TI

The quest for low-power devices: 2012

TI creates a new low power (lower! as if the old MSP430 wasn’t already a true low power mcu) mcu product line.  The announced power reduction is @50% with new “Wolverine” MCU platform.

– Active power as low as 100 µA/MHz
– <400 nA standby with RTC and brown-out protection
– 250× less energy per bit with FRAM
– < 7 µs to wake to active from standby

The announced “Wolverine” keeps as many peripherals as can be in a “retention” mode using a power gating controller. Only the needed modules are fully powered, the idle ones are in “retention” mode.

The FRAM unified memory architecture can also lower all the memory power requirements.

It looks like one of the big energy efficient boards but at a nano-scale.

The announced voltages are 0.9v(GO, TI, GO!), 1.8v, 3.3v and 5v.

Product line page: http://www.ti.com/ww/en/mcu/wolverine/

RSS EETimes Semiconductor News
  • Comment on Chip Startups for AI in Edge and Endpoint Applications by David-Dorval February 22, 2020
    Hello, very nice article about edge and end point AI. At STIMIO, where we built entire sensors, MCU and telecom solutions, we are very concerned about energy consumption of our products that run on batteries, often very costly to change. And end-point AI is definitely the solution to extend device as long as firmware update […]
  • Comment on The Gremlins in the Machine by dt_hayden February 21, 2020
    I think the lack of such a button comes from a combination of extreme linear thinking and a bit of arrogance. Along the lines of "Why would you possible need an oops! button, just hit the buttons you want.". Here in the cube farm, we have a home grown software application, developed by at least […]
  • Comment on Chip Startups for AI in Edge and Endpoint Applications by ckachris February 21, 2020
    Great article Sally! Besides performance, one critical point for the adoption and success of AI chips is the easy integration with high-level ML framework like Spark ML, Scikit-learn, pytorch, tensorflow etc. Chip vendors must provide high level APIs that will allow easy deployment, scaling, resource management and task scheduling in a familiar way as it […]