General Description
The MSU30x2 is a monolithic talking microcomputer that can memorize voice up to 60/90/120 seconds using MOSEL qualified coding method(MPCM). Its an integration of traditional 4-bit microcomputer and voice chip with minimal external components. LCD driver and miscellaneous interface are provided for versatile applications. With more than bit 10K ROM/RAM inside, this chip meets every intelligent novelty. Customer requested function and voice data will be built in by changing masks during fabrication.

Hardware Features
■ Low current consumption
■ Maximal function with minimal cost
■ Current output could drive 8 ohm speaker with a transistor, Vout could drive buzzer directly.
■ The voice content is stored up to 120 seconds at 6 KHz (B0000h) and can be separated to 256 sections.
■ Duration of each section can be different and is multiples of 100h.
■ Duration of section with appended memory-less mute is up to 40 seconds (100000h).
■ Each trigger can access a sentence, up to 256 sentences could be access. 1024 entry count are provided.
■ Working at 2.4V through 6.0V
■ Precise voice sample rate 8 KHz is provided.
■ Auto ramp up and ramp down.
■ Halt mode is provided. Very low current consumption at Halt mode.
■ LCD driver provided, can drive up to 75 segments
■ Built-In clock generator
■ Built-In doubler, halver, tripler
■ Internal program ROM : 1024 x 15 bits;
■ Internal program RAM : 64 x 4 bits
■ Internal stack RAM: 4x10 bits.
■ Two 4-bit input ports
■ Two 4-bit input/output ports
■ One 4-bit output port

Software Features
■ 76 instructions, in 39 mnemonics
■ 4-level subroutine nesting (also used for interrupt)
■ Two external factors (INT, S&M) for interrupt
■ Two internal factors (timer, divider) for interrupt