TI launches first 64bit C2000 real time microcontroller
Cette publication existe aussi en Français
Texas Instruments (TI) has launched two series of its C2000 real-time microcontrollers including a 64bit version with custom logic and one with a neural processor.
The 200Mhz F29H85x series is built with three new 64-bit C29x digital signal processor cores with a context aware memory protection unit and custom logic, while the 150MHz TMS320F28P55x series is the first TI C28x real-time microcontroller with an integrated neural processing unit (NPU), enabling fault detections with high accuracy and low latency.
TI will show both devices at electronica 2024 in Munich, Germany, this week.
Designers are increasingly looking for a single MCU with more flash memory, larger computational capabilities and more integrated functionality. This need is particularly evident as vehicles shift from discrete solutions to integrated solutions, where a single chip is capable of handling various computational and housekeeping functions within the powertrain. TI’s new C29 core uses a 64bit architecture to provide more than double the real-time signal-chain performance of the C28 core.
The F29H85x also includes diagnostics and error-checking mechanisms to support ISO26262 design flows and IEC61508 automotive and industrial safety standards up to ASIL D and SIL 3. These MCUs also provide cybersecurity capabilities, with a fully isolated hardware security module that protects systems from unauthorized access and cyberthreats. Proprietary safety and security unit provides runtime safety and security without a performance penalty using a context-aware memory protection unit for hardware isolation of CPU tasks, with freedom from interference and self-test processes.
The C29x supports 32-bit and 64-bit floating- and fixed-point signal-processing running from 4MB of on-chip flash or RAM and also has trigonometric maths instructions to speed up common algorithms key to real-time control systems.
The C29x CPU1 and CPU2 cores can be put in lockstep for detection of permanent and transient faults while Logic Power-On Self-Test (LPOST) and Memory Power-On Self-Test (MPOST) provide start-up detection of latent faults. Safe interconnects provide fault detection between the CPU and the peripherals. An ADC safety checker compares ADC conversion results from multiple ADC modules without additional CPU cycles and a Waveform Analyzer and Diagnostic (WADI) can monitor multiple signals for proper operation and take action to ensure a safe state is maintained.
The device architecture features a Safe Interconnect (SIC) for end-to-end code and data safety, with CPU-based ECC protection for all memories and peripheral endpoints. Hardware Security Manager (HSM) provides security support with secure key and code provisioning in untrusted factory environments, and supports Firmware-Over-The-Air updates of HSM and host application firmware, with A/B swap capability and rollback control. SSU (Safety and Security unit) enables superior run-time safety and security features.
This can be used create safety isolation (Freedom From Interference) among the threads running on same CPU or different CPUs. The SSU features a context-sensitive MPU mechanism that automatically switches access permissions in hardware based on currently executing thread or task. This eliminates software overhead, enabling real-time code performance without compromising system safety. The SSU provides multi-user debug authentication, and also supports Live Firmware Update (LFU) and FOTA for application firmware updates with A/B swap and rollback control.
Two 16-bit Analog-to-Digital Converters (ADC) and three 12-bit ADCs have up to 80 analog channels as well as an integrated post-processing block and hardware oversampling. Two 12-bit buffered DACs and twenty-four comparator channels are available. Thirty-six frequency-independent PWMs, all with high-resolution capability, enable control of multiple power stages, from 3-phase inverters to advanced multilevel power topologies.
The PWMs have been enhanced with Minimum Dead-Band Logic (MINDL), Diode Emulation (DE), and Illegal Combo Logic (ICL) features. The Configurable Logic Block (CLB) allows the user to add custom logic and potentially integrate FPGA-like functions into the C2000 real-time MCU.
The NPU in the TMS320F28P55x series offloads the execution of the neural network model from the main CPU, achieving five to 10 times lower latency than software implementations to enable faster, more accurate decision-making.
The model that runs on the integrated NPU learns and adapts to different environments through training, helping systems achieve greater than 99% fault detection accuracy to enable more informed decision-making at the edge. TI’s complete AI toolchain, which includes models that are optimized and tested for specific applications, helps engineers with any level of experience complete the AI model development process.
“Engineers increasingly strive to design automotive and industrial applications that are more energy efficient and can make faster decisions, driving a need for more scalable processing power, expanded memory, and on-chip safety and security enablers,” said Amichai Ron, senior vice president, TI Embedded Processing. “For decades, our C2000 MCUs have been enabling manufacturers to measure and efficiently process real-world data. Today’s introduction of enhanced real-time performance and edge AI in our C2000 portfolio will empower engineers to solve more complex problems and reach higher levels of system efficiency, safety and sustainability.”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
