5744 RCCU problem

取消
显示结果 
显示  仅  | 搜索替代 
您的意思是: 

5744 RCCU problem

2,748 次查看
liujinhang
Contributor III

hi ,

we want some help about 5744 RCCU fault reason getting.

in attachment program, we made a fault flag like below:

163702_163702.bmprccu.bmp

we want to know the reason of the RCCU flag.

once again give you trouble and thank you.

Original Attachment has been moved to: RCCU-ERROR.zip

标签 (1)
10 回复数

2,266 次查看
李丽丽
Contributor I

When developing a projectI found the configruation register FCCU_NCFS_CFG2,FCCU_NCFS_CFG3,FCCU_NCFS_CFG4 are not defined in the basic software.

 

In the basic software,only NCFSC0-NCFSC31,now I want to configure NCFSC44,how to do?

0 项奖励
回复

2,267 次查看
petervlna
NXP TechSupport
NXP TechSupport

Hi,

I have checked your attached program and I cannot find any issue with RCCU faults.

The RCCU fault is set when microcontroller executions does not match safety lake execution.

NCF[11]  - RCCU_0a (Core redundancy mismatch: interface (other than D-MEM or DMA) out of lockstep.

There is also a note:  This fault is not generated in debug mode because the RCCUs are disabled in debug mode. See Core lockstep and RCCU disablement in debug mode for more information.

But according to my tests RCCUs are active even in debug mode. I have reported this some time ago, but I have no feedback from design yet.

To get an good overview on this fault I recommend you to read:

http://cache.nxp.com/files/32bit/doc/app_note/AN5259.pdf?fsrch=1&sr=3&pageNum=1 

Try to run your program without debugger (then the RCCU errors should not appear). When your program reach while loop attach debugger and check the results of FCCU.

Peter

0 项奖励
回复

2,267 次查看
chshli
Contributor II

Hi Peter,

About the above problems,do you have some feedback from design ?

Thanks

0 项奖励
回复

2,267 次查看
petervlna
NXP TechSupport
NXP TechSupport

Hi,

According to design RCCUs are disabled in debug mode.

As I cannot reproduce your issue, and here it is working correctly, how far you get with this issue?

Peter

0 项奖励
回复

2,267 次查看
chshli
Contributor II

Hi Peter,

According to the printscreen,NCF_S0 = 0x00000800,that is to say the NCF[11] is set,and we do reproduce the issue with the attached program in normal mode.We just want to know what leads the RCCU faults.

Thanks

0 项奖励
回复

2,267 次查看
petervlna
NXP TechSupport
NXP TechSupport

Hi,

Can you see this issue in stand alone mode (no debugger attached)?

Here is explanation of RCCU faults:

1.1.     Core redundancy mismatch: interface (other than D-MEM or DMA) out of lockstep

This fault is triggered whenever the redundancy checker (RCCU) detects a mismatch between any output of the safety lake and equivalent output of the original lake except D-MEM and DMA.

The root cause can be a permanent or transient fault either in the safety or original lake which propagated to any of the outputs of the lake.

pastedImage_1.png

1.2.     Core redundancy mismatch: D-MEM array interface out of lockstep

This fault is triggered whenever the redundancy checker (RCCU) detects a D-MEM mismatch in any outputs of the safety lake with the outputs of the original lake.

The root cause can be a permanent or transient fault either in the safety or original lake which propagated to any of the outputs of the lake.

pastedImage_2.png

1.3.     Core redundancy mismatch: DMA array interface out of lockstep

This fault is triggered whenever the redundancy checker (RCCU) detects a DMA mismatch between any output of the safety lake and equivalent output of the original lake.

The root cause can be a permanent or transient fault either in the safety or original lake which propagated to any of the outputs of the lake. The DMA array itself is not replicated, but DMA array interface (in other words DMA controller) is replicated.

pastedImage_3.png

0 项奖励
回复

2,267 次查看
brucemckenney
Contributor III

Are these fragments from a public document? I'm working with the MPC5643L, and I don't see anything like this in the RM or SM.

Yes, I can sorta-kinda guess from the (succinct!) text in the FCCU section what's going on, but what you've posted is still enlightening.

0 项奖励
回复

2,267 次查看
petervlna
NXP TechSupport
NXP TechSupport

Hi Bruce,

This is explanation from my application note dealing with FCCU fault sources:

http://cache.nxp.com/files/32bit/doc/app_note/AN5259.pdf?fsrch=1&sr=3&pageNum=1 

But this one is aimed for MPC5744P - which is new version of MPC5643L.

The safety mechanism is unfortunately different.

If you have any doubts about safety mechanisms on MPC5643L create a new question on forum and I can help you.

Peter

2,267 次查看
brucemckenney
Contributor III

Thank you, this is very interesting reading.

My reading of the 5643L vs 5744P Reference Manuals is that the RCCU mechanisms are actually rather similar. There is the delayed Lock-Step, but everything else seems to work the same. Indeed, it seemed that the 5744P is the "indicated upgrade" for the 5643L. Are there things I'm not seeing? (I'm not an expert at this.)

PS: Sorry for hijacking this thread, but it was Right There (:-)).

0 项奖励
回复

2,267 次查看
petervlna
NXP TechSupport
NXP TechSupport

Hi,

Yes and no.

In some way MPC5744P is update of MPC5643L.

Safety reaction modules like FCCU are however different. All faults are non-critical and user giver them priority based on design requirements. Connection between modules is also different.

Strategy of replicated safety area is similar.

However reporting and reaction on faults is more advanced in MPC5744P.

Peter

0 项奖励
回复