2015年3月26日 星期四

MSP430 goes 32bit with Cortex-M4F

MSP432 Texas Instruments has introduced an ARM Cortex-M4F version of its MSP430 microcontroller family, taking it from 16 to 32bits, with no extra power draw.


“The integrated DSP engine and floating-point core of the ARM Cortex-M4F core enable a multitude of high performance applications, such as signal conditioning and sensor processing, while maintaining performance headroom for product differentiation,” said TI.


Using its ARM architectural licence to modify the core, and its own 90nm low-power chip process, TI claims this the lowest production Cortex-M4F yet. Running at up to 48MHz, power is 95µA/MHz active power and 850nA in stand-by. The process is flash only, so there will be no FRAM versions.


Called the MSP432 family, most peripherals are shared with the MSP430 family, although there is a new ADC – 14bit resolution (13.2 ENOB) at 1Msample/s consuming 375µA at full speed.


On EEMBC’s ultra-low power ULPBench, it scores 167.4 “outperforming all other Cortex-M3 and M4F MCUs on the market”, said TI.


What about Ambiq’s near threshold and sub-threshold design? This has lower figures.


MSP432 But it is not in production, replied a TI spokesman – pointing out that TI made samples of a sub-threshold MCU available six years ago and could not get process variation under control.


Power comes through a choice of on-chip regulators – and LDO or dc-dc converter. Operating voltage is 1.62 to 3.7V, so it can be connected to lithium primary coin cells, but not Li-ion cells which peak at 4.2V when charged.


RAM is split into eight separately-powered banks to allow power saving at 30nA per bank.


Flash is up to 256kbyte in dual banks for simultaneous read and write.


Encryption for data and code is covered by (AES) 256 hardware.


Design support includes evaluation boards, evaluation software, and plenty of documents from day-one.







from News http://ift.tt/1E3Rjee

via Yuichun

沒有留言:

張貼留言