Hackster is hosting Hackster Holidays, Ep. 6: Livestream & Giveaway Drawing. Watch previous episodes or stream live on Monday!Stream Hackster Holidays, Ep. 6 on Monday!

Aspinity Launches AML100 Analog TinyML Chip, Promises 95 Percent Power Saving

Designed to let high-power digital parts stay asleep, this power-saving tinyML chip does its work in the analog domain.

Analog machine learning specialist Aspinity has launched the first product in its analogML range, the AML100, which it claims offers always-on tinyML projects the promise of a 95 percent reduction in power draw — boosting battery life considerably.

"We've long realized that reducing the power of each individual chip within an always-on system provides only incremental improvements to battery life," claims Tom Doyle, founder and chief executive of Aspinity. "That’s not good enough for manufacturers who need revolutionary power improvements. The AML100 reduces always-on system power to under 100μA, and that unlocks the potential of thousands of new kinds of applications running on battery."

The AML100 is made up of an array of independent configurable analog blocks (CABs), each of which can be turned to a variety of tasks including sensor interfacing and machine learning. Key to its capabilities is its operation in the analog, rather than digital, domain, which Aspinity claims allows the system to intelligently reduce the volume of data being shuffled around a hundredfold.

That, in fact, is how the chip delivers on its claimed 95 percent power savings: By figuring out what data are important before anything is converted to digital, the chip can keep higher-power digital components like microcontrollers asleep — dramatically boosting battery life for always-on tinyML projects.

The company has confirmed two evaluation kits, for security and voice-activation projects. (📹: Aspinity)

The initial AML100 supports four analog sensors, including microphones and accelerometers, and includes analog compression capabilities claimed to boost the accuracy of wake-word engines. Aspinity is targeting a range of use-cases for the part, including voice-first systems, security, predictive and preventative maintenance, and biomedical monitoring.

The chip is sampling today, Aspinity has confirmed, with volume production for general availability planned for the fourth quarter. The company has also launched two evaluation kits: EVK1 with glass-break and alarm tone detection for security applications and the EVK2 for voice detection with pre-roll collection and delivery.

More information on the AML100 and analogML is available on the Aspinity website.

Gareth Halfacree
Freelance journalist, technical author, hacker, tinkerer, erstwhile sysadmin. For hire: freelance@halfacree.co.uk.
Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles