contribute to a part of the voltage drop, that is, the PN node voltage U, especially LED welding and wiring harness connection will produce additional voltage differences under different currents, combined with actual test experience data to consider the relevant impact and introduce quadratic terms to further optimize the parameter model:

[0042] ;

[0043] The total voltage drop, that is, the PN node voltage U, is modeled as , where n is the number of LED particles,

[0044] The final model is used to fit the LED.

[0045]

[0046] Among them, the parameters a, b, and c are obtained by the laboratory through regression analysis by measuring the corresponding values ​​of voltage and current of each single LED at different temperatures. In practical applications, for each specific LED, the coefficients a, b, and c are fixed values.

[0047] Connect the above multiple LEDs into a system, measure the corresponding values ​​of voltage and current at different temperatures, and obtain the parameters d, e, f through regression analysis.

[0048] When the model formula is completed, the current I and the voltage drop, i.e., the PN junction voltage U, used by the current system are substituted into the model formula to obtain the PN junction temperature T corresponding to the LED.

[0049] For LEDs that are directly mounted on a heat sink or used directly but not mounted on a PCB, this method does not require a sensor, thereby resolving the problem of inaccurate testing and greatly increased complexity in the prior art solution of placing a temperature measuring board specifically near such LEDs.

[0050] For light distribution systems that have no space to place temperature sensors in the surrounding space, this method does not require sensors, ensures heat dissipation and reduces the size of the optical system, which is conducive to the trend of miniaturization of vehicle lights.

[0051] After the model is obtained, the output voltage of the LED constant current power supply is detected by using the detection circuit or the output voltage detection of the headlight driver itself to obtain the current I value and voltage U value of the actual electronic system, and then substitute them into the above model. , and obtain the value of the LED junction temperature T.

[0052] First, obtain the LED voltage signal, LED current signal and driver ambient temperature data;

[0053] The LED voltage signal, LED current signal and driver ambient temperature data are processed by moving average to remove interference and impurities and optimize data values;

[0054] Then, the LED voltage signal data after moving average processing and the LED current signal data after moving average processing are substituted into the function of the model formula to obtain the value of the junction temperature T;

[0055] Combined with the driver ambient temperature data after moving average processing, a further corrected T value is obtained.

[0056] In application, after obtaining the model formula of this embodiment, it is necessary to further combine the actual electronic system to obtain the current I value and the voltage U value. This implementation method uses a detection circuit to detect the output voltage of the LED constant current power supply. The traditional digital car light driver itself has its own output voltage detection. After some adjustments in actual use, the relevant signal can be directly used to obtain the current I value and the voltage U value, and then substituted into the model of this embodiment above. , and obtain the value of the LED junction temperature T.

[0057] Specific control detection methods such as Figure 1 :

[0058] First, the LED voltage signal, LED current signal and driver ambient temperature data are obtained. In order to remove interference and impurities, the LED voltage signal, LED current signal and driver ambient temperature data are processed by moving average respectively. The continuous sampled data is regarded as a queue with a fixed length of N (N is set to 6 in this embodiment). After a new data sampling, the first data of the above queue is removed, and the remaining N-1 data are moved forward in sequence, and the new sampled data is inserted as the tail of the new queue, and then the arithmetic operation is performed on this queue to obtain the average value. Since this process is mobile and real-time, the processed data can greatly reduce interference and impurities and obtain optimized data values. Then, the LED voltage signal data after moving average processing and the LED current signal data after moving average processing are substituted into the function of the model formula of this embodiment to obtain the value of the junction temperature T. When the above last step is further combined with the driver ambient temperature data after moving average processing, a further corrected T value can be obtained.

Claims

1. A method for detecting LED overheating without using a sensor, characterized in that: The following steps are included: First, the original semiconductor theory formula can be obtained: Where: U: PN junction voltage, I: current, T: junction temperature, k: Boltzmann constant C: parameters related to chip manufacturing, Eg: bandgap width, q: electron charge U: PN junction voltage, I: current, T: junction temperature, k: Boltzmann constant C: parameters related to chip manufacturing, Eg: bandgap width, q: electron charge, Let a=k / q, b=1 / C, c=keg / q, Then U LED (I, T) = aTln(bI) + cT; Next, considering that various metal and semiconductor resistors in actual applications will contribute to a part of the voltage drop, that is, the PN node voltage U, and LED welding and wiring harness connection will produce additional voltage differences under different currents, combined with actual test experience data to consider the relevant impact and introduce quadratic terms to further optimize the parameter model: U wire (I, T)=dIT 2 +eIT+fI; The total voltage drop, that is, the PN node voltage U, is modeled as follows, where n is the number of LED particles. The final LED is fitted using the following model: IN total (I, T)=aTln(bI)+cT+dIT 2 +eIT+fI, Among them, the parameters a, b, and c are obtained by the laboratory through regression analysis by measuring the corresponding values ​​of voltage and current of each single LED at different temperatures. In practical applications, for each specific LED, the coefficients a, b, and c are fixed values. Connect the above multiple LEDs into a system, measure the corresponding values ​​of voltage and current at different temperatures, and obtain the parameters d, e, f through regression analysis. When the model formula is completed, the current I and voltage drop, i.e., PN junction voltage U, used by the current system are substituted into the model formula to obtain the PN junction temperature T corresponding to the LED; After the model is obtained, the output voltage of the LED constant current power supply is detected by using the detection circuit or the output voltage detection of the headlight driver is used to obtain the current I value and voltage U value of the actual electronic system, and then substitute them into the above model U total (I, T) = aTln(bI) + cT + dIT 2 +eIT+fI, get the value of LED junction temperature T; First, obtain the LED voltage signal, LED current signal and driver ambient temperature data; The LED voltage signal, LED current signal and driver ambient temperature data are processed by moving average to remove interference and impurities and optimize data values; Then, the LED voltage signal data after moving average processing and the LED current signal data after moving average processing are substituted into the function of the model formula to obtain the value of the junction temperature T; Combined with the driver ambient temperature data after moving average processing, a further corrected T value is obtained.

2. The method for detecting LED overheating without using a sensor as claimed in claim 1, characterized in that: For LEDs that are directly mounted on a heat sink or used directly and not mounted on a PCB, this method does not require a sensor.

3. The method for detecting LED overheating without using a sensor as claimed in claim 1, characterized in that: For light distribution systems where there is no space in the surrounding space to place a temperature sensor, this method does not require a sensor.