Captured to LMZ10503 internal compensation wiki page

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.

LMZ10503 internal compensation

Other Parts Discussed in Thread: LMZ10503
I've designed TI LMZ10503 switching regulator into a new design.  I have the data sheet and app note SNVA417C that further elaborates on how to compute compensation values for this switcher.  This switcher is also supported under Webench.
I've plugged all of the formulas into a spreadsheet so as to compute new compensation values depending on my particular configuration.  I expected to verify my spreadsheet by reproducing some of the sample values provided in the tables in both documents.
My spreadsheet correctly reproduces the example with real numbers in the data sheet.  However, the data sheet does not include the compensation component formulas.  The compensation formulas are only in the app note and there is no example with real numbers.  However, they do provide tables of a limited number of pre-computed values.
The problem is that I am unable to reproduce the compensation value table with the formulas provided in the app note.  If I can't reproduce the tables, then I can not be sure that the values are correct for my new design with the components I'm using.  I need assistance to identify this inconsistency.
  • Sevrum,

    In the calculation of the table and Webench tool, certain amount of Cout derating is taken into account.

    The tables provided assume Cout derate to about 80%. Webench calculation has a built-in curve for Cout derating based on voltage rating and Vout. Derated output capacitance is used in compensation calculation.

    The values provided in the tables were all tested on the bench.

    -Yang

  • Can you take a look at the calculations shown here and see if you can help with the questions? There are differences in the webench and calculations. Thanks,

    5756.LMZ10503.xls

  • The attached excel file cannot open up. Would you attach it again please?

    Yang

  • Sevrum, we are looking into it. We still have warnings when opening up the attachment.

  • Sorry Guys...please try this one

    4743.LMZ10503.xls

  • Hello Sevrum,

    Please excuse the delay in our responses. I am re-attaching the spreadsheet that you have attached from before just to clarify if this is exactly what you sent. For some reason, we are seeing messages that say the data might be corrupted. Also, in the spreadsheet, you say that there are formulae, but we do not see those and the related information either.8666.LMZ10503 (1).xls

    That being said, I understand that you are seeing discrepancies between app note, datasheet and Webench. The LMZ10503 employs the voltage mode control and compensating the voltage mode control method can be a little challenging. The calculations in Webench are exactly the same as those in the app note. But the example values in the tables on the App note are further tweaked to get the best operation in the design. The equation for Ccomp has an approximation in it which yields in first pass values which then would need to be tweaked to get the desired operation.

    The voltage mode control on the LMZ10503 results in the double pole caused by the L and Cout. To compensate this, two zeros are placed near the double pole. Half of the compensation is inside the IC and the zero resulting from the internal compensation is located at about 17KHz. Based on the selection of Cout, the external Ccomp value is then calculated based on the approximated equation in the app note. Going forward from there, the upper feedback resistor is then calculated. Now depending on design performance, Rcomp is selected based on the ESR zero. Whether it is required or not will depend on the design performance.

    The value of the output capacitor is calculated using the inequality on page 11 of the datasheet. This same calculation is used in Webench. The max limit of the output cap is ideally then chosen on the type of application. Since Webench does not have exact knowledge of the end application, the max value of the capacitance is limited to twice the minimum value. The examples in the datasheet are chosen based on bench validation and best observed performance. You could as such use the tabulated values from the app note and be able to see good system performance. You could also simulate the circuit in Webench. Webench has warning messages that say that you should not use component values outside the bound, but then it does not stop you from trying to do so and simulating the result. You could create "Alternate part number" in the BOM section and use those values in the simulation to see the performance. 

    In a nut shell, the calculations will yield good first pass numbers and the application notes and the datasheet have component values that have been tested on the bench for optimal performance. The derating effect of the ceramic capacitors is taken in consideration in the Webench software. The derating information is gathered from the cap vendors and/or vendors's simulation software and is then used to calculate the actual capacitance. I hope that this information is helpful. Please feel free to write to us if you have more questions regarding this.

    Best Regards,
    Akshay