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.

CSD95372AQ5M: Switching performance questions

Part Number: CSD95372AQ5M

Hi, TI experts

1. This device claims 2MHz switching capability, how to conclude this value? Due to all the internal latency inside(e.g. driving delay, dead time)?

If possible, I’d like to have a formula, or a figure showing how to calculate.

2. During high-frequency switching dead-time, I'd like to learn if a Schottky diode is inside to bypass body diode, to optimize reverse recovery and voltage ringing stress?

As you know if a normal silicon body diode is used for switching, the reverse recovery too serious to be applied in a high-frequency CCM buck converter.

  • Hi,

    Thank you for your interest in this device.

    To answer your questions:

    1.The max switching frequency is based on the design limits and device performance. There is no simple formula to calculate.

    2. We don't use Schottky diode, we have internal circuit to optimize the reverse recovery and ringing reduction. The design details are not for public disclosure. I cannot share it on the open forum, hope you can understand.

    Best,

    Qingquan

  • Thanks very much for reply.

    Sounds like monolithically integrated Schottky-like(or performance effective) diode, for example you know INFINEON OptiMOS5 technology.

    Of course I totally understand detailed design information should be confidential, actually I'm not interested in the details at all, I'm just interested in general idea and general combinations, for better understanding the principle.

    Could I try to understand that, where is the frequency limitation, more related to internal driver parameters, or more related to power FET performance? A general answer is okay.

  • There is tradeoff when operating at very high frequency. I don't see much interests above 2MHz at this point, due to higher power loss not only in power FET but also inductors.  So to answer your question, it's not simply the driver or FET limitation, but overall system consideration also comes into play.

  • Thanks for your kind answers, Qingquan.

    However, maybe I didn't quite explain my questions well, please allow me to ask more about the frequency limitation data marked for this kind of device.

    I definitely understand that fsw is a trade-off result of system-level design consideration, like time latency limitation in each session, component power loss, overall efficiency, size, EMI, thermal, cost, control architecture/stability, etc...

    I'm not asking this actually--- as I don't believe the switching performance range in datasheet should include all these system-level performance parameters into play, because you never see a switching frequency spec in a normal power FET datasheet, but you can find this spec in DrMOS datasheet. And I think this must be a calculated or experiments-based data, not coming from mind only. This is the reason, why I ask this question.

    I only care about the switching capability(w/o considering system performance) of power stage itself only, is the driver inside, or power FET turn-on/turn off timing, limiting the fsw spec?

  • To add some information, I know presently in typical application, not much interests are put for pushing fsw higher than 1MHz in poL/VRM.

    I'm just doing some technical research for high-frequency power converters, and the end equipment might not be a 12V-1V buck converter---we might take power stage as a general half-bridge topology for all kinds of power conversion, for example like LLC resonant converter.

    Consequently, figuring out why you mark 2MHz as a spec in datasheet, helpful to mu understanding and makes sense to me.

    Appreciate it for all of your comments.

  • Hi,

    it's trade-off as I mentioned. You might be able to operate above that frequency but the performance cannot be guaranteed. We have both driver and FET in the power stage, so we need to considered the overall performance in the typical applications.

    I understand you only care about switching capability, however we need to consider more when design the product, it's almost impossible for us to recommend max frequency only based on switching capability. Hope you can understand.

  • I can understand, no problem, but academic spirit just makes me hard to accept your explanations here.

    I can share with you about my thoughts below for technical discussion only. For fair comparison, just take below 60A power stage as example.

    ①CSD95481 is obvious next-gen product of CSD95372B, system-efficiency and thermal are much better, proving that FOM(figure of material) performance is much better than previous.BUT, the max switching spec keeps the same, still 1.25MHz.

    ②On the other hand, please look at CSD95372A and CSD95372B,  B has IOUT functions and circuits, but A doesn't.

    From FET and system-level performance, A.ver is poorer than B.ver, but surprised to see A.ver 2MHz, higher than B.ver 1.25MHz. 

    My conclusion:

    It is obvious to me, this spec is not closely related to FET parameters, and also system efficiency/thermal performance. It should be related to other internal circuits performance and parameters, like IOUT.

    But I can understand the trade-off and difficulty of giving a reasonable value for it, appreciate it for your answers again.

  • Hi,

    Thank you for your understanding. What you mentioned is basically correct, it's not just FET spec, but considered overall performance as power stage. We cannot share the internal circuits information on open forum as it requires NDA agreement.

  • Understood, but I think understanding this spec doesn't need any details of internal parameters.

    What we've discussed is basically enough to me, thank you again. We can close this ticket.