• State TI Thinks Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 1 answer
  • Subscribers 65 subscribers
  • Views 540 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.

LM21215A: Output noise issue

JH Shin
Guru 11090 points
Part Number: LM21215A
Other Parts Discussed in Thread: TPSM82866A, TPS62870, TPSM82866C

Tool/software:

Hello,

My customer was using the LM21215AMHE-1 under the following conditions.

  - INPUT : 5V, OUTPUT : 0.9V, LOAD : 6A

They designed the circuit around the end of 2022, referring to the TI WEBENCH results below.

Recently, during board production, an issue occurred where noise was generated at the output of LM21215A as shown below.

When they ran TI WEBENCH again this month, they found differences in the component values ​​below:

By simply changing the FILTER value connected to the COMP pin, the output noise was reduced as shown below.

Q1. What is the root cause of this issue?

Q2. Could you please explain why there was a difference in the TI WEBENCH results?

Thank you.

JH

  • 0
    TI__Expert 4478 points

    Hello JH,

    I will forward your question to a colleague, who is familiar with the history of this device. The answer may come early next week.

    Regards, Werner

  • 0
    TI__Guru**** 150870 points

    Hi JH,

    I've added the Webench folks to this thread.  Webench, was there an update done to this model in the past few years?

    JH, this is a very old device.  Nowadays, we have much simpler 6A devices.  Would TPSM82866A, TPSM82866C, or TPS62870 work for you?

    Thanks,

    Chris

  • 0 in reply to Chris Glaser
    Guru 11090 points

    Hello Chris,

    Thank you for your reply.

    The customer may wish to consider using the product presented above in their new design. This product is currently in production and we need immediate response from TI on this issue.

    Regards,

    JH

  • 0 in reply to JH Shin
    TI__Intellectual 2350 points

    Hi ,

    I hope this message finds you well.

    As part of our investigation, we are comparing the previous design with the current one. Our initial findings indicate changes in the BOM components, and we are delving deeper to identify any other discrepancies.

    We will compile all our findings and provide a comprehensive update by the end of this week. Your patience and support during this process are greatly appreciated.

    Thank you for your understanding.

    Best regards,
    Sateesh A G

  • 0 in reply to Sateesh Goudappanavar
    TI__Expert 5820 points

    Hi JH Shin,

    We verified LM21215A Webench model, there are no updates done to the model in the past few years.
    We compared the shared 2022 Webench design with current Webench design for condition VinMin=5V,VinMax=5V,Vout=0.85V and Iout=6A following are our observations.

    Apart from compensation components change we observe the change in inductor ,input capacitor and output capacitor selected between the two designs.
    The inductor, input capacitor and output capacitor getting selected has a dependency on the switching frequency.
    Change in switching frequency might cause the change in above components as well as change in compensation components like Cc1 and Cc2.

    Change in switching frequency between current Webench design and shared 2022 Webench design might be causing the discrepancy.

    If possible kindly share the Webench design report for 2022 Webench design with us, this will help us to confirm our above observations.

    Change in compensation components might probably  be improving the stability due to which the noise in output could be reduced.But  and  can comment more on this.
    Kindly revert back to us for any further queries.

    Thanks and Regards,
    Umayal

  • 0 in reply to Umayal Ramanathan
    TI__Guru**** 150870 points

    Thanks for looking into this, Umayal.

    Yes, JH's waveform appears to be instability, which was fixed by choosing new compensation components.  Yes, if the switching frequency is different, this would result in a different inductor and capacitors, which would require different compensation.

    Chris

  • 0 in reply to Umayal Ramanathan
    Guru 11090 points

    Hi Umayal,

    Thank you for your reply.

    Please refer to the SAV files provided by the customer below.

    WBDesign_6A.zip

    All they can set in TI WEBENCH are the input power (MIN & MAX), current, output power, and operating temperature as shown below.

    They are asking for a clear explanation as to why the above RC values ​​have changed even though nothing has changed in TI WEBENCH.

    Regards,

    JH

  • 0 in reply to JH Shin
    TI__Guru**** 150870 points

    Hi JH,

    Do you have the inputs that were given to create each of these designs?  There are more options, beyond the simple Vin, Vout, etc, in the Design Consideration section.

    Umayal, the 2022 design shows a SYNC clock attached to the device.  This is not present in the 2024 design.  What would cause Webench to add this to the schematic?  Would the customer need to click a box for clock sync?

    Chris

  • 0 in reply to Chris Glaser
    Guru 11090 points

    Hi Chris,

    The customer said they did not change any additional settings other than the four settings above(Vin, Vout, Current, Temp).

    Below is the result of the customer also checking the 8A and 10A conditions in 2022.

    The results for this condition showed no additional circuitry on the SYNC pin.

    WBDesign_8A_10A.zip

    Thanks,

    JH

  • 0 in reply to JH Shin
    TI__Expert 5820 points

    Hi JH Shin,

    We re-verified LM21215A Webench model and confirm there are no updates done to the model since 2017.

    Following are the three possible reasons to usually observe discrepancy between the designs on Webench.

    1.   Webench BOM components selection is dynamic. Webench selects the most suitable and optimized components with respect to cost   ,footprint and efficiency for a given input condition.
    When new components are added to our database ,incase the newly added components suits the best for the application then it could be selected. Hence we might observe the change in BOM components.

    2. User has the feasibility to customize the Webench design by editing the component limits or create the custom components.

    3. Change in "Configuration Options" can alter BOM components accordingly. 


    But to exactly find the root cause for this issue we require the Webench design report for "2022 Webench design" as shown in image below.

    OR Kindly share the Webench design link for "2022 Webench design" with us.

    .svg files shared by customer will not provide us details of other operating conditions like switching frequency or "Configuration Options" setting used by the design. Webench design report  or  Webench design link will provide us the details of other operating conditions and "Configuration Options" setting.

    Webench design report  or  Webench design link is required to investigate this issue further.

    Thanks and Regards,

    Umayal

  • 0 in reply to Umayal Ramanathan
    Guru 11090 points

    Hi Chris, Umayal

    Thanks for your kind reply.

    Could you explain why the output noise is improved when only the COMP pin connection value of the 2024 Webench design is changed on a board designed with reference to the 2022 Webench design of LM21215A?

    The customer wants to know what is the root cause of this symptom.

    Regards,

    JH

  • 0 in reply to JH Shin
    TI__Guru**** 150870 points

    Hi JH,

    The 'noise' appears to be instability.  So, changing the compensation improved the stability, it seems.

    And stability is not just dependent on what is shown in the Webench design.  It also depends on everything that is connected to the output, including items on other pages of the schematic.

    We always recommend that customers validate their designs, including measuring the loop stability in their system.  Did they measure this on either design?

    Chris