• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 98 subscribers
  • Views 387 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.

AM5716: About reference schematics (decoupling capacitor)

Atsushi Kadowaki
Prodigy 120 points
Part Number: AM5716
Other Parts Discussed in Thread: TMDXIDK5718

Dear Experts,

We are using AM5716.
To determine the decoupling capacitors for each power supply, We referred to the schematics for the AM571x evaluation board TMDXIDK5718 (file name: AM571X_INDUSTRIAL_EVM_3M0000_PDF_REV1_3A.pdf).

We ran into an errata i893 near the end of the design and made quality improvements for VDD_MPU.
At that time, we found SitaraTm Processor Power Distribution Networks: Implementation and Analysis (file name: sprac76d.pdf) and used it as a reference.

There was a difference between these two files about the decoupling capacitors.
We have two questions.
[Q1] Is it wrong to refer to the evaluation board when designing a circuit?
[Q2] sprac76d.pdf is listed in the WEB "Technical documentation", but it is not marked with a star.
Is it not possible to sort the technical documentation not marked with a star in order of importance?
It would be helpful if you could tell me because We want to proceed smoothly with the development of high-performance processors.

star = Top documentation for this product selected by TI

Best regards,
Kadowaki

  • 0
    TI__Guru 52545 points

    The Power Distribution Guide (sprac76) include the requirements for the power distribution.  These are listed as Max Reff, Decoupling Cap Max LL (loop inductance), and Max Target impedance.  Power simulations from your design should meet these requirements and will help drive if you decoupling capacitors are sufficient.   

    Note the same requirements table in the application note provides EXAMPLE decoupling capacitors.  However - decoupling recommendations is design dependent, as decoupling efficiency is impacted by PCB design.  The recommendation is to start with the application note example (or the EVM example), and then let the power simulations drive if more capacitors are required or if some can be removed. 

    I don't know the exact difference between the application note example capacitors and the EVM.  I would start with which-ever has more capacitors (to be conservative).

  • 0 in reply to Robert Eschler
    Prodigy 120 points

    Mr. Robert Eschler-san,

    Thank you very much for your quick reply.
    Please let me continue.

    Below is a comparison table between sprac76d and AM571x EVB.
    Only VDD_MPU is extracted.
    The last line is the capacitor that we decided on by running a simulation that includes the PCB.

    [Q3]The EVB has more types of capacitors, but the total is almost the same. In such a case, which one should I refer to at the circuit design stage?
    Also, please let us know your opinion on [Q2].


    Best regards,
    Kadowaki

  • 0 in reply to Atsushi Kadowaki
    TI__Guru 52545 points

    Again - the simulation results can help drive the capacitor selection.  If there is a 'spike' in the impedance at a particular frequency, then a specific capacitor value may target that frequency.  However - having too many different capacitor values can also cause anti-resonance (and oscillate).  For our designs - we usually stick to just a few different capacitance values.   Thus it is OK to have fewer 'different' capacitor values but just more of them.

  • 0 in reply to Robert Eschler
    Prodigy 120 points

    Thank you for your response.

    I have found that the optimization method is preferable by referring to various documents and simulating.
    I understand that it is better to use fewer different capacitance and more capacitors of the same capacitance to prevent anti-resonance.

    Let me ask one more last question.
    [Q4]What are the conditions and philosophy behind the design of the capacitors (type and number) on the evaluation board?
    Is this the result of simulation? 
    This is the question for the future.


    Best regards,
    Kadowaki
  • 0 in reply to Atsushi Kadowaki
    TI__Guru 52545 points

    For the EVMs I am involved with, yes.  We will run power simulations on the primary power rails and ensure the target impedances are met.  We will optimize the capacitors but not too close to minimal as we understand customers copy our designs and their implementation may not be as good as ours.  For this particular EVM (TMDXIDK578) - unfortunately I don't have the background.  The EVM designer(s) are not longer with TI.

    As mentioned, the EVM decoupling would be a good starting point as that design has been tested.

  • +1 in reply to Robert Eschler
    Prodigy 120 points

    Thank you very much.
    we understand.