The AML100 has been designed to reduce always-on system power by 95%, allowing manufacturers to extend the battery life of devices or migrate walled powered always-on devices to battery – opening up new classes of products for voice-first systems, home and commercial security, predictive and preventative maintenance, and biomedical monitoring.
By minimising the quantity and movement of data through a system it is possible to efficiently reduce power consumption, but many always-on devices don’t have that capability. Instead, they continuously collect large amounts of natively analogue data as they monitor their environment and digitise the data immediately – in the process, wasting system power processing data that are mostly irrelevant to the application.
The AML100 delivers substantial system-level power-savings by moving the ML workload to ultra-low-power analogue, where the AML100 can determine data relevancy with a high degree of accuracy and at near-zero power. Consequently, the AML100 is the only tinyML chip that is capable of intelligently reducing data at the sensor while the data is still analogue and keeps the digital components in low power mode until important data is detected, eliminating the power penalty of digitisation, digital processing, and transmission of irrelevant data.
“We’ve long realised that reducing the power of each individual chip within an always-on system provides only incremental improvements to battery life,” said Tom Doyle, founder and CEO, 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 heart of the AML100 is an array of independent, configurable analogue blocks (CABs) that are fully programmable within software to support a wide range of functions, including sensor interfacing and ML. This versatility is able to deliver a tremendous advantage over other analogue approaches, which are rigid and only address a single function.
The AML100, however, is highly flexible, and can be reprogrammed in the field with software updates or with new algorithms targeting other always-on applications.
The precise programmability of the AML100’s analogue circuits also eliminate the chip-to-chip performance inconsistencies typical of standard analogue CMOS process variation, which has severely limited the use of highly sophisticated analogue chips, even when the inherent low power of analogue makes it better suited for a specific task.
Key Features of the AML100
- Consumes less than 20µA when always-sensing
- Intelligently reduces quantity of data by up to 100x while the data are still in analogue
- Features field-programmable functionality to address a wide range of always-on applications
- Leverages patented analogue compression technology for preroll collection to maintain accuracy of wake word engine in voice-enabled devices
- Supports 4 analogue sensors in any combination (microphones, accelerometers, etc.)
- Available in 7mm x 7mm 48-pin QFN package
The AML100 is currently sampling with key customers with volume production planned for Q4 2022.
