• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 2 answers
  • Subscribers 65 subscribers
  • Views 619 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

Can TPS2373 work with UPoE PSE?

chentsu chung
Mastermind 6180 points
Part Number: TPS2373
Other Parts Discussed in Thread: TPS23756, TPS2372

Hi Team,

Our customer use UPoE PSE injector as source of TPS2373, but this cannot work.

Further more, this UPoE injector work well with TPS23756.

1. For UPoE sourcing, TPS2373 should work as class4 PD, am I right?

2. Where TI recommend that I can look into?

Best regards,

C.T.

  • 0
    TI__Expert 5615 points

    C.T.,

    Have you, or your customer, already reviewed our available collateral on this topic?

    E2E FAQ:  

    Training Video:  

    Please let me know if/when you have and if having any additional concerns or questions.

    Thank you for your local support and promotion of PoE products from Texas Instruments,

  • 0 in reply to Thomas Lewis
    Mastermind 6180 points

    Hi Thomas,

    1. Yes, This training mention that "need to configure BT's PDs first-class resistor to class 4", so is this means just make sure CLSA of TPS2373 is 63.4Ohm, and CLSB don't care?

    2. We tried to modity both CLSA and CLSB to 63.4Ohm, and it's worked. Is this the only way to do?

    3. Customer use ST's BT solution to test with UPoE, it's work without any modification, does TI has any ideal?

    4. Or any other way to make TPS2373 work with UPoE without modifying CLSA/CLSB?

    Thanks you very much,

    C.T.

  • 0 in reply to chentsu chung
    TI__Expert 5615 points

    C.T., My answers are inline:

    1. Yes, This training mention that "need to configure BT's PDs first-class resistor to class 4", so is this means just make sure CLSA of TPS2373 is 63.4Ohm, and CLSB don't care?  

    {Answer:  Yes, that is correct.  A nonstandard UPOE switch won't do more than two fingers. This makes CLSB irrelevant (can be any class resistor or unpopulated). }

     

    2. We tried to modity both CLSA and CLSB to 63.4Ohm, and it's worked. Is this the only way to do?

    {Answer: For UPoE applications, CLSB can be unpopulated.  However, if the PSE is HDBT (PoH) non-standard, both CLSA AND CLSB must be 63.4 ohm in order for it to work correctly.}

    3. Customer use ST's BT solution to test with UPoE, it's work without any modification, does TI has any ideal?

    {Answer: The ST datasheet says that it’s single signature PD. They are probably doing LLDP (instead of physical-layer classification) which our BT PD’s can do for UPOE as well.

    Said differently, ST's solution is a single PD interface with dual bridge and the only way it can do UPOE is by LLDP which is a big investment in software and a data path.  Our solution is more flexible, as it can support "forced" (physical layer) UPOE (with 2 x TPS2373) or LLDP UPOE (1 x TPS2373).}

    4. Or any other way to make TPS2373 work with UPoE without modifying CLSA/CLSB?

    {Answer:  Yes.  The customer could implement an LLDP approach - but that is a rather significant investment decision.}

     

    Please let me know if having any additional concerns.

    Thanks,

  • 0 in reply to Thomas Lewis
    Mastermind 6180 points

    Hi Thomas,

    Here is the datasheet of this PSE injector, please help to check it for PSE type:

    PD-9501GR.pdf

    1. Because TPS2373 needs to be modified CLSB then can sink power,can I expect that the injector is not UPoE?

    2. If ST's PD is not set both CLSA/CLSB same as class4, then this injector won't be PoH, am I right?

    3. From your previous answer3, you mention that ST's solution is single PD with dual bridge, but TPS2373 is single PD interface as well, am I right?

    Sorry I don't get the difference.

    4. If ST and TI's PD are the same(single PD interface), then they should have same performance, why ST's can sink power but TI's needs modification?

    Thank you very much!

    Best regards,

    C.T.

  • 0 in reply to chentsu chung
    TI__Expert 5615 points

    C.T.,

    Answers inline:

    Here is the datasheet of this PSE injector, please help to check it for PSE type:

    PD-9501GR.pdf

    {Answer:  They are claiming it to be Type 2, IEEE 802.3at, compliant...AND capable of sourcing 60W of PoH power.  PoH is MicroChip's old non-compliant, 'Type 3', solution.  They released these solutions before the IEEE 802.3bt standard was ratified and it would very likely fail interoperability testing for the Type 3 logo.  I would have the customer ask MicroChip for the EA PoE logo for this end equipment.}

    1. Because TPS2373 needs to be modified CLSB then can sink power,can I expect that the injector is not UPoE?

    {Answer:  Correct - this PSE injector does not utilize UPoE. To my knowledge, none of the MicroChip injectors do...all of them are PoH/HDBaseT implementations prior to 2019 releases.  Moving forward, I suspect that MicroChip will follow the new standard instead.  Our PD solutions can work well with PoH injectors if following our guidance in video training.}

     

    2. If ST's PD is not set both CLSA/CLSB same as class4, then this injector won't be PoH, am I right?

    {Answer:  The injector is always acting the same way - as a PoH injector (not UPoE).  The behavior will not change no matter what is connected to it.  With that said, the PSE will not send Type 4 power if the ST PD communicates to the PSE injector that it can only accept Type 2 power via the resistor settings on CLS1/CLS2 pins (4/5).}

    3. From your previous answer3, you mention that ST's solution is single PD with dual bridge, but TPS2373 is single PD interface as well, am I right?

    Sorry I don't get the difference.

    {Answer:  You got it right.  Their solution integrates the dual bridge.}

    4. If ST and TI's PD are the same(single PD interface), then they should have same performance, why ST's can sink power but TI's needs modification?  {Answer:  Without greater understanding of the full ST-based board, I cannot accurately help guide you.  Maybe it utilizes the data path to communicate (LLDP)?  I will notify the PD applications engineer to review this thread and advise further if relevant.}

     

    Thanks for your continued local support,

  • 0 in reply to Thomas Lewis
    Mastermind 6180 points

    Hi Thomas,

    About you mention before "I will notify the PD applications engineer to review this thread and advise further if relevant"

    May I know it's any further update? Thank you very mach.

    Best regards,

    C.T.

  • 0 in reply to chentsu chung
    TI__Mastermind 40770 points

    Hi CT,

    There are only two ways to do UPOE. Through LLDP using a single signature PD or through Forced UPOE using a dual signature PD.

    If only using a single TPS2372 or TPS2373 then the PD must use LLDP to go to higher power. The PSE should still at least turn ON to Type 1 power first then negotiate higher power through the data lines. 

    The PD can still request Type 2 power with a class 4 resistor however, the PSE should still not allocate higher power unless it is negotiated. UPOE still requires mutual identification.

    Are you saying the injector is not turn ON at all to 48V? Can you send a waveform of VDD-VSS (or VDD-RTN) on the PD?

    My suspicion is that this injector is trying to do more than UPOE and possible trying to do other non-standard solutions.

    Either way, you should try to be compliant to the .bt standard as moving forward, this will be more interoperable PoE solution.

  • 0 in reply to Darwin Fernandez
    TI__Expert 5615 points

    CT, It has been nearly 1 week since our last reply.

    If we don't see a reply from you within the next 24-hours, I am going to close the thread.

    You can always reopen it by simply replying.  Let us know if needing any additional help.

    Thanks,

  • 0 in reply to Thomas Lewis
    Mastermind 6180 points

    Hi Thomas, Darwin,

    Thanks for your great help, my customer still not provide the waveform so I will colse this thread.

    Best regards,

    C.T.