Announcement

Collapse
No announcement yet.

SDRAM and Flash Length tunning/matching

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • SDRAM and Flash Length tunning/matching

    Hello there,
    this is my first time in this forum so here it goes.
    I am currently working on a project with STM32F4 MCU and the following SDRAM( W989D2DBJX6I @ 166 MHz) and NAND flash (MX30UF4G16AB-XKI) memories .
    Everything is running at 1,8 V

    The Board Stackup is as follow:
    L1 Signal
    L2 Vcc
    L3 Signal
    L4 Signal
    L5 Vss
    L6 Signal
    Now, due to space constraints the SDRAM and Flash are placed quite close to the MCU and therefor each other.
    Longest net from Flash is around 56mm and from SDRAM 37mm. I also added series resistors close to the MCU for both Data and Add Bus .

    So the issue is: I was able to match the length from NAND memorywithout much troubles. But the SDRAM since its so close to the MCU and due to its orientation the Data Lines 16..31 became quite short(~20mm) compared to the rest of the bus nets. This nets are not shared with the NAND flash.
    I was able to match the length by bringing the signals to L1 (upper Signal layer) and add some accordeons("zigzags") but I am afraid that this might bring more issues than benefits. Specially being one of the outside layers.
    The board layout is pretty much finished, just this issue remains. I know the best would be to start the routing from the SDRAM from zero by perhaps change the component orientation to vertical , but that would mean redo mutch of the work :-(.
    What would you recommend in such situation?

    Much appreciated for your help.

    Regards,
    Pedro


  • #2
    Hi,
    You're sure that stm32f4 works at 1.8 volt perfect.I am working with 3.3 volt with this ic.

    Comment


    • #3
      Hi ,
      according to ST Datasheet "Table 18. Limitations depending on the operating power supply range" its is possible to run it at 168MHz and between 1,7V to 2,1V.
      http://www.st.com/resource/en/datasheet/stm32f479ai.pdf
      I must admit I haven't tested it yet. In our application I will give the option to change to 3.3V as well in case we need more "MIPS".
      Greetings,
      Pedro

      Comment


      • #4
        Originally posted by pedlaf View Post
        Hi ,
        according to ST Datasheet "Table 18. Limitations depending on the operating power supply range" its is possible to run it at 168MHz and between 1,7V to 2,1V.
        http://www.st.com/resource/en/datasheet/stm32f479ai.pdf
        I must admit I haven't tested it yet. In our application I will give the option to change to 3.3V as well in case we need more "MIPS".
        Greetings,
        Pedro
        yes,you said right

        Comment


        • #5
          Honestly, I do not know It is very hard to say - the layout may work just fine (depends how it looks). Normally, I would try to avoid to go into this kind of situation from beginning, now it may be hard to try to improve it.

          As you correctly noted, what the tracks on L1 will be doing, you do not know. I am not sure how the L2 looks (how good reference plane it is), but layout on L1 in this case may influence EMC/EMI results.

          What to do? Depends how much time you have. If you are in hurry, you may try to build your board as it is. In the worse case, you may slow down the memory (?) but your software guys may still be able to use the board for software development. And later, if you also make improved version of your board, it could be interesting to compare EMC/EMI tests between them.

          If you feel bad and you are not sure, then re-do the layout.

          Comment


          • #6
            Hi there,
            many thanks for your input. I guess is like you said is a though one. I will keep the current design for the first prototype and later on make a second one with the improvements.
            Make it a great week.
            Greetings,
            Pedro

            Comment

            Working...
            X