View AB20-4_4641201.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

application brief ab204 replaces an 1149 4 thermal management considerations for superflux leds thermal management is critical in the design of led signal lamps because temperature affects led performance and reliability. the following section details the effects of temperature on leds. in addition, thermal measurement techniques of led signal lamps and recommended design practices for proper th ermal management are covered. table of contents importance of thermal management for high-power led assemblies 2 temperature induced effect s on led light output 2 change in dominant wavelength (color) as a function of junction temperature 2 temperatureinduced failures of leds 3 thermal modeling of led assemblies 4 thermal resistance of led automotive signal lamps 4 junctiontoambient thermal resi stance measurement procedure 5 junctiontoambient thermal resistance measurement 5 estimating junctiontoambient thermal resistance 6 evaluating junction temperature and forward current 6 light output and forward current 7 derating example cases 7 recommended design practices for proper thermal management 8 pcb design 8 maximum metallization 8 led spacing 9 lamp housing design and mounting of the led array 9 circuit design 10 current control 10 power dissipation 10 switching power supplies 11 ambient temperature compensation 11 appendix 4a 12 alternate junctiontoambient thermal resistance measurement procedure 12
importance of thermal management for highpower led assemblies temperature induced effects on led light output the junction temperature of the led affects the devices luminous flux, the color of the device, and its forward voltage. junction temperature can be affected by the ambient temperature and by selfheating due to electrical power dissipation. the equation for luminous flux as a function of temperature ( c) is given below: v (t 2 ) = v (t 1 )e Ck ? tj where: v ( t 1 )= luminous flux at junction temperature t 1 v ( t 2 )= luminous flux at junction temperature t 2 k = temperature coefficient ? t j = change in junction temperature ( t 2 t 1 ). typical temperature coefficients for various high brightness leds are listed in table 4.1. the degradation of flux as a function of increasing temperature for a typical redorange, absorbingsubstrate (as) or transparent substrate (ts) alingap le d is shown in figure 4.1. note, luminous flux has been normalized at 25 c. this graph shows the profound affect that temperatures within the normal operating guidelines can have on luminous flux. as shown, an increase in the junction temperature of 75 c can cause the level of luminous flux to be reduced to onehalf of its room temperature value. from this, it is clear that temperature effects on luminous flux must be accounted for in the design of a led assembly. table 4.1 temperature coefficient for high-brightness led materials. led material type temperature coefficient, k as alingap, red-orange 9.52 x 10 -3 as alingap, amber 1.11 x 10 -2 ts alingap, red-orange 9.52 x 10 -3 ts alingap, amber 9.52 x 10 -2 figure 4.1 luminous flux versus ambient temperature for a typical red-orange as/ts alingap led when operated at a constant current.
3 change in dominant wave-length (color) as a function of junction temperature the junction temperature of leds also affects their dominant wavelength , or perceived color. the equation for dom inant wavelength, d , as a function of temperature is: where: d ( t 1 )= dominant wavele ngth at junction temperature t 1 d ( t 2 )= dominant wavele ngth at junction temperature t 2 a rule that is easy to remember is the dominant wavelength will increase one nanometer for every 10 c rise in junction te mperature. in most designs of red automotive signal lamps, this change in color is not important because the allowed color range is very large (approximately 90 nm). however, for some amber automotive signal lamps, this color shift can be a concern and should be accounted for where the allowed color ranges are small (approximately 5 to 10 nm depending on the region al specifications). temperature-induced failures of leds leds are typically encapsulated in an optically clear epoxy resin. at a certain elevated temperature, known as the glass transition temperature, t g , these epoxy resins transform from a rigid, glasslike solid to a rubbery material. a dramatic change in the coefficient of thermal expansion (cte) is generally associated with the t g . the t g is calculated as the midpoint of the temperature range at which this change in cte occurs, see figure 4.2. to avoid catastrophic failure of led packages, the junction temperature, t j , should always be kept below the t g of the epoxy encapsulant. lumileds specifies a maximum junction temperature, t j (max) , which is below the t g of the epoxy encapsulant used. for superflux leds, t j (max) = 125 c. if the t j (max) is exceeded, the cte of the epoxy encapsulant will permanently and dramatically change. a higher cte causes the epoxy encapsulant to expa nd and contract more during temperature changes. this causes more displacement of the wire bond within the led package, resulting in a premature wearout and breakage of the wire. wire bond breakage results in an open failure. figure 4.2 expansion-temperature relationship for clear, epoxy, led encapsulants.
4 thermal modeling of led assemblies thermal resistance of led automotive signal lamps thermal resistance is associated with the conduction of heat, just as electrical resistance is associated with the conduction of electricity. defining resistance as the ratio of driving potential to the corresponding transfer rate, thermal resistance for conduction can be defined as shown in the equation below: _ where: r = thermal resistance between two points ? t = temperature difference between those two points q x = rate of heat transfer between those two points the thermal resistance of an led signal lamp (junctiontoambient th ermal resistance, or r ja ) is made up of two pr imary components: the thermal resistance of the led package (junctiontopin ther mal resistance, or r jp ) and the thermal resistance of the lamp housing (pin toambient thermal resistance, or r pa ). these two components of thermal resistance are in a series configuration, therefore: r jp + r pa = r ja (led emitter) (lamp housing) (led signal lamp) this is shown graphically in figure 4.3. assuming all the electrical power is dissipated in the form of heat (approxi mately 5to10% of the power is dissipated optica lly), the equation for junctiontopin thermal resistance ( r jp ) of an led can be written in the form of the equation below: where: p = the total electrical power into the led ( i f * v f ) for led lamp assembli es, the equation for junctiontoambient thermal resistance, r ja , of an individual led within the assembly can be written as: where t j = ? t j + t a . as can be seen from this equation, in order to determine r ja of an led within a lamp assembly, the rise in juncti on temperature, and the electrical power into the device mu st be determined. the electrical power into the led under test can easily be determined by multiplying its forward current and forward voltage. the rise in junction temperature can be determined by measuring the change in forward voltage of the led under test. figure 4.3 graphic representation of the components of thermal resistance.
5 junction-to-ambient thermal re sistance measurement procedure a simple method for measuring the r ja of a lamp assembly is possible by assuming the r jp of the device under test (dut) is of a typical value. by making this assumption, only the pin-to-ambient thermal resistance, r pa , needs to be measured to calculate the r ja of the lamp ( r ja = r j p + r pa ). this simplified procedure for measuring r ja is described below: step 1: assume the r jp of the led emitter is that shown in the data sheet (typical r ja for hpwa-xx00 = 155 c/w, and for hpwt-xx00 = 125 c/w). step 2: pick one led within the assembly to be used as the dut. the hottest led in the assembly should be chosen, for example an led in the middle of the assembly and next to a resistor. step 3: solder a small thermocouple (approximately 0.25 mm in diameter) onto one of the cathode leads of the dut near the top surface of the pcb. large thermocouples, which can alter the thermal properties of the dut, should be avoided. step 4: assemble the modified pcb into the lamp housing such that the thermocouple wires are extending outside the lamp. step 5: energize the entire lamp assembly at the design voltage for a minimum of 30 minutes. this will allow the lamp assembly to thermally stabilize. step 6: measure the pin temperature of the dut along with the ambient temperature in the room. step 7: calculate the r pa of the lamp assembly using the following equation: tp - ta r pa = p where the power, p , into the dut is calculated by multiplying the heating/design current by its corresponding forward voltage. step 8: calculate the r ja of the lamp assembly by adding the r jp of the emitter from step 1 to r pa from step 7. junction-to-ambient thermal resistance measurement these sections give detailed instructions on how to perform th ermal resistance measurements on led assemblies. the first method described in the box above, junction toambient thermal resistance measurement procedure, allows for simple measurements to be made on lamp a ssemblies without an elaborate test setup. the second method presented, estimating junctiontoambient thermal resistance, eliminates the need for measured thermal resistance. this type of estimation is ideal for early evaluations, where an actual prototype and/or test equipment is not available. an alternate method for measuring thermal resistance is provided in appendix 4a. this method monitors th e change in forward voltage of the led to determine the change in junction temperature and thermal resistance. this method requires an elaborate test setup and precise measurements. this technique is commonly used by lumileds lighting. lumileds will evaluate the thermal resistance of led assemblies and sign al lamps upon request. please contact your local applications engineer for information.
6 table 4.2 typical r ja values for the classes of led lamp assemblies led lamp classification typical r ja (c/w) class 1 325 class 2 400 class 3 500 class 4 650 estimating junction-to-ambient thermal resistance the procedures described in junction-to-ambient thermal resistance measurement procedure are accurate methods for determining the r ja of an led within a plastic lamp assembly. however, in some cases, the time and/or equipment may not be available to perform such testing. in these cases, an educated estimate may be the best method available. lumileds has developed some basic classifications of led lamp assemblies and corresponding r ja estimates. below is an explanation of the different classes, and the r ja estimates. class 1: single row of leds with the current-limiting resistors/drive circuitry located off of the pcb, either in the wire harness assembly or on a separate pcb. class 2: single row of leds with the current-limiting resistors/drive circuitry located on the same pcb as the leds. this is the most common situation for led chmsl assemblies. class 3: multiple rows, or an x-y arrangement, of leds with the current-limiting resistor s/ drive circuitry located off of the pcb, either in the wire harness assembly or on a separate pcb. class 4: multiple rows, or an x-y arrangement, of leds with the current-limiting resistor s/ drive circuitry located on the same pcb as the leds. this is the most common situation for led rear combination lamp applications. table 4.2: lists the typical r ja values for each class of led lamp assembly listed above. these are only estimates and should not be used for detailed, worst-case analyses. evaluation junction temperature and forward current the primary concern when evaluating the thermal characteristics of an led assembly is to ensure that the ju nction temperature of the leds is kept below the specified maximum value (125 c for superflux leds). there are three factors which determine junction temperature: 1) ambient temperature, 2) r ja , and 3) power into the led. below is a sample junction temperature calculation, which illustrates how these three factors interact: t j = ( r ja . p led ) + t a t j = ( r ja . i f led . v f led ) + t a typical values for t a(max ) are shown in table 4.3. to determine the worstca se, highest junction temperature, this equation becomes: t j max = ( r ja . p led max ) + t a max t jmax = ( r ja . i f max . v f max ) + t a max t jmax 125 c
7 lumileds plots these curves for different values of r ja along with their intersection with the maximum drive current of 70 ma, and their intersection with the maximum ambient temperature of 100 c and includes this graph in all led data sheets. this graph is typically referred to as the derating curves . the derating curves for hpwtxx00 devices, are shown in figure 4.4. derating curves for hpwaxx00 devices are provided in the superflux led technical data sheet. refer to sidebar derating example cases for further explanation. light output and forward current the relationship between light output and forward current for different thermal resistances is shown in figure 4.5. for led assemblies with low thermal resistances ( r ja = 200 c/w), the relative flux increases al most proportionally to the forward current. however, for led assemblies with high thermal resistances ( r ja = 600 c/w), the relative flux can actually decrease as forward current is increased. for assemblies with high r ja , a great deal of heating occurs resulting in high junction temp eratures. in these cases, the effects of increasing junction temperature can offset the effects of increasing forward current. proper thermal management and drive current selection is critical to maximizing the performance of leds. derating example cases case 1 class 1 led chmsl consider an led chmsl application using 12 hpwt mh00 leds in a row, with a cu rrent limiting resistor in the wire connector. the auto manufacturer has specified a maximum ambient temperature of 75 c. from table 4.2 the thermal re sistance can be estimated as r ja = 325 c/w. using figure 4.4, the maximum allowable forward current through each led is 55 ma at t a (max) = 75 c. case 2 class 4 led rear combination lamp (rcl) consider an led rcl application using 36 hpwtmh00 leds in a 6x6 pattern, with the drive circuitry on the same pcb as the leds. the auto manufacturer has specified a maximum ambient temperature of 75 c. from table 4.2 the thermal re sistance can be estimated as r ja = 650 c/w. using figure 4.4, the maximum allowable forward current through each led is 30 ma at ta(max) = 75 c. as can be seen from these simplified sample cases, the r ja has a major impact on junction temperature, and thus maximum allowable forward current. the different applications using the same led have a difference in maximum forward current of nearly 2:1. a more detailed determination of maximum forward current is presented in application brief 203 electrical design considerations for superflux leds .
8 recommended design practices for proper thermal management pcb design proper pcb design can reduce the r ja of a led lamp assembly, an d thus reduce the junction temperature of the leds. listed below are some recommended practices for the design of led pcbs. maximum metallization conventional pcb design involves connecting various points on the board with traces of sufficient width to handle the current load. this process is usually visualized as adding traces to a blank pcb. for led pcbs, this process should be reversedvisualized as removing metal only where needed to form the electrical circuit. large metal pa ds surrounding the cathode leads of the leds are ideal. very little heat is conducted through the anode leads of the led, so additional metallization surrounding these leads does not help. table 4.3 typical t a (max) values for automotive signal lamps application typical t a (max) (c) exterior-mounted signal lamp 70 interior-mounted chmsl 80 interior, head-liner mounted chmsl 90 figure 4.4 graph of hpwt-xxoo derating curves. figure 4.5 relative luminous flux vs. forward current. figure 4.6 led chmsl pcb with proper metallization and component placement.
9 the resistors should be located in a remote portion of the pcb (away from the leds), on a separate pcb, or in the wire harness if possible. if this is not possible, the resistors should be distributed evenly along the pcb to distribute the heat generated. in ad dition, the traces from resistors to metallized areas surrounding cathode leads on the leds should be minimized to prevent resistors from heating adjacent leds. this can be accomplished by thinning down these tr aces, or by having metallized areas contacting the leds and resistors only contact the anode leads of the led. a portion of an led chmsl pcb depicting the design concepts di scussed is shown in figure 4.6. led spacing most of the electrical power in an led is dissipated as heat. tighter led spacing provides a smaller area for heat dissipation, resulting in higher pcb temperatures and thus higher junction temperatures. the leds should be spaced as far apart as packaging and optical constraints will allow. most chmsl applications use only a single row of leds at spacing greater than 15 mm which is ideal, as opposed to many amber turn signal applications which use a tightly spaced (less than 10 mm) xy array of leds. lamp housing design and mounting of the led array led lamp housings should be designed to provide a conductive path from the backside of the pcb to the lamp housing. this is typically accomplished by mounting the backside of the pcb directly to the lamp housing such that they are contacting one another across the entire length of the pcb. th is mounting scheme can be improved by applying a thermally conductive pad between the pcb and the lamp housing. the thermally conductive pad conforms to the features on the backside of the pcb and provides a larger contact area for conduction. often the pcb is mounted to the lamp housing on top of raised bosses. in this case, the area for conduction into the lamp housing is reduced to the contact area on the top side of the bosses, greatly reducing its effectiveness. another common configuration mounts the pcb along its top and bottom edges to slots in the side of the lamp housing. again, the area for conduction into the lamp housing is reduced to the contact areas of the slots, which reduces the effectiveness of conduction. if the pcb is mounted in such a way that conduction to the lamp housing is not effective (trapped air is a very poor conductor of heat), then allowances for convective cooling should be made. the most common technique to take advantage of natural convec tion is to put holes in the top and bottom side of the lamp housing to allow for vertical air flow over the pcb. however, where the lamp housing must be sealed for environmental reasons, this type of approach is impractical.
10 circuit design circuit design can help control the junction temperature of the leds in two important ways: 1) minimize fluctuations in the drive current (power input), and 2) dissipate a minimum amount of heat, or dissipate heat in such a way as to minimize its effect on the leds. current control an ideal drive circuit will provide the same current to the leds even as ambient temperatures and battery voltages vary. inexpensive, simple current control circuits can be designed to accomplish this task. a schematic of such a circuit is shown in figure 4.7. current control circuits are often too expensive and unnecessary for led chmsl applications. the most common led chmsl drive circuit consists of a current limiting resistor(s) and a silicon diode for reverse voltage protection in series with the leds. in this circuit design, the input current into the leds varies as the battery voltage changes. the current control characteristics of this type of circuit improve as larger resistor/s are used with fewer leds in series. however, circui ts with fewer leds in series will have greater heat generation in the drive circuit. figure 4.8 graphs the forward current provided to the leds vs. the input battery voltage for resistor circuits with three, four, and five leds in series. for more information on picking the optimum design current, and led drive circuit for your application, please reference application brief 203 electrical design considerations for superflux leds. power dissipation if the led drive circuit is in a remote location relative to the leds (in the wire harness or on a separate pcb), then the power dissipated by the drive circuit does not affect the junction temperature of the leds. drive circuit heating is a concern when the drive circuit is on the same pcb as the leds. drive circuit power dissipation, and thus heat generation is inversely proportional to the number of leds in series. circuits with fewer leds in series will have greater heat generation in th e drive circuit. for most automotive app lications in which the battery voltage is approximately 13 v, lumileds recommends configuring four leds in series. four leds in series is a good compromise between forward current control, heat generation, and minimum turnon voltage for the led array. figure 4.7 schematic of a current control circuit for led automotive lamp applications. figure 4.8 led forward current vs. battery voltage for circuits of two, three, four and five leds in series with a current limitin g resistor.
11 switching power supplies current sources, which operate efficiently over a wide range on input voltages, can be designed using pulsewidth modulation (pwm) circuitry. such circuits have the advantage of low heat dissipation, and large input voltage compliance. this type of power supply is traditionally used in applications where electrical efficiency and heat dissipation are of critical importance, such as a laptop computer. due to their widespread adoption in other applications, the cost of components has decreased, and their availability has increased, making this an interesting alternative for driving led arrays. a block diagram of a simple switching current source is shown in figure 4.9. the pwm module varies the pulse width based on the input and feed back voltages. the feedback voltage is proportional to the current through the led array, where voltage is measured directly above a small fixed resistance connected to ground. the filter circuitry is used to smooth out the output voltage of the pwm / transistor switch. with minor modifications, this type of circuit can be used to drive multiple led arrays and a variety of drive circuits. ambient temperature compensation drive circuitry can be designed which compensates for increasing ambient temperature by decreasing the forw ard current to the led array. this allows the la mp designer to drive the led array at a higher forward current by reducing the amount of current derating. temperature compensation is achieved by incorporating temperature sensitive components into the drive circuitry, such as positive temperature coefficient (ptc) resistors. an example of the resistance vs. temperature characteristics of a ptc resistor is shown in figure 4.10. it can be seen that the resistance of such a device radically increases when the body temperature of the ptc resistor reaches the switching temperature. by designing a drive circuit such that the switching temperature occurs at a temperature less than t a(max) , full current derating is not necessary. consider the case in which the switching temperature of the ptc resistor is achieved at an figure 4.9 led driver module for automotive lighting applications. figure 4.10 resistance-temperature curve for ptc resistor.
12 ambient temperature of 50 c at the maximum input voltage. the forward current at t a < 50 c is 55 ma, and due to the increase in resistance the forward current at t a > 50 c is 30 ma. in such a case, the maximum junction temperature will be achieved at 50 c, therefore, 50 c can be used as t a(max) in the current derating calculations. an example of a current control circuit using temperature compensati on is shown in figure 4.11. appendix 4a alternate junction-to-ambient thermal resistance measurement procedure step 1: pick one led within the assembly to be used as the dut. the hottest led in the assembly should be chosen, for example an led in the middle of the assembly and next to a resistor. step 2: electrically isolate the dut from the rest of the circuit by cutting the appropriate copper traces on the printed circuit board (pcb). step 3: solder long thin wi res onto one cathode lead and one anode lead of the dut. these wires should be long en ough to extend outside the lamp housing once it is reassembled because they will be used to apply the heating current and to measure the ? v f of the dut. step 4: complete the original circuit of the pcb assembly by attaching a dummy led onto the pcb to take the place of the isolated dut. this can be accomplished by soldering long, thin wires to one cathode lead and to one anode lead of an led, which is of the same type as the dut. next solder the other end of these wires directly to the pcb in such a way as to have this dummy led take the place of the dut in the circuit. step 5: assemble the modified pcb into the lamp housing such that the dummy led and the dut wires are extending outside the lamp. step 6: measure the initial v f of the dut at a very low test current. this te st current should be low enough such that it causes a minimum amount of heating (1 ma is recommended). step 7: energize the entire lamp assembly at the design voltage, and dut at the design current for the individual leds for a minimum of 30 minutes. this will allow the lamp assembly to thermally stabilize. figure 4.11 current control circuit using temperature compensation.
13 step 8: measure the v f of the dut at the heating current ( v f heating). step 9: turn off all power to the lamp, and immediately ( 10 ms) remeasure the v f of the dut at the test cu rrent selected in 6). step 10: calculate the ? t j of the dut by dividing the ? v f ( ? v f = v f (step 6) v f (step 9) ) by the appropriate factor in table 4.3. step 11: calculate the power, p , into the dut by multiplying the heating/design current by its corresponding v f heating as determined in step 8. step 12: calculate r ja using the values of ? t j and p calculated in steps 10 and 11. lumileds can provide the r ja measurements of led lamp assemblies as described above as a service to its led customers. table 4.3 ratios of the change in forward voltage vs. the change in junction temper ature for high-brigh tness led materials led material type ? v f / ? t j ( mv / c) as alingap -2.0 ts alingap -2.0
14 company information lumileds is a worldclass supplier of li ght emitting diodes (leds) producing billions of leds annually. lumileds is a fully integrated supplier, producing core led material in all three base colors (red, green, blue) and white. lumileds has r&d deve lopment centers in san jose, california and best, the netherland s. production capabilities in san jose, california and malaysia. lumileds is pioneering the highflux led technology and bridging the gap between solid state led technology an d the lighting worl d. lumileds is absolutely dedicated to br inging the best and brig htest led technology to enable new applications and markets in the lighting world. lumileds www.luxeon.com www.lumileds.com for technical assistance or the location of your nearest lumileds sales office, call: worldwide: +1 408-435-6044 us toll free: 877-298-9455 europe: +31 499 339 439 a sia: +65 6248 4759 fax: 408-435-6855 email us at info@lumileds.com lumileds lighting, llc 370 west trimble road san jose, ca 95131 ? 2002 lumileds lighting. all rights reserved. lumileds lighting is a joint venture between agilent technologies and philips lighting. luxeon is a trademark of lumileds lighting, llc. pr oduct specifications are subject to change without notice. publication no. ab204 (sept2002)

▲Up To Search▲

Price & Availability of AB20-4

	To Download AB20-4 Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .