Application of voice chips in smart meters

Introduction

With the continuous improvement of intelligence, the application of intelligent instruments is becoming more and more extensive. In order to make the instrument device more powerful and more complete, we added a voice system to the instrument to increase the intelligence of the instrument.

The schematic diagram of the voice system is shown in Figure 1, which is divided into two parts: recording and playback. The system uses a single-chip microcomputer to collect data, and after processing, it is converted into a judgment signal to determine which sound the voice chip plays. At the same time, the single-chip microcomputer provides a control signal to the voice chip to make it work normally. A power amplifier circuit is connected to the output end of the voice chip to make the speaker sound loud enough.

Figure 1 Schematic diagram of the voice system

1. Implementation of recording function

1.1 ISD25 series voice chip

The ISD25 series voice chip [2] is a special product of Winbond. This chip adopts the technology of storing analog data directly in semiconductor memory without A/D or D/A conversion. Therefore, it can reproduce voice, music, tone and effect sound very realistically and naturally, avoiding the quantization noise and "metallic sound" caused by quantization and compression of general solid-state recording circuits.

The ISD25 series voice chips have the following features: easy to use, single-chip storage, convenient recording and playback; high sound quality, natural sound; 60/75/90/120s multiple time levels are available; manual operation/microcontroller control compatible; edge trigger or level trigger can be used when playing; can play in a loop; automatic power saving control, static current is 1μA when saving power; information storage does not require backup power supply; rich addresses, can perform multi-terminal information processing; on-chip information can be stored for 100 years; storage unit can record repeatedly 100,000 times; built-in clock source; playback can be controlled by single-chip microcomputer programming; single power supply operation; PDIP, SOIC and TSOP packages.

1.2 Recording circuit and its principle

The recording circuit is shown in Figure 2. ISD25120 recording can be operated manually or by single-chip microcomputer control. Here we use the manual method. The address generation circuit is used to control the address input of A8~A2. The chip select signal generation circuit is used to control the duration of each recording segment. After /CE becomes low, recording operation is allowed. The chip latches the state of the address line at the falling edge of /CE. When recording, the address terminal provides the starting address, and recording continues until /CE becomes high.

Figure 2 Recording circuit

1.3 Storage of Voice Signals

ISD25120 can process multiple segments of information, which is divided into 600 segments, and the storage time of each segment is 0.2s. Therefore, the storage time corresponding to each address of ISD25120 is 0.2s, and it is difficult to achieve such accuracy by manual operation. Due to the limited I/O ports of the microcontroller, in order to reduce the I/O ports used as address signals during playback, we set the lowest two addresses to zero (A1A0=00) and the highest address to zero (A9=0), so that the minimum accuracy of manual control is 0.2×4=0.8s.

Before recording, the initial address of each segment is allocated first, and the reserved storage time of each segment is about 2s longer than the actual storage time to prevent overflow. When the information ends is set during recording. As long as the rising edge of the /CE terminal arrives, the recording stops. At this time, the ISD25120 chip will set an information end flag EOM in an independent internal EEPROM unit. When the playback is triggered by the /CE pulse, the playback continues until the EOM position. Each row of the ISD chip storage array can be addressed independently, and 4 EOM positioning points are evenly arranged in each row. Since the addressing time of each row is 0.2s, the resolution of EOM is 50ms. In this way, the maximum delay from the end of the information to the output of the EOM signal is 50ms. The rising edge of EOM actually marks the end of the information, so the voice continues to be output from the chip when EOM is at a low level, and stops at the rising edge.

2 Implementation of the playback function

2.1 ST7 general-purpose microcontroller

ST's microcontroller [3] adopts XFLASH technology, EEPROM electrically erasable technology, low power technology, high reliability technology, anti-interference technology, etc., which makes the microcontroller have obvious advantages in performance, structure, and quality. It has the advantages of a wide voltage operating range, rich on-chip peripherals, the same memory structure as a general computer, a variety of programmable low-power modes, and easy development.

We choose the ST7LITE29 microcontroller, which has 8KB of program memory, 384B of RAM, and 256B of data EEPROM with read-out protection function.

2.2 Playback circuit and its principle

The playback circuit is shown in Figure 3.

Figure 3 Playback circuit

The single-chip microcomputer ST7LITE29 collects data through the PB0, PB1, and PB2 ports, processes the collected data, and generates variables to determine which recording the voice chip reads. The single-chip microcomputer sends the corresponding initial address to ISD25120 based on these variables, and then uses PB3 to output a low pulse to /CE, allowing the voice chip to read the specified recording and send the sound through the output terminal (SP+, SP-) to the power amplifier circuit for amplification, and then send the amplified sound signal through the speaker. [page]

2.3 Power amplifier circuit

If the speaker is directly connected to the sound output terminal (SP+, SP-) of ISD25120, the sound produced is too small. Therefore, a power amplifier circuit should be connected to the sound output terminal and then connected to the speaker to make the sound loud enough.

The power amplifier circuit is mainly implemented using MC34119P[4]. MC34119P is an audio amplifier from MOTOROLA. It can increase the voltage swing of the sound output terminal at a low operating voltage (minimum 2V) to achieve the purpose of amplifying the sound. The features of MC34119P are as follows: ① Wide operating voltage range (2～16V); ② Low quiescent current (2.7mA) when the operating voltage is provided by the battery; ③ Power saving control terminal; ④ Very low quiescent current (65μA) when power saving; ⑤ Wide load resistance range (greater than 8Ω); ⑥ Output power consumption of 32Ω speaker is 250mW; ⑦ Low total distortion value (0.5%); ⑧ Gain of sound bandwidth can be adjusted from less than 0dB to more than 46dB; ⑨ Only a small number of external components are required.

The power amplifier circuit is shown in Figure 4.

We take C1=5μF, C3=0.1μF, R1=3.3K, R2=50K, R3=50K.

The voltage gain is

The adjustable range of voltage amplification is about 30 to 60.

The best sound amplification effect is obtained when Av = 54.

2.3 Software Implementation

The flow chart is shown in Figure 5. The program is implemented in Cosmic C and compiled on SofTec Microsystems inDART for ST7.

3 结束语

在仪表中加入语音系统，使仪表装置的功能更强、效率更高、适用性更好。这种设计方法已经运用到棒球速度测试中，它可以读出击球速度并且告知怎样调整击球姿势，极大地方便了新手学习棒球。

参考文献
1 PM50 系列智能语音芯片.中青世纪,2003
2 ISD2560/75/90/120 SINGLE-CHIP,MULTIPLE-M-ESSAGE,VOICE RECORD/PLAYBACK DEVICE 60-,75-,90-,AND 120-SECOND DURATION. Revis-ion 1.0. Winbond Electronics Corp, May 2003
3 张迎新,雷文,陈胜,迟明华等编著.ST7 通用单片机原理及应用.北京航空航天大学出版社，2004
4 MC34119 LOW POWER AUDIO AMPLIFIER. Re-vision 1. Motorola, 1996