This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA4VM-Q1: I want to use PCIE1(4 lanes) communicate between the two TDA4VH. How should I asign PCIe relate control signal such as wake#/perst#/prsnt#/clkreq# in this condition?

Part Number: TDA4VM-Q1
Other Parts Discussed in Thread: TDA4VH

Dear TI,

I want to use PCIE1(4 lanes) communicate between the two TDA4VH in our L3 ADS DCU.  These two TDA4VH formed as a double redundancy structure just as I draft in above picture.One TDA4VH's fault behave can not affect the other nomarl operation is the double redundancy structure.So Can you tell me how should I asign PCIe relate control signal such as  wake#/perst#/prsnt#/clkreq# in this condition?

Thanks!

  • Hello, 

    There are different ways and details to achieve redundant systems so below notes are just one "view" for your reference. I will start by first listing some guessed redundant functions/restrictions:

    Requirement of the redundancy system:

    1. will each side access other side's address space during functional (i.e., ) execution?

    2. will each side access other side's address space for diagnostics (i.e., compare results, synchronization)?

    With these considerations, we may have the following notes:

    1. topology - you may either use one side as RC and other side as EP, or, both sides as EP. Where RC-EP configuration would be more standard. 

    2. refclk - instead of letting RC to supply refclk to EP, you can configure both sides to use their internal refclk.

    3. PERST - again if you like to use more standardized RC-EP model and offshelf drivers, you can have it. otherwise, you can just start EP first and then RC, similar to the SDK demo. 

    4. Software - you may need to hijacker PCIe link retraining to safety handling function. 

    regards

    jian