Optimizing High-Voltage Common Mode Rejection Performance of the ACPL-x6xL Ultra-Low Power Optocouplers White Paper By: YEE Chee Weng, PENG Jia Abstract The ACPL-x6xL products are optically coupled ultra-low power 10 MBd digital CMOS optocouplers that operate with low power consumption, especially when the light emitting diode (LED) operates at low current. When a current is applied to drive the LED, common mode rejection (CMR) can be affected by leakage paths inherent to the LED. This paper describes the optimization of CMR performance with split limiting resistors at the LED. Introduction Common mode noise is often seen at the system application level where there is a difference in the ground levels of an isolating component’s input control circuitry and output control circuitry. This is especially true when a ground line is floating (device ground connected to a common line). In the ACPL-x6xL family, the common mode rejection (CMR) specification indicates the ability to reject common mode noise. This is also known as common mode transient rejection (CMTR). CMTR describes the maximum tolerable rising/falling rate of a common mode voltage (given in volts per microsecond, V/s). The CMTR specification includes the amplitude of the common mode voltage (VCM) that can be tolerated. The common mode voltage slew rate that the optocoupler can tolerate and hold the correct output state is referred to as common mode transient immunity (CMTI). Common mode noise can be coupled to the optocoupler output by external circuitry. Common mode noise, especially in a high electromagnetic interference (EMI) environment, can adversely affect the output state of the optocoupler through a conductive medium, primarily capacitive and inductive parasitics, Metallic printed circuit board (PCB) tracks that operate at high frequency can couple charge to the LED input pin or to the optocoupler output pin through parasitic capacitors between adjacent metal tracks. It is often difficult to identify the root cause of common mode noise or interference that is introduced by the circuit/system/application or by other forms of external factors that couple noise. When the source of common mode noise is identified, corrective measures are easy to implement by adding decoupling capacitors or filters to the system, or by adding some form of shielding. HVCMR Measurement CMR Measurement Figure 1 shows the experimental test setup to measure the Avago Technologies ACPL-M61L’s high-voltage common mode rejection (HVCMR). Three production samples were randomly selected for the measurements. VDD1 is the transmitter supply voltage used to turn on the LED. Limiting resistor R1 is connected to the LED anode and R2 is connected to the LED cathode. R1 and R2 connected in this common-mode fashion, rather than a single resistor, enhances CMR performance. Figure 2 shows VCM applied to both grounds. The main purpose is to monitor glitches at the output of the Avago ACPL-M61L optocoupler. VCM is applied at a specific frequency in the form of pulses and at a common-mode voltage level (VCM). The optocoupler output will be logic low (VOL) with LED turned on. When the LED is turned off by shorting the anode and cathode to ground, the optocoupler output changes to a logic high (VOH). ACPL-M61L R1 VDD2 VO AMP VDD1 C1 R2 GND1 GND2 VCM Figure1. HVCMR measurement setup On the output side of the ACPL-M61L, there is a photodiode to sense light from the LED. The photodiode signal is amplified by an output amplifier (AMP). VDD2 is the supply voltage of the detector. Decoupling capacitor C1 is between VDD2 and GND2 and filters power supply line noise. A common supply voltage (VCM) is applied between the two grounds, GND1 and GND2. Transient rising edge VCM is adjusted to a particular amplitude, and the rising/ falling edge is adjusted to a particular slope until dipping glitches appear at the output during the common mode voltage’s rising and falling edges as the LED is turned on or until peaking glitches appear at the output during the common mode voltage’s rising and falling edges when the LED is turned off. Under certain circumstances, glitches may be present only at the falling edge of VCM or vice-versa. To determine CMR, two adjustments can be made to VCM. The first adjustment is to lessen the slope of the rising/ falling edge of VCM until glitches disappear completely. The second adjustment is to reduce the VCM level until the glitches disappear completely. CMRH is the slew rate measurement when the optocoupler output has a logic high (VOH) with no dipping glitches. Likewise, CMRL is the slew rate measurement when the optocoupler output has a logic low (VOL) when the peak glitches just disappear. Slew rate is defined in Equation 1. VCM Slew Rate = V/s (Eq. 1) t VCM : Common mode voltage t : Time rate of transient rising edge or transient falling edge Transient falling edge $VCM VCM $t VOH LED off Dip Peak VOL LED on Figure 2. CMR test with transient ramp-up and ramp-down VCM 2 Parasitic Capacitance Effect CMR Performance vs. Resistor_ratio Figure 3 shows the optocoupler’s parasitic capacitors, Ccathode and Canode. Ccathode is larger than Canode by the C_ratio factor. Both the LED cathode and the photodiode anode have larger areas than other functional blocks. Their ground planes (GND1 and GND2) and lead frames are large, therefore contributing to higher capacitance. 5 V operating voltage ACPL-M61L R1 VDD2 Ianode VO VDD1 Canode C1 For Avago ACPL-M61L optimal CMR performance, a Resistor_ratio value of 1.6 is recommended, as shown in Figure 4 (dotted line). This is to minimize Ianode leaking through parasitic capacitor Canode and maximize the LED bias current. Likewise, more current flows through R2 than parasitic capacitor Ccathode. To minimize LED bias current from being diverted away, the Resistor_ratio must be chosen correctly to match the parasitic capacitor impedance characteristics of Ccathode and Canode. 40 Ccathode 35 R2 GND2 Icathode VCM Figure 3. HVCMR measurement setup (parasitic capacitances shown) On the other hand, Canode is parasitic capacitor between the LED anode and the photodiode anode, but the LED anode’s lead frame is smaller than the photodiode anode’s ground plane and lead frame. Therefore, Canode is smaller than Ccathode. The C_ratio is defined in Equation 2. Ccathode C_ratio = (Eq. 2) Canode Canode is between the LED anode and the GND2 ground. Ccathode is between the LED cathode and the GND2 ground. The parasitic capacitors leak current away from the current flowing through the LED during AC operation. Ccathode and Canode behave as an open circuit during DC operation, and there is no DC current leakage. Though resistors R1 and R2 are usually recommended to have equal value, the parasitic capacitors Ccathode and Canode cannot be neglected since the two terminals no longer balance in common mode. Different R1/R2 ratios are used. Since Ccathode is bigger than Canode, the impedance of Ccathode is smaller than the impedance of Canode. More current (Icathode) will leak through Ccathode and less current (Ianode) will leak through Canode during AC operation. In order to minimize the leakage current flowing through Canode and Ccathode, R1 and R2 are placed at the anode and at the cathode of the LED respectively. Their resistance ratio factor (Resistor_ratio) is defined by Equation 3. R1 Resistor_ratio = (Eq. 3) R2 R1 is the limiting resistor between VDD1 and LED anode R2 is the limiting resistor between LED cathode and GND1. 3 CMR kV/Ms GND1 33.3 33.8 33.3 24.4 25.6 24.2 30 25 20 15 10 5 0 1.2 Test conditions LED = 2 mA LED = 1.6 mA VDD1 = 5 V, VDD2 = 5 V R1 = R2 = ±1% Tolerance 1.3 1.4 1.6 1.5 Resistor ratio 1.7 1.8 Figure 4. CMR performance vs. Resistor_ratio (R1 and R2 ±1% tolerance, VDD1 = 5 V, VDD2 = 5 V) With a Resistor_ratio value from 1.4 to 1.6, a CMR of more than 33 kV/s (blue graph ) is achieved with a 2.0 mA LED bias current. With a smaller 1.6 mA LED drive current, good CMR performance of more than 24 kV/s (red graph ) is achieved. R1 + R2 = RTotal (Eq. 4) R1 = 1.6 (Eq. 5) R2 Substitute Eq. 5 into Eq. 4, R R2 = Total (Eq. 6) 2.6 R1 and R2 can be calculated with Equations 4 to 6. In order to achieve excellent CMR performance, R1 = 1.1 k and R2 = 680 are chosen to produce 2 mA of LED drive current (blue graph ), while R1 = 1.4 k and R2 = 910 are chosen for 1.6 mA of LED drive current (red graph ). These values of resistors are widely available and are highly recommended to have ±1% tolerance. The printed circuit board should have good shielding to minimize EMI that may affect CMR performance. R1 –10% R2 +10% 35 kV/μs 30 kV/μs 25 kV/μs 20 kV/μs R1 ±1% R2 ±1% R1 –10% R1 +10% R1 +10% R2 –10% R2 +10% R2 –10% 34.9 kV/μs 33.9 kV/μs 30.9 kV/μs 34.9 kV/μs 25.6 kV/μs 70 60 50 CMR kV/Ms Figure 5 shows ACPL-M61L CMR performance when R1 and R2 have a ±10% tolerance. It can be seen that given the same degree of tolerance in the resistor values, the Resistor_ratio remains 1.6 and the optocoupler CMR performance remains optimized. However, when the resistors do not rise or fall with the same tolerance (example R1 has a -10% tolerance but R2 has a +10% tolerance), CMR performance is compromised because the Resistor_ratio changes from the optimum value of 1.6. 58.7 58.7 58.7 58.7 58.7 57.6 58.7 58.7 58.7 55.4 40 30 Test conditions LED = 2 mA LED = 1.6 mA VDD1 = 3.3 V, VDD2 = 3.3 V R1 = R2 = ±1% Tolerance 20 10 0 1.2 1.3 1.4 1.6 1.5 Resistor ratio 1.7 Figure 6. CMR performance vs. Resistor_ratio (R1 and R2 ±1% tolerance, (VDD1 = 3.3 V, VDD2 = 3.3 V) 15 kV/μs R1 –10% R2 +10% R1 ±1% R2 ±1% R1 –10% R2 –10% R1 +10% R2 +10% 10 kV/μs 5 kV/μs 70 kV/μs 0 kV/μs 60 kV/μs 58.7 kV/μs 58.7 kV/μs 58.7 kV/μs 50 kV/μs 40 kV/μs 58.7 kV/μs 58.7 kV/μs Test conditions : LED = 2 mA LED = 1.6 mA 1.8 VDD1 = 5 V VDD2 = 5 V R1 +10% R2 –10% 30 kV/μs Figure 5. CMR performance versus R1 and R2 tolerance, (VDD1 = 5 V, VDD2 = 5 V) 3.3 V operating voltage The recommended Resistor_ratio value of 1.6 can be applied to a 3.3 V operating supply voltage, as shown in Figure 6. With a Resistor_ratio range of 1.3 to 1.7, even a lower LED drive current of 1.6 mA is able to exhibit superior CMR performance. Both the blue graph and red graph show a CMR of more than 50 kV/s. With a 3.3 V input supply voltage to bias the LED, a lower total resistance value (R1+R2) is needed to drive the LED. More current is concentrated to bias the LED and optimize the CMR performance. Similar to the 5 V operating voltage condition, ACPL-M61L CMR performance is consistent when R1 and R2 with ±10% tolerance are chosen for a 3.3 V operating supply voltage. As shown in Figure 7, a Resistor_ratio value of 1.6 applies when R1 and R2 tolerance changes relatively. Again, CMR performance is not compromised as long as the recommended Resistor_ratio value is set with ±1% tolerance resistors. For product information and a complete list of distributors, please go to our web site: 20 kV/μs 10 kV/μs 31.9 kV/μs 0 kV/μs Test conditions : LED = 2 mA LED = 1.6 mA VDD1 = 3.3 V VDD2 = 3.3 V Figure 7. CMR performance vs. R1 and R2 tolerance, (VDD1 = 3.3 V, VDD2 = 3.3 V) Conclusion Common mode noise occurs at the system application level where there is a difference in the ground level of the optocoupler’s input control circuitry and output control circuitry. It is difficult to identify the root cause of common mode noise or interference. With ultra-low power operation, the ACPL-x6xL family’s CMR performance can be optimized through application of split limiting resistors at the LED. www.avagotech.com Avago, Avago Technologies, and the A logo are trademarks of Avago Technologies in the United States and other countries. Data subject to change. Copyright © 2005-2012 Avago Technologies. All rights reserved. AV02-3053EN - May 25, 2012