Guidelines for Improving the RF Immunity of Audio Amplifiers

By Kymberly Schmidt Maxim Integrated Products

F immunity, or RF it to identify the pins susceptible to RF. Pick a This article discusses design susceptibility, is frequency of interest in a system: for example, and measurement tech- Rquickly becoming 2.4 GHz in a WLAN application. Antenna the- niques to optimize the RF as important a design ory states that a trace length of ~1.2 inches (3 immunity performance consideration as PSRR, cm) is a quarter wavelength (λ/4) at 2.4 GHz, of audio amplifier ICs THD+N, and SNR in the and would be a highly efficient antenna. To audio portions of cellular compute this length at other frequencies, use phones, MP3 players and notebook computers. Bluetooth is proliferating as a wireless-serial- l = c / (4f) cable replacement for headsets and micro- phones in mobile applications. Wireless LAN where l = length in meters, c = 3×108, f = fre- (WLAN), using the IEEE 802.11b/g protocol, is quency in Hz. practically standard in PC and laptop comput- Cut a 1.2 in. wire and solder it directly to a ers. The TDMA multiplexing scheme found in pin on the IC. Measure the IC’s RF immunity GSM, PCS and DECT technologies remains a performance at the frequency of interest, 2.4 considerable RF nuisance. Today’s dense RF MHz ±10% in this case. Remove the 1.2 in. environment raises concerns regarding an wire and solder it to a different pin on the electronic circuit’s susceptibility to RF and amplifier. Repeat the RF measurement RF’s impact on the integrity of the overall sys- remembering that it is important to ensure tem. The audio amplifier is a system block that the test setup is identical for each test that can be susceptible to RF. run. Continue in this fashion until the 1.2 in. An audio amplifier can demodulate the RF wire has been placed at every amplifier pin carrier and reproduce the modulated signal and an RF measurement recorded at the fre- and its harmonic components at the output. quency of interest. Finally, measure the IC’s Some of the frequencies fall into the audio RF immunity performance without the 1.2 in. baseband, producing unwanted audible noises wire antenna connected to the pins. (MAXIM’s at the system’s speaker or headphone output. test setup is described later in this article.) To avoid this problem, a system designer must This last test run provides a baseline for fully understand the limitations of the select- the amplifier’s performance. Compare this ed amplifier IC and its respective PCB layout. result to the previous series of test runs. The This article will guide designers to optimize comparison will identify the amplifier pins the RF immunity performance of an audio most susceptible to RF demodulation. Given amplifier at the board level. this data, a PCB can be optimized to reduce the amount of RF energy coupled into the Finding the Source of RF Noise amplifier pins. The key to a successful layout (i.e., high immunity to RF) is first to identify the source The Role of the Capacitor of RF noise coupling. If an evaluation kit is Take, for example, the BIAS pin of the available for the selected audio amplifier, use selected IC. Assume that the BIAS pin

46 High Frequency Electronics High Frequency Design RF IMMUNITY

Figure 1 · (a) Nonideal capacitor model; (b) impedance profile of the Figure 2 · RF immunity test results of nonideal capacitor model. the MAX9750C speaker amplifier; noise floor = –94.4 dBV. exhibits poor RF immunity at the fre- Controlling Noise at the Input Pins quency of interest. The first and most An audio amplifier’s input pins RF immunity performance. obvious PCB consideration would be will always be a source of RF noise Expensive methods such as LC fil- to limit the trace length from the coupling. Ensure that the input trace ters on RF-susceptible amplifier pins BIAS pin to the decoupling capacitor. lengths are shorter than λ/4 for the or low-ESR capacitors added to the If the trace length is optimized and system’s RF signal. A ‘quiet ground application diagram can also be RF demodulation is still a concern, plane’ can also be implemented to implemented at board level. These consider adding a small bypass reduce the amount of RF noise cou- methods are very effective, but costly. capacitor (on the order of 10 pF to pled into the input pins. When laying If the source of RF noise is identified 100 pF) to GND at the amplifier pin. out the PCB, flood a ground plane and understood, expensive solutions The capacitor’s impedance profile can around each input trace of the IC. can be avoided. create a notch filter at the system’s This ground plane will aid in isolat- most sensitive frequency (in this ing any high-RF signals from the Summary case, 2.4 MHz). Refer to the input pins of the selected audio Poor RF immunity performance of impedance profile of a capacitor amplifier. an audio amplifier will impact the model (C1) in Figure 1(b). The MAX9750 combination audio integrity of the overall system design. If C1 were an ideal capacitor, the power amplifier and headphone If the problem is identified and impedance would decrease as the fre- amplifier provides an example of how understood, corrective action can be π quency increased (XC =1/[2 fC]). to control the effects of RF noise. taken to ensure that audible RF Ideal capacitors do not, however, exist Engineering evaluation has identi- demodulation is avoided. In general, in a practical environment. The fied nine pins on the MAX9750 that keep the input, output, bias, and impedance of a nonideal capacitor demonstrate susceptibility to RF: power supply traces shorter than 1/4 model (Figure 1(b)) reaches a mini- INL, INR, BIAS, VOL, BEEP, OUTL+, wavelength for the system’s RF sig- mum at self-resonance [1] and then OUTR+, OUTL–, and OUTR–. A 33 nal. If a higher level of RF immunity begins to increase with frequency. pF capacitor added at the BIAS pin is required, connect a small-value The inductive component takes over improved RF immunity performance capacitor to ground directly at the π (XL = 2 fL) at frequencies greater on average by 3.6 dB. A 3x reduction pin of the IC (even when a larger than f0. To make matters worse, the of input trace length and a ground value capacitor is already connected PCB trace will likely add to the series plane flood around the left, right, and at the pin). Utilize ground plane inductance. This behavior may pro- PC-beep input pins further improved floods near highly susceptible ampli- duce disappointing results if the the RF immunity performance of the fier pins. Finally, physically place capacitor is selected for a filter oper- MAX9750 (Figure 2). high-RF-energy system blocks away ating at frequencies near or above its Figure 2 demonstrates the typical from highly susceptible audio ampli- self-resonance. However, the capaci- RF immunity performance of the fier pins. Equipped with this general tance value can be selected so the MAX9750 IC. External factors such understanding, the unwanted audi- self-resonance minimum will shunt as antenna strength, cable length, ble demodulation “buzz” can usually high-frequency components to GND. speaker type, etc., can also affect the be minimized to an acceptable level.

48 High Frequency Electronics High Frequency Design RF IMMUNITY

Description of Maxim’s the side of the chamber provide RF Immunity Test System power to the DUT and a means to To obtain accurate and repeatable connect an output monitor. In this test results, the device under test case, a digital multimeter (configured (DUT) must be exposed to a known to report units in dBV) monitors the RF field strength. Maxim has devel- demodulated, 1 kHz signal amplitude oped a test method for quantifying in real-time. As the RF carrier wave repeatable RF susceptibility results frequency is varied between 100 MHz by use of an anechoic test chamber, and 3 GHz in predetermined steps, signal generator, RF amplifiers, and a the DMM measurements are record- field-strength sensor. ed. A sample 800 MHz to 1,000 MHz Figure A represents a typical test sweep is reported in Figure C. setup. The DUT in this case is an Figure A · RF immunity test circuit If Maxim’s RF test system is not operational amplifier. The amplifier’s for a typical operational amplifier. readily available (this is the case in non-inverting input is shorted to most situations), there are alterna- GND, using a 1.5 in. wire loop to sim- tives for testing the RF immunity ulate a PCB trace length. A standard tion type supplied by the signal gen- performance of an audio amplifier. 1.5 in. input trace length is selected erator, and the power delivered by the The first alternative utilizes com- so that RF immunity performance RF power amplifier. The modulated monly stocked lab equipment: a high can be compared across several signal is fed to the appropriate power frequency generator, an antenna, an Maxim amplifiers. Note that an input amplifier. The amplifier output is Audio Precision (AP) unit, and a com- trace from the DUT to the input monitored with a directional coupler puter (PC). The high frequency gener- source may act as an antenna. The and a power meter. The subsequent ator simulates the energy present in amplifier’s output is loaded with the RF field is uniformly applied inside an RF rich environment. The genera- rated impedance. Next, the amplifier the test chamber. tor’s high frequency signal is fed to an is placed in a test chamber, which At Maxim, the DUT is placed in antenna positioned near the DUT subjects its contents to a near-uni- the center of the shielded chamber. A (audio amplifier). The DUT’s output form field strength, and a modulated field-strength sensor consistently is then fed to the input of the Audio RF carrier is applied to the test measures a 50 V/m [2] electric field Precision which is controlled by the chamber. The demodulated signal is strength close to the DUT. An RF car- PC. A continuous FFT is recorded monitored at the amplifier’s output rier wave with its frequency varied while the peak voltage at the enve- while Maxim’s RF Test System simu- between 100 MHz and 3 GHz is 100% lope or modulation frequency is lates an RF-rich environment. amplitude modulated with an audio logged. A higher peak voltage at the Figure B illustrates Maxim’s RF frequency of 1 kHz. Access ports on modulation frequency signifies poor Test System. The test chamber is housed in a Faraday cage, isolating the DUT from external fields. The complete test set-up is comprised of the following equipment:

• Signal generator: SML-03, 9 kHz to 3.3 GHz • RF power amplifier: 20 MHz to 1,000 MHz, 20 W • RF power amplifier: 1 GHz to 3 GHz, 50 W • Power meter: 25 MHz to 1 GHz • Parallel wired cell (test chamber) • Electric field-strength sensor • Computer (PC) • DMM (dBV meter)

The PC sets the range of frequencies, modulation percentage, modula- Figure B · Setup for Maxim’s RF immunity test method.

50 High Frequency Electronics test can simply include a GSM cell in audio design support for the phone, operating at frequencies MultiMedia Business Unit. between 800 MHz and 2,000 MHz Kymberly has a bachelor’s degree in with a 217 Hz envelope frequency, in electrical engineering from UCLA. a basic listening test. The cell phone Notes can be placed close to the DUT while 1. At self-resonance, the capaci- a call is placed. If the RF rejection of tive and inductive impedances cancel the selected amplifier is poor, the 217 each other out leaving only a resis- Hz GSM modulation frequency can tive component. The self resonance is potentially be demodulated and your given by ears can monitor the RF rejection of the audio amplifier. 1 f = Once a specific test method is 0 2π LC selected, it is important to ensure the Figure C · MAX4232 test results. test environment and test conditions 2. A typical are well-documented, so they are mobile phone operating in the 1.8 identical for all tested amplifiers. GHz to 2.0 GHz region can emit up to performance of the audio amplifier. This will allow for repeatable and a 100 V/m peak field strength The frequency of the generator can be accurate test measurements which (depending on the distance and orien- varied while the continuous FFT is can be used for comparison. tation of the device to the field- recorded. The changing frequency strength sensor). A standard 50 V/m makes it easier to identify the sensi- Author Information field strength was selected at Maxim tive frequency band of the selected Kymberly Schmidt is a Strategic as it provides useable data from vari- audio amplifier. Applications Engineer at Maxim ous audio amplifiers with varying Alternatively, an RF immunity Integrated Products. She specializes degrees of RF immunity.