CI-D0XGS01J Module Datasheet¶
Module introduction¶
Overview¶
This module is a general-purpose, portable, low-power and high-performance voice recognition module developed for low-cost offline voice application solutions. The model is CI-D0XGS01J, and the main chips are CI1301, CI1302, and CI1303. It can realize the recognition of up to 300 offline command words (different models support different numbers of words).
The module has the following characteristics:
The module is small in size and 30mm in length and width × 40mm, working voltage 3.6V-5.5V, on-board power amplifier, with one microphone, one speaker, one 5V power supply and UART interface. The module can be used directly by inserting the microphone and speaker for power supply, or it can be directly connected to the product main control board through the connector, powered by the 5V power supply of the product main control board, UART communication or GPIO control, without welding. The module includes two 3.5mm screw holes for convenient fixation and installation.
- The main chip of the module supports offline neural network calculation, single microphone noise reduction and enhancement, 360 degree omnidirectional pickup, which can suppress environmental noise and ensure the accuracy of speech recognition in noisy environments. Offline speech recognition using this module does not rely on the network, has small time delay, high performance, and can achieve a high recognition rate of more than 97%. The 10 meter long distance recognition can achieve a response time of up to 0.2S.
- The module can be applied to products with energy consumption level requirements and battery products.
- The module adopts industrial devices with high reliability.
| Module selection | Less than 100 local command words | Less than 200 local command words | Less than 500 local command words |
|---|---|---|---|
| Single microphone offline voice module with socket | CI-D01GS01J | CI-D02GS01J | CI-D03GS01J |
Introduction to module main chip¶
CI1301, CI1302 and CI11303 are AI chips dedicated to speech processing, which support local speech recognition, Chinese, English, Japanese and other global languages. They can be widely used in household appliances, lighting, toys, wearable devices, industry, automobiles and other product fields to achieve voice interaction and control and various intelligent voice solutions.
CI1301, CI1302&CI1303 integrate the brain neural network processor BNPU V3 and CPU core developed by Chipintelli. The main frequency of the system can reach 220MHz. SRAM up to 640KByte is built in, PMU power management unit and RC oscillator are integrated, dual channel high-performance low-power Audio Codec and multi-channel UART, IIC, IIS, PWM, GPIO, PDM and other peripheral control interfaces are integrated. The chip only needs a few peripheral devices such as resistors and capacitors to realize various hardware solutions of intelligent voice products, with high cost performance.
For more details about CI1301, CI1302&CI1303 chips, please click the following link:
☞CI1301&CI1302&CI1303 Chip Datasheet
Module Application Scenario¶
The module can be used as a combination of voice recognition front end+customer hardware main control board scheme, or as a single chip main control module for lamps, toys and other schemes. In application, external microphone and loudspeaker shall be connected, and the power shall be supplied by external 5V power supply.
The CI-D0XGS01J module supports less than 300 offline speech recognition command words, which can be used for electric fans, heating tables, clothes dryers, small appliances, toys, lighting and other products that require less command words.
Module specifications¶
Physical drawing of module¶
The physical module is shown in Figure 4. The speech recognition module is mounted on one side. The main IC includes speech recognition chip CI1302/CI1303, power amplifier, etc. The sound is input from a single microphone and sent to the amplifier driver horn for playing after being recognized by the voice IC. The maximum driving power of the amplifier is 1.5W@8 Ω and 2W@4 Ω.
Module dimension drawing¶
As shown in Figure 5, the module is rectangular in shape and 30 ± 0.3mm in size × 40 ± 0.15 mm, the PCB thickness is 1.6 ± 0.2 mm, and the module height is 7.6 ± 0.4 mm. Users can design the structure according to this size.
Definition of module hardware interface¶
This module has the following functional interfaces:
- Double wire single microphone interface, 2.0mm spacing bus port is used. To ensure good speech recognition effect, it is recommended to use a microphone with sensitivity of - 32 ± 3dB and signal to noise ratio ≥ 65dB. Please click ☞refer to microphone device for more information;
- Double wire single speaker interface adopts 2.5mm spacing bus port. In order to ensure good voice broadcast effect, it is recommended to use speakers with cavities. For more information, please click ☞refer to speaker device;
- Four wire power supply and UART interface adopt 2.5mm spacing bus port. Please refer to Figure 6 for the pin sequence. The UART pin in this interface can also be configured as a GPIO port in addition to serial communication.
The function description of all external pins of the module is shown in Table 2:
| Pin number | Pin name | I/O type | IO drive capacity | IO power on default state | Function definition |
|---|---|---|---|---|---|
| 1 | 5V | P | - | - | 5V power supply |
| 2 | GND | P | - | - | ground signal |
| 3 | UART_ TX | IO,T+U | 4mA | IN,T+U | 1.UART0_ TX 2.PB5 |
| 4 | UART_ RX | IO,T+U | 4mA | IN,T+U | 1.UART0_ RX 2.PB6 |
| 5 | MICL - | - | - | - | - |
| 6 | MICL+ | - | - | - | Positive pole of microphone |
| 7 | SKPL - | - | - | - | speaker output |
| 8 | SKPL+ | - | - | - | Horn output |
Some symbols in the above table are described as follows:
I input
O output
IO bidirectional
P power or ground
T+D tristate plus pull-down
T+U tristate plus pull-up
OUT power-on defaults to output mode
IN power-on defaults to input mode
Module electrical characteristic parameters¶
| Parameter | Condition | Minimum Value | Typical Value | Maximum Value | Unit | Remarks |
|---|---|---|---|---|---|---|
| Module supply voltage | / | 3.6 | 5 | 5.5 | V | NOTE1 |
| Module broadcast status current (normal volume) | 4 Ω 3W horn | / | 70 | / | mA | NOTE2 |
| Module working current | / | / | 40 | / | mA | NOTE3 |
| Monitor state current in quiet environment | 5V power supply | / | 35 | / | mA | / |
| Chip IO interface voltage | / | 3 | 3.3 | 3.6 | V | / |
| Module UART interface voltage | / | 4.5 | 5 | 5.5 | V | / |
NOTE 1: 5V is the typical power supply voltage of the module. If the input voltage exceeds 5.5V, the module may be damaged
NOTE2: The maximum current of the module in the broadcast state can reach 250mA. According to the principle of double margin, it is required to provide a group of power supply with a driving capacity of 500mA for the module
NOTE3: The typical value is silent during the test. The maximum value is the recognition and broadcast state during the test
Module temperature and humidity parameters¶
The temperature and humidity parameters of CI-D0XGS01J are shown in Table 4.
| Parameter | Minimum Value | Typical Value | Maximum Value | Unit | Remarks |
|---|---|---|---|---|---|
| Module working ambient temperature | - 40 | 25 | 85 | ° C | / |
| Module storage environment temperature | - 40 | 25 | 100 | ° C | / |
| Module storage humidity | 0% | / | 5% | RH | / |
Module application¶
Module power on and start¶
When using the module, connect the speaker and microphone, input the 5V power supply to the module through the power supply base, and the module will start after power on. When power on is normal, the speaker will broadcast a prompt audio. At this time, there will be print information on the UART port. The user can connect the UART port to the computer with USB serial port debugging aid and other tools. If the printed information is seen in the serial port debugging software in the computer, the module will start successfully, as shown in Figure 7. Note that the UART interface of the module is a 5V level high-speed serial port. It is no longer necessary to perform level conversion when interfacing with a 5V level system.
The power amplifier chip on the module is powered by 5V power supply. The rated power supply current of 500mA shall be guaranteed for 5V power supply. The power supply shall be stable and the ripple shall be within 30mV.
Module Default Command Word¶
If it is a user’s mass production module, the firmware of the command term specified by the user will be burned before leaving the factory. If not specified by the customer, the module will have its own default firmware with default command words for user testing, as shown in the following figure:
Module default serial port communication protocol¶
The default firmware of this module supports serial port communication protocol for communication with the upper computer. The serial port protocol is extensible and has the following characteristics:
- Complete transmission package, including header and footer, length, verification, message type, and message serial number.
- Support variable length commands for easy expansion.
- Message type (command, notification, reply).
- Command message, configurable, reply to ACK. The notification message has no ACK.
- The message format will be the same as that of the bootloader upgrade, which is distinguished from the bootloader protocol through the header.
- The default baud rate is 9600.
- Note: Only UART0 interface is reserved for the module, and UART0 interface is the print output interface by default. If UART0 is required as the above serial port protocol interface, the code must be modified, and the modification method can be implemented by referring to the serial port protocol part of the documents in the ☞CI130X chip SDK.
- Supported commands: query the protocol version number, query the system version number, set the volume (the volume rating is defined in user_config. h), play the local broadcast, reset commands, etc. The specific protocol format is shown in the following figure:
Example 1:
A5 FC 07 00 A0 91 18 01 55 E0 01 00 00 1B 9B 02 FB is resolved as follows:,
A5 FC：head
07 00: Valid data is 7byte
A0: This is the command word information
91: The command number is 0x91 (this data content is command word data)
18: Packet serial number, the 0x08th outgoing data of this serial port, which is continuously accumulated
01 55 E0 01 00 00: unique data of the current command word
1B: Command word threshold
9B 02: Cumulative Sum
FB: End data
Note: If only command words and thresholds are concerned in the application, only 7 valid data in the blue part can be concerned.
Example 2:
A5 FC 02 00 A3 9A 17 00 B1 05 02 FB is resolved as follows:
A5 FC ：head
02 00: Valid data 2byte
A3: currently notification data
9A: The command number is 0X9A (the data content this time is the voice module content change)
17: This serial port sends data for the 0x07th time, and the value is continuously accumulated
00 B1: Valid data. (This data indicates entering the wake-up state)
05 02: Cumulative Sum
FB: End data
Note: Change the data to notification data, and the user can choose to use this information according to the situation.
For more content analysis data, please refer to the serial port protocol part in ☞CI130X chip SDK. The following figure is a reference screenshot of protocol data:
Software development¶
If the default firmware of the module cannot meet the user’s needs, the user can develop the software by himself and modify the command words, broadcast voice, serial communication protocol and other functions of the module.
The software development process mainly includes the following steps:
- SDK development kit material download
- Model making (language model+acoustic model)
- Speech synthesis
- The command word information table is associated with the audio file
- Firmware packaging
For detailed development process, please click ☞CI130X chip SDK.
Firmware burning¶
Preparation before burning¶
Before burning the module, users need to prepare the following items:
- Modules to be burned
- USB to serial port tool
- Firmware burning tool (pack_update_tool. exe)
- Firmware information (files in *. bin format)
- Microphone with 2.0mm pitch
- 2.5mm pitch horn
- Several DuPont lines
Hardware connection and burning¶
As an example of the USB to serial tool shown in the above figure, the power, ground and serial port transceiver pins of the USB to serial port need to be connected with the corresponding pins of the module before burning (note that the RXD and TXD of the USB to serial port correspond to the UART0_TX and UART0_RX of the module respectively). The connection method is shown in the following figure. The wiring diagram of module burning and the silk screen diagram on the back of the module are shown in the figure below.
Open the firmware burning tool (this tool can find PACK_UPDATE_TOOL.exe in the CI130X_SDK tools directory of the SDK development kit), select the corresponding model according to the chip, click the firmware upgrade button, select the prepared firmware file, and find the serial port number assigned to the USB serial port tool by the corresponding computer. After the preparation is ready, the chip will enter the programming mode through instant power on, and the firmware can be downloaded at this time. If the USB to serial port tool is not recognized on the computer, please install the corresponding driver on the computer.
Function test after burning¶
After the firmware is successfully burned, it is recommended to perform a functional test on the module to verify whether the firmware is successfully burned. During the function test, plug the microphone and loudspeaker into the module to be tested, power on and observe whether it can be normally powered on and broadcast, and use the wake-up word and command word to test whether it can be normally wakened up and recognized. If it can work normally, the module functions normally and the burning is successful; Otherwise, the burning fails, and further investigation is required.
Possible problems and solutions in use¶
This chapter lists the problems that may be encountered in the use of some modules and the corresponding solutions.
- The module cannot burn and update firmware.
After the above problems occur, please check the following operation points:
- Whether the serial port pins are connected correctly, whether the TX and RX are connected reversely, whether the USB to serial port tool drive at the computer end is normal, and whether the PC burning tool has selected the correct serial port number;
- If the above two points are correct, the module cannot be burned, and a multimeter is required to measure whether the module power supply voltages of 5V, 3.3V and 1.1V are correct. Refer to the figure below for each hardware measurement point. If there is a problem with the voltage or crystal, please replace the module or repair the module hardware. If there is no problem in the above inspection, please contact our technical support personnel for help.
- The module is burned and there is no broadcast after power on.
After the above problems occur, please check the following operation points:
- Confirm whether the burning firmware matches the board;
- Confirm that the horn is correctly connected and the power supply is normal;
The oscilloscope is used to measure the voice output test point of the main chip. If there is no output, check whether the firmware is correct. If there is output, check whether the power amplifier components on the module are welded abnormally. If the power amplifier is abnormal, replace it and test again. The measuring points are shown in the figure below. If there is no problem in the above inspection, please contact our technical support personnel for help.
- After the module is burned, the command word is announced but not recognized after power on:
- Check whether the connection between the microphone and the socket is intact;
- Check whether the positive and negative pole directions of the microphone are consistent with the markings on the module board and are not inserted reversely;
- Use a multimeter to measure whether the MICBIAS pin corresponding to the main chip is about 2.8V, and use an oscilloscope to measure whether the microphone input pin has an input voice waveform (the voltage of each oscilloscope cell is adjusted to the 100mv gear). If the signal is normal, consider whether the firmware is correct, and if the signal is abnormal, observe whether the board hardware has physical damage. The measuring points are shown in the figure below. If there is no problem in the above inspection, please contact our technical support personnel for help.
Other application precautions¶
Because the ESD level of CI1302 and CI1303 chips is high and the module is designed to facilitate user expansion, ESD devices are only designed in the microphone position on the module. ESD devices can be added to products with high ESD requirements. Users are recommended to wear anti-static bracelets or anti-static gloves and finger cots during inspection and welding. Please reserve ESD protective devices at the corresponding backplane connector to ensure the quality and reliability of the product.
When using, pay attention not to connect the microphone, speaker and power supply serial port incorrectly, and pay attention to prevent short circuit at the back test point.
Note that the serial port of this module is a 5V level standard. Please use a 5V level serial port for communication. The user can use the USB to serial port tool to debug the developed software. During debugging, add the serial port printing command to the corresponding position in the SDK software, generate firmware after compilation and burn it, and then debug and verify.
Manufacturing Guide, Storage and Packaging Ordering Information¶
Production Guide¶
The module adopts an integrated interface with terminals, which is simple and convenient to produce. The microphone, speaker and power communication terminal can be used by inserting them into their corresponding terminals. The board is designed as a fool proof interface, and there is no possibility of incorrect insertion between the three terminals. Please wear electrostatic gloves and electrostatic bracelets during the insertion operation, and use appropriate force to ensure that the connector is inserted in place. Open the vacuum electrostatic packaging bag before starting the assembly.
Storage conditions¶
The module is vacuum packed, so the requirements for storage conditions are not high, and it can be stored in a non condensing atmospheric environment of<40 ℃/90% RH. The humidity sensitivity level MSL of the module is level 3. After the vacuum bag is unsealed or leaks, please control according to the humidity sensitivity level 3.
Packaging Ordering Information¶
| Product model | Packaging method | Number of modules per pallet | Number of modules per package | Number of modules per box |
|---|---|---|---|---|
| CI-D01GS01J CI-D02GS01J CI-D03GS01J |
Tray+electrostatic bag+carton | 40pcs | 10 trays 400pcs in total | 3 bags 1200pcs in total |
Procurement and technical support¶
If the user wants to purchase our product samples, please click on ☞Batch Purchase to obtain more information.
If you want to obtain technical support, please log in to ☞Chipintelli Speech AI Development Platform.