• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 62 subscribers
  • Views 183 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.

TPS82740B: phase instability impossible due to microcontroller decoupling capacitors?

John Mamish
Prodigy 30 points
Part Number: TPS82740B
Other Parts Discussed in Thread: TPS62840

Hi,

I am using TPS82740B in a new design. On page 21, section 10.2.2.2, it says "Care should be taken that the total effective capacitance present at the output does not exceed 10µF in order to guarantee loop stability." The microcontroller that I'm using calls for a total of about 15uF of distributed decoupling capacitance.

Does the sentence in the datasheet refer only to output capacitors placed right next to the TPS82740B, or will the loop stability be affected by my microcontroller's decouping caps? If the microcontroller is only a few centimeters away from the TPS82740B, will the parasitic power plane inductance between the regulator and the micro be enough to mitigate this issue?

Thanks,

John

  • 0
    TI__Guru*** 144470 points

    Hi John,

    Yes, as it says, all capacitors count.  The IC can't tell which schematic page they are on--all are electrically present to it.

    Note that the D/S refers to the capacitance, which is the actual amount of decoupling present.  This reduces the nominal values of all the caps to account for their DC bias, etc.  It's not clear if your 15 uF refers to capacitance or nominal capacitor values.  15uF nominal could result in less than 10 uF effective capacitance.

    As well, point of load regulators like the TPS82740 are best placed as close to their loads as possible.  A few centimeters is far away on a PCB.  The TPS82740 includes a 4.7uF (nominal) output cap, if that can help you meet your 15uF total.

    We also are sampling the lower Iq TPS62840, in case it is of interest.

  • 0 in reply to Chris Glaser
    Prodigy 30 points

    Hi Chris,

    Thanks for your response!

    >The IC can't tell which schematic page they are on--all are electrically present to it.
    I realize the the IC can't tell which schematic page the caps are on.  I was wondering how the parasitic inductance of my power network (~10nH) would affect the loop stability, and whether it would mitigate the higher capacitive load. Furthermore, I imagine that, for many buck regulators, an externally placed LC filter would sufficiently decouple the feedback network from the load. Designing such an LC filter, however, requires switching supply know-how beyond my scope. I was hoping that TI had some guidelines for putting in such a filter.

    >Note that the D/S refers to the capacitance, which is the actual amount of decoupling present.  This reduces the nominal values of all the caps to account for their DC bias, etc.  It's not clear if your 15 uF refers to capacitance or nominal capacitor values.
    Thank you, that's a great point. I'm just talking about my microcontroller in this question, but once other chips are added, I think that the nominal capacitance will be about 25uF. With my DC operating point and expected capacitor derating, this puts me well above 10uF.

    >As well, point of load regulators like the TPS82740 are best placed as close to their loads as possible.
    It seems like I missed this major point. Thank you for educating my about that. I'm lead to think that the 82740 might not be right choice for my application.

    >We also are sampling the lower Iq TPS62840, in case it is of interest.
    Thanks for the suggestion. Do you know when it will enter production? What sample quantity can I get?

    Keep making great chips and providing outstanding engineer support!

  • 0 in reply to John Mamish
    TI__Guru*** 144470 points

    Hi John,

    There is a transient pspice model for the TPS82740B, which you can use to assess your LC filter. http://www.ti.com/product/TPS82740B/toolssoftware#model

    If you place the other loads on the load switch output, you get another 10uF of allowed Cout.

    Yes, all power supplies provide the best regulation when placed closest to the load.  Perhaps you can share some of the decoupling caps by placing the oads close to each other as well?

    TPS62840 should be in full production very soon.  You can order samples now.  I'm not sure of any sample limit.

  • 0 in reply to Chris Glaser
    Prodigy 30 points

    Thanks Chris!

    >There is a transient pspice model for the TPS82740B, which you can use to assess your LC filter.

    Thanks! That's great!

    >If you place the other loads on the load switch output, you get another 10uF of allowed Cout.

    I appreciate the suggestion. I might end up doing so; and I might be able to meet my requirements that way. I really like this regulator.

    >TPS62840 should be in full production very soon.  You can order samples now.  I'm not sure of any sample limit.

    I'll definately look and see if it's appropriate for my application.

    Thanks again for all your help.