P A T E N T P LIN D O C #: 1592 E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES T H E I N F I N I T E P O W E R O F I P N N O VA T I O N R E L I M I N A R Y DESCRIPTION The LXM1592/93 series of floating output drive CCFL (Cold Cathode Fluorescent Lamp) Inverter Modules are specifically designed to drive large LCD displays (11.3" and larger), which are used in notebook computers. These new inverters were specifically designed to reduce the leakage currents from the lamp to the reflector or the metal frame of the panels. The floating output architecture of these inverters also permits a much wider dimming range when compared to nonfloating designs, and an additional 10% efficiency improvement is realized. Both the LXM1592 and LXM1593 are fully customizable (electronically and mechanically) to specific customer requirements. The modules convert unregulated DC voltage from the system battery or AC adapter directly to high-frequency, highvoltage sine waves required to ignite and operate CCFL lamps. The module design is based on a proprietary Linfinity IC that provides important new performance advances. Remarkable improvements in efficiency and RF emissions result from these single stage resonant inverters, featuring a patent pending Current Synchronous, Zero Voltage Switching (CS-ZVS) topology. CSZVS produces nearly pure sine wave cur- rents in the lamp, enabling maximum light delivery, while reducing both conducted and radiated noise. This topology simultaneously performs two tasks including line voltage regulation and lamp dimming through lamp current regulation. These two functions are performed in a single power stage made up of a pair of lowloss MOSFETs. The MOSFETs drive a low current resonant circuit that feeds the primary of a high voltage transformer with a sinusoidal voltage. Required L and C values in the resonant circuit are such that very low-loss components can be used to obtain higher electrical efficiency than is possible with previous topologies. Two module versions are available. The half-bridge LXM1592 provides peak efficiency when operated at input voltages above 7 volts. The LXM1593 achieves higher efficiency at input voltages above 4.5V with its full-bridge drive circuit. The modules are equipped with a dimming input that permits full range brightness control from an external potentiometer, and a sleep input that r educes module power to a few microwatts in shut-down mode. Each module features output open and short circuit protection. S H E E T ■ FULLY FLOATING OUTPUT ■ 35% MORE LIGHT OUTPUT AT 2.5 WATTS ■ GREATER EFFICIENCY THAN GROUNDED OUTPUT DESIGNS ■ 4.5V TO 30V INPUT VOLTAGE RANGES ■ VERSATILE BRIGHTNESS CONTROL INPUT ■ 3 MICROAMP SLEEP CURRENT ■ OUTPUT SHORT CIRCUIT PROTECTION AND AUTOMATIC OVER VOLTAGE LIMITING ■ 8mm MAX HEIGHT, NARROW FOOTPRINTS ■ MINIMIZE THERMOMETER EFFECTS ■ MINIMIZE LAMP TO PANEL LEAKAGE CURRENT A P P L I C AT I O N S ■ 11.3" LCD PANELS AND LARGER ■ NOTEBOOK AND SUB-NOTEBOOK COMPUTERS ■ PERSONAL DIGITAL ASSISTANTS ■ PORTABLE INSTRUMENTATION ■ AUTOMOTIVE DISPLAYS ■ DESKTOP DISPLAYS ■ AIRLINE ENTERTAINMENT CENTERS BENEFITS ■ ULTRA-HIGH EFFICIENCY, LINE VOLTAGE REGULATION AND SLEEP MODE EXTEND COMPUTER BATTERY LIFE ■ COOL OPERATION PERMITS CLOSE PROXIMITY TO LCD PANEL WITHOUT DISPLAY DISTORTION FLOATING OUTPUT A RCHITECTURE ■ SMOOTH, FULL-RANGE BRIGHTNESS CONTROL GIVES YOUR PRODUCT A HIGH QUALITY IMAGE High Voltage Transformer CCFL CS-ZVS Inverter A T A K E Y F E AT U R E S PRODUCT HIGHLIGHT DC VIN D ■ LOW EMI / RFI DESIGN MINIMIZES SHIELDING REQUIREMENTS ■ NARROW, LOW-PROFILE STANDARD MODULES FIT INTO MOST LCD ENCLOSURES ■ SINGLE-SIDED PCB SAVES EXPENSIVE HIGH VOLTAGE INSULATING TAPES M O D U L E O R D E R I N F O R M AT I O N HALF-BRIDGE DRIVE FULL-BRIDGE DRIVE LXM1592-xxxxx-zz LXM1593-xxxxx-zz See instructions inside for completing module part number. F O R F U R T H E R I N F O R M AT I O N C A L L ( 7 1 4 ) 8 9 8 - 8 1 2 1 Copyright © 1996 Rev. 0.2 6/96 11861 WESTERN A VENUE , G ARDEN G ROVE , CA. 92841 1 P AT E N T P E N D I N G PRODUCT DATABOOK 1996/1997 LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T A B S O L U T E M A X I M U M R AT I N G S (Note 1) Input Supply Voltage (VIN)........................................................................................ LXM1592 = -0.3V to 30V / LXM1593 = -0.3 to 7.0V Output Voltage, no load .............................................................................................................................. Internally Limited to 1700VRMS Output Current .............................................................................................................................................. 7.0mARMS (Internally Limited) Output Power ........................................................................................................................................................................................ 4.5W Input Signal Voltage, (SLEEP and BRITE Inputs) .................................................................................................................. -0.3V to 6.5V Ambient Operating Temperature, zero airflow ........................................................................................................................ 0°C to 60°C Storage Temperature Range ................................................................................................................................................... -40°C to 85°C Note 1. Exceeding these ratings could cause damage to the device. All voltages are with respect to Ground. Currents are positive into, negative out of the specified terminal. R E C O M M E N D E D O P E R A T I N G C O N D I T I O N S (R.C.) This module has been designed to operate over a wide range of input and output conditions. However, best efficiency and performance will be obtained if the module is operated under the condition listed in the 'R.C.' column. Min. and Max. columns indicate values beyond which the inverter, although operational, will not function optimally. Parameter Symbol Input Supply Voltage LXM1592 LXM1593 Output Power Brightness Control Input Voltage Range Lamp Operating Voltage Lamp Current - Full Brightness Operating Ambient Temperature Range VIN PO VBRITE VLAMP IOLAMP TA Recommended Operating Conditions Min. R.C. Max. 7 4.5 12 30 6.5 4.2 2.5 650 6.5 60 2.5 0.8 240 Units 500 5 0 V V W V VRMS mA RMS °C ELECTRICAL CHARACTERISTICS Unless otherwise specified, these specifications apply over the recommended operating conditions and 25°C ambient temperature for the LXM1592/1593. Parameter Symbol Test Conditions LXM1592/1593 Min. Typ. Max. Units Output Pin Characteristics Full Bright Lamp Current Minimum Lamp Current Lamp Start Voltage Operating Frequency IL (MAX) IL (MIN) VLS fO VBRITE = 2.5 V DC, SLEEP = Logic High VBRITE = 0.8 V DC, SLEEP = Logic High 0°C < TA < 60°C VBRITE = 2.5VDC, SLEEP = Logic High, VIN = 12V 5.9 6.2 2.0 6.5 mA mA RMS VRMS KHz -1000 2.6 nADC VDC VDC 50 5.5 0.8 100 VDC VDC µADC 2.50 2.60 VDC µADC 3 92 90 10 µADC % % 1200 70 Brightness Control Input Current Input Voltage for Max. Lamp Current Input Voltage for 50% Lamp Current IBRITE VC VC VBRITE = 0V DC IO (LAMP) = 100% IO (LAMP) = 50% 2.4 -200 2.5 1.25 SLEEP Input Input Logic 1 Input Logc 0 Input Current VIH VIL IIN 2.2 0 VSLEEP = 0 - 5VDC Voltage Reference Output Voltage Output Current V REF IREF 0 < IREF < 500µA 2.40 500 Power Characteristics Sleep Current Electrical Efficiency (calculated values) 2 IIN (MIN) η VIN = 5VDC , SLEEP = Logic 0 LXM1592, VIN = 12VDC, IO (LAMP) = 5mARMS LXM1593, VIN = 5VDC, IO (LAMP) = 5mARMS Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T FUNCTIONAL PIN DESCRIPTION Conn. CN1 Pin Description CN1-1 CN1-2 VIN Input voltage. (+4.5 to +30VDC ) CN1-3 CN1-4 GND Power supply return. CN1-5 SLEEP Logical high on this pin enables inverter operation. Logical low removes power from the module and the lamp. A floating input is sensed as a logical low and will disable inverter operation. If not used, connect SLEEP through a 33kΩ resistor to VIN or directly to any voltage between 2.5 and 5.5V. CN1-6 BRITE Brightness control input. Apply 0.9 to 2.5 volts DC to control lamp brightness. Lamp current varies linearly with input voltage. Open circuit or 2.5V gives maximum brightness. CN1-7 AGND Brightness control signal return. For best results do not run VIN power supply current return through this pin. CN1-8 VREF Reference Voltage Output. 2.5V @ 500µA max. For use with external dimming circuit. CN2-1 LAMP HI High-voltage connection to high side of lamp. Connect to lamp terminal with shortest lead length. Do not connect to ground. CN2-2 LAMP LO High-voltage connection to low side of lamp. Connect to lamp terminal with longer lead length. Do not connect to ground. CN2 Copyright © 1996 Rev. 0.3 6/96 3 P AT E N T P E N D I N G PRODUCT DATABOOK 1996/1997 LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A LY S I S I N F O R M AT I O N INTRODUCTION LIGHTING CHARACTERISTICS OF CCFLs The duration of time that it takes for the light output to stabilize must first be determined before any meaningful measurements can be made. This is important when trying to maintain consistency between measurements, and is also important in minimizing the required testing time. Several factors affect the light output of the CCFL’s, such as operating current waveshape and frequency, proximity of the lamp to conducting surfaces, inverter output configuration, and ambient temperature, among other things. In addition, the newer lamps have very small diameters and operate at higher gas pressures. It appears that this makes these lamps electrically more unstable. In order to determine the time required to reach steady state for a particular lighting system in this test, a completely automated data acquisition system has been set up that is capable of taking light output data at uniform time intervals. The power supply, the ammeter and voltmeter are all controlled by the computer. The photometer’s RS-232 port is connected to the RS-232 port of the computer. Figure 1 shows a block diagram of this setup. With this setup, the calculation of the power input and efficiencies is greatly simplified, because automation and data gathering consistency are assured. 4 LCD Panel Power Supply Photo Sensor Middle of Screen V CCFL A Inverter J1803 Photometer Instrument Controller (Light out in Nits) Tektronix J17 Desktop Computer Parallel Port RS-232 FIGURE 1 — MEASUREMENT SETUP 100 samples are taken from a system at 3 second intervals consisting of the following: 1. Lamp type: 560Vrms Operating at 5-6mA. 2. Inverter type: Half-bridge floating output CS-ZVS inverter at 10V input. 3. Lamp is housed in a 11.3” active matrix LCD panel. 4. The panel is laid flat on a desk with the photometer placed at the center of the panel. The result of these measurements is shown in Figure 2. This figure shows the initial turn-on profile of the lamp under specific environmental conditions. 155 135 VIN = 10V 115 PIN = 3.72W Nits This section discusses some general topics in testing and evaluating Cold Cathode Fluorescent Lamps (CCFL) along with the inverters that drive them as they are used in active and passive matrix LCD displays. In particular, this discussion will concentrate on the testing of the Current Synchronous Zero Voltage Switching Inverter. The past two years have seen a rapid change in the types of available LCD displays, as well as their lighting and inverter systems. Significant strides have been made in the light transmission efficiencies of the optical systems in addition to efficiency gains in their lighting and inverter systems. At the same time, some of these improvements, especially in the reflector and lamp housing systems, now pose difficulties when driving these lamps. The discussion which follows will examine lighting characteristics of the CCFL’s and experimental data which can be used to determine the duration of time that it takes for the light output from the CCFL to stabilize. In addition, light output efficiency calculation methods will be presented that can help sort out various efficiency claims from different inverter manufacturers. As part of the following discussion, the parasitics of the CCFL/Panel system will be modeled and SPICE simulations will reveal the current profile in the lamp. Finally, actual performance data will be presented comparing non-floating versus floating secondaries, with an analysis of this data. 95 75 55 35 3 103 203 303 Time - (sec) FIGURE 2 — INITIAL TURN-ON CHARACTERISTICS OF THE CCFL IN A HIGH-EFFICIENCY 11.3" LCD PANEL Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) LIGHTING CHARACTERISTICS OF CCFLs (continued) Eff = 3.8 3.78 3.76 3.74 Nits / Watt The rapid increase of light output during the first few seconds of this test is due to the fact that mercury vapor inside the CCFL reaching steady state concentration. The continual increase of light output from the lamp at a slower rate is a result of the thermal time constants of the system. Essentially, as the lamp gets warmer, it tends to become more efficient. Figure 3 shows the light output efficiency of the system as calculated by using the following formula. Light Out (Nits) Power In (Watts) 3.72 3.7 3.68 3.66 3.64 40 3.62 3.6 35 3 103 203 303 Time - (sec) Nits / Watt 30 FIGURE 4 — INVERTER POWER INPUT PROFILE DURING INITIAL TURN ON 25 20 15 10 3 103 203 303 Time - (sec) FIGURE 3 — LIGHT OUTPUT EFFICIENCY PROFILE DURING INITIAL TURN ON Figure 3 clearly shows the increase in efficiency as the lamp in the panel is self heating. This graph also shows that 303 seconds is not a sufficient amount of time for this system to reach a steady state. Figure 5 shows what the required amount of time is for this system to reach a steady state. Figure 4 shows the inverter power input profile during initial turn on. It is interesting to note that, when the inverter is first turned on, the input power is lower. This is a result of the higher impedance of the lamp. It takes a finite amount of time for the mercury to fully vaporize, thereby reducing the impedance of the lamp and permitting it to reach a steady state in terms of power. Figures 5 and 6 show the above-mentioned system at a slightly different input power taken at a different time than the previous graphs. The light output and efficiency data is probably different because of a different ambient temperature. The sampling interval for these graphs was set at 10 seconds. 172 152 132 Nits 112 92 72 52 32 0 200 400 600 800 1000 Time - (sec) FIGURE 5 — LIGHT OUTPUT VERSUS TIME AT INITIAL TURN ON, 10sec SAMPLING PERIOD Copyright © 1996 Rev. 0.3 6/96 5 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) LIGHTING CHARACTERISTICS OF CCFLs (continued) 3.95 43 38 3.9 3.85 28 Watts Nits / Watt 33 23 3.8 18 3.75 13 8 0 200 400 600 800 3.7 10 1000 12 14 16 18 20 Time - (sec) Input Voltage - (V) FIGURE 6 — LIGHT OUTPUT EFFICIENCY VERSUS TIME AT INITIAL TURN ON, 10sec SAMPLING PERIOD FIGURE 8 — POWER INPUT VERSUS INPUT VOLTAGE Based on the graphs of Figures 5 and 6, it can be determined that this system reaches steady state in approximately 17 minutes. INVERTER INPUT VOLTAGE CONSIDERATIONS Almost all power conversion devices lose some efficiency when operated at voltages beyond their nominal values. In order to investigate the effect of input voltage variation on the light output efficiency, the input voltage to the inverter has been varied from its minimum to its maximum operating condition. The results of this effort are shown in Figures 7 and 8. The total input power increases by approximately 200mW when input voltage is increased from 10V to 20V. This corresponds to a light output efficiency change of 5.8%. This efficiency degradation is, of course, smaller when the input voltage is from 10V to 14V, as is the case in most systems, where the efficiency change is only 2.8%. The higher input voltages depicted in Figures 7 and 8 correspond to operating from an AC power source. Figure 9 shows the light output regulation versus input voltage. This graph shows the excellent light output (line) regulation characteristic of a CS-ZVS inverter with the floating output. The total line regulation is only ±0.23% because of this the purity of the lamp drive current as well as the true load current sensing capability of this circuit. 42.5 156.2 42 156 PIN = 3.8W 155.8 41 Nits Nits / Watt 41.5 40.5 40 155.4 39.5 155.2 39 10 12 14 16 18 20 Input Voltage - (V) FIGURE 7 — LIGHT OUTPUT EFFICIENCY VERSUS INPUT VOLTAGE 6 155.6 155 10 12 14 16 18 20 Input Voltage - (V) FIGURE 9 — LIGHT OUTPUT VERSUS INPUT VOLTAGE LIGHT OUTPUT REGULATION Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) INVERTER INPUT VOLTAGE CONSIDERATIONS (continued) All the information discussed thus far can be very useful when trying to design the power subsystem. The minimal decrease in efficiency of the CS-ZVS inverter enables the system designer to have a relatively wide operating input voltage range without a significant efficiency penalty. The inverter is normally designed for the minimum battery voltage. Efficiency is optimized when the minimum operating voltage is as close as possible to the nominal operating voltage. A FEW WORDS ON NEW LCD PANEL DISPLAYS Significant efficiency improvements have been made to the optical systems of newer, larger LCD panels, panels that are typically 11.3" inches and larger. However, these improvements, including improvements in the lightpipe, the reflector and the CCFL itself, have caused increased leakage currents from the lamp to the reflector and/or panel's metal frame. This condition results in degraded light output and reduced dimming ranges, when used with backlight inverters equipped with nonfloating (or grounded) high voltage sides. Further compounding the leakage current problem is an increase of the operating voltages of CCFL's, with some lamps requiring as high as 650VRMS to operate. In a non-floating or grounded inverter, the output of the high-voltage transformer is referenced to ground, permitting leakage currents to circulate between the panel, the system ground and the inverter ground. In order to address these leakage currents, a new inverter configuration has been designed by Linfinity, which uses a floating output drive, coupled with Linfinity's patent pending CS-ZVS technology. Generally speaking, in a floating output drive, the high-voltage side of the inverter transformer is not referenced to ground and, therefore, interrupts the path of the leakage currents, preventing them from flowing into the system ground. Because the Linfinity LXM1592 and LXM1593 are configured with a unique COUT Reflector CS-ZVS Inverter DC VIN CP1 CCFL T1 CPN RS combination of Linfinity's floating output architecture and CSZVS technique, they significantly reduce the leakage currents from the lamp to the reflector of the metal frame of the panel, further improving the efficiency of these newer inverters over non-floating, or grounded designs. The LXM1592 and LXM1593, equipped with the Linfinity floating output drive architecture, yield an additional 10% improvement in light output and also permit a wider dimming range, resulting in a more uniformlylighted, as well as more efficient and brighter panel. Linfinity's floating output drive scheme, which currently is the only design which senses the secondary side lamp current, achieves very accurate lamp current regulation and, as such, is unique and superior even to other floating output implementations. SIMPLIFIED MODELING OF THE CCFL-PANEL SYSTEM Non-Floating Configuration Figure 10 shows the electrical configuration of a non-floating drive. In this system, the CCFL current is being sensed with a resistor referenced to the inverter ground. The panel, along with the reflector, is also electrically “tied” to the inverter ground. A “thermometer effect” (or brightness gradient) is created when the parasitic capacitance and the reflector are diverting useful current from the lamp to ground. This effect is very intense in some of the newer panels because the reflector is metal or metal-coated plastic or is situated very close to the panel itself. An additional side effect of this leakage is a marked reduction of efficiency. The following experiment was performed in order to quantify this capacitance. A lamp was broken at both cathode ends and an AWG#18 bus wire was inserted through the tube. This assembly was then placed in the cavity of a metal reflector and the capacitance was measured using a standard RLC bridge. The measured parasitic capacitance was approximately 15pF. Normally this capacitance is distributed along the length of the tube. Also, the lamp wiring formed a parasitic capacitance with the metal frame, which in this case was about 14pF. With the above information, a simple discrete distributed electrical model was constructed to help analyze the system. This electrical model of the non-floating configuration is shown in Figure 11. The parasitic shunt capacitors shown as CP1-CPN produce a current gradient across the length of the lamp that results in the “thermometer effect” that exhibits itself as a brightness gradient. In extreme cases, this exhibits itself as partial lighting of the lamp with the “hot” side of the lamp being the brightest. FIGURE 10 — NON-FLOATING OUTPUT CONFIGURATION. CP1 THROUGH CPN REPRESENT DISCRETIZED DISTRIBUTED PARASITIC CAPACITANCE Copyright © 1996 Rev. 0.3 6/96 7 P AT E N T P E N D I N G PRODUCT DATABOOK 1996/1997 LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D S A T A H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) SIMPLIFIED MODELING OF THE CCFL-PANEL SYSTEM (con't.) 1 IL Resistive Element RP1 RPN CP1 Resistive Element Current Current through Resistive Element Normalized to 6.132mA 0.995 0.99 0.985 0.98 0.975 1 11 6 Discrete Distributed Electrical Model of the Lamp In order to study the effects of the parasitic capacitance, the lamp was divided into 20 identical segments, consisting of both resistive and capacitive elements. Assuming a full brightness operating impedance of 100KΩ, each individual resistive element would be 5KΩ and each capacitance would be 0.75pF. The circuit then was solved by using a circuit simulator, such as SPICE. The capacitance of the wiring in the non-floating drive was inconsequential and was ignored. The accuracy of the above model is thought to be limited because of the nonlinear nature of the lamp impedance along the lamp length as a result of the thermometer effect (resulting in impedance modulation). Figure 12 shows the result of SPICE simulation on the 20 element model. Impedance was adjusted for 600V and 6mA operation. This graph shows the variation of the current flow in the resistive elements of the lamp that produces light output. Furthermore, it shows that the current is higher at the “hot” end of the lamp by 2.4%. The effect of this is minimal light nonuniformity from one end of the lamp to the other. This is also apparent in the real circuit. 21 FIGURE 12 — CURRENT GRADIENT ASSUMING 20 ELEMENT ANALYSIS. FULL BRIGHTNESS Figure 13 shows the result of SPICE simulation on the above 20 element model when the lamp is dimmed to 1/3 brightness. Impedance was adjusted in this case for 600V and 2mA operation. The current differential in this case was 20%. The consequence of this is that the brightness change from one end of the lamp to the other will likely be more than 25%, a variance which is clearly visible to the human eye. 1 Normalized to 2.36mA 0.98 Resistive Element Current FIGURE 11 — NON-FLOATING OUTPUT CONFIGURATION. ELECTRICAL MODEL OF LAMP THAT SHOWS THE CURRENT PROFILE AS A RESULT OF THE PARASITIC SHUNT CAPACITANCE 16 Resistive Element 0.96 0.94 0.92 0.9 0.88 0.86 0.84 0.82 1 6 11 16 21 Resistive Element FIGURE 13 — CURRENT GRADIENT ASSUMING 20 ELEMENT ANALYSIS. ONE-THIRD BRIGHTNESS LEVEL 8 Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) SIMPLIFIED MODELING OF THE CCFL-PANEL SYSTEM (con't.) CH COUT 0.999 Normalized to 6.036mA 0.998 0.997 0.996 0.995 0.994 Reflector DC VIN CP1 CCFL CS-ZVS Inverter T1 1 Resistive Element Current Figure 14 shows the electrical configuration of a floating drive. In this system, the CCFL current is being sensed either in the primary side of the high-voltage transformer or at the secondary side. The panel and the reflector are electrically connected to the primary side (of T1) inverter ground. 0.993 1 6 11 16 21 Resistive Element CPN FIGURE 16 — CURRENT GRADIENT WITH FLOATING DRIVE MODELING CL CH & CL Are Wiring To Panel Capacitances FIGURE 14 — FLOATING OUTPUT CONFIGURATION Figure 16 shows the result of SPICE simulation on the circuit of Figure 15, again with a 20-element model. Current through Resistive Element IL The total current deviation in this case is 0.7%. Although this a small deviation, it is expected that in a real physical circuit, the difference would be higher as a result of other unmodelled parasitics and lamp non-linearities. Figure 17 shows the simulation results for the dimmed case of the floating drive. The total current deviation in this case is 6.1%. Thus, the floating drive introduces a smaller brightness gradient than the non-floating drive, resulting in a more uniformly lighted panel. 1 Resistive Element RP1 RPN CH CP1 CPN Discrete Distributed Electrical Model of the Lamp CL Resistive Element Current 0.99 Normalized to 2.105mA 0.98 0.97 0.96 0.95 0.94 1 6 11 16 21 Resistive Element FIGURE 15 — FLOATING OUTPUT CONFIGURATION ELECTRICAL MODEL OF LAMP Copyright © 1996 Rev. 0.3 6/96 FIGURE 17 — CURRENT GRADIENT WITH FLOATING DRIVE MODELING 9 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) NON-FLOATING INVERTER MEASUREMENTS 152 132 3.8 3.75 Watts A series of measurements were made to determine the performance of the previously used inverter with non-floating output. The exact same components were used to make the comparison. The light out versus time test was first run to determine at which point the light output reaches a steady state. Figure 18 shows the results of this test in graphical form. 3.7 3.65 Nits 112 3.6 0 92 200 400 600 800 1000 Time - (sec) 72 FIGURE 20 — CURRENT GRADIENT WITH FLOATING DRIVE MODELING 52 32 0 200 400 600 800 1000 FINAL ANALYSIS, FLOATING VERSUS NON-FLOATING LAMP DRIVE Time - (sec) FIGURE 18 — LIGHT OUT VERSUS TIME AT INITIAL TURN ON FOR NON-FLOATING INVERTER The light out efficiency versus time curve then has been calculated by using the light out and the power input data (Figure 20). The results of this effort are shown in Figure 19. As expected, light out efficiency improves as the lamp warms up reaching a maximum value of 38.86 Nits/Watt. 39 34 Nits / Watt 29 24 19 14 9 0 200 400 600 800 1000 Table 1 summarizes the performance differences between the floating and non-floating drive configurations evaluated in the testing discussed above. As has been seen, the performance gains strongly depend on the physical configuration of the lamp and the reflector assembly. One of the panels that was tested exhibited higher leakage, along with a significant improvement in the dimming range and the thermometer visual effect. The improvement in efficiency between floating versus non-floating in this case was 20%. Of course, as is usually the case, improvements in performance often come with an increase in circuit complexity. While inverters employing floating output design tend to be somewhat more complex, because of the advantage realized through their use, they often are the most optimum or only viable way to drive the latest generation LCD panels. The Current Synchronous Zero Voltage topology employed in the inverters manufactured by Linfinity Microelectronics is used in floating designs that exhibit “True Current Sense”, i.e. a reflection of the actual lamp current is being sensed to provide superior lamp current regulation. This can be contrasted with the average primary current sense of Buck-Royer oscillator- based inverters. It is noteworthy that the efficiency gains of the singlepower stage CS-ZVS topology compared to the double power stage Buck-ROYER combinations is more that 20%. Time - (sec) FIGURE 19 — LIGHT OUT EFFICIENCY VERSUS TIME AT INITIAL TURN ON FOR NON-FLOATING INVERTER 10 Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T E C H N I C A L / A N A L Y S I S I N F O R M AT I O N (continued) TABLE 1 Parameter Floating Non-Floating Input Power Light Output Light Out Efficiency Percent Max-Min Current Difference Because of Parasitics (note 1) 3.175 Watts 154.3 Nits 41.5 Nits/Watt 0.7% 6.1% at at Full Bright 1/3 Dimmed To 50% of full Brightness Current 3.726 Watts 144.8 Nits 38.86 Nits/Watt 2.4% 20% at at Full Bright 1/3 Dimmed To 65% of full Brightness Current Dimming Range (note 2) Note 1. Note 2. Improvement over Non-Floating -0.3% 6.56% 6.8% 242% at Full Bright 227% at 1/3 Dimmed This refers to the max and min currents in resistive elements of the 20 element analysis. The parasitics used did not pertain exactly to the 11.3" LCD panel used to make the measurements. The results are provided for comparison purposes. It is expected that the parasitics of the panel used to make the light measurements are lower than those depicted. Dimming range here is defined as the point where a visible "thermometer" effect just takes place. SUMMARY Several new ways for testing CCFL’s and inverters have been presented. The emphasis throughout these tests has been on how to make fair comparisons. To that end, a method has been presented that makes certain that the light output has reached steady state with all inverters tested, thus guaranteeing fair comparisons. The lamp/reflector parasitics were modeled and the lamp current profile was calculated based on these models. This gave insights on the effect of the parasitics either when the lamp is at full brightness or dimmed. Both the non-floating and floating inverter designs were considered and analyzed. The use of a floating lamp architecture resulted in approximately a 6.8% improvement in light output efficiency when it was compared to a non-floating design. The dimming range with the floating drive was also better by more than 15%. The comparison between the floating versus the non-floating drive designs were presented in tabular form for easy evaluation. Copyright © 1996 Rev. 0.3 6/96 11 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T T Y P I C A L A P P L I C AT I O N S AC/DC Adapter VIN VREF LAMP HI R1 100k Sytem Battery (NiCd, NiMH, etc.) BRITE LXM1592/93 3.3 or 5V From Power Management Logic SLEEP CMOS or TTL R2 56k LAMP LO AGND GND CFL TUBE Inside LCD Panel FIGURE 21 — NOTEBOOK SYSTEM APPLICATION CMOS or TTL gate From Power Management Logic VIN SLEEP VREF R1 BRITE Lamp Current (%) = LAMP HI CFL TUBE LXM1592/93 R2 LAMP LO AGND VBRITE x 100 VREF R1 = 100k typical, 5k minimum R2 = Value optional to determine lowest brightness setting R2 = 0.5 R1 minimum GND Longest Lead FIGURE 22 — POTENTIOMETER BRIGHTNESS CONTROL & SLEEP MODE +5V S1 "Increase" +5V U1 47k N.O. UC From Logic (optional) D N.O. S2 "Decrease" 47k DC RH VREF RW RL BRITE R1 AGND • S1 & S2 are momentary push buttons, normally open contacts. • U1 = 100k digital pot. • R1 value optional for choosing dimming range: for R1 = 100k, Range = 100% to 50% for R1 = 50k, Range = 100% to 33% • R1 minimum value is 0.5 RPOT. LXM1592/93 +5V FIGURE 23 — NONVOLATILE DIGITAL BRIGHTNESS CONTROL 12 Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D S A T A T Y P I C A L A P P L I C AT I O N S R1 100k BRITE A Logic Inputs R3 402k (continued) • Drivers are open collector. • Run separate ground from driver chip to LXM1592/93 AGND to prevent noise pickup. VREF R2 150k H E E T • This scheme may be expanded to 3 or 4 bits for more resolution. Lamp Current (% Full On) = AGND LXM1592/93 B A B Lamp Current 0 0 1 1 0 1 0 1 52% 60% 80% 100% VBRITE x 100 VREF FIGURE 24 — LOW COST DIGITAL BRIGHTNESS CONTROL 1kHz Standard CMOS Gate 100k 20k 1N914 Brightness (% Full On) = VREF BRITE 1µF VBRITE x 100 VREF AGND LXM1592/93 Input Duty Cycle Lamp Current * 100% 50% 0% 100% 70% 33% * 100% DC is a logic HI at CMOS gate input. FIGURE 25 — PWM BRIGHTNESS CONTROL Copyright © 1996 Rev. 0.3 6/96 13 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT DRIVE , C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D S A T A H E E T COMPLETING THE MODULE PART NUMBER LXM159 Module Type = = - Half-Bridge Drive (7.0V to 30V Battery Voltage) Full-Bridge Drive (4.5V to 6.5V Battery Voltage) Nominal Input Voltage (4 thru 11 NiMH cells) = = = = 4.8 VDC 6.0 VDC 7.2 VDC 8.4 VDC 5 6 7 8 = = = = 9.6 VDC 10.8 VDC 12.0 VDC 13.2 VDC 9 0 = = 5 6 7 8 = = = = 8.1 VDC 9.0 VDC 9.9 VDC 10.8 VDC 9 0 = = RSVD * RSVD * 450 VRMS 500 VRMS 550 VRMS 600 VRMS 9 0 = = RSVD * RSVD * 900 VRMS 1000 VRMS 1100 VRMS 7 8 9 = = = 1200 VRMS 1300 VRMS 1400 VRMS = = RSVD * RSVD * = = = = 4.5 VDC 5.4 VDC 6.3 VDC 7.2 VDC M Minimum Input Voltage 1 2 3 4 X X X X R X - Z Z RSVD * RSVD * P 1 2 3 4 X L 2 3 X E 1. Choose either the half or full-bridge version by determining your operating conditions. See Recommended Operating Conditions Table. 2. Choose the nominal input voltage you will be using, that is, the voltage where you want efficiency to be highest. Selections are in 1.2V increments to match the 1.2V/cell potential of NiCd and NiMH batteries. If a different type of power source is being used, select the closest nominal voltage. 3. Choose the minimum input voltage where full lamp brightness is needed. For convenience, selections are in 0.9V increments, corresponding to end of discharge potential for NiCd and NiMH cells. Your selection need not correspond to the number of cells selected for nominal voltage input. 4. Specify lamp running voltage. 5. Specify maximum lamp start voltage. 6. Specify lamp running current. Nominal Lamp Operating Voltage 1 2 3 4 = = = = 250 VRMS 300 VRMS 350 VRMS 400 VRMS 5 6 7 8 = = = = 1 2 3 = = = 600 VRMS 700 VRMS 800 VRMS A Maximum Lamp Start Voltage 4 5 6 = = = Nominal Lamp Operating Current at Full Brightness = = = = 2 mARMS 3 mARMS 4 mARMS 5 mARMS S 1 2 3 4 5 6 7 8 = = = = 6 mARMS 7 mARMS 10 mARMS 12 mARMS 9 0 Reserved Mechanical Configuration Factory Assigned RSVD = Reserved for Special Requirements 14 Copyright © 1996 Rev. 0.3 6/96 PRODUCT DATABOOK 1996/1997 P AT E N T P E N D I N G LXM1592/LXM1593 F LOATING O UTPUT D RIVE, C USTOMIZABLE CCFL INVERTER MODULES P R E L I M I N A R Y D A T A S H E E T PHYSICAL DIMENSIONS LXM1592 4.61 (117) 2.60 (66.0) 1.75 1.75 Ø 0.069 (1.75) Location Hole 2-1 0.41 (10.5) L1 1-1 1 T1 1 Connector CN-1 Ø 0.118 (3.0) Mtg. Hole Grounded Both Sides 0.512 (13.0) Connector CN-2 0.287 (7.30) Max. Warning!! High Voltage Present 0.039 (1.0) All dimensions in inches (mm) CN-1 = JST P/N: 05FMS-1.0SP CN-2 = JST P/N: SM02-(8.0)B-BHS-1-TB Copyright © 1996 Rev. 0.3 6/96 15