• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 3 answers
  • Subscribers 63 subscribers
  • Views 377 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.

LM25143: 8-phase operation with 4 devices

Taylor Ives
Prodigy 10 points
Part Number: LM25143
Other Parts Discussed in Thread: UCD3138A

Tool/software:

Hello,

We are looking to design a 8-phase converter for very high output currents. In the datasheet it notes this is possible if we provide our own SYNC signals that are appropriately phase shifted (0, 45, 90, 135)

Is there a preferred method for this with low component count? We considered a CPLD to implement a chain of flip-flops clocked at a higher speed and divided down but wondered if there is a better method or specialist device for simplicity.

Thanks

  • 0
    TI__Guru 72622 points

    Hi Taylor,

    The phases of the master and first slave run  at 0°, 180°, 90° and 270°, so you would need a 45° phase shift for the other two ICs.

    What is the spec here (Vin range, Vout, Iout, Fsw) and do you really need 8 phases?

    Regards,

    Tim

  • 0 in reply to Timothy Hegarty
    Prodigy 10 points

    Hi Timothy,

    Thanks for your reply. The application in question is high-powered lasers for the medical industry.

    Vin 22V-26V, Vout 15V, Iout Max 250A, Fsw 1MHz.

    Considering 8 phases to limit individual phase current to around 30A for higher reliability. Output voltage ripple and load transient response are also key factors that should be improved the higher number of phases.

    Thanks,

    Taylor

  • 0 in reply to Taylor Ives
    TI__Guru 72622 points

    Understood, thanks Taylor. 30A/phase at 1MHz is a bit high with 40V FETs. I recommend 400kHz if possible.

    --

    Tim

  • 0 in reply to Timothy Hegarty
    Prodigy 10 points

    Hello again,

    We are looking for a high power density and as such were considering the integrated smart switches provided by TI but I understand your comment about the FET switching speeds. We could adjust the design but we want to maximise efficiency. Could GaN FETS benefit our application?

    The original question was more on the external SYNC clocking and a good method of producing the required signal with the correct phase offset. Does TI have a recommended solution?

    Finally, we have started to consider the advantages and disadvantages of a software defined controller and the potential benefits ot could provide in terms of being able to tune the individual tolerances and performance of each phase to better balance the stresses involved. Such as using the UCD3138A. What are your opinions on this?

    Thanks,

    Taylor

  • 0 in reply to Taylor Ives
    TI__Guru 72622 points

    Hi Taylor,

    I doubt GaN will offer much benefit at this input voltage (24V). Given the high current, you may not want to make it too dense from a thermal management perspective.

    We have a circuit for generating clock signals with a particular phase shift, but I will need some time to dig it up.

    I'm not an expert on digital control - you may want to submit a separate E2E post for that. I doubt it will make the solution smaller though.

    Regards,

    Tim