topic Re: SAFE-Mode, Error Analysis: MPC5643L in MPC5xxx

SAFE-Mode, Error Analysis: MPC5643L

nilsschroeder — Tue, 13 Jan 2015 16:40:03 GMT

The MCU on our board starts at SAFE-Mode, on the Eva-board the Code switches the current mode from DRUN to RUN1 without problems. The Code previously worked on our board. Which functions can we test? Can the voltage supply be the problem? Programming over the JTAG-interface is possible and the it looks like the Debugging is working. The FCCU seems to switch the MCU in SAVE-Mode. Should the xosc switch by the code be possible in SAVE-Mode? Witch procedure is recommended for an error-analysis. Where can I find detailed informations.

Re: SAFE-Mode, Error Analysis: MPC5643L

petervlna — Wed, 14 Jan 2015 13:54:37 GMT

Hi Nils.,

Nils Schroeder wrote:
The Code previously worked on our board. Which functions can we test? Can the voltage supply be the problem? Programming over the JTAG-interface is possible and the it looks like the Debugging is working. The FCCU seems to switch the MCU in SAVE-Mode. Should the xosc switch by the code be possible in SAVE-Mode? Witch procedure is recommended for an error-analysis. Where can I find detailed informations.

[Peter] – Yes, it can be problem. If you are using for example BIST self-test after power on than Leopard can cause the safe mode. So first of all make sure you are in Leopard datasheet spec and also make sure your power supply is strong enough to deliver the requested current in proper time.

To exclude the weak power supply issue you can apply external power supply to your board.

[Peter] – The first thing to do is to check the RGM registers DES and FES in order to identify the root cause of SAFE mode.

Are you in safe mode directly after reset? Place breakpoint or while(1) loop at the start of your code to make sure no other code is executed after reset and check the Mode entry for Mode.

Share with me the concept of RGM[FES], RGM[DES], and FCCU[CFSx], FCCU[NCFSx] registers.

Do not forget that you need to read FCCU via OPx bit set.

All the informations required for SAFE mode analyses are in MPC5643L reference manual. However this require wide experiences with this micro to fully understand how to find the issue.

Re: SAFE-Mode, Error Analysis: MPC5643L

nilsschroeder — Tue, 20 Jan 2015 13:37:00 GMT

Hello, thanks for the advises. The device is directly in SAFE-Mode after reset. With JTAG debugging we read the following register contents:

RGMFES 0xd480

RGMDES 0x8000

we figured out a flash, ECC or lock-step error event has occured (FFLECCRCC) is set to 1.

FCCU_STATE tells us the FCCU is in normal state. Sweitching to config state by writing key for OP1 and setting OPR in FCCU_CTRL to 1 results to Poeration statur Aborted.

We tried to switch the FCCU mode to OP 9 or 10 (FCCU.CTRL.B.OPR=9/10). But the [CFSx] and [NCFSx] Register dont change.

FCCUCFS0 0x00000000

FCCUCFS1 0x00000000

FCCUCFS2 0x00000000

FCCUCFS3 0x00000000

FCCUNCFS0 0x00000000

FCCUNCFS1 0x00000000

FCCUNCFS2 0x00000000

FCCUNCFS3 0x00000000

Are there any steps to do to run operations of the FCCU?

How can we investigate the problem reportet by FFLECCRCC.

Thank you.

Regards

Re: SAFE-Mode, Error Analysis: MPC5643L

petervlna — Tue, 20 Jan 2015 14:15:42 GMT

[Peter] – Ok, the micro can transition directly from RESET to SAFE mode via mode entry module.

RGM[DES] seems to me ok -> power on reset. If you are not handling POR flag in your application SW this flag will remain set.
RGM[DES] – Here I see the issue. First indication is flag
1. F_FCCU_SAFE -> FCCU detected fault and forced the micro to SAFE mode.
2. F_FL_ECC_RCC -> Flash, ECC event or RCC error.

To exclude RCC error it is good to know the mode of microcontroller.
1. Is it Lock-step mode?
2. Or Parallel Decoupled mode
To exclude the Flash error I can send you my example code which is also handling FCCU.

But I have a feeling that the safe mode is caused by your SW as there is not issue with voltage.

Re: SAFE-Mode, Error Analysis: MPC5643L

nilsschroeder — Tue, 20 Jan 2015 16:30:12 GMT

Hi,

the microcontroller is in Lock-step mode (verfied by project settings and bit in System Status Register)

Our SW works on the Evalboard but not on a self designed board. This board design already worked before with the same SW.

Power Supply and soldering was already checked. As written before JTAG and debugging is working. So we think the SW

have be correct and there might be an HW Fault detected by the controller.

Is there a way to distinguish between Flash,ECC,RCC faults to come closer to the origin of the fault?

Did we understand the procedure to read the FCCU (N)CF status registers right?

We would be very grateful if you could send us your example code.

Thanks again!

Regards,

Nils

Re: Re: SAFE-Mode, Error Analysis: MPC5643L

petervlna — Wed, 21 Jan 2015 11:16:42 GMT

[Peter] –

Hi Nils,

Attached is my example code for dealing with safe mode and correct startup sequence.

I have a feeling that this code will run fine even on your board.

I did not saw your startup code, but the problem can be here.

Before we will proceed let me know if my example is working as expected.

Tip: In the LSM the core registers must be initialized by 64-bit. If not then as soon as you use one of them the redundancy checker will trigger F_FL_ECC_RCC flag and force reset and put micro in SAFE mode.

Check also my startup file (startup.c) for comparism.