CN104535913A - Heat testing method and system for LED assembly with built-in temperature detection function - Google Patents

Abstract

The invention discloses a heat testing method and system for an LED assembly with the built-in temperature detection function. According to the heat testing method and system, when the LED multi-chip assembly is designed, a proper layout is found by building a heat balance equation of nodes and optimizing the positions of chips on the assembly with an improved genetic method so that the highest temperature of the assembly can be lowered; meanwhile, a plurality of heat sensors are designed in the assembly and arranged in high-transient-temperature areas in the assembly respectively so that the temperature change conditions inside the assembly can be detected; a driving circuit, connected with the heat sensors and provided with the heat control function, of the LED assembly is designed, when the temperature of the assembly exceeds a given maximum value, power supply of the assembly is adjusted, and the temperature of the assembly can be lowered. By means of the heat testing method and system, the supplied current can be self-adaptively adjusted according to the temperature condition of the assembly, the temperature of the assembly is maintained within a normal range, and therefore serious influences, caused by the excessively high temperature, on LEDs are avoided.

Description

There is heat testing method and the test macro of the LED component of built-in temperature detection

Technical field

The invention belongs to the field tests of LED chip, particularly a kind of heat testing method and test macro with the LED component of built-in temperature detection.

Background technology

Light emitting diode (Light Emitting Diode, LED) is the extremely competitive novel energy-conserving solid state light emitter of one.LED from birth so far, has achieved true color and high brightness, and creates white light LEDs on the basis of blue led and purple LED, thus achieves the once leap that the mankind throw light in history.Compare with fluorescent light with incandescent lamp, LED has more characteristic and advantage, has efficient, energy-conservation, and environmental protection, the advantage such as the life-span is long, volume is little, be used widely in fields such as backlight, automotive lighting, Special Work illumination, general illuminations.LED is the high-quality light source meeting energy-conserving and environment-protective in current known light source most, along with further developing of LED, may challenge the leading position of incandescent lamp, fluorescent light, Halogen lamp LED etc.

The photon release of LED comes from the transition of electronics at energy interband.For LED, the electric energy inputted approximately only has 10% to 30% to be converted into luminous energy, and remaining energy is then converted into heat energy, therefore will produce higher heat flow density on a very little LED chip area, such as, produce up to 10 ⁶w/m ²heat flow density.If heat can not be dispersed in environment, the junction temperature of LED chip will raise.Along with the rising of junction temperature, the light output of chip will constantly reduce.

Particularly in recent years, LED is the same with other electron devices, constantly to miniaturization and high-power future development, high heat flux becomes inevitable, if can not dispel the heat in time, then device temperature is too high, the overall performance of LED will be had a strong impact on, luminescence efficiency and serviceable life are reduced, even can cause the damage to the linkage interface of chip and mechanical stress etc., thus the inner structure of chip is produced destroy.Therefore, not only need the output performance considering light when designing LED, and need junction temperature and the heat radiation of considering chip.The reliability of LED and performance thereof depend on whether have good thermal design and whether take good cooling measure to a large extent.

To the use of LED, in general illumination, people need high-power LED light source particularly LED white light source.The implementation method of large power white light LED mainly contains the following two kinds: one is the direct single high-power LED chip of encapsulation, such as commercially existing 1W, 3W even large power white light LED of 5W; Two is form more high-power LED multi-chip module by encapsulating multiple low-power LED.Therefore, a LED multi-chip module is made up of multiple LED chip, and it is that two or more LED chip is connected on a common circuit substrate, and realizes the connection of each chip chamber.Here to LED chip wherein, by adopting different connection in series-parallel combinations, various different rated voltage and electric current can be realized, improve Integral luminous usefulness, reduce costs.In practice in order to be reduced at the complicacy in process, often suppose that the power of each LED chip comprised in a LED multi-chip module is identical.The present invention is directed the LED multi-chip module of general structure, the LED chip that it has different capacity by some formed, and the power of each LED chip wherein can be identical, also can be different.For for purpose of brevity, below by LED multi-chip module referred to as LED component.

To single LED chip, along with the increase of working current can produce certain heat, to the change of LED chip junction temperature be caused, and the performance of LED is had an impact, such as, cause the efficiency step-down etc. of forward voltage drop change, colour temperature change, wavelength shift, opto-electronic conversion.Similarly, to the LED component be made up of multiple LED chip, along with the increase of working current will produce a large amount of heats, and performance such as luminous Strong degree and the luminescence efficiency etc. of LED component are had an impact.

Therefore, a kind of heat testing method and test macro of effective LED component is needed badly.

Summary of the invention

Primary and foremost purpose of the present invention is that the shortcoming overcoming prior art is with not enough, provides a kind of heat testing method with the LED component of built-in temperature detection.Namely the LED component that the present invention is directed to is LED multi-chip module.

Another object of the present invention is to provide described and realize the above-mentioned test macro with the heat testing method of the LED component of built-in temperature detection.

Object of the present invention is achieved through the following technical solutions: a kind of heat testing method with the LED component of built-in temperature detection, as shown in Figure 1, comprises the steps:

(1) layout optimization of LED component: to a kind of LED component of given initial configuration, by setting up the thermal balance equation of each node, and the position using a kind of Revised genetic algorithum residing in LED component to each chip is optimized, find out suitable layout, to reduce the maximum temperature of LED component;

(2) design of built-in thermal sensor: calculate the LED component after step (1) layout optimization, calculates the transient temperature of each node location of LED component, obtains the node that transient temperature is higher; In LED component, design several thermal sensors, and they are placed on the node region that in LED component, transient temperature is higher respectively, to detect the temperature variations of LED component inside;

(3) Automatic adjusument of LED component temperature: design has the driving circuit of the LED component of heat control function, wherein, is connected by the thermal sensor in step (2) with driving circuit; The LED component temperature that thermal sensor monitors sends driving circuit to, when the temperature of LED component has exceeded given maximal value, driving circuit just regulates supply electric current to LED component, namely supply electric current is reduced, the junction temperature of each chip is reduced, thus the temperature of whole LED component is reduced.

The thermal balance equation of described each node is specific as follows: for internal node, and thermal balance equation is such as formula shown in (A); For straight border node, thermal balance equation is such as formula shown in (B); To the turning boundary node with 90 °, thermal balance equation is such as formula shown in (C);

$\frac{U_{i+1,j}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+\frac{U_{i-1,j}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+\frac{U_{i,j+1}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+\frac{U_{i,j-1}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+V_{i,j}+(U_0-U_{i,j})\·{\mathrm{\θ}}_0\·\mathrm{\η}=0---\left(A\right)$

$\frac{U_{i-1,j}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+\frac{U_{i,j+1}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+\frac{U_{i,j-1}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+V_{i,j}+(U_0-U_{i,j})\·{\mathrm{\θ}}_0\·\mathrm{\η}+(U_1-U_{i,j})\·{\mathrm{\θ}}_1\·\mathrm{\η}=0---\left(B\right)$

$\frac{U_{i-1,j}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+\frac{U_{i,j+1}-U_{i,j}}{\mathrm{\τ}\·\mathrm{\ξ}}+V_{i,j}+(U_0-U_{i,j})\·{\mathrm{\θ}}_0\·\mathrm{\η}+2\·(U_1-U_{i,j})\·{\mathrm{\θ}}_1\·\mathrm{\η}=0---\left(C\right)$

Wherein, i represents the row in LED component, and j represents the row in LED component; U represents temperature, as U _{i+1, j}represent the temperature of node (i+1, j), by that analogy; τ=1/ ρ, ρ are the heat transfer coefficients between node; ξ is the ratio of the sizing grid size of transverse axis and y direction, i.e. ξ=ε/η, ε is the sizing grid size of transverse axis, and η is the sizing grid size of the longitudinal axis; (τ ξ) is the thermal conduction resistance between node (i+1, j) and node (i, j); U ₀for the air themperature in the external world; θ ₀for perpendicular to the coefficient of heat transfer between the air of orientation substrate and node (i, j); U ₁for the environment temperature of boundary direction; θ ₁for the coefficient of heat transfer between the air of boundary direction and node; V _i,jfor the heat that node (i, j) produces within the unit interval;

The interlace operation of the improved adaptive GA-IAGA described in step (1) can select the one in three kinds of crossover locations such as a point of crossing, two point of crossing and three point of crossing, have employed respective choosing with adjustment mode to produce new individuality individual component value to each crossover location simultaneously;

The specific implementation step of the improved adaptive GA-IAGA described in step (1) is as follows:

With 1 to n, each node is numbered, represents with body one by one the position (i.e. node) that each chip is residing on assembly; To a given individual z, use corresponding thermal balance equation group and solve, draw the temperature value at each node location place in the steady state, thus the maximum temperature values obtaining all nodes (being designated as U _max); Being defined as just when H (z) of individual z:

$H\left(z\right)=\frac{30}{1+U_{\max }}$

Form population by multiple individuality, by the evolution of population, obtain preferably individual, more excellent individuality here refers to a kind of placement scheme with the LED component chips of less maximum temperature;

Represent the evolutionary generation of population with parameter k, represent that kth is for population with B (k), represent the number of the individuality in population B (k) and the scale of population with N, use X _irepresent i-th individuality in population B (k), X _i(i=1,2 ..., N);

Step 1: the value of putting parameter k is 0, i.e. k=0;

Step 2: generate initial population B (0), method is: random generation individuality, and they are placed in initial population; Calculate each the individual X in initial population B (0) _ijust when H (X _i), i=1,2 ..., N;

Step 3: perform and select operation: to the individual X in current population B (k) _iand X _j, calculate the number that the value of respective components in their coding is not identical, use w here _ijrepresent; Computing function C (w _ij) value: if w _ij> σ, then C (w _ij)=0; If 0<w _ij≤ σ, then C (w _ij)=1-w _ij/ σ, σ is the normal number of a setting here.Be calculated as follows function L (X _i) value:

$L\left(X_i\right)=H\left(X_i\right)/\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}C\left(w_{\mathrm{ij}}\right)$

L (X _i) value as individual X _inew for H (X _i), by L (X _i) value be assigned to H (X _i);

Calculate each individual X _iselect probability R (X _i) as follows:

$R\left(X_i\right)=H\left(X_i\right)/\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}H\left(X_j\right)$

According to each individual X _iselect probability R (X _i), N time is carried out to the individuality in current population B (k) and selects operation, therefrom select individual composition one pairing storehouse Ω;

Step 4: perform interlace operation: carry out random pair to the individuality in pairing storehouse Ω, method is: individual to each in pairing storehouse Ω, selects not identical with it individuality randomly and matches with it from remaining N-1 of pairing storehouse Ω individuality.To obtained N number of pairing, interlace operation is carried out to each pairing and generates a new individual Y _i, method is as follows: use X _uand X _vrepresent two individualities in a pairing; One of first the scope of the number F of stochastic generation point of crossing, number F is between 1 to 3, i.e. F=1,2 and 3; Number F according to generated point of crossing sets crossover location;

To F=1, i.e. the situation of a point of crossing, then between 1 to N stochastic generation round values as crossover location; Then the individual component value at crossover location of exchange pairing, and the value of other one-component is adjusted;

To F=2, i.e. the situation of two point of crossing, then the round values that stochastic generation two is different between 1 to N is as crossover location; Then first crossover location of exchange pairing individuality and the value of second crossover location, and the value of other two components is adjusted;

To F=3, i.e. the situation of three point of crossing, then the round values that stochastic generation three is different between 1 to N is as crossover location; Then the individual X of exchange pairing _uand X _vthe value on first crossover location, second crossover location and the 3rd crossover location, and the value of other three components to be adjusted;

By as upper type the N number of new individual Y that generates _i(i=1,2 ..., N) and replace original N number of individuality in population B (k);

Step 5: carry out mutation operation: stochastic generation integer γ between 1 to N; γ individuality is selected randomly from current population B (k); Individual to each in selected this γ individuality, select one-component randomly and make a variation to the value of this component, concrete grammar is that the value of the component that the value of this component and another are selected at random exchanges; Former γ the individuality that γ generated in this way new individual replacement is selected from population B (k);

Step 6: calculate each the individual X in current population B (k) _ijust when H (X _i), i=1,2 ..., N;

Step 7: judge whether end condition meets, if do not meet, then put k:=k+1, forward step 3 to; Otherwise forward step 8 to;

Step 8: export in current population B (k) just when being that maximum individuality.

The number of the thermal sensor described in step (2) sets according to the number of LED chip in LED component;

The circuit structure of the thermal sensor described in step (2) is as follows: comprise a bipolar transistor, operational amplifier, analog to digital converter, a resistance etc.; Wherein, bipolar transistor, resistance, operational amplifier are connected successively with analog to digital converter, namely the base signal line of bipolar transistor is connected with an input end of operational amplifier after being connected with a resistance, the output terminal of operational amplifier is connected to the input end of an analog to digital converter, and the output of this analog to digital converter is the value with temperature correlation; Because the signal in the base stage of transistor is very faint, therefore use operational amplifier that this signal is amplified, spread out of again afterwards;

Driving circuit described in step (3) adopts pulse width modulation to design, and it is ON time by changing power switch pipe or changes dutycycle closing time, and makes output voltage reach stable by the adjustment of dutycycle.Concrete performing step is: to primary voltage, after a kind of effect of bleeder circuit, delivers to a kind of end of oppisite phase of error amplifier, to be compared and amplify by error amplifier to its difference, produces error signal; Compared by a kind of triangular wave that error signal and a kind of oscillator produce by pwm comparator, a kind of square wave of final generation variable duty ratio is to control conducting and the cut-off of power tube again.

The structure of described driving circuit is as follows: comprise reference voltage module, light-adjusting module, constant current source module, storer, control module; Control module is connected with reference voltage module, light-adjusting module, constant current source module and storer etc. respectively, and constant current source module is connected with storer;

The whole driving circuit that described reference voltage module is LED component provides voltage reference value accurately, the accuracy of this reference voltage can have influence on the accuracy of driving circuit institute output current, simultaneously the performance characteristic of the temperature value of the whole LED component also relevant temperature of side light LED component when this reference voltage.Described reference voltage module is preferably the X60008 voltage source that Xicor company produces, this voltage source can generate 10 continuously adjustable accurate voltage-references, its temperature coefficient is 1ppm/ DEG C, absolute initialization precision is ± 5mV, supply electric current is 800nA, there is current source and the inverse current of 10mA, and there is the sequential short circuit electric current of 80mA.

Described light-adjusting module comprises the oscillator unit, sawtooth wave generation unit, comparator unit and the Date Conversion Unit that connect successively; This module realizes the change to the forward current in LED component.Here, the square-wave signal produced by oscillator is fed to sawtooth wave generation unit; Sawtooth wave generation unit with the charge and discharge of square-wave signal, square wave is converted into sawtooth wave by constantly control capacitance, and its is sent into the negative-phase input of comparator unit; Comparator unit amplifies the faint difference of input end signal, makes the output of light-adjusting module can produce corresponding pulse width modulating signal according to the difference of dim signal; Date Conversion Unit is that the signal from comparator unit is carried out shaping and conversion process, sends the control module of driving circuit afterwards to.

Described constant current source module is used for driving the constant current of LED component, comprises the current source circuit, signal conditioning circuit and the feedback control circuit that connect successively; Wherein current source circuit is a kind of D/A converting circuit, and it is the voltage for generation of certain numerical value under the effect of control module; Signal conditioning circuit is for amplifying the value of this voltage; The working current of feedback control circuit required for LED, by the size using signal conditioning circuit to adjust the magnitude of voltage that current source circuit exports adaptively.

Described memory module is used for storing temperature value, test data etc. that thermal sensor gathers.Described memory module is SDRAM storer, is preferably the HY57V641620 chip of Hynix company.The memory capacity of this chip HY57V641620 is 4 groups × 16M position (8M byte), and operating voltage is 3.3V, and has 16 bit data width.

Described control module is used for the courses of work such as supply to the collection of temperature, drive current, light modulation or toning and controls.Described control module is preferably single-chip microcomputer; Be more preferably the single-chip microcomputer that model is PIC16F887; Harvard's bus structure that this single-chip microcomputer adopts data bus to be separated with instruction bus; frequency of operation can reach 20MHz; instruction cycle is 200ns; there is the analog to digital converter on 10, eight tunnel; sampling period is 20 μ s, and has the anti-interference functions such as electrification reset, low-voltage reset, house dog protection.

Described thermal sensor is connected with the control module of described driving circuit by the output terminal of its operational amplifier.

Realize the test macro with the heat testing method of the LED component of built-in temperature detection, comprise the driving circuit of thermal sensor and LED component; Thermal sensor is connected with the driving circuit of LED component.

Realize the test macro with the heat testing method of the LED component of built-in temperature detection, also comprise: determine the module of each node temperature of LED component and optimize the module of LED component layout.Determine that the module of each node temperature of LED component is for determining the temperature of each node after the temperature of each node in LED component initial configuration and LED component structure optimization; After the temperature value of node each in LED component initial configuration being input to the module optimizing LED component layout, can be optimized LED component layout; The temperature of each node after LED component structure optimization can determine the placement location of thermal sensor;

The module of the described each node temperature of determination LED component, be for by thermal balance equation to each node calculate temperature value in LED component.

The module of described optimization LED component layout is for the placement scheme by using a kind of Revised genetic algorithum to obtain the LED component chips with less maximum temperature.

Principle of the present invention:

The thermal characteristic of LED mainly comprise the generation of heat, the conduction of heat, convection of heat and heat the aspect such as to disperse.The LED thermal parameters that usual people pay close attention to mainly contains junction temperature, thermal resistance and case temperature etc.

Thermal resistance is defined as along the ratio of the temperature difference on device heat passage with the heat-dissipating power on passage.For single led, usually get the temperature difference between certain reference point on chip PN junction and heat passage.Following mode is adopted to the calculating of thermal resistance: R=(T _j-T _x)/P _h.Here R represents thermal resistance, T _jthe junction temperature of the device under test of test condition when stablizing, T _xthe reference temperature of designated environment, P _hit is the dissipated power of device under test.

To junction temperature T _jcalculate by such as under type: T _j=T _j0+ △ T _j, △ T _j=μ × △ T _e.Here T _j0the initial junction temperature before device under test does not apply heating power, △ T _jthe variations injunction temperature amount owing to being applied with caused by heating power, △ T _ebe the variable quantity of temperature sensitive parameter value, μ represents △ T _jwith △ T _ebetween the constant of relation.

To single led, when it being placed to the chip substrate enterprising enforcement used time, the structure adopted comprises following multilayer from the inside to surface: PN junction, metab, chip substrate, heating radiator etc.Therefore, the path representation that single led heat can be transmitted is: PN junction-metab-chip substrate-heating radiator-environment.Thermal resistance overall for LED can be regarded as the result of the thermal resistance series connection of each layer, when not having heating radiator, the entire thermal resistance from PN junction to environment is each layer thermal resistance sum, and it can be expressed as: R _t=R _p+ R _b+ R _s, wherein R _trepresent the entire thermal resistance of whole conduction process, R _prepresent the thermal resistance between metab and PN junction, R _brepresent the thermal resistance between metab and chip substrate, R _srepresent the thermal resistance between chip substrate and environment.

To the LED component be made up of multiple LED chip, owing to there is multiple thermal source (each LED chip is a thermal source) in assembly here, the specification of each thermal source is also not quite similar, and also there is the situation of heat trnasfer between each thermal source, therefore complicated than single led to the thermal characteristics of LED component.When calculating the thermal resistance of LED component, be the circuit structure diagram namely connected according to the connected mode of each LED chip in assembly, and the thermal resistance of each LED chip, adopt and ask the calculation procedure of the all-in resistance of resistor network to carry out.Such as, if in LED component, the specification of each chip is consistent, and they encapsulate on one substrate in parallel, if do not consider the effect of the heat trnasfer between chip, now just can simplify the computation process of the thermal resistance of LED component, and obtain the thermal resistance R of LED component _zfor:

$\frac{1}{R_z}=\frac{1}{R_{p1}}+\frac{1}{R_{p2}}+\mathrm{\&Lambda;}+\frac{1}{R_{\mathrm{pn}}}$

Wherein, n is the number of the LED chip in assembly; R _p1, R _p2..., R _pnrepresent each chip thermal resistance separately.

Usually, assuming that by the chip mount in LED component on the substrate of one piece of rectangular area.Substrate divides impartial grid, and has supposed that a chip can be placed in the position of each net point or node (i, j).Here i=1,2 ..., s; J=1,2 ..., t.Table 1 be s=5 and t=5 time the situation of node, now on substrate, 25 chips can be placed at most.In assembly, each LED chip can produce certain heat, and a part for these heat passes to contiguous unit, and another part is taken away by air by convection current.Through the regular hour, produce in heat, between heat transfer and convection current cooling effect, can thermal equilibrium be reached.

Table 1

(1,1)	(1,2)	(1,3)	(1,4)	(1,5)
					(2,1)	(2,2)	(2,3)	(2,4)	(2,5)
(3,1)	(3,2)	(3,3)	(3,4)	(3,5)
					(4,1)	(4,2)	(4,3)	(4,4)	(4,5)
(5,1)	(5,2)	(5,3)	(5,4)	(5,5)

The temperature of each node is solved below with thermal balance equation.Each chip in LED component can reach stable temperature within given a period of time, and each chip, as a micro unit, can derive the thermal balance equation of each node according to micro unit principle of energy balance.To an internal node (i, j), assuming that be a unit length perpendicular to the third dimension direction of substrate, in the unit interval from the heat of node (i+1, j) delivery node (i, j) be then:

$W_{i+1,j}=\frac{U_{i+1,j}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}$

Wherein, U _{i+1, j}and U _i,jthe temperature of node (i+1, j) and node (i, j) respectively; τ=1/ ρ, ρ is the heat transfer coefficient between node here; ξ is the ratio of the sizing grid size of transverse axis and y direction, i.e. ξ=ε/η, and ε is the sizing grid size of transverse axis here, and η is the sizing grid size of the longitudinal axis; (τ ξ) is the thermal conduction resistance between node (i+1, j) and node (i, j).Similarly, can obtain respectively in unit interval interior nodes (i-1, j), (i, j+1), the heat expression of delivery nodes (i, j) such as (i, j-1).

To LED component, be convection heat transfer according to the third dimension direction perpendicular to substrate, then the heat perpendicular to orientation substrate delivery node (i, j) is:

W ₀＝(U ₀-U _i,j)·θ ₀·η

U in above formula ₀for the air themperature in the external world, θ ₀for perpendicular to the coefficient of heat transfer between the air of orientation substrate and node (i, j).Node (i, j) is as a thermal source node, and it can produce certain heat within the unit interval, and this heat is designated as V _i,j.At steady state, can obtain according to principle of energy balance:

W _i+1,j+W _i-1,j+W _i,j+1+W _i,j-1+V _i,j+W ₀＝0

Namely

$\frac{U_{i+1,j}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+\frac{U_{i-1,j}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+\frac{U_{i,j+1}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+\frac{U_{i,j-1}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+V_{i,j}+(U_0-U_{i,j})\&CenterDot;{\mathrm{\&theta;}}_0\&CenterDot;\mathrm{\&eta;}=0$

As above thermal balance equation is for internal node, as follows with the situation of the turning boundary node with 90 ° to the straight border node in reality:

1. to the situation of straight border node, if all nodes of jth row are border, then to node (i, j), have (i-1, j), (i, j+1), (i, j-1) etc. import heat to node (i, j); Do not have (i, j+1) and import heat to node (i, j), but now boundary direction can import heat to node (i, j), this heat is:

W ₁＝(U ₁-U _i,j)·θ ₁·η

U in above formula ₁for the environment temperature of boundary direction, θ ₁for the coefficient of heat transfer between the air of boundary direction and node.Therefore, to straight border node, at steady state, have according to principle of energy balance:

$\frac{U_{i-1,j}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+\frac{U_{i,j+1}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+\frac{U_{i,j-1}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+V_{i,j}+(U_0-U_{i,j})\&CenterDot;{\mathrm{\&theta;}}_0\&CenterDot;\mathrm{\&eta;}+(U_1-U_{i,j})\&CenterDot;{\mathrm{\&theta;}}_1\&CenterDot;\mathrm{\&eta;}=0$

Above the situation that all chip nodes of jth row are border is illustrated, can processes similarly other straight border node.

2. to the situation of turning boundary node with 90 °, if node (i, j) be this boundary node, the all nodes that then now can be used as the i-th row are all nodes of border jth row is simultaneously border, by above to the result of straight border node situation, the thermal balance equation can at steady state with the turning boundary node of 90 ° is:

$\frac{U_{i-1,j}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+\frac{U_{i,j+1}-U_{i,j}}{\mathrm{\&tau;}\&CenterDot;\mathrm{\&xi;}}+V_{i,j}+(U_0-U_{i,j})\&CenterDot;{\mathrm{\&theta;}}_0\&CenterDot;\mathrm{\&eta;}+2\&CenterDot;(U_1-U_{i,j})\&CenterDot;{\mathrm{\&theta;}}_1\&CenterDot;\mathrm{\&eta;}=0$

To the analysis of various micro unit node above use, micro unit thermal balance equation is listed respectively to each LED chip in whole LED component, thus a composition system of linear equations, by solving this system of equations, draw the temperature value at each node location place in the steady state.

To the LED component be made up of n LED chip, when the position (node at place) that these chips are residing is in assembly not identical, then the temperature value at each node location place is also different in the steady state.Use a kind of Revised genetic algorithum to find out each chip correct position residing on assembly below, namely find out suitable layout, and the temperature of assembly is optimized, to reduce the maximum temperature of assembly.The calculation process of this algorithm as shown in Figure 2.

First, the method for encoding to individuality is as follows.With 1 to n, each node is numbered, represents with body one by one the position (node) that each chip is residing on assembly.Such as, body z=(53921764810) represents one by one: chip 5 is the positions being placed on node 1, and chip 3 is the positions being placed on node 2, and chip 9 is the positions being placed on node 3, and the rest may be inferred.

To a given individual z, use corresponding thermal balance equation group and solve, draw the temperature value at each node location place in the steady state, thus the maximum temperature values obtaining all nodes (being designated as U _max).Being defined as just when H (z) of individual z:

$H\left(z\right)=\frac{30}{1+U_{\max }}$

Secondly, form population by multiple individuality, by the evolution of population, obtain preferably individual, more excellent individuality is here a kind of placement scheme with the LED component chips of less maximum temperature.

Represent the evolutionary generation of population with parameter k, represent that kth is for population with B (k), represent the number of the individuality in population B (k) and the scale of population with N, use X _irepresent i-th individuality in population B (k), X _i(i=1,2 ..., N).Here N is the constant of a setting, such as N=40.

Provide a kind of a kind of Revised genetic algorithum (abbreviation algorithm 1) asking more excellent individuality below, its computation process is as follows: a kind of chip layout in LED component is regarded as body one by one, and encodes to this individuality.Simultaneously individual z just when being defined as H (z).

Step 1: the value of putting parameter k is 0, i.e. k=0.

Step 2: generate initial population B (0), method is: the random individuality produced in N number of initial population; Calculate each the individual X in initial population B (0) _ijust when H (X _i), i=1,2 ..., N.

Step 3: perform and select operation.To the individual X in current population B (k) _iand X _j, calculate the number that the value of respective components in their coding is not identical, use w here _ijrepresent.Computing function C (w _ij) value: if w _ij> σ, then C (w _ij)=0; If 0<w _ij≤ σ, then C (w _ij)=1-w _ij/ σ.Here σ is the normal number of a setting, such as 1≤σ≤8.Be calculated as follows function L (X _i) value:

$L\left(X_i\right)=H\left(X_i\right)/\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}C\left(w_{\mathrm{ij}}\right)$

L (X _i) value as individual X _inew for H (X _i), by L (X _i) value be assigned to H (X _i).

Calculate each individual X _iselect probability R (X _i) as follows:

$R\left(X_i\right)=H\left(X_i\right)/\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}H\left(X_j\right)$

According to each individual X _iselect probability R (X _i), N time is carried out to the individuality in current population B (k) and selects operation, therefrom select individual composition one pairing storehouse Ω.

Step 4: perform interlace operation.Carry out random pair to the individuality in pairing storehouse Ω, method is: individual to each in pairing storehouse Ω, selects not identical with it individuality randomly and match with it from remaining N-1 of pairing storehouse Ω individuality.

To obtained N number of pairing, interlace operation is carried out to each pairing and generates a new individual Y _i, method is as follows: use X _uand X _vrepresent two individualities in a pairing; One of first the scope of the number F of stochastic generation point of crossing, number F is between 1 to 3, i.e. F=1,2 and 3; Number F according to generated point of crossing sets crossover location.

To the F=1 i.e. situation of a point of crossing, then between 1 to N stochastic generation round values as crossover location; Then the individual component value at crossover location of exchange pairing, and the value of other one-component is adjusted.Such as, to the individual X of pairing _u=(43921765810) and X _v=(56731024198), if crossover location is 3, then the new individual Y generated _ifor Y _i=(43721965810).

To the F=2 i.e. situation of two point of crossing, then the round values that stochastic generation two is different between 1 to N is as crossover location; Then first crossover location of exchange pairing individuality and the value of second crossover location, and the value of other two components is adjusted.Such as, to the individual X of pairing _u=(67139421058) and X _v=(14310895276), if crossover location is 3 and 7, then the new individual Y generated _ifor Y _i=(67319451028).

To the F=3 i.e. situation of three point of crossing, then the round values that stochastic generation three is different between 1 to N is as crossover location; Then the individual X of exchange pairing _uand X _vthe value on first crossover location, second crossover location and the 3rd crossover location, and the value of other three components to be adjusted.Such as, to the individual X of pairing _u=(13748105692) and X _v=(10283694517), if crossover location is 2,4 and 7, then the new individual Y generated _ifor Y _i=(12738104695).

By as upper type the N number of new individual Y that generates _i(i=1,2 ..., N) and replace original N number of individuality in population B (k).

Step 5: carry out mutation operation.Stochastic generation integer γ between 1 to N; γ individuality is selected randomly from current population B (k); Individual to each in selected this γ individuality, select one-component randomly and make a variation to the value of this component, concrete grammar is that the value of the component that the value of this component and another are selected at random exchanges.Such as, to individual X _u=(92385610417), first, from individual X _uthe one-component of middle Stochastic choice is its 4th component, and the value of this component is 8; Next, then from individual X _uin the one-component selected at random be its 9th component, the value of this component is 1; The new individual Y then generated _ifor Y _i=(92315610487).

Former γ the individuality that γ generated in this way new individual replacement is selected from population B (k).

Step 6: calculate each the individual X in current population B (k) _ijust when H (X _i), i=1,2 ..., N.

Step 7: judge whether end condition meets, if do not meet, then put k:=k+1, forward step 3 to; Otherwise forward step 8 to.

Step 8: export in current population B (k) just when being that maximum individuality.Whole algorithm terminates.

The end condition of algorithm 1 is: the iterations carried out when algorithm is greater than 5000 i.e. k>5000, or the individuality in nearest continuous two generations population B (k-1) and B (k) does not change, and the individuality namely in them is identical.

The more excellent individuality obtained by algorithm 1, just corresponds to a kind of placement scheme with the LED component chips of less maximum temperature.

To LED component, by after the optimization of carrying out chip layout, just design multiple (being set as m here) thermal sensor in assembly, with the temperature variations of detection components inside.To choosing of the value of m, mainly with the number of LED component chips about: if the number of chip less (being not more than 3), then the value choosing m is less, and the value of such as m generally can be taken as 1 to 2; If the number of chip more (being greater than 3), then the value choosing m is just corresponding comparatively large, and the value of such as m generally can be taken as more than 3.

The design of described thermal sensor is the use of to the following characteristic of bipolar transistor: the magnitude of voltage of the base stage of bipolar transistor changes along with the change of environment temperature.The present invention is connected described thermal sensor with the driving circuit of LED component, this driving circuit can control the heat that LED component produces.Due to LED component, the luminosity controlling it its essence is the luminous flux controlling it and export, therefore can control its luminosity by the forward current controlling LED component, the size simultaneously by reducing supply electric current reduces the junction temperature of LED component and the heat of generation.

The present invention has following advantage and effect relative to prior art: heat testing method and the test macro with built-in temperature detection provided by the present invention, the adaptive size of current regulating supply LED can be carried out according to the temperature conditions of LED component, thus can avoid due to temperature too high brought LED overall performance is caused at such as luminescence efficiency and serviceable life etc. have a strong impact on.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the heat testing method of the LED component with built-in temperature detection.

Fig. 2 is the calculation flow chart that a kind of improved adaptive GA-IAGA used in the present invention is optimized LED chip layout.

Fig. 3 is the structural drawing of the driving circuit of LED component of the present invention.

Fig. 4 is the structural drawing of LED component Thermal test system provided by the invention.

Embodiment

Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited thereto.

Embodiment 1

The step with the heat testing method of the LED component of built-in temperature detection provided by the invention is as follows: the layout optimization of (1) LED component: when designing LED component, first the position by using a kind of Revised genetic algorithum residing on assembly to each chip is optimized, find out suitable layout, to reduce the maximum temperature of assembly; (2) design of built-in thermal sensor: design multiple thermal sensor in assembly, with the temperature variations of detection components inside; (3) Automatic adjusument of LED component temperature: design thermal control circuit in the driving circuit of LED component, with the temperature of monitoring assembly, when the temperature of assembly has exceeded given maximal value, then the power supply of LED is adjusted to the size namely reducing supply electric current, the junction temperature of each chip is reduced, thus reaches the object that the temperature of whole LED multi-chip module is reduced.

Below for the assembly be made up of 9 LED chips, specific embodiment of the invention step is described.This 9 LED chip called after L1, L2 ..., L9, the heating power of each chip is followed successively by: 0.16, and 0.16,0.34,0.82,1.15,0.25,0.25,0.47,0.34.

Step (1): the chip mount in LED component on the substrate of one piece of rectangular area.Substrate divides impartial grid, and on the position of each net point or node (i, j), has placed chip, here an i=1,2,3; J=1,2,3 (as shown in table 2).Each LED chip in assembly can produce certain heat, and a part for these heat passes to contiguous unit, and another part is taken away by air by convection current.

Table 2

(1,1)	(1,2)	(1,3)
			(2,1)	(2,2)	(2,3)
(3,1)	(3,2)	(3,3)

Step (2): to a kind of given about LED chip layout in assembly, set up the thermal balance equation of each node, namely micro unit thermal balance equation is listed respectively to each LED chip in whole LED component, and form a system of linear equations.

Step (3): adopt the system of linear equations of over-relaxation iterative method to second step gained to solve, draw the temperature value at each node location place in the steady state.

Step (4): a kind of chip layout in LED component is regarded as body one by one, and this individuality is encoded.Simultaneously individual z just when being defined as H (z).

Step (5): use algorithm 1 above to calculate and obtain a more excellent individuality, this more excellent individuality corresponds to a kind of placement scheme with the LED component chips of less maximum temperature.Calculate in algorithm 1 each individuality of population just when time, employ the result of step (3).

Step (6): to the assembly be made up of 9 LED chips, through step (2) as above to step (5) process after, the one obtained preferably placement scheme is (379182465), as shown in table 3.

Table 3

L3	L7	L9
			L1	L8	L2
L4	L6	L5

Step (7): by step (6) obtain on the basis of layout, calculate the transient temperature of each node (i, j) position on substrate, i=1 here, 2,3; J=1,2,3.Now calculated transient temperature value is three higher node locations is (2,2), (3,1) and (3,3).

Step (8): design three thermal sensors, and the region they being placed on respectively node (2,2), (3,1) and (3,3) place.Here, thermal sensor circuit structure primarily of as lower part form: bipolar transistor, operational amplifier, analog to digital converter, a resistance etc.Base signal line, the resistance of bipolar transistor are connected successively with the input end of operational amplifier, and the output terminal of operational amplifier is connected with the input end of analog to digital converter, and the output of this analog to digital converter is the value with temperature correlation.Here because the signal in the base stage of transistor is very faint, therefore use operational amplifier that this signal is amplified, spread out of again afterwards.

Step (9): the driving circuit of design LED component.The function of this driving circuit is when the temperature of assembly has exceeded given maximal value, then the power supply of LED is adjusted to the size namely reducing supply electric current, the junction temperature of each chip is reduced, thus the temperature of whole LED component is reduced.The structure of the driving circuit of LED component is as shown in Figure 3: control module is connected with reference voltage module, light-adjusting module, constant current source module and storer respectively, and constant current source module is connected with storer.

Step (10): by step (1) as above to step (9), just complete the design of chip layout to LED component and test structure, just can build the Thermal test system of LED component on this basis, the structure of this test macro as shown in Figure 4.

As follows to the concrete implementation step (job step) of this test macro:

(1) open the switch of driving circuit, switch on power.

(2) parameter of the control module in driving circuit to reference voltage module, light-adjusting module, constant current source module etc. carries out initialization.

(3) driving circuit produces the electric current of certain numerical value, and is applied to the input end of LED component.

(4) LED component sends the light of some strength, and along with producing certain heat.

(5) thermal sensor in LED component detects the heat produced, and the temperature value of LED component is sent to the control module of driving circuit.

(6) control module calculates the maximal value of the temperature value that each thermal sensor transmits, when this value exceeded one preset value time, then the power supply of LED component is adjusted to the size namely reducing supply electric current, the junction temperature of each chip can be reduced like this, and reduce the temperature of whole LED component, thus the temperature of LED component is maintained in normal scope.

Above-described embodiment is the present invention's preferably embodiment; but embodiments of the present invention are not restricted to the described embodiments; change, the modification done under other any does not deviate from Spirit Essence of the present invention and principle, substitute, combine, simplify; all should be the substitute mode of equivalence, be included within protection scope of the present invention.

Claims (9)

1. there is a heat testing method for the LED component of built-in temperature detection, it is characterized in that comprising the steps:

(1) layout optimization of LED component: to a kind of LED component of given initial configuration, by setting up the thermal balance equation of each node, and the position using a kind of Revised genetic algorithum residing in LED component to each chip is optimized, find out suitable layout, to reduce the maximum temperature of LED component;

(2) design of built-in thermal sensor: calculate the LED component after step (1) layout optimization, calculates the transient temperature of each node location of LED component, obtains the node that transient temperature is higher; In LED component, design several thermal sensors, and they are placed on the node region that in LED component, transient temperature is higher respectively, to detect the temperature variations of LED component inside;

(3) Automatic adjusument of LED component temperature: design has the driving circuit of the LED component of heat control function, wherein, is connected by the thermal sensor in step (2) with driving circuit; The LED component temperature that thermal sensor monitors sends driving circuit to, when the temperature of LED component has exceeded given maximal value, driving circuit just regulates supply electric current to LED component, namely the size of supply electric current is reduced, the junction temperature of each chip is reduced, thus the temperature of whole LED component is reduced.

2. the heat testing method with the LED component of built-in temperature detection according to claim 1, it is characterized in that: the interlace operation of the genetic algorithm described in step (1) is the one in selection point of crossing, two point of crossing and three point of crossing, respective choosing with adjustment mode to produce new individuality individual component value be have employed to each crossover location simultaneously.

3. the heat testing method with the LED component of built-in temperature detection according to claim 1, is characterized in that:

The circuit structure of the thermal sensor described in step (2) is: bipolar transistor, resistance, operational amplifier are connected successively with analog to digital converter.

4. the heat testing method with the LED component of built-in temperature detection according to claim 3, is characterized in that:

The structure of the driving circuit described in step (3) is as follows: comprise reference voltage module, light-adjusting module, constant current source module, memory module and control module; Control module is connected with reference voltage module, light-adjusting module, constant current source module and memory module respectively, and constant current source module is connected with memory module.

5. the heat testing method with the LED component of built-in temperature detection according to claim 4, is characterized in that: described reference voltage module is X60008 voltage source, and described memory module is SDRAM storer, and described control module is single-chip microcomputer.

6. realize the test macro with the heat testing method of the LED component of built-in temperature detection described in any one of Claims 1 to 5, it is characterized in that: the driving circuit comprising thermal sensor and LED component; Thermal sensor is connected with the driving circuit of LED component.

7. test macro according to claim 6, is characterized in that: also comprise the module determining each node temperature of LED component and the module optimizing LED component layout; Determine that the module of each node temperature of LED component is for determining the temperature of each node after the temperature of each node in LED component initial configuration and LED component structure optimization; After the temperature value of node each in LED component initial configuration being input to the module optimizing LED component layout, can be optimized LED component layout; The temperature of each node after LED component structure optimization can determine the placement location of thermal sensor.

8. test macro according to claim 7, is characterized in that: in described determination LED component, the module of each node temperature is by using thermal balance equation to come each LED chip accounting temperature value.

9. test macro according to claim 7, is characterized in that: the module of described optimization LED component layout is the placement scheme by using a kind of Revised genetic algorithum to obtain the LED component chips with less maximum temperature.