• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 6 answers
  • Subscribers 63 subscribers
  • Views 423 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.

TPS23731: PSR calculations

Carlo
Expert 7715 points
Part Number: TPS23731
Other Parts Discussed in Thread: TPS23754, TPS23734, PMP22301

Hi team,

Our customer is interested in the TPS23731, however, they would like to know how to calculate the PSR specifications, such as transformer, bias and compensation network?


Regards,
Carlo

  • 0
    TI__Mastermind 21320 points

    Hey Carlo,

    We do not have any application notes on these topics, but we do have reference designs. Email me your customer Vout/output power and I can give a suggestion. Thanks!

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • 0
    Prodigy 40 points

    Hi. Vout 24V/25W solution. I have used TPS23754 and other PoE controllers before, but not with PSR. This time I am planning on using TPS23734 in flyback mode with sync sec rectification and active clamping for improved efficiency. But I noted that one of the new features in this controller is PSR, and thus I was thinking that I could get rid of the opto-feedback. If I understand correctly, using PSR could further improve efficiency and reduce comps (opto feedback). I asked about the TPS23731 because it was promoted as PSR controller, but I see that the TPS23734 has the same PSR features (and some more related to GAT2). I do not need very close regulation of the 24V. (23 - 26 would be fine). I was planning on making the board for optional mounting in order to test different cases, such as PSR and opto-fb. I enclose the DRAFT schematics at this stage. My end-goal is to have the circuity small and efficient (>90%).

    The above text is copied from my previous reply, but the above is the essence of my question.Power PoE - opto-fb - 2021-02-04.pdfPower PoE - psr - 2021-02-04.pdf

  • 0 in reply to Dag Vedaa
    TI__Mastermind 21320 points

    Hello, 

    I can share a design with Carlo for you to review. 

    You are correct that the TPS23734 is the same as TPS23731, but it has an additional gate drive. 

    I would not advise a sync flyback for 24V output. The Vds is going to need to be 200V+. The FET will need the lowest gate charge you can find, so it will need to be expensive. Additionally the gate drive will need to be the best you can find. Even with all of this, you still might need to snub the FET which will cost efficiency.

    Sync flybacks also typically have shoot through issues with high output voltages. I cannot recommend this topology for this output voltage. It will just be costly if you do. If cost is not a consideration, then i would recommend a driven sync flyback that uses a pulse transformer to control the gate drive. You will want the best. 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • 0 in reply to Michael Pahl
    Prodigy 40 points

    Hi Michael,

    Thank you for your feedback.

    You mention that you can share a design me to review. I can not see any attachment, or is it to come?

    Secondly, you write about what you can not advise, but what can you advise or recommend for this design?

    You mention that the Vds for sync flyback is going to need to be 200V+, but with the active flyback (ref my first reply and schematics) it will be far below 150V, or am I missing/misunderstanding something? The gate charge of the transistor selected is 7nC, should it be lower than this? 

    You do mention that you may recommend a driver sync flyback that uses a pulse transformer to control the gate drive (secondary gate I assume), and that is what I have on the opto variant on my schematics.

    Also my original question was about PSR option for this design, but I can not find any more info regarding this.

    All in all, what is your topology recommendation for PoE+, 25W/24V output, overall efficiency > 90%? Important design factors are small size, high efficiency. Cost is relevant, but the not most important driver.

  • 0 in reply to Dag Vedaa
    TI__Mastermind 21320 points

    Hello,

    I said I would send the design to Carlos, the original poster of this thread. 

    I would recommend a diode flyback. 

    I was talking about the secondary FET - the one that makes it a sync vs a diode. The primary will probably be 150V as you said. But the secondary will need to be 200V (where as in a 5V application it can be 30V-40V). 7nC is fine. 

    Please look at our reference designs below for 24V out: you will see they are all diode flybacks besides two ACF's. The ACF's have higher efficiency, but require more space, and cannot do PSR. The TPS23731 could do PSR for one of these designs, but I am not sure it can get above 90%. 

    PMP22301 gets above 90% for the converter, and you could copy the secondary here. But that is not a guarantee.  

     

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • 0 in reply to Michael Pahl
    Prodigy 40 points

    Hi,

    Yes, the secondary side FET will need to be high voltage - thank you for pointing out.

    But adjusting for this, have a secondary FET @200V and transformer gate drive via GAT2, I expect this would give better efficiency than the diode flyback. Though the difference in efficiency would probably be less than for lower voltage outputs. I will make room for the diode flyback design for the proto.

    From my understanding, PSR design would generally give lower overall efficiency, correct?

    Regarding the PMP22301 design, I have seen this before, and I assume it could be upgraded from TPS23754 to TPS23734. One of the "issue" I have with this is that I can not find any documentation regarding the transformer, and I can not find the transformer as off-the-shelf item. Do you have spec for this, or similar OTS item? Would the Coilcraft POE300F-24LD or POE40Q-24ED be acceptable substitution? Or other suggestion?

  • 0 in reply to Dag Vedaa
    TI__Mastermind 21320 points

    Hello,

    Yes it will be higher but avoiding shoot through will be difficult. 

    PSR should not affect the efficiency.

    The transformer will not support PSR. No PSR transformers are off the shelf. My suggestion is to work with your preferred transformer vendor, or use Linkcom or Wurth to define a new one. 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • 0 in reply to Michael Pahl
    Prodigy 40 points

    Hi,

    Thank you for your feedback.

    I was hoping that some more details could be provided regarding the transformer in question (app note of similar) and example calculations of the parameters. Working with a transformer vendor to make a custom transformer could be fine but will take time and unfortunately probably be too costly for this particular project with limited volume.

    I have checked some other  manufacturer's design that are using available transformers, and also have provided the guidelines for correct transformer selection. Maybe the TPS23734 is not just correct for this design this time.

    Again, thank you for your time and feedback.

  • 0 in reply to Dag Vedaa
    TI__Mastermind 21320 points

    Hello,

    No matter where you go, a primary regulation transformer will require a new definition unless the vendor has already designed one. So even if you calculate it on your own, you will have to get it custom made. PSR requires a special transformer construction method and therefore they are usually not readily available "off the shelf."

    If you want to do active clamp forward, where you can use off the shelf transformers, please see the following application note that details both the transformer spec and the DCDC design:

     

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments