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.

DP83TC811R-Q1: RGMII interface timing budget

Part Number: DP83TC811R-Q1
Other Parts Discussed in Thread: DP83TC811

Hello Colleagues,

I'm trying to understand the timing budget of the RGMII link of the DP83TC811.

My layout can be considered perfect, all lines matched less than 1 cm. (Both RX and TX) At least this can be excluded from the equations.

The clock is 25MHz, data lines (RX and TX) are SDR and the control lines (RX_CTRL, TX_CTRL) are DDR as specified in the datasheet, and RGMII specification.

The clock period is 40 ns.

First I would like to figure out how to configure clock delay.

My current setup is:

PHY: TX Shift mode (internal delay) RX Shift mode (internal delay)

MAC:TX Aligned mode (internal delay) RX Aligned mode (internal delay)

I think this setup is correct, but when doing the measurements the RX shifting will be visible, and the TX shifting will happen in the PHY.

When the setup is:

PHY: TX Shift mode (internal delay) RX Aligned mode (no internal delay)

MAC:TX Aligned mode (no internal delay) RX Shift mode (internal delay)

I'll see the bus aligned, but I will never be sure, if my setup is correct, since I cannot measure the internal signals in PHY or MAC.

(Let's assume, that I don't know anything about the SW, at any point of time. I can be sure about what the oscilloscope is showing to me. )

Then it could be more convincing to set as follows:

PHY: TX Aligned mode (no internal delay) RX Shift mode (internal delay)

MAC:TX Shift mode (internal delay) RX Aligned mode (no internal delay)

I'll see the clock delayed, so I can be sure, that at least before the receiver blocks (PHY - TX, MAC - RX), the clocks are correctly delayed.

However at this low speed, I could add delay for both RX and TX on both MAC and PHY side , it wouldn't cause any problem. (assume 2 ns in MAC, and 2ns in PHY )

PHY: TX Shift mode (internal delay) RX Shift mode (internal delay)

MAC:TX Shift mode (internal delay) RX Shift mode (internal delay)

Then I'll see both RX and TX clock delayed, but there will be even more inside MAC-RX and PHY-TX.

What do you think?

If you could do any of the above mentioned setups, which one would you prefer and why?

Many thanks!

  • Hi Atiila,

    Typically as a starting point I would either pick MAC or PHY as the shifted mode and then the opposite as aligned as you mentioned. 

    My question for you is what type of issues are you seeing when configuring in all these different setups? Perhaps setting up a table would be easier to show what kind of results you're seeing as I'd expect my set up suggestion would have proper communication between MAC and PHY.

    Thanks,

    Cecilia

  • Hi Cecilia,

    At the moment, I don't have any issue, except I need to define to the software guys how to set up the link correctly.

    Now I'm sure, that I understand the meaning of aligned and shifted mode, so after reading the datasheet of the SoC and PHY I can define the setup.

    What do you think, would it work if I set up the delayed mode for both RX and TX, both MAC and PHY? (Maybe it would be even robust at this low speed?)

    Thanks,

    Attila

  • Hi Attila,

    I would suggest having only one of the devices in delayed mode and the other in aligned. Your MAC should be programmable and allow you to change the modes if possible. The DP83TC811 at least gives you this option to try out the different modes through software or hardware. 

    Thanks

    Cecilia