Speech Synthesis LSI Series Ver.3 - Rohm

Speech Synthesis LSI SeriesVer.3.0

C o n n e c t w i t h I o T

*LAPIS Semiconductor study

Proven customer support provided for over 40 years

P

*

Supports

external memory

Qualified under

the AEC-Q100

automotive

standard

Built-in

long playback

memory

Built-in

short playback

memory

Speech Synthesis LSI series

LAPIS Semiconductor

Memory

Amp

ADPCM2HQ-ADPCMSynthesis

Simple

operation High audio

quality

Greater space

savings

Multiple

functions

In addition to leveraging sound technologies ranging from

audio compression/decompression and acoustic processing

to digital filters and amps, LAPIS Semiconductor's wide

lineup of speech synthesis LSIs integrate peripheral

technologies such as P2ROMTM and flash memory.

The broad portfolio includes series with built-in ROM ideal for

voice guidance/alarm/message/sound playback along with

high reliability types optimized for automotive applications.

Our focus on ‘enhanced sound quality’ has been

well-received by customers for many years.

We deliver solutions and support that meet customer

demands while providing added value.

*LAPIS Semiconductor study

Product development that takes intoaccount the needs of the times

Wide lineup ranges from simpleto high performance types

Proven customer support provided for over 40 years

*

LAPIS Semiconductor speech synthesis LSIs allow users to immediately add high quality voice functionality

to virtually any applicationLAPIS Semiconductor

Speech Synthesis LSIs feature:

LAPIS Semiconductor


LAPIS Semiconductor


t

s Simple

operationSimple

operation r

is High audio

qualityHigh audio

quality

r

is Greater space

savingsGreater space

savings

◆Enables easy audio playback by sending a simple command from the MCU

◆48kHz Max sampling

frequency (middleware

typically from 8kHz to 12kHz)

◆Built-in high performance

DAC and filter

◆Highly integrated design

from speech synthesis

engine to speaker amp

reduces the number of

external parts

◆Fewer external parts

reduces the failure rate

LAPIS Semiconductor


s Multiple

functionsMultiple

functions

◆Multiple built-in audio-related functions. Equipped with fail-safe functions including playback

sound detection for automotive use

•Simultaneous multi-channel playback function

•ROM editing functionand more...

Multiple functionsHigh fidelity/

high compressiondecoder

Simultaneousvocalization

Editingfunction

Speed andpitch conversion

Onboardwriting

All-in-one design

Decoder(hardware)

Amp

FilterMemory

Comprehensive support from LSI selection to mass productionFull services provided, including sound creation

Customer Tasks

Select the LSI series/ROM capacity

Create a phrase list

Create original audio

Voice analysis and editing

Complete the ROM data

STEP1. STEP3.

STEP4.

STEP5.

STEP2.

LAPIS Semiconductor and/or Customer Tasks

Select the LSI based on set requirements

Write before shipment

Applications

Speech? Soundeffects?

Melody?

PlaybackTime

Address

1 Good MorningGood Afternoon2

3

Speech

Studio recording

Editing

Volume adjustment

Mixing

Equalizing

Melody/sound effect creation

Existing source

Ad

dit

iona

l rec

ord

ing

(in

som

e ca

ses)

Add

ition

al re

cord

ing

(in s

ome

case

s)D

emo/

corr

ectio

n

05

Automotive-grade

ML2253x series

Flash Memory ROM Capacity

The ML2253x series of automotive-grade 4 channels speech synthesis LSIs

with built-in Flash memory are optimized serial audio interfaces and voice data.

Features include HQ-ADPCM, 16bit D/A converter, and low-pass filter for

achieving high sound quality, along with an integrated 1.0W monaural speaker

amp that enables direct speaker drive.

A failure detection function is also built in capable of detecting malfunctions.

All functions necessary for audio output are integrated on a single chip, making

it possible to easily add voice functionality by simply adding this LSI. *The maximum external memory that can be connected (there is no internal memory)

ML2253x series

　　to 2017 2018 2019 future from 2020

ParameterAdvantages of ROHM’s

New ProductsConventional Product

ML22Q573

New ProductML22Q53x/ML22530

Audio Interface

Flash memory rewrite fromMCU

ROM Capacity

Internal speaker amp

Simultaneous soundgeneration (mixing) function

Playback abnormality detection

Speaker disconnect/Short-circuit detection

Operating Temperature

None

None

4Mbit (Flash)

1W Class AB(8Ω, DVDD=5V)

4 channels 4 channelsEnables 4 channels simultaneous

playback of in-vehicle notification sounds

None LSI failure detection function added

Short-circuit detection only Enhanced functionality

–40°C to +105°CSupports high temperature operation:

+105°C

Serial Audio Interface

Available

16/4/2Mbit

1W Class AB(8Ω, DVDD=5V)

Available

Speaker disconnect/Short-circuit detection included

–40°C to +105°C

Supports audio playback from internal memory and external audio input

Rewrite possible after shipment

Expanded memory variety

Provides Class AB operation withoutconcern for radiation noise

Thermistor

Speech Synthesis LSIs

ML225xx series

ML223xx series

ML22594 (6Mbit)ML22Q553 (4Mbit)ML22(Q)573 (4Mbit)ML22572 (2Mbit)

ML22Q374 (696Kbit)ML22Q394 (696Kbit)

ML22Q535 (16Mbit)ML22Q533 (4Mbit)ML22Q532 (2Mbit)ML22530 (External ROM)

onalityyImproved funnncctioctionaaannceand performaaamanc

Enhanced failure detectionEnhanced failure detection▶Integrated failure detection functions

(Playback/Clock abnormality detection)

▶Built-in fail-safe functions(Short circuit/High-temperature/

Command transfer error detection)

▶Additional I/F (Serial audio I/F(I2S)etc.)

▶AEC-Q100 qualified

High performance・High fidelityHigh performance・High fidelity▶Guaranteed +105°C operation

▶High audio quality/high compression algorithms

▶Built-in fail-safe functions (High-temperature/Command transfer error detection)

▶Internal 1.0W speaker amp

Simple・CompactSimple・Compact▶Guaranteed +85°C operation

▶Internal 1.0W Class D speaker amp

▶Simple and Compact

Ideal for long audio playback in automotive applications

ML22Q532

ML22Q533

ML22Q535

ML22530

2Mbit

4Mbit

16Mbit

128Mbit*

06

Enables rewriting to internal/external Flash memoryfrom the MCU via SPI I/F

The ML22Q53x and ML22530 allow data to be

rewritten to the internal Flash and external

memory, respectively, via the MCU’s SPI I/F.

In addition, audio data can be rewritten in

the field even after product shipment.

Equipped with a variety of fail-safe functions

Multiple built-in fail-safe functions,

such as for detecting playback

abnormalities, speaker pin

short-circuits, internal LSI

temperature, supply voltage drops,

and MCU command transfer errors,

make it possible to monitor for

a variety of LSI failure modes.

Short-circuit detectionPlayback abnormality detection

Output becomes H when groundfaults occur between the SP pins

(output normally L)

Output becomes H duringabnormal playback

High temperature error detection

Speech Synthesis

LSIs

Detection

Circuit

ThermistorThermistor

ERR outputCPU

Output becomes H at high temperatures(output normally L)

Command transfer error detection

Speech Synthesis

LSIs

ERR outputCPUSPI

I/FComparator

SI

SCK

SO

Inputcommands

twice

If the commands are differenttwo times, the output becomes H

(output normally L)

ERR outputCPU

SP+ pin

SP– pin

Short-circuitdetection

Short-circuit

Short-circuit

Feature

1

Feature

2

Enhanced failure detection

High performance・High fidelity

Simple・Compact

Rewriting audio data

ML22Q53x

MCU

Internal

Flash

memory

SPI I/F Pin

SPI I/F Pin

Rewritable

MCU

External

Flash

memoryRewritable

ML22530

Playbackcheck

MCU

Logic Analog

Provides fault detection fromthe input to the speaker pins

07

ML2257x series ML223xx series

Built-in speaker amp ensures stable operation at +105°C

Backside die pad featuring superior heat dissipation ensures high temperature operation

Automotive-grade speech synthesis LSIs are offered featuring

a guaranteed operating temperature range of –40°C to

+105°C. Integrated memory and speaker amp, together with

support for high temperature operation, ensure worry-free

use in automotive applications.

The exposed backside die pad enables heat dissipation through

the substrate, making it possible to suppress temperature rise

within the LSI (including the speaker amp) and ensure stable

performance even at an ambient temperature of +105°C.

Applicable Models: ML22572, ML22(Q)573, ML22Q553, ML22594

Guaranteed +105°Coperating temperature

Backside Die Bond

Mold Resin

Die BondPaste

Frame

Device Chip

Spacer ChipDie BondPaste

Thermal dissipation to the substrate suppresses temperature rise

Terminal Terminal

Data expansion possible by connecting external memory to the internal ROM

Memory can be expanded by connecting external memory

up to 128Mbit to the internal 6Mbit Mask ROM. Storing

common sound data in the internal ROM and using external

memory (capacity depends on the model) to store extended

data facilitates customer development and inventory

management.

Applicable Models: ML22594

Advantages of external memory

MCU

Japanese

Wear yourseatbelt

English

Speech Synthesis ICs

Buuu-n

Peee peeeTick tick(blinker sound)Tick tick

(blinker sound)

Peee peeeWear yourseatbelt

Peee peee

Smooth

replacement of

extended data

Common Phrases

External memory

Thermistor

The ML2257x series features a guaranteed operating

temperature of +105°C Max and integrates a 4 channels

mixing function along with speaker amp.

Multiple fail-safe functions required for automotive

applications are also built-in, including detection of

the internal LSI temperature, supply voltage drops,

and MCU command transfer errors.

MaskROM Flash Memory ROM Capacity

Pin compatibility

*Integrated 6Mbit allows external memory to be connected

Integrated oscillator and Class D amp reduces component count

The built-in high accuracy 4.096MHz+1.5% (–10°C to +50°C)

oscillator and Class D amp require almost no peripheral

parts and enable audio playback by simply connecting to

a speaker. In addition, the compact package reduces

mounting area.

Reduced mounting area

Conventional

MCU Speech Synthesis LSIs

Class D AmpOscillator

Speaker

LAPIS Semiconductor



Speaker

Built-in

ass D AmOscill

Reduces the number of

components

from 5 to 3


components

from 5 to 3

Ideal for short audio playback in automotive applications

Feature

1

Feature

2

Feature

3

Automotive-grade Speech Synthesis LSIs

2Mbit

4Mbit

4Mbit

6Mbit*

692Kbit

692Kbit

—

ML22Q573

ML22Q553

—

ML22Q374

ML22Q394

ML22572

ML22573

—

ML22594

—

—

シートベルトを締めて下さい

シートベルトを締めて下さい

08

Tick tick(blinker sound)

Fail-safe functions ensure high system reliability

Worry-free fail-safe functions

4 built-in error flags for detecting

speaker pin short-circuits,

internal LSI temperature,

supply voltage drops, and

MCU command transfer errors

make it possible to notify

the control MCU of

abnormalities via the ERR pin.

Enables simultaneous playback of up to 5 channels (4 channels audio playback + 1 channel external input). What’s more, the volume of each channel can be independently adjusted, providing greater ease-of-use.

5 channels mixing and independent volume adjustment

The ML22572, ML22(Q)573, ML22Q553, and ML22594

integrate a mixing function that enables simultaneous

playback of 4 internal and 1 external channels.

This makes it possible to play multiple sound effects, audio,

and external audio input at the same time. The volume of

each channel can be independently adjusted. And users can

play mixed sounds at a certain volume without changing

the volume of the original audio.

Applicable Models: ML22572, ML22(Q)573, ML22Q553, ML22594

GAIN

GAIN

channel5

SPAmp.

DAC

16bitDAC

VOL

ADPCM

channel1

CPU

CPU I/F

Independent volume control in 3 stages

channel2

channel3

channel4

Peeepeee


Seatbelt warning

Buuu-n

Seatbelt

warning

Buuu-n

Peeepeee

MIXING

AIG AVOL

External input

Short-circuit detection

Applicable Models: ML22Q553, ML22594, ML22Q374, ML22Q394

High temperatureerror detection

SpeechSynthesis LSIs

DetectionCircuit

ThermistorThermistor

ERR outputCPU

Output becomes Hat high temperatures(output normally L)

Applicable Models:ML22572, ML22(Q)573

Supply voltagedrop detection


Vref

ComparatorComparator

CPUERR output

When the voltage drops belowthe set voltage the output

becomes H(output normally L)


Command transfererror detection


ERR outputCPUSPI

I/FComparator

SI

SCK

SO

Inputcommands

twice

If the commands are differenttwo times, the output

becomes H(output normally L)


Flash memory types support onboard writingOnboard writing for each customer based on destination

The ML22Q573, ML22Q553, ML22Q374, and ML22Q394

are Flash memory types that enable onboard writing. This

allows audio data to be uploaded while mounted on

the customer’s board. Incorporating a speech synthesis LSI

on board facilitates inventory management, making it possible

to customize voice data according to the destination. In addition,

the flexible design enables rewriting before shipment as well

as audio data to be updated at service centers.

Note: Onboard write data must be purchased from a write manufacturer

Applicable Models: ML22Q573, ML22Q553, ML22Q374, ML22Q394

Shipment to English-speakingcountries

Shipment to Chinese-speakingcountries

Shipment to Korean-speakingcountries

Inventory management ofsystem boards by mounting

speech synthesis LSIs Hello!

你好。

Speech

Synthesis LSIs

Onboard program

System board

System board

System board

ERR outputCPU Short-circuit

detection

Short-circuit

Short-circuit


components

from 5 to 3

Feature

4

Feature

5

Feature

6

SP+ pin

SP– pin

Output becomes Hwhen ground faults occur

between the SP pins(output normally L)


09

ML226xx series (Built-in Flash Memory)

Flash Memory The ML226xx series is equipped with 4Mbit to 32Mbit

Flash memory. The large memory capacity is ideal for long

audio playback such as voice guidance, while Flash

memory makes rewriting of audio data possible.

ROM CapacitySerial I/FI2C I/F

ML228xx/7xx series

ML226xx series

e

ML22Q626/666 (32Mbit)

ML22Q625/665 (16Mbit)

ML22Q624/664 (8Mbit)

ML22Q623/663 (4Mbit)

ML22825/865/725/765 (16Mbit)

ML22824/864/724/764 (8Mbit)

ML22823/863/723/763 (4Mbit)

ParameterAdvantages of ROHM’s

New ProductsConventional Product

ML2282x/ML2286x

New ProductML22Q62x/ML22Q66x

ROM Types

ROM Capacity

Playback mode

Max Phrase Count

Internal speaker amp

Speaker disconnect/Short-circuit detection

Speaker Amp Supply

P2ROM™

4/8/16Mbit

4bit ADPCM28bit straight PCM

8bit non-linear PCM16bit straight PCM

1,024 4,096 4x higher phrase count

0.7W Class AB(8Ω, DVDD=5V)

Built-in high efficiency Class D amp with higher Max output

None Enhanced functionality

Same potential as DVDD3V MCU/5V speaker amp

combo possible

Flash Memory

4/8/16/32Mbit

HQ-ADPCM4bit ADPCM2

8bit straight PCM8bit non-linear PCM16bit straight PCM

1.0W Class D/AB(8Ω, DVDD=5V)

Available

Different potential as DVDD

Enables rewriting to Flashmemory from the MCU via SPI

Expanded memory variety

Utilizes market-provenhigh fidelity, high compression

HQ-ADPCM

　　to 2017 2018 2019 future from 2020

▶Built-in flash memory (can be rewritten from MCU*)

▶Built-in high fidelity, high compression algorithm(HQ-ADPCM)

▶Internal 1.0W speaker amp (Class AB/D)

▶Selectable internal/external oscillator

▶Integrated fault detection function(Speaker disconnect detection, Short-circuit detection)

▶Built-in P2ROMTM

▶4bit ADPCM2

▶Internal 0.7W speaker amp (Class AB)

▶Speed · pitch conversion

Ideal for long audio playback in industrial and consumer applications

High performance · High fidelityHigh performance · High fidelity

Long playback Speech Synthesis LSIs

onalityyImproved functiom ctionaand performancen manc

*Only ML22Q62x/22620 can be rewritten from MCU.

ML22Q663

ML22Q664

ML22Q665

ML22Q666

ML22Q623

ML22Q624

ML22Q625

ML22Q626

4Mbit

8Mbit

16Mbit

32Mbit

10

Enables rewriting to internal Flash memory from the MCU via SPI I/FThe ML22Q62x allows data to be rewritten to

the internal Flash memory via the MCU’s SPI I/F

In addition, audio data can be rewritten in the field

even after product shipment.

Built-in high output, high efficiency Class D ampThe ML22Q62x and ML22Q66x

integrate a high efficiency, high

output 1W Class D amp. Class D

amps provide greater efficiency and

generate less heat than conventional

Class AB amps.

Built-in speaker fault detection functionBoth the ML22Q62x and ML22Q66x are

equipped with failure detection functions for

detecting speaker disconnections and

short-circuits. This allows for immediate

notification of speaker failures. Speaker output

is stopped when a short-circuit is detected to

prevent overcurrent.

High performance · High fidelity

Feature

1

Feature

2

Feature

3

Rewriting audio data

ML22Q62x

Disconnection/short-circuit detection function

Speaker Output Power vs. Efficiency

100

80

60

40

20

00 1,000

Speaker Output Power (mW)

Effi

cie

ncy (%

)

2,000

Class DClass ABAve.

efficiency improved

40%

Short-circuit

Disconnection

Internal

speaker amp

Disconnectiondetection

Short-circuitdetection

Error Output

MCU SPI I/F Pin

RewritableInternal

Flash

memory

11

ML228xx series ML227xx series (Built-in P2ROMTM)

Development TAT Comparison

Audio DataCreation

6 to 8 weeks

Prototype Mass Production

3 to 4 weeks

6 to 8 weeks

Audio DataCreation

Mass Production

CarefulDevelopment

Mass Production

ShortenedDevelopment

Small to Large-ScaleProduction

9 to 12 weeks (shortened)

1/4 the time of Mask ROMMass production

in approx. 1 month

LSI Development

The ML228xx series includes a mixing function that

provides simultaneous playback of 2 internal and 1 external

channels while enabling the volume to be adjusted for

each sound. It is also possible to prioritize each channel.

For example, users can mix and use external voice calls

while generating voice guidance and sound effects using

the speech synthesis LSI. In addition, since the speaker

amp function can be used independently for external

voice data, it can be shared within the system as an amp

to reduce costs.

Integrated mixing function for 2 internal and 1 external channelsExternal input is also possible, allowing the speaker amp to be used independently

3 channels mixing and independent volume adjustment

Applicable Models:

ML22863, ML22864, ML22865, ML22823, ML22824, ML22825

GAIN

GAIN

channel3

SPAmp.

DAC

16bitDAC

VOL

ADPCM

channel1

CPU

CPU I/F

channel2

Ping pong

GoodAfternoon

MIXING

There’s a visitor.

AIG AVOL

留守メニュー再生

通　話

モニター

GoodAfternoon

External input

Independent volume control in 3 stages

The ML228xx and ML227xx integrate

LAPIS Semiconductor’s proprietary P2ROMTM.

The LSIs can be put in standby mode during assembly,

then written to with audio data after receiving an order,

significantly shortening both sampling creation and mass

production TAT.

Large ROM capacity is ideal for long audio playback such

as voice guidance.

P2ROM™

Without pitch/Speed conversion

With pitch/Speed conversion

Without pitch/Speed conversion

With pitch/Speed conversion

ROMCapacity

Serial I/FI2C I/F

P2ROMTM achieves shorter TAT Reduces inventory and delivery risks

The ML228xx and ML227xx integrate LAPIS Semiconductor’s

proprietary P2ROMTM memory technology for storing

audio data. This provides ROHM the capability to

immediately ship prototype samples and shorten mass

production TAT for customers by half. Development TAT

can be shortened to around one month, about a quarter

that of Mask ROM. P2ROMTM enables one-time writing in

the customer’s environment, making it possible to verify

the sound quality on the spot without taking additional

time. (Write shipment only from LAPIS Semiconductor for

mass production)

Ideal for long audio playback in industrial and consumer applications

Feature

1

Feature

2


components

from 5 to 3

Mask ROMProducts

P2ROMTM

Products

No prototype required

There’s a visitor.

Ping pong

ML22863

ML22864

ML22865

ML22763

ML22764

ML22765

ML22823

ML22824

ML22825

ML22723

ML22724

ML22725

4Mbit

8Mbit

16Mbit

Long playback Speech Synthesis LSIs

12

ML2256x series ML223xx series

The ML223xx series includes a disconnection detection

function that can provide notification (i.e. via LED) in the event

no sound is output due to disconnection between the speaker

amp and speaker. Passing current between SPP and SPM

enables detection through High/Low determination. This

ensures worry-free operation even in the event of speaker

disconnection.

Applicable Models: ML22Q374, ML22Q394


The built-in high accuracy 4.096MHz+1.5% (–10°C to +50°C)

oscillator and Class D amp require almost no peripheral

parts and enable audio playback by simply connecting to

a speaker. In addition, the compact package reduces

mounting area.


Disconnection detection function


100

80

60

40

20

00 1,000


Effi

cie

ncy (%

)

2,000

Class DClass AB

The ML2256x and ML223xx series are equipped with 692Kbit to

4Mbit ROM. Both series include pin-compatible Flash memory

and Mask ROM types, and share similar characteristics.

Developed based on Flash memory products, mass production

can be carried out with either Mask ROM products or Flash

memory shipped blank (unwritten).

Select the best format according to application needs.

Integrated Class D amp ensures high efficiency output

Class D speaker amps are capable of 1.0W (Max) output,

feature low heat generation even at high (audio) volumes,

and minimize heat dissipation measures. In addition,

efficiency is excellent at low battery power, reducing current

consumption and prolonging battery life.

ROM CapacityFlash MemoryMaskROM

Ideal for short audio playback in industrial and consumer applications

Feature

1

Feature

2

Feature

3

Reduced mounting area

Conventional



Speaker

LAPIS Semiconductor


Oscillator

Speaker

Built-in

Oscill


components

from 5 to 3


components

from 5 to 3

Integrated oscillator and Class D amp reduces component count

Built-in disconnection detection function eliminates the need to worry about speaker failure

Ave. efficiency improved

40%

Pin compatibility

DisconnectionDisconnectiondetection CPU

SPP

SPM

ERR

ERR output

SPAmp.

ERR output

H Disconnection

L Connection

Detection requestCPU

I/F

Detect speaker

failure

ML22562

ML22563

—

—

—

ML22Q563

ML22Q374

ML22Q394

2Mbit

4Mbit

692Kbit

692Kbit

Short playback Speech Synthesis LSIs

13

ML226xx series ML224xx series

Easily enable voice generation from external memory

Controlling external memory allows ROHM Speech Synthesis LSIs to

play back audio data without placing a load on the MCU. Voice

generation is easily achieved by inputting control commands of only a

few bytes from an MCU.

High fidelity, high compression HQ-ADPCM makes it possible to reduce memory capacity

The ML226xx series adopts the newly developed high fidelity, high

compression voice algorithm HQ-ADPCM that provides a clearer

sound and richer audio range while reducing data size compared with

conventional ADPCM, providing a more comfortable listening

experience. For example, melodies and sound effects that are degraded

with ADPCM and cannot be played with uncompressed PCM can be

compressed with HQ-ADPCM, ensuring superior sound quality while

reducing memory capacity. Up to 80% compression is possible vs

4bit ADPCM and 20% compared with 16bit ADPCM.

HQ-ADPCM enables high fidelity audio compression

Comparison of Data VolumeRequired for 30s Playback

16bitPCMApprox.

7.7Mbit Approx.

3Mbit

HQ-ADPCM

Approx.

1.9Mbit

4bitADPCM

Speech

Melody

Sound

Effects

Conventional

Reduced approx. 68％Reduced approx. 68％

16bitPCM

CD quality sound and data

With HQ-ADPCM

With Conventional ADPCM

Sound qualityHigh

Required data volumeLarge

Image

nd data

HQ-ADPC

onventtion

ConventionalConventional

HQ-ADPCMHQ-ADPCM

Built-in speaker amp

External Flash memory type

MCU

B



Serial Flash Memory

(Voice Data)

Play/Stop commands only

Thank you for

your purchase

No need to send audio data via MCU

SpeakerAmplifiers

Select the capacityaccording to

playback time(128Mbit or less)

Applicable Models: ML22620, ML22660

General-PurposeMCU

The ML226xx and ML224xx

series support external serial

memory up to 128Mbit.

Ideal for long audio playback.

Applicable Models: ML22620, ML22660


Integrated Class D amp ensures high efficiency output

Class D speaker amps are capable of 1.0W (Max) output, feature low

heat generation even at high (audio) volumes, and minimize heat

dissipation measures. In addition, efficiency is excellent at low battery

power, reducing current consumption and prolonging battery life.

Ideal for long audio playback using external memory in industrial and consumer applications

Feature

1

Feature

2

Feature

3

Speech Synthesis LSIs control external memory, facilitating voice generation

100

80

60

40

20

00 1,000


Effi

cie

ncy (%

)

2,000

Class DClass AB

Ave. efficiency improved

40%

Enablescompression ofmelodies andsound effects

ML226xx series

ML224xx series

to 2017 2018 2019 future from 2020

ML22620

ML22660

ML22420

ML22460

▶Internal 0.7W speaker amp (Class AB)

▶4bitADPCM2

▶4ch mixing

Supports external memory

▶Built-in high fidelity, high compression algorithm (HQ-ADPCM)

▶Internal 1.0W speaker amp（Class AB/Class D）▶Selectable internal/external oscillator

▶Integrated fault detection function (speaker disconnect detection, short-circuit detection）

High performance · High fidelityHigh performance · High fidelity


Improved functionalitym ctional yand performancen manc

14

ML610Q300 series

Reduced approx. 68％

Conventional

HQ-ADPCM


Achieves audio output functionality on a single chip■ Low-pass filter reduces noise

■ Audio algorithms

4bit ADPCM2, 8bit PCM (Non-linear/Straight),

16bit PCM (Straight)

■ Equipped with 1W (Max) speaker amp at 5V

■ Supply voltage range: 2.2V to 5.5V

Highperformance

with low powerconsumption

High outputspeaker amp

3V/5Vcompatibility

High fidelityaudio

playback

High performance 8bit CPU

U8 core■ Original RISC-type CPU

■ 3-stage pipeline structure enables 1 instruction

per clock operation

■ 8bit operation features equivalent performance

to 16bit MCUs

Easily enable audio playback

■ Audio playback functionality is achieved by

simply setting audio data in the register

■ Integrating 4-word FIFO into the audio register

makes it possible to extend the interval

between audio data request interrupts

LAPIS Semiconductor MCUs with speech output function incorporate an oscillator, EEPROM, speech synthesis LSI, speaker amp, and other peripheral components, contributing to set miniaturization.

Speaker

OscillatoratorOscilla

EEPROM

SpeakerSensor


SpeakerAmp

General-PurposeMCU

General-PurposeMCU

Sensor Audio Playback8bit MCUs

Space-saving monolithic designreduces failure riskConventional systems require 5 chips

Fewerexternal

parts

ML610Q300 series Elements Integrating peripheral components saves space and reduces failure risk

The ML610Q300 series of MCUs incorporate an original RISC

type 8bit CPU U8 core.

Compared to conventional middleware audio playback, audio

quality is superior, control is easier, and MCU performance can

be maximized.

Pipeline processing

1 Clock 1 Command

Clock

Fetch InstructionA

Decode

Execute

None Included Flash Memory P2ROM™

LCD Driver ROM Capacity

Ideal for audio playback using a single chip in industrial and consumer applications

Audio Playback Conditions

16bit PCM at 16kHz

Audio data request interrupt time

250µsec

Feature

1

Feature

2

Feature

3

InstructionA

InstructionB

InstructionB

InstructionB

InstructionC

InstructionC

InstructionC

InstructionA

ML610Q304

ML610Q360

—

—

—

ML610Q380

96KByte

160KByte

128KByte

None

16Mbit

None

(Audio data storage)

Audio playback 8bit MCUs

High performance · High fidelity

15

Development Support Tools

[Speech LSI Utility]

CD (Software [included])

PC

Amp

Speaker Dynamic speakers 8Ω and above

Power supply cable [included]

SDCB

A reference board is

mounted on the SDCB

Waveform/Phrase editing

Listening evaluation

Device evaluation

Various referenceboardsSpeech LSI Utility

SDCK (Sound Device Control Kit)

The Edit ROM function allows multiple phrases to be played in succession. The following functions can be configured using the Edit ROM function.

• Continuous playback (The specified number of continuous playbacks is unlimited, depending only on the memory capacity.)

• Silence insertion function (20ms to 1,024ms)

The Edit ROM function enables efficient use of the audio ROM memory capacity. Also, phrases can be played back continuously in a single

process without sending commands each time with an MCU, significantly reducing CPU load.

[Phrase Configuration Example using the Edit ROM Function] [ROM Data Conversion]

Speech Management Area

The weather

will be sunny Rain

today Tomorrow’s

weather will be

Edit ROM Area (Blank)

Phrase 1 The weather will be sunny

Phrase 2 The weather Rain today

Phrase 3 Tomorrow’s weather

Phrase 4 Tomorrow’s weather

Phrase 5 The weather Tomorrow’s weatherSilence

●A development tool for creating, writing, and listening to ROM data stored in LAPIS Semiconductor’s speech synthesis LSI.

●Enables a series of evaluations utilizing both hardware (SDCB and reference board) and software (Speech Synthesis LSI Utility)

●Optional starter kit makes it easy to begin development using LAPIS Semiconductor’s speech MCUs

SDCB (Sound Device Control Board)

The SDCB (Sound Device Control Board) enables

writing and evaluation to OTP/Flash by connecting

to a reference board that integrates LAPIS

Semiconductor’s speech synthesis LSI.

Target Device Reference BoardConnecting to the SDCB (Sound Device Control

Board) allows users to write and playback audio

data. In addition, usage and evaluation is possible

via MCU control by connecting to the customer's

system board.

Speech LSI UtilitySpeech LSI Utility is a tool for operating the SDCB.

Everything from waveform editing and ROM data

creation to sample writing and evaluation are

enabled.

SDCK (Sound Device Control Kit)

The Edit ROM function reduces memory capacity along with CPU load

today

will be sunny today

Rain today

will be sunny today Rain today

■ Facilitates operation from waveform editing to audio creation

■ Compatible with all models■ Compact SDCB (70×90mm)

USB cable

[included]

16

65.5 81.9 98.9 131.1 262.1 327.7 395.7 524.3 327.7 409.6 494.6 655.4

16.4 20.5 24.7 32.8 65.5 81.9 98.9 131.1 163.8 204.8 247.3 327.7

4.1 5.1 6.2 8.2 16.4 20.5 24.7 32.8 20.5 25.6 30.9 41.0

2.0 2.6 3.1 4.1 8.2 10.2 12.4 16.4 10.2 12.8 15.5 20.5

Memory size(Ex.)

■ HQ-ADPCM features a compression rate around 5x higher than that of 16bit PCM (uncompressed)

16bit PCM [Sec.](Uncompressed)

ADPCM2 [Sec.](Compressed)

HQ-ADPCM [Sec.](Compressed)

16Mbit(2MByte)

4Mbit(512KByte)

1Mbit(128KByte)

512Kbit(64KByte)

Relationship BetweenCompression Rateand Playback Time (With 16bit PCM=1)

1(16bit PCM)

4×(ADPCM2)

5×(HQ-ADPCM)

16kHz

8kHz

12.8kHz

10.7kHz

16kHz

8kHz

12.8kHz

10.7kHz

16kHz

8kHz

12.8kHz

10.7kHz

Get started as soon as you open the box

Standalone audio evaluation is possible without connecting to a PC

Easily start program development using the included program development support tool and sample program

Utilize the included audio data conversion tool to incorporate audio data into the control program

Speech MCU Development Starter Kit

Audio playback time for each memory size and sampling frequency

Easy!

Easy!

Easy!

●ML610Q304 Reference Board

●ML610Q304 Demo Board

●nanoEASE

●Speaker

●Cable Set

●USB Memory (Includes the following)

ML610Q304 Speech MCU Starter Kit Package

Audio Output Function Type

Cable Set

USB Memory

Reference Board

nanoEASE

8Ω Dynamic Speaker

Demo Board

ML610Q304 Speech MCU Starter KitMain Contents

Complete set of program development support tools

Complete set of audio data conversion tools

Audio scale playback sample program

17

Product Specifications

Segment Type 8bit MCU with Built-In LCD Driver

Automotive-grade Speech Synthesis LSIs (P.05 to 08)

Audio Playback 8bit MCUs (P.14)

Short Playback Speech Synthesis LSIs (P.12)

Standard 8bit MCUs (Industrial-grade)

Part No.OperatingFrequency Oscillator

4.096MHz4.000MHz

4.096MHz

4.096MHz

4.096MHz

ROM Capacity(bit) No. of Phrases

MaxPlayback Time

Flash2M

Flash4M

Flash16M

Mask ROM2M

Mask ROM4M

Flash4M

Flash4M

Flash692K

External128M Max

OperatingTemperature

(°C)

–40 to +105

–40 to +105

–40 to +105

–40 to +85

TQFP48

SSOP30

SSOP30

SSOP30

OperatingVoltage

(V)CPU I/F

SP AmpOutput (W)/

Class

MixingCount

(Internal)Other

2.7 to 3.6 or3.3 to 5.5

2.7 to 5.5

4.5 to 5.5

2.0 to 5.5

4,096

1,024

30

90sec.*1

192sec.*1

13min.*1

109min.*3

1.0/Class AB

1.0/Class AB

1.0/Class AB

1.0/Class D

4ch

4ch

4ch

1ch

Fail-safe

MCU rewrite*7

Failure detection

Speaker pinshort detection

function


function



functionI2C

I2C

Clock syncserialI2C

Clock syncserial

Clock syncserial

Clock syncserial

1,024*5

(Internal 512, External 512)

Mask 6M*4

External128M Max

Internal 304sec.*1

External 109min.*3

*5 Total of 512 phrase internal Mask and 512 phrase external memory

*6 1024 phrases (1 bank) x 4 banks

*7 In the case of clock sync serial

*1 Max playback time with HQ-ADPCM at 6.4kHz sampling frequency

*2 Max playback time with ADPCM2 at 6.4kHz sampling frequency

*3 Max playback time with HQ-ADPCM and external memory (128Mbit Max) at 6.4kHz sampling frequency

*4 Internal 6Mbit Mask can be connected to external memory (128Mbit)

ML22Q532

ML22Q533

ML22Q535

ML22530

ML22572

ML22573

ML22Q573

ML22Q553

ML22594

ML22Q374

ML22Q394

InternalExternal

External

External

Internal

99sec.*1

202sec.*1

28sec.*2

Mask2M

Mask4M

Flash692K

Flash4M

4.096MHz

4.096MHz

–40 to +85

–40 to +85

SSOP302.7 to 5.5

2.0 to 5.5

1,024

30

99sec.*1

1.0/Class AB

1.0/Class D

4ch

1ch

Fail-safe



function

Clock syncserial

Clock syncserial

ML22562

ML22563

ML22Q563

ML22Q374

ML22Q394

External

Internal

202sec.*1

28sec.*2

Package



MaxPlayback Time


(°C)

OperatingVoltage

(V)CPU I/F

SP AmpOutput (W)/

Class

MixingCount

(Internal)Other Package

TSSOP20

TSSOP20

TSSOP20

TSSOP20

ML610Q380 2.2 to 5.532.768kHz(Built-in RCoscillation/

Crystal oscillation)

8.192MHz 0.122µs30.5µs

2.0µA Flash 128K — Flash ROM 2K 7 4 34–40 to +70

ML610Q304 2.0 to 5.532.768kHz(Built-in

RC oscillation)8.192MHz 0.122µs/

30.5µs2.7µA –40 to +85 Flash 96K 2K Flash ROM 1K 1 3 11

ML610Q360 2.2 to 3.6 32.768kHz(Crystal oscillation) 8.192MHz 0.122µs/

30.5µs1.7µA –40 to +85 Flash

P2ROM™160K 3K P2ROM™ :

16Mbit2K 8 3 29

Part No.

Operating Conditions ROM/RAM Functions/Features

Functions/Features

OperatingVoltage

(V)

Operating Frequency (Max) MinInstruction

Execution Time

MinInstruction

Execution Time

CurrentConsumption(Typ@HALT)


(°C)ROM Type

ROM Capacity(Byte)

Data FlashCapacity

(Byte)

Audio DataStorage Area

RAM Capacity(Byte)

Port

Low-Speed Clock High-Speed Clock Input OutputInput/Output

Input/Output

Part No.

Operating Conditions ROM/RAM

OperatingVoltage

(V)

Operating Frequency (Max) CurrentConsumption(Typ@HALT)


(°C)ROM Type

ROM Capacity(Byte)

Data FlashCapacity

(Byte)Audio Data

Storage AreaRAM Capacity

(Byte)Port

Low-Speed Clock High-Speed Clock Input Output

18

Supports external memory Speech Synthesis LSIs (P.13)

Long Playback Speech Synthesis LSIs (P.09 to 11)

I2C

I2C

I2C

I2C

4.096MHz4.000MHz

4.096MHz

External128M Max

External128M Max

–40 to +70

–40 to +85

TQFP32WQFN32

SSOP30

Failure detection

—

—

2.7 to 3.6 or3.3 to 5.5

2.7 to 5.5

4,096

1,024

109min.*3

87min.*2

1.0/Class D

0.7/Class AB

4ch

4ch

MCU rewriteFailure detection

Failure detection

—

Clock syncserial

I2C

Clock syncserial

ML22620

ML22660

ML22420

ML22460

InternalExternal

External

4.096MHz4.000MHz

4.096MHz

Flash4M

Flash8M

Flash16M

Flash32M

–40 to +70

–40 to +85

TQFP32WQFN32

TQFP32WQFN32

SSOP30

SSOP30

SSOP30

SSOP30

2.7 to 3.6 or3.3 to 5.5

4,096

192sec.*1

397sec.*1

13min.*1

27min.*1

192sec.*1

397sec.*1

13min.*1

27min.*1

161sec.*2

325sec.*2

11min.*2

161sec.*2

325sec.*2

11min.*2

161sec.*2

325sec.*2

11min.*2

161sec.*2

325sec.*2

17min.*2

1.0/Class AB,Class D

1.0/Class AB,Class D

0.7/Class AB

0.7/Class AB

0.7/Class AB

0.7/Class AB

4ch

4ch

2ch

1ch

1ch

2ch

MCU rewriteFailure detection

Speed and pitch conversion

Speed and pitch conversion

Clock syncserial

Clock syncserial

Clock syncserial

ML22Q623

ML22Q624

ML22Q625

ML22Q626

ML22Q663

ML22Q664

ML22Q665

ML22Q666

ML22823

ML22824

ML22825

ML22723

ML22724

ML22725

ML22863

ML22864

ML22865

ML22763

ML22764

ML22765

InternalExternal

4.096MHz4.000MHz

Flash4M

Flash8M

Flash16M

Flash32M

–40 to +702.7 to 3.6 or

3.3 to 5.5

2.7 to 3.6 or4.5 to 5.5

4,096

1,024*6

InternalExternal

External

P2ROMTM

4MP2ROMTM

8M

P2ROMTM

16M

4.096MHz –40 to +852.7 to 3.6 or

4.5 to 5.51,024*6External

P2ROMTM

4M

P2ROMTM

8M

P2ROMTM

16M

4.096MHz –40 to +852.7 to 3.6 or


P2ROMTM

4M

P2ROMTM

8M

P2ROMTM

16M

4.096MHz –40 to +852.7 to 3.6 or


P2ROMTM

4M

P2ROMTM

8M

P2ROMTM

16M



MaxPlayback Time


(°C)

OperatingVoltage

(V)CPU I/F

SP AmpOutput (W)/

Class

MixingCount




MaxPlayback Time


(°C)

OperatingVoltage

(V)CPU I/F

SP AmpOutput (W)/

Class

MixingCount


6(16bit×3)

4(16bit×2)

8(16bit×4)

8bit Timer

8bit Timer

Functions/Features

Package CodePWM WDT ADC (Method)

Serial Port SupplyVoltage

Detection

SupplyVoltage

Detection

LCD DriverExternalInterruptFactor

ExternalInterruptFactor

Speaker Amp Output(W)/Class Other

I2C SSIO UART

Functions/Features

Package CodePWM WDT ADC (Method)

Serial PortLCD Driver

Speaker Amp Output(W)/Class Other

I2C SSIO UART

— 1 10bit×3(Sequential)

Master/Slave

×12 Half duplex

×1— — 9

1.0(5V operating voltage)/

Class D

Audio playback/ADPCM decoder/

internal speaker amp

P-VQFN28-0505-0.50/P-SSOP30-56-0.65

P-WQFN32-0505-0.50

— 1 12bit×4(Sequential)

— 2Half duplex×2(Full duplex×1/Half duplex×1)

VLS×1 — 70.5

(3V operating voltage)/Class AB


internal speaker ampP-TQFP64-1010-0.50

16bit×2 1 10bit×8(Sequential)

Master×1

2 Half duplex×2

BLD×1 96 dot Max24seg.×4com.

50.6

(5V operating voltage)/Class AB


internal speaker ampP-QFP80-1414-0.65

Catalog No.62B7240E 02.2020 PDF © 2020 ROHM Co., Ltd.

January 1st, 2020.

Documents

Speech Synthesis LSI Series Ver.3 - Rohm