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.

DP83848YB: Questions regarding timing specs of the DP83848YB

Part Number: DP83848YB

A customer has an inquiry on the timing specs of the DP83848YB.  See below.

I have some questions regarding timing specs in the DP83848YB datasheet. 1) The timing specs T2.4.1 and T2.5.1 for TX_CLK and RX_CLK high/low time of 16ns to 24ns, could this vary this much from cycle to cycle or would it be the same from cycle to cycle? For example if the current cycle is 16ns, could the next cycle be 24ns? 2) Also, I don't see any timing specs on period jitter for TX_CLK and RX_CLK. Would it be the same as the input jitter on the oscillator connected to X1? For example, if the oscillator we are using has 20psec of peak to peak period jitter, should we therefore add 20ps of period jitter to TX_CLK and RX_CLK? 3) Also, the setup, hold, and output delay timing specs for MII, do these timing specs already account for period jitter on TX_CLK and RX_CLK, assuming we are meeting the 25MHz oscillator spec of less than 800psec of input jitter?

  • Hi Brandon,

    Here are the answers to your questions:
    1. TX_CLK and RX_CLK duty cycle is outlines within the IEEE Specification 22.2.2.1 and 22.2.2.2.
    The requirement is for a nominal 20ns high/low with variation allowed from 14ns to 26ns. The DP83848YB will exceed this requirement by 2ns on both high and low side. You will be well within the specification. When you actually look at it on the scope and in simulations, our TX_CLK and RX_CLK do not swing with this much variation cycle-to-cycle.
    2. You will not see enough jitter on TX_CLK or RX_CLK to cause setup/hold failures. There is enough margin for the 25MHz clock at 100BASE-TX and 2.5MHz at 10BASE-T. You would see issue on the MDI (cable side) before you get close to seeing timing issues with the MAC. The XI jitter spec will ensure you maintain low enough jitter for the MDI to work properly and pass the jitter spec for PMA.
    3. Setup/hold and output delay will give you more than enough margin. Again, if you see any significant amount of jitter on the MII, you will not be able to link on the MDI. I have never seen any jitter issues experienced on the MII. The device will break on the MDI before the MII.

    Kind regards,
    Ross
  • Hi Ross: Thank you for the reply. We shared with customer. Please see his followup below.

    Hello Naser, I have an additional question regarding the DP83848YB that I would like to add. The datasheet does not specify any dynamic overshoot spec for the inputs above the VCC or below ground. It only has the DC Input Voltage (Vin) spec of -0.5V to VCC+0.5V in the Absolute Max Ratings.

    A lot of datasheets will have an additional dynamic overshoot spec of something like -2.0V to VCC+2.0V for 2nsec. Can you provide a spec like this for this chip?

    Using your IBIS model we are seeing about -0.7V for less than 2nsec which is exceeding the DC input spec, and need to determine is this is OK since it is less than 2nsec and not DC.

    Regards,
    Naser