Deploying Edge AI to Detect Photovoltaic Arcing with the TMS320F28P55x Real-Time Microcontroller

Arcing in photovoltaic systems is a fire risk. Over time, electrical connections corrode causing arcing that can remain undetected for long periods of time. Today's arc detection systems rely on FFT algorithms but, due to the broad and changing nature of electrical arcing noise, detection rates can be as low as 60%.

Texas Instruments has built a demonstrator for this application around their lastest TMS320F28P55x real-time microcontroller. In addition to its powerful C2000 core it features a neural processing unit (NPU). With some clever machine learning, it proves to be a much more accurate approach to arc detection, offering much sought after innovation in these renewable energy applications.
 
 

Motor Control for Safety-Critical Industrial and Automotive Applications with TI's New F29H85x Series Real-Time Microcontrollers

Building on the well-known C2000 family of real-time microcontrollers, Texas Instruments have launched the F29H85x series. These feature their latest C2900 processor with very long instruction word (VLIW) architecture capable of executing eight instructions per clock cycle. 

The series targets functional safety applications in automotive (ASIL D) and industrial (SIL3). In the superset device, the three cores can be partitioned as a single core and two lock-step cores. The peripherals and interfaces include CAN, CAN-FD, SENT, and others along with picosecond accurate PWMs that are highly integrated with the ADC block. This makes them ideal for power converters and FOC motor control. 

In the demo at electronica, two FOC-controlled motors only required 10% of the total available processor performance, leaving plenty of cycles for other parts of an embedded application.