KT148A Chip User Manual

Document Status： Document Identification： Publicly Available

[] Draft Current Version： V5
[√] Officially Released Author： Qingyue Electronics
[ ] Under Revision Completion Date： 2023-12-26

1
 Version History

Version Date Reason

V1.0 2021/12/12 Internal Beta Version and Optimized Version

1.Corrected descriptive errors in the documentation.

2.Enhanced the description of chip power consumption -- see section 1.3 for
V1.1 2022/05/12 details.
3.Added new frequently asked questions -- see section 8 for details.
V2 2022/07/04 1.Updated some descriptive errors, formatting, and images.

1.Updated some descriptive errors, formatting, and images.

V3 2022/08/24 2.Updated the FAQ compilation in section 8.
1.Updated the formatting of the FAQ compilation in section 8.
V4 2023/03/06 2.Updated some descriptive issues; no changes, deletions, or additions have been
made to the functionality.
1.Updated the FAQ compilation.
V5 2023/12/26 2.Added F4 command, primarily to address the insufficient of 233 voice address
issue, see "section 4.5" for details.

2
 『 Contact US 』for Technical Support and Sales Service
/*********************************************************************

EMail: [email protected]

WhatsApp: +86 159 0279 7635

WeChat: JAJB-Go

Tel: +86 159 0279 7635

*********************************************************************/

- Feel free to call for consultation at any time.

- Welcome to contact us anytime to obtain documentation or communicate.

Linkedin: https://www.linkedin.com/company/blevoice/
Blogger: https://blevoice.blogspot.com
Tumblr: https://www.tumblr.com/blog/blevoice
TikTok: https://www.tiktok.com/@blevoice
[BLE Voice Store]: https://www.aliexpress.com/store/1104656113

1. To request samples, please search for key words such as "KT6368A" directly on
AliExpress, enter our official store [BLE Voice Store] for solutions.
2. For each product, through rigorous development, debugging, and testing processes,
we provide comprehensive technical documentation including user manuals, FAQs,
schematics, Datasheet, Reference Code, debugging guides, critical notes and ect.
3. As our products expand into broader applications and industries, all related
technical resources are continuously updated to ensure relevance.
4. Full technical packages for every product are permanently hosted on Google Drive,
offering free lifetime access, and the corresponding download links are highlighted on
the detail pages of each product in our AliExpress store.
5. While our documentation may appear on third-party platforms, forums, or
websites (shared by enthusiasts, partners, or industry peers), the latest versions are
exclusively available on Google Drive.
 Content

1. OVERVIEW .............................................................................................................................................................................. 4

1.1 I NTRODUCTION ..................................................................................................................................................................... 4

1.2 F UNCTIONAL S UMMARY ........................................................................................................................................................ 4
1.3 C HIP P OWER CONSUMPTION D ESCRIPTION ............................................................................................................................ 4

2. HARDWARE PARAMETER DESCRIPTION .............................................................................................................................5

2.1 P IN DESCRIPTION ..................................................................................................................................................................5

2.2 GPIO E LECTRICAL C HARACTERISTICS ..................................................................................................................................... 6
2.3 GPIO I NTERNAL CHARACTERISTICS ........................................................................................................................................ 6
2.4 R EFERENCE CIRCUIT .............................................................................................................................................................. 7

3. ONE-WIRE COMMUNICATION PROTOCOL ......................................................................................................................... 7

3.1 COMMUNICATION FORMAT ....................................................................................................................................................7

3.2 COMMUNICATION COMMANDS ..............................................................................................................................................8
3.3 E XAMPLE WAVEFORMS FOR O NE -W IRE S ERIAL COMMUNICATION ...........................................................................................9
3.4 D ETAILED E XPLANATION OF THE START S IGNAL FOR O NE -W IRE S ERIAL COMMUNICATION ........................................................ 9

4. REFERENCE EXAMPLE CODE .............................................................................................................................................. 10

4.1 COMMUNICATION E XAMPLE : F2—LOOP P LAYBACK .............................................................................................................. 11

4.2 COMMUNICATION E XAMPLE : F3—COMBINATION P LAYBACK ................................................................................................. 11
4.3 COMMUNICATION COMMAND F0 DESCRIPTION [U LTRA -LOW POWER M ODE ] ....................................................................... 12
4.4 COMMUNICATION D ETAILS AND N OTES ................................................................................................................................13
4.5 COMMUNICATION COMMAND F4 DESCRIPTION [V OICE A DDRESS E XTENSION ] [O NLY FOR CUSTOM V ERSIONS ] ...................... 13

5. USER VOICE REPLACEMENT ............................................................................................................................................... 13

5.1 S TEP 1 — G ENERATE AUDIO F ILES (R ECOMMENDED FORMAT: WAV) ...................................................................................13

5.2 S TEP 2 — CONVERT A UDIO F ILES TO F1A F ORMA ...............................................................................................................14
5.3 S TEP 3 — D OWNLOAD AUDIO VIA S ERIAL TOOL .................................................................................................................. 15

6. DETAILED STATEMENTS ...................................................................................................................................................... 17

7. ORDERING INFORMATION ................................................................................................................................................. 18

7.1 R EFERENCE S CHEMATIC .......................................................................................................................................................18

7.2 PACKAGE D IMENSIONS ........................................................................................................................................................18

8. FREQUENTLY ASKED QUESTIONS (FAQ) ........................................................................................................................... 19

3
1. Overview

1.1 Introduction
The KT148A is a 32-bit DSP voice chip in a standard SOP8 package. It comes with a built-in 420KByte voice
storage space, supporting a maximum voice length of 420 seconds, multiple voice segments, and can directly
drive a 0.5W speaker. It also allows users to replace the voice content themselves.

The current advantages of this chip are as follows:

1.High cost-performance ratio. Compared to traditional OTP chips, the improved manufacturing process
significantly reduces costs, and the chip can be reprogrammed.
2.Users can replace the voice content themselves via the serial port on a computer, without the need for other
expensive tools.
3.The chip has sufficient built-in space to pursue higher sound quality effects.

1.2 Functional Summary

1. 32-bit DSP voice chip, 5MHz internal oscillator, PLL up to 120MHz.
2. Operating voltage 2.0~5.5V; note that there are detailed hardware differences when supplying below 3.3V.
3. Standby power consumption is less than 25uA; there is also a low power mode of 1.5uA.
4. 16-bit PWM output, can directly drive an 8 ohm / 0.5W speaker.
5. Provides tools for users to convert audio formats themselves, which is very simple for all users.
6. Supports up to 200 seconds at 16KHz and up to 400 seconds at 8KHz.
7. Customers can replace the voice content directly via the serial port using the tool software on the computer.
8. Supports one-wire serial port control, and various protocols can be customized according to needs.
9. Supports 224 segment addresses, and can be expanded for more needs; see section 4.5 for details.
10. Has hardware IIC interface, UART interface, and can customize various functions for customers.
11. The chip has the ability to be reprogrammed, so there is no need to worry about incomplete testing leading to
batch issues at any time.

1.3 Chip Power Consumption Description

State Detailed Description
Playback State 1.The power consumption during playback depends on the external speaker load, generally
maintained between 150mA and 200mA.
2. The chip itself consumes about 8mA during playback.
Idle State 1. When the chip is not playing or has finished playing, it will automatically enter standby
mode.
2. The chip's own power consumption is about 25uA, which is sufficient for general application
scenarios.

4
Ultra-Low 1.The chip also has a command to enter ultra-low power consumption, i.e., shutdown state,
Power State with a power consumption of 2uA.
2. For details, see section 4.3 of the manual.

2. Hardware Parameter Description

The chip adopts an SoC solution, integrating a 16-bit MCU and a dedicated aDSP for audio decoding. It uses
hardware decoding, which further ensures system stability and sound quality. The compact package size also
makes it easy to embed into other products.

2.1 Pin Description

Pad
Pad Name ATTR Description
No
1 GND Power Power Ground (GND)
Busy Signal: Low during playback, high when idle.
2 PB9 I/O Multiplexed with the serial port download, connected to the RX pin of the
MCU or USB-to-TTL converter.
[ONELINE-DATA]: One-wire serial input.
3 PB1 I/O Multiplexed with the serial port download, connected to the TX pin of the
MCU or USB-to-TTL converter.
4 PB0 I/O KEY Button: Test IO—trigger playback by grounding.
Non-polarized, supports 8 ohms 0.5W.
5 SPK1 Out Note: Pull low during power-up to enter the voice download state. See
section 5 for details.
6 SPK2 Out Non-polarized, supports 8 ohms 0.5W.
7 VDDIO Power Internal LDO Output, Maximum 100mA.
8 VBAT Power Power Positive

5
1、Power Supply Notes：--- Refer to section 8.23 of the FAQ for details
(1)It is recommended to supply 5V to the chip. Higher voltage increases the driving power of the speaker,
resulting in louder sound.
(2)For lithium battery scenarios [3.3V--4.2V], connect the positive terminal directly to Pin 8 (VBAT).
(3)For dry battery scenarios, connect the positive terminal directly to Pin 8 (VBAT).
(4)For ultra-low voltage scenarios, such as 2V--3V coin cell batteries, short Pin 7 and Pin 8, then connect to the
positive terminal.
2、Quick Test Instructions：
==》Pin 4 is the test pin. Trigger it to ground using a button or tweezers, and it will play the next voice.
==》After connecting the speaker and powering on, you can test by grounding Pin 4.了

2.2 GPIO Electrical Characteristics

IO Input Characteristics
Symbol Parameter Min Typ Max Unit Test Condition
VIL Low-Level Input Voltage -0.3 - 0.3*VDD V VDDIO=3.3V
VIH High-Level Input Voltage 0.7VDDIO - VDDIO+0.3 V VDDIO=3.3V
IO Output Characteristics
Symbol Parameter Min Typ Max Unit Test Condition
VOL Low-Level Output Voltage - - 0.33 V VDDIO=3.3V
VOH High-Level Output Voltage 2.7 - - V VDDIO=3.3V

2.3 GPIO Internal Characteristics

Port General High Internal Internal Comment
Output Drive Pull-Up Pull-Down
Resistor Resistor

PB0,PB1 8mA 64mA 10K 60K 1、PB0 & PB1 default pull down
2、internal pull-up/pull-down
PB9 8mA 10K 60K
resistance | accuracy ±20%

6
2.4 Reference Circuit

Note: After testing the actual sample, if no voice replacement is needed, the button and resistor on Pin 5 can be
omitted.

3. One-Wire Communication Protocol

The chip currently supports a one-wire serial communication protocol, using high and low pulse widths to
determine 0 or 1.

3.1 Communication Format

1. Pull the DATA line (ONELINE) low for 6ms, then send 8-bit data, starting with the least significant bit (LSB)
followed by the most significant bit (MSB).
2. Use the ratio of high and low pulse widths to represent each data bit.

Note:High pulse must come before the low pulse.

7
Recommended pulse widths: 200us (high) : 600us (low).
Note that valid ranges: Narrow pulse = [150us--400us], Wide pulse = [500us--1000us].
Note that use a 3:1 or 1:3 pulse ratio to ensure communication stability.
Note that ensure the initial 6ms low pulse. During initial debugging, use an oscilloscope or logic analyzer to
verify the timing, with an acceptable error of around 10%.
1. 0x00H to 0xDFH: Voice address codes or function codes.
2. 0xFEH: Voice stop code. Sending this command stops voice playback.
3. The interval between codes must be greater than 20ms.

3.2 Communication Commands

NO. Code Voice Table

0 0x00 Reserved
1 0x01 Voice 1 (voices can interrupt each other)
2 0x02 Voice 2
3 0x03 Voice 3
...
15 volume levels: 0xE0 is the minimum volume (0), 0xEF is the maximum volume
(default maximum volume on first power-up).
After sending the volume command, the KT148A will automatically save and apply
224 0xE0~0xEF
the setting. Upon power cycling, it will revert to the default maximum volume.
(Note: To adjust volume, send the volume code first, wait 20ms, then send the voice
code.)
Shutdown.Enter ultra-low power mode = 1.7uA (see section 4.3 for details). After
225 0xF0 entering this state, to control the chip again, pull the DATA pin low for 40ms to wake it
up first, then you can send commands to control as normal.
Loop Command — See section 4.1 for details.
242 0xF2 If you need to loop a specific voice, send the voice code first, wait 20ms, then send the
loop code.
Continuous Playback — See section 4.2 for details.
Format: F3H + Voice Address A, Voice Address B, Voice Address C, ...
During playback of Address A, subsequent codes will not interrupt. After finishing A,
243 0xF3 it will play B, then C, and so on.
A 20ms delay is required between F3 and the voice addresses. A maximum 40ms
delay is required between one set of continuous addresses and the next (where "F3 +
voice address" is considered one set).
244 0xF4 Voice Address Extension Command — See section 4.5 for details.
Re s et Co mman d,t he ch ip wil l re se t imme di ate ly upon re ce iv in g this
245 0xF5
comm and.
Sto p Co mm a nd . If re c e i ve d du ri n g pl ayb ac k , th e cu rre nt voi c e wi ll st op
254 0xFE
im me d i ate l y. Th e vo i ce wil l al s o st op au to mat i ca ll y afte r fi ni s h i ng pl ayb a ck .

8
3.3 Example Waveforms for One-Wire Serial Communication
These waveforms are provided to assist customers in debugging, as there are limited debugging techniques for
one-wire serial communication.
Data Reference Waveforms from Logic Analyzer
Corresponding to
0x01

Waveform Meaning: Start signal + 0000 0001 (LSB first).

Corresponding to
0x03

Waveform Meaning: Start signal + 0000 0011 (LSB first).

Corresponding to
0x04

Waveform Meaning: Start signal + 0000 0100 (LSB first).

Key Points:
1. The start signal must not be shorter than 6ms. Strictly adhere to this when writing the program, as the start
signal is very important.
2. If issues arise, use debugging tools,use debugging tools immediately—preferably a logic analyzer.
3. For details on the start signal, refer to section 3.4.

3.4 Detailed Explanation of the Start Signal for One-Wire Serial Communication
In practical use, some customers may still have questions about the start signal. Here, we provide a detailed
explanation.
The start signal serves two purposes:
1. First Purpose: Command Start Data Recognition,The KT148A chip internally configures the ONELINE IO pin
with a weak pull-up and capture functionality.
==》When a falling edge is detected, it indicates that data is arriving. The chip then waits for a rising edge, and if
the low pulse duration meets the 6ms requirement, the subsequent data is considered valid. This is the
underlying principle.
2. Second Purpose: Wake-Up Function for the KT148A Chip,The chip has a mechanism to enter standby mode
after 5 seconds of inactivity. Therefore, when data arrives via the one-wire serial interface, the chip first wakes
up and then receives the data.

Start Signal Low Pulse Duration,The low pulse duration of the start signal must be between 5.5ms and 6.5ms.It
is recommended to use 6ms, with an acceptable error margin of 10%.
9
Test Details:
Can trigger playback

Cannot trigger playback

in standby mode

Can trigger playback

4. Reference Example Code

/*******************************************************************************
- Function Description: One-wire data transmission [Note: Use an oscilloscope or logic analyzer to check pulse
width during debugging]
- Module: Internal
- Parameters: dat = Data to be sent
- Return: None

First, pull the DATA line (ONELINE) low for 6ms, then send 8-bit data, starting with the least significant bit (LSB)
followed by the most significant bit (MSB).
********************************************************************************/
#define IO1_HIGH() JL_PORTA->DIR &= ~BIT(6);JL_PORTA->OUT |= BIT(6); //PA6 output high level
#define IO1_LOW() JL_PORTA->DIR &= ~BIT(6);JL_PORTA->OUT &= ~BIT(6); //PA6 output low level

void oneline_send_one_data(u8 dat)

{
u8 i = 0 ;
IO1_LOW() ;/*Pull the bus low first*/
udelay(5000) ;/*Delay 6.26ms --- This is the start signal for communication*/
CPU_SR_ALLOC();//Disable interrupts (chip-specific, interface may vary)
OS_ENTER_CRITICAL();//Disable interrupts
for(i=0 ; i<8 ;i++){
if(dat & 0x01){
IO1_HIGH() ;
udelay(500);/*Delay 600us*/
IO1_LOW();
udelay(170) ;/*Delay 200us*/
}else{

10
 IO1_HIGH() ;
udelay(170);/*Delay 200us*/
IO1_LOW();
udelay(500) ;/*Delay 600us*/
}
dat = dat >> 1 ;/*Send the most significant bit first, followed by the next, and so on*/
}
IO1_HIGH() ;/*Pull the pin high when idle*/
OS_EXIT_CRITICAL();//Enable interrupts
/*Note: Ensure the signal pin is pulled high during chip initialization. */
}
Example of function call：
oneline_send_one_data(0x05);//Corresponds to playing the 5th voice segment

4.1 Communication Example: F2—Loop Playback

void oneline_test_CMD_F2(void){
oneline_send_one_data(0x01) ;
udelay(20) ;/*Delay 25ms*/
oneline_send_one_data(0xF2) ;
udelay(20) ;/*Delay 25ms*/
} Effect: The first voice segment will loop playback continuously. To stop the playback, send the FE
command.

4.2 Communication Example: F3—Combination Playback

void oneline_test_CMD_F3(void){
oneline_send_one_data(0xF3) ;
udelay(20) ;/*Delay 25ms*/
oneline_send_one_data(0x01) ;
udelay(20) ;/*Delay 25ms*/
oneline_send_one_data(0x02) ;
udelay(20) ;/*Delay 25ms*/
oneline_send_one_data(0x03) ;
udelay(20) ;/*Delay 25ms*/
} Effect: The chip will play voice segments 1, 2, and 3 in sequence. To stop playback, send the FE command.
During combination playback, if new combination playback data is received, the current playback will stop, and
the new commands will be executed.
==> For the F3 command, a maximum of 30 voice segments can be combined.Example Waveform:

11
The test waveform corresponds to sending 0xF3 0x01 0x02 0x03 0x04, which triggers combination playback of
voice segments 1, 2, 3, and 4.
The delay between each data packet is 27.46ms.
1. The chip has a wide tolerance for recognizing the F3 combination playback command.
2. After receiving the F3 command, the chip waits for 100ms. If a voice command (e.g., 0x01) is received within
this 100ms, it will wait another 100ms. If no further commands are received within 100ms, it will start
processing the combination playback.
3. Recommended: A delay of 25ms between commands is sufficient, as shown in the logic analyzer example
above.
Video Demonstration Link: https://haokan.baidu.com/v?vid=17612806804625327051,For more details, refer to
the following documentation:

4. After receiving all commands, the chip will start combination playback. The interval between each voice
segment is approximately 50ms.

4.3 Communication Command F0 Description [Ultra-Low Power Mode]

1. Specifically used to make the chip enter ultra-low power mode. After playback ends, the chip automatically
enters standby mode with a power consumption of 25uA.
2. If even lower power consumption is required, send the F0 command. This can be done in any state.
3. After receiving the F0 command, the chip will enter shutdown mode, with power consumption maintained at
around 1.7uA.
4. To make the chip receive commands again, follow these steps:
- Pull the dat signal low for 10ms (this wakes up the chip).
- Then pull it high for 30ms to wait for the chip to initialize.
- A total of 40ms is sufficient, after you can then send the corresponding commands.
5. During debugging, you can observe the BUSY indicator. After entering ultra-low power mode, the chip
behaves as if powered off:
- RAM is cleared.
- All IO pins are in a high-impedance state.
- The ONELINE pin is weakly pulled up, waiting for a falling edge to wake up.
- After wake-up, the BUSY pin will output a fixed high level.
6. After waking up from ultra-low power mode, the volume setting will reset to the default maximum state,
similar to a power cycle.

12
4.4 Communication Details and Notes
1. After the control MCU powers on, the communication IO (i.e., the ONELINE pin) must be set to high-level
output as soon as possible.
2. It is recommended to connect a 1K resistor in series between the MCU and the KT148A chip to adapt to
different voltage levels.
3. During playback, the voltage on Pin 2 (PB9) of the chip is approximately 3.1V. Once the chip enters idle state,
this IO voltage drops to around 2.8V due to the chip's internal mechanism for entering low-power standby mode
when idle. In standby mode, the IO pin is set to a high-impedance state.
4. After powering on, the KT148A chip requires approximately 40ms to be ready to receive commands. Therefore,
it is recommended to wait at least 50ms before sending commands to the KT148A.

4.5 Communication Command F4 Description [Voice Address Extension] [Only for

Custom Versions]
The F4 command is introduced to address the issue of insufficient voice addresses when the number of voices
exceeds 223. If such a large number of voices is not required, this section can be ignored.

Example:
1. To play voice at address 291, send: F4 01 23 【 01 = 0x01 (high byte)】 【23 = 0x23 (low byte) 】
==》F4 is the identifier, and 01 23 forms 0x123 = 291, representing the 291st voice.
2. To play voice at address 1176, send: F4 04 98
==》send: F4 04 98，04 98 forms 0x498 = 1176, representing the playback of the 1176th voice.
3. The F4 command must be exactly 3 bytes long (no more, no less). After receiving the complete F4 01 23
command, the chip will immediately execute the playback.
4. After receiving the F4 command, the chip will wait for 100ms. If a voice command like 0x01 is received within
this 100ms, it will wait another 100ms.
==》If the second address data (0x23) is received, it means the full address (0x0123) has been received, and the
chip will immediately play the voice at address 291.
== 》 If no address data is received within 100ms, the communication is considered failed (e.g., only 0x11 was
received).
5. If the voice address does not exceed 233, this extension command is not needed.

5. User Voice Replacement

The chip supports voice replacement via a serial port. The required tools are as follows:
1. PC Audio Compression Software — Provided by us, or you can record your own.
2. PC Download Tool Software — Provided by us.
3. USB-to-TTL Serial Tool — Recommended to use the CH340G USB-to-TTL module.

5.1 Step 1 — Generate Audio Files (Recommended Format: WAV)

Refer to the following document path for details:

13
5.2 Step 2 — Convert Audio Files to F1A Forma

1. It is recommended to use the WAV format for the original file. Use our provided tool software, and the output
format can only be F1A or F1B.
14
2. The sampling rate and bit rate can be flexibly chosen. Generally, a 16K sampling rate and 6K bit rate are
recommended,and the effect will be very good. If space allows, feel free to experiment. If there are many voice
files, a 8K sampling rate and 8K bit rate are also effective.
3. Add the file, select the output directory.Finally, simply click "Convert" to proceed.A progress bar will be
displayed during the process.
4. The software will convert the original audio file into an xxx.f1a format file. Note that this file cannot be played
on a computer; it can only be downloaded into the chip for decoding and playback. The total size of the
converted file must not exceed 420Kbyte (note: this refers to the size after conversion).

Important Details:
1. Some customers reported conversion failures on XP systems. It is recommended to use Windows 7 or
Windows 10.
2. When selecting the format, only F1A can be chosen. The sampling rate can only be 8K or 16K; 32K is not
supported.
3. The final generated file must have the .F1A extension. Other formats are not acceptable.
4. If the computer reports a virus, ignore it. Please disable the annoying 360 or any antivirus software, as they
may block the process due to lack of "protection fees". If the software is missing from the folder, it may have
been deleted by antivirus software.
5. If an F1B format file is generated, it can still be used normally. This occurs if you select a 32K sampling rate,
which produces larger files with better quality, but this is unnecessary.

5.3 Step 3 — Download Audio via Serial Tool

15
1. You must short Pin 5 to ground before powering on the chip. You will hear a "beep" from the speaker, after
which you can release the short.
2. At this point, the chip automatically enters download mode. It is recommended to use the CH340G
USB-to-TTL module.
3. Connect the serial module to the computer:USB-to-TTL TX to chip Pin 2,USB-to-TTL RX to chip Pin 3,then
connect the ground (GND) pins.
4. Click "Open Serial Port" on the host software, with the baud rate fixed at 115200. import the file, and click
"Download All." Wait for the process to complete.
5.Note that after the download is complete, you can trigger Pin 4 to ground to test and listen to the downloaded
audio file.

A complete test setup is shown in the actual image above.

16
6. Detailed Statements
 Development Prerequisites
KT series products will provide comprehensive development templates, drivers, and application documentation
to facilitate user experience. However, users are expected to be familiar with the hardware platform of their
designed products and have knowledge of C programming.

 EMI and EMC

The mechanical structure of KT series chips inherently results in EMI performance differences compared to
integrated circuit designs. The EMI performance of KT series chips meets the requirements of most applications.
If users have special requirements, they must consult with us in advance.
The EMC performance of KT series chips is closely related to the design of the user's baseboard, especially the
power circuit, I/O isolation, and reset circuit. Users must fully consider these factors when designing the
baseboard. We strive to improve the electromagnetic compatibility of KT series chips but do not provide any
guarantees for the EMC performance of the user's final application products.

 Right to Modify Documentation

Shenzhen Qingyue Electronics Co., Ltd. reserves the right to modify the documentation related to KT series
products at any time without prior notice.

 ESD (Electrostatic Discharge) Protection

Some components of KT series products have built-in ESD protection circuits. However, in harsh environments, it
is recommended that users implement additional ESD protection measures in the baseboard design, especially
for power and I/O circuits, to ensure stable operation.
When installing KT series products, ensure to discharge static electricity from your body for safety, such as by
wearing a properly grounded anti-static wristband or touching a grounded metal object like a water pipe.

17
7. Ordering Information
By default, products are shipped in tubes, with 100pcs per tube, 10K per box, and 100K per carton.

7.1 Reference Schematic

7.2 Package Dimensions

The package is consistent with common SOP8 chips, such as the LM358 and other SOP8-packaged chips. For
details, refer to the chip specification document.

18
8. Frequently Asked Questions (FAQ)
If you randomly set up a circuit yourself, we do not provide technical support. It is recommended to use our
test baseboard to filter out issues first. Otherwise, strange problems may arise, making it impossible to
provide support. Additionally, if you encounter issues, please refer to the documentation and the FAQ below
first.

Question 1 1. What is the KT148A? What functions does it have? What are its features?

Answer 1. The KT148A is a dedicated voice playback chip with a built-in 420-second voice storage space.
It can directly drive a speaker, supports repeated burning, and allows users to replace voices
themselves. It features simple one-wire serial control, low cost, and high cost-performance
ratio.
2. It supports one-wire serial control, interrupt playback, and combination playback, making it
ideal for applications like gaming machines and amount announcements.
3. It provides PC software for users to download voices, with no limit on the number of
segments. The software is included in the resource package.
4. It has ultra-low standby power consumption of 20uA and ultra-low shutdown power
consumption of 2uA. It does not support external amplifiers. If an external amplifier is needed,
consider the KT142A or KT404A chips.

Question 2 2. What is the operating voltage of the KT148A? What is the current? What is the control
method? What size speaker does it support?
Answer 1. The KT148A operates at a voltage range of 2.0V--5.2V. Higher supply voltage results in greater
speaker power, better sound quality, and louder volume.
2. The optimal operating voltage for the KT148A is 4.6V. If using a 5V power supply, it is
recommended to connect a diode to slightly reduce the voltage. Diodes such as 4001 or 4148
can be chosen.
- In this case, connect a 105 capacitor between Pin 7 and ground (refer to the reference
schematic).
- Important: Pin 7 must not be left floating.
3. If applications with a power supply below 3.6V, we recommend shorting Pins 7 and 8 of the
KT148A together. This allows the chip to operate between 2.0V---3.6V. At this point, the power of
the speaker is definitely not as high as 5V, but the effect is still acceptable.
4. The standby current is 20uA, and the operating current is approximately 8mA. If an external
speaker is connected, the current will depend on the speaker's power.
5. The control method is one-wire serial communication. For details, refer to Section 3 of the
manual. It is recommended to use a logic analyzer for debugging, as it is the simplest approach.

Question 3 3. Why is there no response after I solder the KT148A to the PCB?
Answer 1. This issue is somewhat vague, so multiple checks are needed to identify the problem.
19
 2. First, check if the chip's power supply is normal. Specifically, verify if the input voltage at Pin 8
is within the range of 2.0V--5.2V.
3. Next, check if Pin 7 outputs a voltage of 3.3V. If the voltage exceeds 3.7V, the chip may have
been damaged and will need to be replaced.
4. Then, check Pin 2, which is the BUSY pin. It outputs a low level during playback and a high level
when idle. If idle for more than 5 seconds, it will automatically enter standby mode and become
high-impedance.
5. The chip also has a test pin, Pin 4. Triggering it to ground will play a sound. This IO is provided
for customer testing purposes.

Question 4 4. How to download voice files into the KT148A chip? What is the order of copying?
Answer 1. Refer to Section 5 of the manual for detailed steps. The process is divided into 3 steps. For
more information, see: `\99_KT148A Voice Chip Complete Data Package_V5_20231212\02_User
Manual_UART Version\KT148A-SOP8 Voice Chip User Manual_V5.pdf`.
2. The copying order follows the sequence in which files are loaded in the host software. The file
at the top is assigned Number 1, and the subsequent files follow in order.

Question 5 5. Does the KT148A support 3 buttons corresponding to different voice playbacks?
Answer 1. Refer to another document in the resource package: `20240524_KT148A Chip IO Trigger
Function Description_Supports 3 Channels_V14.pdf`.

2. Note: When requesting samples, be sure to contact customer service and specify
"KT148A_V14_xxxx_Button Version". This is a separate version.

Question 6 6. Does the KT148A support specific functional development? What resources does the chip
have?
Answer 1. The KT148A chip is packaged in SOP8, with only 3 usable IO pins. These pins can be used as
inputs or outputs, for AD sampling, or to simulate IIC communication. Either is acceptable.
2. Therefore, some simple applications, such as controlling flashing lights or sampling voltages,
can be customized and developed.

20
Question 7 7. Can the protocol for updating the built-in voice via serial port be provided for the KT148A?
Answer 1. The process of updating the voice in the KT148A chip is relatively complex, so it requires a
certain level of technical capability. Otherwise, after-sales support becomes too troublesome and
overwhelming. If the protocol is absolutely needed, a $1 ,000 deposit is required, which will be
refunded after accumulated 50K units are shipped in bulk.
2. For those with this requirement, please refer to another solution we offer:
`30_KT803C_KT804C_Serial Port Voice Download Complete Resource_V4_20220510`. Search
online for more details using relevant keywords.
3. Ultimately, the main issue is the significant after-sales pressure, especially from customers
who purchase a few samples and demand extensive support. It creates a heavy burden.

Question 8 8. What are the ESD data, operating temperature, and packaging details for the KT148A?
Answer 1. The chip's contact discharge rating is 4kV. Air discharge has not been tested.
2. Operating temperature range: -40°C to +85°C.
3. The chips are packaged in tubes of 100 pieces each and shipped in tube packaging. [There is a
special requirement for tray packaging, which will be outsourced to a third-party manufacturer
for taping and tray loading.] One box contains 10K units, and one carton contains 80K units.

Question 9 9. Can the volume of the KT148A be adjusted?

Answer 1. Yes, volume adjustment is possible. Refer to the one-wire command instructions in the manual.
2. 15 volume levels, 0xE0 is the minimum volume (0), and 0xEF is the maximum volume (default
on first power-up).
(To adjust the volume, send the volume code first, wait 20ms, then send the voice code.)
Alternatively, the volume can be set directly during power-up. Both are ok.
Note that The volume setting is non-volatile, meaning that each time the volume is set, if the
chip is powered off or enters the ultra-low power consumption state via the F0 command, upon
waking up or being powered on again, the volume will revert to the default maximum state..
3. If the volume is still insufficient after setting it to the maximum, you can adjust the volume at
the audio source. Refer to the document:
`07_Audio Generation,Compression,and Basic Editing – Introduction to Synthesis
Methods_V3.pdf`, located in the 07 folder of the resource package.

Question 10 10. Under the same conditions, with an external VBAT of 5V and 3.3V, will the speaker sound
louder with 5V power supply?
Answer 1. Yes, because the output power of the speaker is directly affected by the supply voltage.
Higher voltage results in greater PWM output power driving the speaker, leading to louder
sound. Conversely, lower voltage results in quieter sound.
2. Additionally, for power supplies below 3.3V, you can short Pins 7 and 8 together. This allows
the chip to operate at 2.2V.
21
Question 11 11. Have you implemented one-wire serial control with CH571F? I’ve been debugging for a
while with no success.
Answer 1. One-wire communication is independent of the main controller used. The issue is likely
related to timing, which is difficult to troubleshoot and is often a firmware issue.
2. First, check if the chip has entered the voice download mode. The condition for entering
download mode is that Pin 5 is pulled low during power-up. Once in download mode, the
one-wire serial IO is used for downloading and no longer functions as a one-wire serial port.
3. Second, verify the timing strictly according to the pulse width specifications provided in our
manual, as shown in the figure below:

4. In summary, many customers have used various MCUs without reporting issues, and the
implementation has been smooth overall.

Question 12 12. Why don't I hear a "beep" from the speaker when I short Pin 5 to ground and power on the
chip?
Answer 1. This situation may occur due to insufficient power supply or the speaker volume being too low
to hear.
2. Since pin 5 is inherently designed to drive the speaker, directly shorting it to ground naturally
results in a lower volume from the speaker.
== 》 Solution: Instead of directly shorting Pin 5 to ground, connect it through a 1K resistor to
ground. This will make the "beep" sound louder during power-up.As shown in the figure below:

==》Pin 5 only needs to detect a low level during power-up to enter download mode. Leaving it
floating puts the chip in normal operation mode.
==》Even if you don't hear the "beep," it's fine. Connect the serial tool and try downloading using
the PC software.
==》In short, experiment a bit, find the pattern, and you'll get the hang of it quickly.

22
Question 13 13. For the KT148A chip, the samples have been tested and are OK. How to handle or import
the built-in voice for mass production?
Answer You can take a closer look at this in one of the documents included in our documentation
package.`KT148A Batch Production Voice Loading Solution_V4.pdf`.

Question 14 14. Why does the host software show that the download fails for the KT148A chip? Is the
download a complete overwrite?
Answer 1. The KT148A has an internal allocation of 420Kbyte space.
2. As long as the total size of the imported file (check by right-clicking and viewing properties)
does not exceed 416Kbyte, it can definitely be downloaded.
3. Each download is a complete overwrite. When the host software initiates a download, the
previous content is automatically erased.
4. If you encounter a download error, close the host software and reopen it to resolve the
issue.

Question 15 15. The host software for KT148A prompts that the file exceeds 350KB. How to handle this?
Answer 1. When generating the final bin file, you may receive a prompt indicating that the file exceeds
350KB. Simply click "OK" and continue generating the bin file.
2. You can ignore the prompt shown in the below figure, as it is designed for compatibility with
other products.
For the KT148A, as long as the imported file does not exceed 420Kbyte, there is no issue.

Question 16 16. After downloading audio to the KT148A, I cannot use one-wire serial control. Before
downloading, one-wire serial control worked fine. How can this issue be resolved?
Answer 1. Please note that once the KT148A enters serial port download mode for voice files, one-wire
control is no longer available.
23
 2. This is because the corresponding IO pin is reassigned to UART functionality,which is the serial
port function.
3. If you need to restore default one-wire control, power cycle the chip and ensure Pin 5 is
floating.It must not be in the ground short circuit state, that is, do not let the chip enter the
download mode, but the normal working mode, Refer to the manual for details.
4. For more information, see the explanation in FAQ 1.12. It describes the principles of chip
downloading, and you'll understand after some study.

Question 17 17. The KT148A is specified to drive an 8-ohm 0.5W speaker, but the sound is too low. Can I
replace the speaker or use an external amplifier?
Answer 1. First, the chip's specifications state it supports an 8-ohm 0.5W speaker. However, But the
difference between a horn and a horn is very big:
(1) Some 8-ohm 0.5W speakers are small in size, resulting in lower volume.
(2) Some 8-ohm 0.5W speakers are larger, producing louder sound.
(3) The chip can also drive 8-ohm 1W, 8-ohm 2W, 4-ohm 0.5W, or 4-ohm 1W speakers, although it
may not fully utilize the speaker's maximum performance.There is no issue of whether it can
drive the speaker or not. Lower impedance speakers require higher current.
(4) You can test with the enclosed speaker we are selling, which produces loud sound and
meets most application requirements.
(5) The chip's internal driver can deliver up to 200mA current. You can measure the actual
current; if it's too low, the speaker load may not be sufficient to achieve the chip's maximum
volume.

2. If the chip's sound quality is acceptable but the volume is too low, you can try adding an
external amplifier. The circuit is as follows:

Note: The amplifier chip must support differential input. Single-ended input is not
recommended. This plan may degrade sound quality.

24
 3. If the sound quality is still unsatisfactory, consider using a chip with DAC output and connect it
to an external amplifier. The amplifier power can be selected based on your needs (e.g., 3W, 5W,
10W).
==》Recommended chips: KT404C, KT404A, KT142C, etc.
==》Recommended amplifiers: HAA8002D (3W), HAA2018 (5W), TPA3110 (10W)

4. Additionally, we recommend the KT142C, a chip with DAC output for external amplification.
It offers excellent performance but comes in an SOP16 package. For more details, search online
for "KT142C voice chip."

Question 18 18. I soldered the KT148A chip onto a test board and tested it. The button can play sound, but
when Pin 5 is grounded and power is connected, there is no "beep" sound. Instead, the
indicator light keeps flashing. After disconnecting Pin 5 from ground, the indicator light stays
on. In both flashing and steady states, I cannot download files. What is the reason for this?
Answer 1. This issue is likely caused by insufficient power supply, causing the chip to continuously
reset.
== 》 The indicator light is connected to the chip's PB9 pin, which should not flash. The only
possibility is that the chip is resetting repeatedly due to insufficient power, making it appear as if
the PB9-driven light is flashing.
2. Troubleshoot as follows:
==》First, if you don't hear the "beep," refer to FAQ 1.12 for circuit modifications.
==》Test the chip on our provided test board.
== 》 Ensure that PB1 and PB9 of the KT148A are disconnected from the MCU or USB-to-TTL
converter. Beacuse some MCU GPIOs have strong driving capabilities, which can backfeed current
into the KT148A, causing it to reset repeatedly.
==》So power the KT148A with 5V, then simply connect the USB-to-TTL TX and RX pins.

On these boards, you may not hear the On these boards, you will definitely hear the "beep"
speaker's "beep" because Pin 5 is because Pin 5 is grounded through a 1K resistor.
directly grounded.

25
Question 19 19. Can the KT148A chip drive headphones? I haven't seen any related documentation.
Answer 1. Since the chip can drive a speaker, it can certainly drive headphones, as headphones are
essentially two small speakers.
2. The circuit diagram for driving headphones is shown below. Essentially, it connects two
speakers in series. Please understand it yourself.
(1) Note that the chip output does not distinguish between left and right channels; both channels
will produce the same sound.
(2) The performance of driving headphones is quite good, comparable to driving a speaker, as
verified by other customers in mass production.
The circuit diagram below also includes headphone insertion detection, where S1 is detected
using an AD (analog-to-digital) method.

Question 20 20. The combination playback function of the KT148A feels slightly discontinuous. There is no
perfect solution, but adjusting the audio source can improve the effect.
Answer 1. For details on the combination playback function, refer to the manual.The recommended
interval between commands is 25ms.

2. The chip's internal processing mechanism is as follows: after receiving all commands for the
voice files need to combination playback, it queues them up and plays them one by one.That is, it
plays file 1, stops, then starts playing file 2, and so on until all files are played.

26
 3.Please note that the 25ms mentioned above refers to the interval between commands. When
the reception is complete, the interval between one voice and the next is 50ms.
==> This 50ms includes the time for the current file to stop decoding, the search and decoding
of the next file to start, and the sound to be produced, which takes approximately 50ms.
4. If users notice discontinuity, they can try the following two methods:
(1) Method 1: Use audio editing software to trim the blank parts at the beginning and end of
the audio source. For specific steps, refer to the following section in the manual.

(2) Method 2: When generating the audio source, try reducing the playback speed of individual
audio files.
== 》 The chip inherently has a delay of about 50ms between finishing one file and starting the
next.
==》If the audio files are played too quickly, the combination playback may appear discontinuous.
== 》 Reducing the playback speed of individual files can mitigate this issue without significantly
affecting the user experience. Some customers have successfully used this approach. This is just a
general suggestion for reference.

For reference, the example for combination playback is located in the following path in the
resource package:It also includes a video demonstration.

Question 21 21. Summary of all information about volume and sound for the KT148A, such as low volume.
You get the KT148A along with an 8-ohm 1W speaker, and testing was fine. However, the
datasheet specifies an 8-ohm 0.5W speaker. When I tested with an 8-ohm 0.5W speaker, the
sound became very low. Should I actually use an 8-ohm 1W speaker?
Answer 1. This issue requires consideration from multiple angles. First, the 8-ohm 0.5W speaker you
chosen a very cheap one, such as a thin-film type?
2. Is the speaker exposed without an enclosure or sound chamber? If so, the sound will naturally
be low.

27
3. If you tested our sample speaker and found the sound acceptable, it means the KT148A's
driving capability is sufficient. The issue lies with the speaker or its enclosure, not the chip.
4. To clarify: The chip KT148A can also drive 8-ohm 1W, 8-ohm 2W, 8-ohm 3W, or 4-ohm 3W
speakers, although it may not fully utilize the speaker's maximum performance.There is no
issue of whether it can drive the speaker or not. Lower impedance speakers require higher
current.
Just like: A speaker may handle up to 1A of current, but if you only supply 200mA. For the
speaker, it can of course bear, but it will feel very easy, can not play its greatest ability. I do not
know this example, can you understand?
==》Larger speakers are also more expensive.
== 》 Many toy products use thin-film speakers because they are cheap. They can still produce
loud sound due to their larger size and well-designed enclosures.
==》For more details, you'll need to study further. We sell chips, not provide full support（not
babysitters）.

5. Here's some additional basic knowledge about speakers and sound.First of all, the size of the
sound depends on the following three factors
Factor 1： The chip's driving capability. For example, the KT148A manual specifies driving
an 8-ohm 0.5W speaker.
== 》 First, the chip's specifications state it supports an 8-ohm 0.5W speaker.
However, But the difference between a horn and a horn is very big:
==》Some 8-ohm 0.5W speakers are small in size, resulting in lower volume.
==》Some 8-ohm 0.5W speakers are larger in size, producing louder sound.
== 》 The chip can also drive 8-ohm 1W, 8-ohm 2W, 4-ohm 0.5W, or 4-ohm 1W
speakers, although it may not fully utilize the speaker's maximum
performance.There is no issue of whether it can drive the speaker or not. Lower
impedance speakers require higher current.
Factor 2： The speaker's design and materials, such as size and whether it's paper-cone or
thin-film.
==》(1) Larger size generally means louder sound.
== 》 (2) Paper-cone speakers are relatively more expensive and produce louder
sound.
==》(3) Thin-film speakers are the cheapest and generally produce lower volume.
Factor 3： The speaker ’s installation environment and enclosure design.
==》For example, if the speaker is held in your hand, the sound will be low because
it's not concentrated.
==》If the speaker is enclosed in a well-designed chamber,voice are concentrated,
the sound will naturally be louder.

28
Question 22 22.I want to ask if this chip can really store 420 seconds of audio? A random 5-second MP3
audio file is already 65k. If that's the case, how large can my MP3 source file be at the
minimum sample rate and bit rate? --- January 12, 2023, 09:44:38
Answer 1. First, the manual clearly states that the chip only recognizes the size of the final generated bin
file [Use the host computer tool to load the file, then click pack, and you can see the total size of
the bin file]. Or after the second step of conversion, the generated F1A file, the total size.
2. The chip has a built-in space of 420Kbyte, which means the final bin file generated by the tool
should not exceed 420Kbyte. There is no limit to the number of segments, but do not attempt to
directly add MP3 files for packing, as the farmat audio will not play.
3. The customer asked about the maximum size of the MP3 before conversion. This question
cannot be answered because the sample rate and bit rate of your MP3 or WAV files, or other
formats, are uncertain [If you don't understand this, please search online yourself].
4. Here is an example of the smallest MP3 format: the song itself is 22.5 seconds long.

(1) If compressed with 16KHZ, 16bit parameters, its size is 48Kbyte. This means that basically, 1
second of audio takes up about 2Kbyte of space.
(2) If compressed with 8KHZ, 8bit parameters, its size is 24Kbyte. This means that basically, 1
second of audio takes up about 1Kbyte of space. It is not recommended to compress to such low
parameters in actual use.
5. For normal files, for MP3 level audio quality, we recommend compressing the file to 32KHZ,
32bit parameters if space allows. If space is really tight, it is recommended to compress to
16KHZ, 16bit. It is definitely not recommended to compress to 8KHZ, 8bit, as this parameter
greatly damages the audio quality.
6. If there are files in other formats, such as WAV or WMA, etc., it is recommended to compress
them to MP3 first, and then convert them. Each time you compress, remember to listen to the
effect on the computer.
7. For audio processing, please refer to "07_Audio Generation,Compression,and Basic Editing –
Introduction to Synthesis Methods_V3". Search for it yourself.
8. Most importantly, remember to read the manual. The download is divided into three steps.
You need to read the materials first before asking questions. Many questions can also find clues
in the materials.

29
Question 23 23.The manual says that when the power supply is greater than 3.5V, pins 7 and 8 should be
separately connected. Does this mean VCC should be directly connected to pin 8, and pin 7
should be left unconnected? However, during my test, the chip only worked properly when pin
7 was connected at 3.7V. I tested this with both a lithium battery and a computer's USB.
Answer This issue can be divided into three key points, which I will explain in detail using the chip's
schematic as an example.

Issue 1: Power supply explanation for the chip

1. The chip's schematic clearly indicates that in the case of a lithium battery power supply, such
as 3.2V to 4.2V, the power must enter pin 8. Additionally, pin 7 of the KT148A cannot be left
floating and must be connected to an external 105 capacitor.
2. If the chip is powered by a coin cell battery, which has a maximum voltage of 3.3V, then pins 7
and 8 of the KT148A should be shorted together, and an external 105 or 106 or larger capacitor
should be connected.The selection of the value mainly depends on the load of the speaker; the
larger the load, the larger the capacitor should be.
3. If the KT148A is powered by a stable 5V or 4.6V supply, the power should enter pin 8 of the
KT148A, and pin 7 should not be left floating as the same; it needs an external filter capacitor.
==> In summary, the internal structure of the KT148A chip includes an LDO. Pin 8 is the input of
the LDO, and pin 7 is the output. The voltage at pin 7 is also the core voltage of the chip and the
high-level voltage of the chip's GPIO. Understanding this makes it simple.
Issue 2: Why does the chip only work when 3.7V is connected to pin 7 of the KT148A when I
test?
1. This is definitely not normal because your 3.7V should be connected to pin 8 of the chip.
2. This issue arises because your 3.7V power supply has weak load capacity. When the KT148A
chip plays audio, it typically requires a momentary current of about 150mA. If the load capacity is
too weak, the voltage will inevitably drop, causing the KT148A to reset.
3. The reason it works when connected to pin 7 is that it happens to be at a critical point, which
is not unusual.
==> In summary, understanding the power supply is very simple: it internally integrates an LDO.
Remember to add a filter capacitor close to the chip externally.
Issue 3: Why does neither the lithium battery nor the USB power supply work properly?
30
 1. We suspect that this issue occurs because the KT148A chip does not have an external filter
capacitor, or it might not be a power supply issue at all, such as entering download mode, etc.
2. For specific issues, we recommend using our test board to verify.
In conclusion, when encountering problems, please first use our test board to check. The chip
itself is very simple, and if your custom-designed board is not working properly, having a
comparison can quickly resolve the issue.

Question 24 24. KT148A has an ultra-low voltage requirement. When using a relatively low-quality 3.3V
coin cell battery for power supply, the voltage can drop to as low as 2.6V. The hardware has
already shorted pins 7 and 8, but it still doesn't work properly. What should I do?
Answer

1. First, KT148A can definitely work at low voltages. There is no doubt about this. The
minimum operating voltage of KT148A can reach 2.2V, but generally, a coin cell battery can
only work down to about 2.6V before it runs out of power.

2. On the hardware side of KT148A, pins 7 and 8 must be shorted, and a 106=10uF capacitor
should be connected to ground.
3. Of course, if it's another type of power supply, please refer to "Question 23" for details on
power handling.

The new version of the chip already defaults to directly supporting coin cell battery operation.
Simply short pins 7 and 8 for power supply, and you can use it without further questions.

Question 25 25. What does the "checksum" marked on the KT148A sample card mean?
Answer 1. The chip is programmable, so there are many firmware versions.
2. How do we distinguish between these firmware versions? This is where the checksum
comes into play.
3. The checksum is unique and generated using our CRC checksum tool.
4. The checksum is essential information to remember. Since there are many customers and
numerous firmware versions, we rely on the checksum as the reference.

31
 5. During batch burning, we also use the checksum to ensure that the correct firmware is
burned without errors.

Please make sure to remember this or include it in the BOM (Bill of Materials).as long as the
checksum is correct, the burned firmware will never go wrong.
Question 26 26. Explanation of the built-in space of the KT148A chip, totaling 420Kbyte, with details on
exceeding the range and size limits.
Answer 1. The smallest storage unit in the chip's built-in space is 4K bytes, which is 4Kbyte.
2. When a user performs a download, the files are stored in the chip's internal space in the
following order:
【 File directory list = 4K 】 --> 【 File 1 - aligned to a multiple of 4K 】 --> 【 File 2 - aligned to a
multiple of 4K】--> and so on until the last file.
3. Example explanation:
(1) If the file 1 that the user needs to download has an actual size of only 0.05Kbyte, when you
download it, this small file will still occupy 4Kbyte of space.
(2) If the file 2 that the user needs to download has an actual size of only 4.05Kbyte, when you
download it, this small file will still occupy 8Kbyte of space.
(3) If the file 3 that the user needs to download has an actual size of only 3.95Kbyte, when you
download it, this small file will still occupy 4Kbyte of space.
4. Therefore, some customers have many files, but each file is very small, which can lead to
insufficient space, as shown below:

This clearly shows insufficient space.

The actual files are only 267K, but the occupied space is 736Kbyte, which exceeds the chip's
maximum capacity of 420Kbyte.
5. The solution to this problem is that the customer must first test and, when ready for mass
32
 production, submit the final audio files to us. These files will then be updated into the chip in
another way, and serial port client downloads will no longer be supported.
==> During the sample stage, we cannot assist with modifications. Please prepare everything
yourself, and we will assist during trial production or mass production.
6. Conclusion:
==> The KT148A chip is not suitable for playing very short music files. It is best to keep
individual files around 4K in size.

Question 27 27. The sound from the KT148A driving the speaker is too low compared to other OTP chips? Is
the sound quality of KT148A poor?
Answer First, the sound from the KT148A chip driving the speaker is not actually low; it is balanced for
sound quality and not intentionally amplified.
==> If during actual testing, you still find the sound a bit low, this is not a problem. You can
contact customer service to request a sample version with higher volume to experience it.
==> To achieve higher volume, there might be a slight sacrifice in sound quality, approximately
10%. However, the sound volume can be increased by about 30%.

Regarding the issue of poor sound quality, there was a previous customer feedback where the
sound from the original prototype was recorded using a phone, then transcoded and
compressed into our chip, and upon playback, the sound quality was reported as poor.

==> This is easy to analyze. Using a phone or other device to record is equivalent to using a very
high sampling rate to record a signal that has already been compressed. After recording,
compressing it again results in a sound quality loss of about 80%.
==> Therefore, determining the quality of sound should not be tested in this manner.
==> For cases like this customer's, it is advisable to synthesize the sound directly online or record
it with a real person.

Finally, it should be noted that the sound quality of KT148A is actually quite good, after all, the
chip's performance is evident, and is at least 10% better than similar OTP voice chips. So, you
need not worry about the sound quality.

Question 28 28. What level does the KT148A chip belong to? Is it commercial or industrial grade?
Answer The KT148A chip belongs to the industrial grade. For details, please refer to the specification
sheet.
Working voltage
VBAT: 2.0v - 5.5v
VDDIO: 2.0v - 3.4v
Operating Temperature: -40°C to +85°C

33
Question 29 29. The KT148A playback is a bit choppy, and there are popping sounds?
Answer 1. The customer's test environment is as follows:

2. Based on this image, a clear hardware issue can be seen: the chip's power supply lacks a
capacitor for filtering.
3. This is crucial. Please refer to the schematic diagram; the capacitor is essential. Without it,
the fluctuations in power supply caused by the speaker can easily lead to instability in the
chip's core operation, resulting in resets, choppy sounds, etc.
4. If such issues occur, it is imperative to test with our test board. Having a comparison will save
you time.
5. Under normal hardware conditions, the KT148A playback is very smooth, with no issues of
choppy sounds or popping noises.

Question 30 30. What is the difference between the button version and the serial port version of KT148A?
Why are there different versions? Please describe in detail.
Answer First, the button version and the serial port version use the same chip, but the firmwares
burned into them are different.
This means that the scenarios for using these two versions are completely different.
However, the method of downloading audio is the same for both versions, please be
aware.

34
 1. Button version: The chip has only 3 IO ports in total, with one IO port corresponding to one
audio segment. It is commonly used in standalone KT148A applications, such as connecting to a
button to play a sound when pressed, or applications like greeting cards that play sound when
opened, doorbells, etc.
2. Serial port version: This version must be paired with an MCU [microcontroller or controller]. In
this setup, KT148A acts as a passive device, and all its operations are executed based on
commands received from the MCU. This makes its usage more flexible and is the most commonly
used scenario.
3. The reason for having different versions is also due to the limited number of pins on the chip,
necessitating the division into versions, a decision made out of necessity.

Question 31 31. The KT148A uses a serial port for downloading. Can I directly use the serial port to send
commands for testing and control during actual testing?
Answer

1. The audio download for KT148A uses a serial port with a baud rate of 115200. The purpose of
this is that the serial port is quite universal.
2. However, during actual debugging, a one-wire serial port [single-wire timing type] is used, as
detailed in the manual.
3. KT148A does not support sending commands via the serial port for control. It's not that it can't

35
 be done, but rather it's unnecessary because using serial port control would require 2 IO ports,
leading to insufficient IO ports on the chip. Additionally, implementing low-power applications
with a serial port is very cumbersome, which is why one-wire control was ultimately chosen.
4. If you insist on using the serial port for control, you can permanently connect a 1K resistor
between pin 5 of the chip and ground to keep the chip in download mode. This way, you can send
commands via the serial port for testing. However, this is not recommended, and technical
support and service will not be provided for this method, please be aware.
5. If you really need this, it can be arranged for an additional fee of $200. However, it is still
recommended to use the one-wire serial port, as it is more stable, user-friendly, and convenient.

Question 32 32. During mass production, do I have to use the KT148A download board to download one by
one?
Answer Rest assured, we have already considered this issue. During mass production, you can provide us
with the packaged audio bin files, and we will batch burn them into the chips. For details, please
refer to the introduction in the documentation package. It's simple and efficient.

Question 33 33. Why can't I download audio to the KT148A test board I bought? I have installed the driver
and carefully followed the three steps, and everything seems fine. What could be the reason for
this issue?
Answer 1. First, the test demo board we provide is functional and capable of downloading audio.
However, users need to have their own CH340 USB-to-TTL module to connect with the KT148A
chip. The connection diagram is as follows:

36
2. Please double-check the port number (COM number) and baud rate selected in your download
software.
The download baud rate is fixed at 115200. To check the COM port number, go to the Device
Manager, as shown below:

(1) If the issue persists, please purchase a complete download board online for comparison. It is
also very affordable, and the time cost of troubleshooting can be high.
(2) However, it is recommended to carefully compare the provided materials and troubleshoot
the issue yourself first. After all, troubleshooting is also a learning process that requires rich
technical skills. Do not easily suspect the manufacturer's products. Since we dare to put it on the
market, we are certainly concerned about any lack of attention to detail that could waste our
customers' time.

37
Question 34 34. Does KT148A support DAC? The external amplifier has poor sound quality. Are there other
chips you can recommend?
Answer 1. KT148A has built-in PWM output and can directly drive a 0.5W speaker. Of course, it can also
drive headphones without any issues.

2. However, it does not support DAC output as the chip does not have this resource. If you need
an application with DAC and an external amplifier, we recommend using the KT142C chip. It
comes in an SOP16 package, and the cost is actually similar. It is also very user-friendly. Refer to
the schematic below:

38
Question 35 35. Why does the KT148A circuit stop playing audio after several power cycles?
Answer

The customer's circuit looks a bit unusual. The communication pin of KT148A is pin 4.
If the audio stops playing after several power cycles, we suspect there are only two possible
reasons:
Reason 1: The chip is damaged, but this probability is extremely low. You can try replacing the
chip to verify.
Reason 2: The SDA-S pin is connected to an external MCU, and this IO port is outputting, causing a
current sink to KT148A.

39
 1. This prevents KT148A from resetting properly.
2. Since KT148A uses a low-voltage reset internally, it should automatically reset upon power-up
and power-down.
3. The solution is to add a resistor between the SDA-S pin and KT148A. A resistor between 100
ohms and 1K should work.
==> This resolves the issue of reverse current.
Other reasons are currently unclear.

Question 36 36. Explanation of the sound volume of KT148A? How to increase the volume? By how much?
Applicable scenarios.
Answer 1. The chip has a built-in Class D amplifier with PWM output, and the power is 0.5W. The sound
volume is actually quite decent.
==》It can be compared to the volume of "Welcome" announcements in supermarkets.
==》It can also be compared to home doorbell products.
==》It can also be compared to products used by children, such as toys.
Of course, the most important factor is that the larger the speaker size, the louder the sound. You
can also compare it with the speaker samples we provide while test.
2. This is suitable for indoor use and applications where the distance is relatively close, such as
devices within about 5 meters of a person.
3. This level of sound volume is not suitable for outdoor devices, such as alarms, megaphones,
etc.
4. Currently, the standard version of KT148A sacrifices some sound volume for better sound
quality. If testing shows that the volume is still slightly insufficient, you can contact us to modify
the firmware. The volume can be increased by about 15%, but this will sacrifice about 15% of the
sound quality.
5. If the product requirement really needs louder sound, we recommend using the KT142C-SOP16
chip, which is similarly priced.
Since it has a DAC, it can be paired with an external amplifier. This way, the sound volume can be
easily adjusted. If you need very loud sound, you can choose a larger amplifier chip, such as 10W,
5W, 3W, etc.

40
Question 37 37. During the process of downloading audio to KT148A, the process stops halfway and displays
a communication timeout.
Answer

Based on the information provided by the customer as shown in the image above:
1. The serial port number on the left is correct, COM5, using a CH340 USB-to-TTL module.
2. The port number and baud rate selected in the download software are also correct.
3. The imported file on the far right is also correct.
4. The issue occurs where the download stops partway through.
Given this information, the customer's setup environment seems correct, but why does the
anomaly occur?
1. It could be that the USB-to-TTL module is unstable, causing errors in the middle of
communication.
2. It could also be due to poor wiring connections; please check carefully.
Other than these, it's hard to think of other reasons, so you'll need to troubleshoot based on your
experience.
In summary, for any issues related to the download process or anomalies, you should first
compare with the download board, which is more efficient than back-and-forth
communication.

41
Question 38 38. When using your KT148A serial port version to play music, it directly causes the
microcontroller to malfunction. After testing the power supply, I found that every time the
music plays, there is a significant fluctuation in the power supply. I checked your manual, which
says the load capacity is insufficient, but when I connected the power supply to your test
board, the microcontroller worked normally again.
Answer This issue is clearly related to the power supply. In fact, a quick online search or a bit of
experience can easily resolve it.
But there's no helping it; the service industry is just that competitive. Understanding is key.
First, the power supply for KT148A is divided into two types, as shown in the figure below: Pay
attention to the power supply description in the upper right corner of the image:

Secondly, the current consumption of the chip depends on the size of the external speaker and
the volume setting.
1. Therefore, the higher the volume, the greater the current consumption. The larger the power

42
 of the speaker, the greater the current consumption.
2. The driving capability of the chip is also limited, with a rough limit of about 200mA.
Finally, the solution is actually quite simple:
1. Make the power traces on your PCB as thick as possible.
2. Add capacitors to the power supply end of KT148A, decoupling capacitors, such as electrolytic
capacitors, 47uF/16V withstand voltage, or others.
3. Keep the traces from KT148A to the power supply as short as possible.
4. Ensure your power supply is sufficient. For example, using DuPont wires to connect to a distant
and weak power source causing large power fluctuations, leading to resets or other anomalies, is
an obvious issue. It's basic knowledge; please brush up on it.
Question 39 39. Why can't the "userdown.bin" file generated by the download and pack function in the host
computer tool be played after downloading? However, importing files one by one allows them
to be played.
Answer 1. The "Pack File" button in the host computer tool software "VoiceCom_V8.exe" is used for mass
production.
2. The KT148A chip does not support directly downloading bin files in the host computer
software.
3. The reason is: The bin file generated by packing is not the original audio file. It contains a lot of
information such as file list information and audio file information. Therefore, it cannot be
directly downloaded and played.
4. It is used for mass production and burning. For details, please refer to folder "09_Mass
Production Guide" in the documentation package.

Question 40 40. I noticed there is a busy status indicator. The parameters state that it is low during playback
and high when idle, but it also goes low a few seconds after being idle. I wanted to use it for
mute control of the amplifier, but it seems unsuitable.
Answer 1. This busy indicator output is only available in the one-wire serial version. Please refer to the

43
 schematic diagram.

2. This busy signal is from pin 2 of the chip, which is a GPIO port and follows the general
characteristics of IO ports.
==》During playback, it outputs low level = 0V
==》When idle, it outputs high level = 3.3V
==》But here is an important point: the chip will enter an ultra-low power state = 2uA after being
idle for 5 seconds, so the busy pin is in a high-impedance state.
It is neither low nor high level, although the indicator light appears to be off.

3. If you want to use it to control an amplifier, it's quite simple. Just add a pull-up resistor
externally, 10K or 22K will do, and handle it flexibly.
4. For applications involving external amplifiers, the KT142C-SOP16 chip is recommended. It
supports DAC output, and the sound quality with an external amplifier is better.
Question 41 41. I would like to ask how long it takes for KT148A to receive a one-wire serial command and
then play the sound. My requirement is for a button prompt sound, and initial testing feels like
there is a delay. How should this be handled?
Answer The total time cost for KT148A from receiving the command to executing and playing the sound is
approximately 15ms + 16ms = 31ms.
== 》 One-wire serial command time [15ms] + finding the file and decoding [16ms] = total time
[31ms]
1. Among these, the one-wire serial reception time is easy to understand, roughly 6ms for the
start + 9ms for the data = 15ms.
2. The time for finding the file and decoding to play is as follows: roughly 16ms.

44
 3. A 30ms time is actually not long, and there is basically no room for further compression. If the
sound seems slow to come out, try to see if there are blank spaces at the beginning of the audio
source itself, and edit it with audio editing software. Specific methods can be found online.
4. At the same time, the one-wire serial version, in theory, reacts faster than the button version
because it saves the time for button debouncing.

Question 42 42. When converting MP3 files, I selected F1A as the output, but the actual file is F1B. What is
the issue?
Answer If an F1B format file appears, it can still be used normally; it's not a problem.

The reason for this is that you selected a "32K" sampling rate. This sampling rate generates larger
files with better sound quality, but it is unnecessary.

In actual use, you can directly choose "8K" or "16K", and that will suffice.

45
 『 Contact US 』for Technical Support and Sales Service
/*********************************************************************

EMail: [email protected]

WhatsApp: +86 159 0279 7635

WeChat: JAJB-Go

Tel: +86 159 0279 7635

*********************************************************************/

- Feel free to call for consultation at any time.

- Welcome to contact us anytime to obtain documentation or communicate.

Linkedin: https://www.linkedin.com/company/blevoice/
Blogger: https://blevoice.blogspot.com
Tumblr: https://www.tumblr.com/blog/blevoice
TikTok: https://www.tiktok.com/@blevoice
[BLE Voice Store]: https://www.aliexpress.com/store/1104656113

1. To request samples, please search for key words such as "KT6368A" directly on
AliExpress, enter our official store [BLE Voice Store] for solutions.
2. For each product, through rigorous development, debugging, and testing processes,
we provide comprehensive technical documentation including user manuals, FAQs,
schematics, Datasheet, Reference Code, debugging guides, critical notes and ect.
3. As our products expand into broader applications and industries, all related
technical resources are continuously updated to ensure relevance.
4. Full technical packages for every product are permanently hosted on Google Drive,
offering free lifetime access, and the corresponding download links are highlighted on
the detail pages of each product in our AliExpress store.
5. While our documentation may appear on third-party platforms, forums, or
websites (shared by enthusiasts, partners, or industry peers), the latest versions are
exclusively available on Google Drive.