抑止TDD noise 的措施
在sysol2 solution中,我们要求对audio信号采用与主地分开的地。例如：用于audio_mic path 的GND_a1靠近MCP(仅有一点与主地连接)，用于audio mic path 的GND_audio_mic接近microphone（不与主地相连）。GND_a2用于audio_spk path 接近 MCP（仅有一点与主地连接）。GND_audio_HP用于audio spk path 接近 earpiece（不与主地相连）。这样可以避免从主地来的干扰噪声直接进入audio path，尤其是TDD burst noise.
在sysol2中, micphone path不是纯净的路径，大多数情况噪声是由此进入。我们就需要注意PCB layout中的audio线是否与其它可能带来干扰的敏感线路并行。同时需要调整优化MICPHONE circuits path和EARPIECE path的一些组件（电容及电阻）的值来抑止TDD noise。将MIC_BIAS线路的滤波措施改为用PI型滤波网络（10UF、100R、22UF）进行滤波，以增加滤波效果。
摆件：MCP靠近射频，因为大约有25根信号线连到射频部分。特别像I&Q,13M时钟线都很关键。IQ信号很容易受到干扰，13MHZ时钟信号会产生EMC问题，这两种线要求尽量让其走线最短。
存储器包括很多高速数据线，所以要求让memory靠近mcp，从而使走线最短。
PMU:最好让PMU靠近电池接口。这样能防止电压下跌和emc。
充电IC需要靠近外部接口，减小V_EXT_CHARGE走线的长度。
32k的晶振需要靠近pmu的晶振输入。
1．音频信号：在基带信号中音频信号是最容易被干扰的信号。它是低频模拟信号。
音频的地要特别的注意（GNDA1,GNDA2,GND_AUDIO_MIC,GND_AUDIO_HP）要用专门的地孔做音频的地。
1.1GNDA1-----PMU的REFGND来的参考地。这个信号和5073的MCP VSS_VB pin(A12) 相连。(保护mic和辅助mic,靠近mcp)
1.2 AGND必须尽可能的靠近bottom connector.
GND_AUDIO_MIC for the normal and auxiliary audio(靠近mic和辅助mic), and GND_AUDIO_HP for the audio amplifier of the hands-free.
1.4 所有的音频信号必须包地保护
1.5 MIC_BIAS, HFAVDD, AVDD 必须远离高速信号和VBAT。
单点接地
GNDA1为音频mic回路
GNDA2为音频speaker回路
PCF50732：
常见问题：
音量方面的问题 音频线需要走宽一点，从而使内阻小一点
回音方面的问题
acoustic指标方面的问题
杂音方面的问题。
MRP Mouth Reference Point
30.1 Sending sensitivity/frequency response
the ratio between the output level, represented by the PCM bit stream at the Digital Audio Interface (DAI) and the input sound pressure in the artificial mouth required to obtain this.
A pure tone with a sound pressure of -4,7 dBPa
送话当量是指当人工嘴在MPR ( Mouth Reference Point ) 上产生 –4.7dBPa从频率100HZ 到 4000HZ的音源送到电话机之听筒送话器上。
30.2 Sending loudness rating
30.3 Receiving sensitivity/frequency response
the ratio expressed in dB between the output sound pressure in the artificial ear and the input level, represented by the PCM bit stream at the Digital Audio Interface (DAI), required to obtain this.
30.4 Receiving loudness rating
30.5 Side tones
30.5.1 Side Tone Masking Rating (STMR)
talker sidetone, expressed by the sidetone masking rating - STMR,
侧音当量是指当工嘴在MPR ( Mouth Reference Point ) 上产生 –4.7dBPa从频率200HZ 到 4000HZ的音源送到电话机之听筒送话器上。同时测量在听筒受话器上的音压相对94dBSPL为0dB参考的量。
30.5.2 Listener Side Tone Rating (LSTR)
how a listener will perceive the background noise picked up by the microphone (listener sidetone rating -LSTR).
背景侧音当量是在仿真量测当电话使用在吵杂环境畤由送话端回授到受话端的量。在MPR ( Mouth Reference Point ) 上产生 -24dBPa( A) 的粉红色噪音( Pink Noise ) 同时量测回授到受话器的量。
The listener sidetone rating defines the effect of interference sound on the voice quality.
The Listener Sidetone Rating (LSTR) is considered a major parameter affecting the user perception of the system.
30.6.1 Echo Loss (EL)
The echo loss is the path loss from the input of the reference speech encoder of the SS to the output of the reference speech decoder of the SS.
either a real time algorithm producing the artificial speech or a pre-recorded tape of artificial speech. Both "male" and "female" artificial speech is required.
第一个10个语音片断输入input of the reference speech encoder，输入信号的The third octave power被计算（允许acoustic echo cancellation）。在第二个10个语音片断后，The third octave power of the echo signal is measured at the digital output of the reference speech decoder of the SS.
The difference between the third octave input power and the third octave output power is entered into the CCITT G.122 TCL algorithm and the acoustic echo loss calculated.
然后重复女声的模拟语音测试，“male”和”female”平均起来给出最终值。
30.7.1 Sending
Idle channel noise:周围环境需要小于30dba
N=0--------63
For RXPGA and TXPGA settings: gain = 20 x log (n/16)
This results in a control range of +6dB to -30dB (and mute).
For RXVOL and SidePGA settings: gain = 20 x log(n/32)
This resulting in a control range of +12dB to -24 dB
[
手机开发]
添加评论 | 评论/阅读(0/418)