msp430 clock

veeraputhiran Thangavel said:

what difference does it make when compare to single clock sourced controller ?

Flexibility.

You can run CPU from one clock (MCLK) while peripherals from another (ACLK, SMCLK). For example clock/watch/timekeeping application usually needs 32kHz oscillator. For single clock microcontroller it would mean that your CPU clock also shall run slooooow using 32kHz clock. This is not the case with msp430 where for clock timer peripheral you can use 32kHz while for CPU have fast 16 MHz or 24 MHz  built-in DCO oscillator clock.

Multiple clock sources provide flexibility in choosing clock sources for different modules within an MCU.

For example, the MSP430G2553 has three clocks:
ACLK: Auxiliary clock
MCLK: Master clock. 
SMCLK: Sub-main clock

If you browse through the MSP430 Family User Guide, you will find that the different peripheral modules like TIMERS, UART, SPI,etc can be driven using any of the above mentioned clock signals, along with the DCO(Digitally Controlled Oscillator)

The reason being is, if you want to achieve a precise low- frequency for a TIMER output, you will need a clock source with low-frequency to do so. In this case, you can go for the ACLK, after soldering the 32kHz crystal that comes with the Launchpad Kit.

And, precise baud rates for UART applications can also be derived from specific clocks. 

There are many other reasons to why we require multiple clock sources, an important one being the Low Power Modes(LPM), as far the MSP Family is concerned.

Regards 

llmars summed this up into a single word beautifully: Flexibility.

Having multiple clocks also gives MSP430 the ability to trade off between performance and power savings. Higher frequency clocks are important when processing at full speed but they can also be turned off in low power modes to reduce current consumption. Looking at the datasheet (http://www.ti.com/lit/ds/symlink/msp430g2553.pdf pg. 23) shows which clocks are turned off in each low power mode (i.e. LPM0: mclk = 0 MHz) and how much this affects current consumption (keep in mind these numbers are estimates, actual current consumption must be measured).

 For more information on MSP430 clocks feel free to check out the user guide: http://www.ti.com/lit/ug/slau144j/slau144j.pdf section 5.

-Nate

veeraputhiran Thangavel said:

What is the need of many clock source in msp430?

veeraputhiran Thangavel said:

what difference does it make when compare to single clock sourced controller ?

For timing, you might need a very precise clock. So you can use a high-precision high frequency crystal and use a lot of power to drive it, just to then scale the frequency down as you only need to run an RTC from it while the CPU (which would need the high frequency) is sleeping most of the time and doesn't need high precision.
So there is  an internal variable speed clock source (the DCO), which isn't very precise, but built-in and cheap, an external crystal input for a high-speed crystal (if required), another one for (usually) a low-power watch crystal and depending on family, even more. Like the very low power low speed oscillator, which is very unprecise, but consumes almost no power. You can pick the clock source that fits your needs best and leave the others off until you need them (if at all). Remember, the MSP series is built for ultra-low power consumption, not for ultra-high coder convenience :)