SFFS277 November 2023 TMS320F280033 , TMS320F280034 , TMS320F280034-Q1 , TMS320F280036-Q1 , TMS320F280036C-Q1 , TMS320F280037 , TMS320F280037-Q1 , TMS320F280037C , TMS320F280037C-Q1 , TMS320F280038-Q1 , TMS320F280038C-Q1 , TMS320F280039 , TMS320F280039-Q1 , TMS320F280039C , TMS320F280039C-Q1
The CPU subsystem has a pair of diverse processing units (C28x and CLA) with different architecture and instruction set. This enables one processing unit to be used for handling the time critical portion code (control CPU) and other processing unit (supervisor CPU) to execute non critical portion of the code, perform diagnostic functions and supervise execution of the control CPU.
In case of identification of fault during diagnostic functions of the supervisor CPU, it can cause the TMS320F28003x MCU to move to a safe state. This concept, “reciprocal comparison by software in separate processing units” acts as a 1oo1D structure providing high diagnostic coverage for the processing units as per ISO 26262-5:2018, Table D.4. The comparison need to be performed several times during a FTTI. Reciprocal comparison is a software diagnostic feature and hence care should be taken to avoid common mode failures. The final attained coverage will depend on quality of comparison (determined by extend and frequency of cross checking). The proposed cross checking mechanism allows for hardware and software diversity since different processors with different instruction set and compiler is used for enabling this. The diversity can be further increased by having separate algorithms being executed in both the cores. In case, failure is identified during reciprocal comparison, NMI can be triggered by software and this in turn will assert ERRORSTS. During the run time, CLA has access to GPIO_Data_Regs that can indicate the error condition on a GPIO pin independent of C28x.