• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 566 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

DP83867IR: Decoupling capacitors and current-mode drive magnetics issues?

Dylan Shekter
Prodigy 60 points
Part Number: DP83867IR

Hi,

I am using a 3rd party board that has a DP83867IRPAP PHY. We occasionally see issues with our product in the form of Ethernet links not coming up or huge power consumption increases (according to an input current monitor) with the PHY increasing ~35 degrees Celcius on on the package (according to a thermocouple).

We have identified 2 issues with the schematic and are wondering the impact of them.

The first and most obvious mistake is that the magnetics integrated in the jack used are for current-mode drivers, and contain a ferrite bead to 1.8v. Of the 2 times I've been able to get logs off of our system after, I saw in both cases the PHY downshifted to 100Mb and the Linux driver complained to check cabling. Not only is there this ferrite bead, but also the center taps of the first set of transformers are all shorted together (and then go to the bead). We are wondering about impact of this mistake. I've seen This forum post and the recommendation not to tie them together. However, I'm still wondering about what we can see.

The other problem is that this board uses the old recommendation of 1 nF capacitors. Can more information be shared on why TI increased the recommended values?

Thank you

  • 0
    TI__Mastermind 39785 points

    Hello,

    DP83867 is a voltage mode driver and has been tested with the configuration listed in the datasheet; no termination resistors required nor center tap shorting. If the device is paired up with a current mode mag-jack, this would violate datasheet guidelines and we have no data for its behavior unfortunately. It is recommended to change the design to a voltage mode mag-jack whose parameters and configurations meet those in the datasheet.

    Regarding 1nF, is this for the 1.0V supply for RGZ? As this is PAP package, this should be 1uF. I would believe that the issue is most likely due to the mag-jack issue than this.

    Sincerely,

    Gerome

  • 0 in reply to Gerome Cacho
    Prodigy 60 points

    Hi Gerome,

    Thank you. Unfortunately we're not in a position to change the jack to a voltage mode jack without a lot of NRE. Just want to confirm when you say

    > If the device is paired up with a voltage mode mag-jack, this would violate datasheet guidelines

    you mean a current mode mag-jack? Even if there's no data, could there be any educated guesses made? We are not the most familiar with this edge case. We can remove the ferrite bead but any information/experience on what happens if the center taps are shorted would be greatly appreciated.

    The 1nF are on all rails, the 1.1v, 2.5v, 1.8v. We have an DP83867IRPAP not an RGZ so we use 1.1v. Taking a look at an old revision of the datasheet, I see that 1 nF was the recommended on all 3 of these rails and isn't anymore. We're certainly focusing more on the mag-jack, but we do want to run to ground why the recommendation changed in case whatever caused the need for change comes up.

    Thank you,

    Dylan

  • +1 in reply to Dylan Shekter
    TI__Mastermind 39785 points

    Hi Dylan,

    Regarding 1nF, this is was initial guidance which has since been updated with new learnings. 

    Regarding clarification, I have updated my previous answer to correct my typo. With current mode mag-jack, team has no information and would violate datasheet guidelines. The best educated guess is that the current mode device on a voltage mode driver would incorrectly bias the line which would disrupt communication.

    Sincerely,

    Gerome

  • 0 in reply to Gerome Cacho
    Prodigy 60 points

    Hi Gerome,

    Thank you very much. We have never seen issues once the link is established, only when bringing up the link. Is that in line with what you're thinking the disruption would be?

    Thank you,

    Dylan

  • 0 in reply to Dylan Shekter
    TI__Mastermind 39785 points

    Hi Dylan,

    As this violates the boundaries of the datasheet, there is no conclusive data through which I can refer to.

    Sincerely,
    Gerome