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.

LM26420-Q1: SW waveforms appear overshoot and undershoot

Part Number: LM26420-Q1
Other Parts Discussed in Thread: LM26420,

Hi LM26420-Q1 When testing the output SW of LM26420, the test found that the SW waveform overshoot and undershoot are very serious, the undershoot reaches 1.05V, the overshoot is 0.7V, the input is 3.3V, and the two outputs are 1.8V and 1.8V respectively. 1.1V, the load is 0.5A~0.6A. Please help to check the schematic, is there any solution to solve the overshoot and undershoot?   

  • Hi,

    - If there is any possibility to use one of our latest dual converters, then I would recommend to do so. You would have a lot of benefits on performance, size, availability, cost, etc.

      TPS6244x-Q1 is the successor of LM26420-Q1. https://www.ti.com/lit/ds/symlink/tps62442-q1.pdf?

    - The schematic looks fine. However, I have some suggestions,

    • The LC filter on the input could cause stability issues if not designed properly. Do you have specific noise requirements for this application? 
    • Please confirm the loop stability with the current BOM. 
    • Feedforward capacitor is not necessary for this device.

    - Switch node ringing is caused by various reasons and most important is to have a good layout practice. Could you share the layout for review?

    - There are other ways to minimize switch node ringing as explained in the below App notes.

         https://www.ti.com/lit/an/slyt465/slyt465.pdf?

         https://www.ti.com/lit/an/slua831a/slua831a.pdf?

    Regards,

    Febin

    • Hi Febin 
    • thanks a lot for your answer
    • The LC filter on the input could cause stability issues if not designed properly. Do you have specific noise requirements for this application? 
    • reply:There is no special requirement for output noise, it is OK not to exceed the EMC limit 
    • Please confirm the loop stability with the current BOM. 
    • reply:fall load:  half load:
    • Feedforward capacitor is not necessary for this device.
    • reply:Can you explain why you removed the feedforward capacitor?

    LAYOUT As follows: Please also help to review :

  • Hi,

    - The application looks stable with the above BOM. 

    - Regarding the feedforward capacitor (Cff), 

    What I meant is that, in the datasheet it was not explicitly mentioned to use an additional external Cff to improve transient response or stability. This device is internally compensated, so it is simple to use and requires few external components. If you can achieve the desired performance without a Cff, then it is also fine. However, the decision is left to the customer.

    - I will provide a feedback on the layout later today.

    Regards,

    Febin

  • Hi,

    I reviewed the Layout and here is my feedback.

    - There is a possibility that the FB pin & SW pin is picking up noise from the additional test points near it. I would suggest to avoid any unwanted test points near noise sensitive pins.

    - The GND connections of R270 and R283 are far away from the GND of the IC. This is not recommended. All the components for the IC should share a common ground point. By flipping them side ways, you could achieve this very easily.

    - Is L201 shielded inductor?

    - Does the EMC limit for your application exceed without the input filter?

    - How many layers does the board have? What does the stack look like?

    - Is there a redundant ground plane in the bottom layer?

    - What is the application? 

    - What is the load to the LM26420-Q1?

    Regards,

    Febin

  • Hi Febin 

    thank you very much ,Below is my answer

    - The GND connections of R270 and R283 are far away from the GND of the IC. This is not recommended. All the components for the IC should share a common ground point. By flipping them side ways, you could achieve this very easily.

    RE:

    - Is L201 shielded inductor?

    RE:L201 is a filter inductor, the π-type circuit filters out the ripple

    - Does the EMC limit for your application exceed without the input filter?

    RE:If there is no attempt to remove it, there should be a margin. There are products of the same type as LM26420-Q1 that pass the EMC

    - How many layers does the board have? What does the stack look like?

    RE:8layer 

    - Is there a redundant ground plane in the bottom layer?

    RE:bott

      middle 

    - What is the application?

    RE:Applied in IVI vehicle cockpit host,1V8 is supplied serializer, tuner, Clock Generator,flash,CVBS . 1V1 is supplied to the serializer,

    - What is the load to the LM26420-Q1?

    RE:1.8V load 570mA, 1.1V load 470mA, both SW waveforms are like this

  • Hi,

    Through which layer is the connection? Components on top layer and GND connection through second layer?

    Regards,

    Febin

  • Hi,

    Do you have any updates?

    Regards,

    Febin

  • Hi Febin

    problem solved

    Adding Snubber can solve

    Add R=2.7ohm, C=4.7nF overshoot is improved

    thanks a lot for answer

    In addition, the grounding spring is used, and the nearby grounding waveform is also improved.

  • Hi,

    That's great! This was explained in the App note shared earlier.

    I will go ahead and close the thread. Please hit 'Resolved ' if you agree.

    Regards,

    Febin