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.

DP83640: Adding POE Capabilities, magnetics will change, want to ensure compatibility

Part Number: DP83640

Hello,

In a custom circuit board, we are currently using the DP83640 PHY with the Delfino F28388D C2000.    Works well.

For our 10/100 magnetics, we are using the JX0011D21BNL.   It looks like this:

We are now designing in POE.   We are looking at something more akin to the JK0-0144BNL which looks like this:

Is there any issue or concern with the compatibility of this magnetics and your PHY? 

Is there another (magnetics) you would recommend and why?

Are there any changes to the (circuit) configuration of the PHY using this topology?   I can supply schematics if there is a need.

Thank you for your support.   It is always appreciated.

Mike

  • Hello Mike,

    Structuraly I see the following differences : 

    1. CTs are shorted on your new magnetic. On your original schematics, were they also shorted on the board? If so then you should be structurally same. 

    2. 22nF caps connected to 75 ohms. If you can have a substitue without 22nF then it should be same as your earlier tested choice. We dont have any data supporting not to use 22nF caps but if you want to go for solution exactly same as your tested one then may be you can do a search for magnetic without 22nF.

    --

    Regards,

    Vikram

  • Vikram,

    The order of the common mode choke and 1:1 isolation transformers seem different between the two.   I may be wrong, but in my research on RJ45 connectors with integrated POE Magnetics,

    • Pulse seem to universally have the following order
    •            cable->   isolation transformer->    common mode choke->    PHY.   
    • Most other manufacturers seem to follow suit on this.   

    Sorry if it was not clear, but my main question was around that.   Is this topology acceptable/compatible with our TI PHY choice?   

    In regard to the caps....obviously the 75-ohm resistors must be ac coupled especially in a POE application.   I agree that the addition of the 22pF caps (to the original 1000pf single cap) should be questioned which is another reason I am reaching out.   Every manufacturer I've looked at seems to have added these in connectors designed for POE applications. 

    In regard to the shorting of the CT's, the F28388D evaluation board from TI has the CT's shorted on the board (to 3.3VDD) as does or our board.   This is required by the PHY.....I would think it would not matter whether the shorting was on the board or on the connector.   But, please Let me know if this concerns TI.

    Overall, I assume TI  has recommendations in regard to integrated magnetics topologies for POE as it clearly has recommendations for non-poe applications.   I can go hack in a connector choice to our existing design and test it, but prefer to start with TI and its experts.

    Thank you for your time.

    Mike

  • Mike,

    We sometimes see better EMC performance on boards with common mode choke towards the PHY side.

    Unfortunately we dont have a list of recommended or a "must-not" have list of RJ-45 connectors for PoE applications. While doing a research on recommendations , I see that in an earlier e2e thread one of my colleague has put together many of the TI's PoE reference designs : 

    https://e2e.ti.com/support/interface-group/interface/f/interface-forum/983127/dp83822i-best-phy-for-tps23881-poe-pse-implementation

    And here is the reference schematic of one of the PoE design with 640 : 

    /cfs-file/__key/communityserver-discussions-components-files/138/DP83848-POE-schematic--snlr019.pdf

    --

    Regards,

    Vikram

  • Hi Vikram,

    Thank you for the link to your colleague Gerome's related posting.   Several of those reference designs match the topology of isolation transformer first, common mode choke second....and have series caps per  75-ohm resistors on the inputs.   That is helpful.

    It would be useful if TI were to consolidate this information into an application note.  

    Thanks again for the help,

    Mike