Difference between EIM and FCCU NCFF

[JRodrigues]

Hello,

I'm currently learning on how to work with some safety concepts on the S32K344-Q257EVB, namely the FCCU, and i wanted to test it and see how it reacts to faults (in terms of changing states in the Finite State Machine (FSM) and checking the FCCU1/FCCU2 signals with an oscilloscope to see if they get asserted or not).

My question is, how should i do that? On the reference manual i have read that the FCCU has some Non-Critical Fault Fake (NCFF) that could be used to achieve that, but there is also an Error Injection Module (EIM) and that is where my confusion begins. What is the difference between both? In which cases should i use one and/instead of the other?

Thanks in advance!

Best regards,

JRodrigues

[petervlna · Accepted solution]

Hello,

the NCFF referring to "All SRAM ECC uncorrectable error ERM_SRAM or ECC cache memory error.." can act as a fault that "happened in the memory and was then reported to the FCCU"?

NCFF will inject the fault only to FCCU fail state machine (FSM). So you can verify the reaction of device in case such fault will come to FCCU.

But NCFF will not inject the ECC error physically to ECC mechanism nor to the memory. For such purpose you will you EIM.

So if you want to verify the reaction of ECC and reaction path on error in ECC mechanism to FCCU NCFSx, you will need to use EIM.

Best regards,

Peter

[JRodrigues]

Hello Peter,

Thanks for the clarification.

Have a good rest of your day!

Best Regards,

JRodrigues

[petervlna]

Hello,

You can use for EOUT tests NCFF which can inject the faults to the FCCU. Usually you use it when you want to verify the FCCU reaction on incoming fault. It however cant physically inject fault to peripherals/ memories. In some cases there is inejction dirrectly on tested peripheral or FCCU can inject fault only to the status register like STCU_ERR and not in the test itself.

However EIM:

The Error Injection Module (EIM) is mainly used for diagnostic purposes. It provides a method for diagnostic coverage of internal memories (for example, system RAM, cache RAMs, and peripheral memories).

EIM enables you to induce artificial errors on error-checking mechanisms of a system, such as ECC for RAM read data and parity bits.

So EIM will create a physical fault on memory and memory then reports fault to FCCU. This way you can test for example if your ECC mechanism is working correctly, etc...

Best regards,

Peter

[JRodrigues]

Hi Peter,

So, in the S32K3xx_fault_map excel file that comes with the reference manual, the NCFF referring to "All SRAM ECC uncorrectable error ERM_SRAM or ECC cache memory error.." can act as a fault that "happened in the memory and was then reported to the FCCU"?

Best Regards,

JRodrigues