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.

BUF18830: Ringing on out pin

Part Number: BUF18830
Other Parts Discussed in Thread: TPS65177

Now the customer is using BUF18830.

Below is the SCH.

3058.SCH.docx

When the picture displayed on the panel is changing, a ringing is observed as below.

Below is the waveform on OUTx pin. (The ringing amplitude is a little different from each other)

Below is the power rail of VS (powered by AVDD of TPS65177)

Could you please give some suggestion about that?

Thank you.

  • Hello Robin,

    Could you please identify the different waveforms (blue/pink/yellow=?)

    Best regards,
    Diaara
  • Hi Diaara,

    I'm really sorry that I forget that.
    The yellow line in all figures is the curve I mentioned. (OUTx and Vs)
    All others curve are other signals for the panel, not BUF18830.
    Thank you.

    Robin
  • Hello Robin,

    It seems that the pink signal is somehow linked to the ringings as your firsts scope plots show. It would be good if we knew which signal it is.
    The ringings on OUTx are most probable due to the ringing on AVDD. You could confirm this by increasing and reducing the output capacitance on AVDD.
    Does the panel not work or have some problem due to these ringings on OUTx?
    Also when you say "Below is the waveform on OUTx pin. (The ringing amplitude is a little different from each other)" do you mean that the ringing is different for each output or it is different for different pictures?

    Best regards
    Diaara
  • Hi Diaara,

    The pink signal is related with the moment when picture is changed. At this time, the load current of AVDD will change from about 400mA to 700mA(Customer told me that). So a voltage spike occurs due to load transient.
    "Below is the waveform on OUTx pin. (The ringing amplitude is a little different from each other)" , I refer to that ringing is different for each output, but not too much.
    Actually we have add the output cap from 10uF*4 (25V rating) to 10uF*4(25V)+ 10uF*5(50V)+47uF(Al eletrolysis).
    Thank you.

    Robin
  • Hello Robin,

    I am sorry but for some reasons my last answer was not published.

    What is the value of AVDD and the part# of the output caps? I think that this increased output capacitance should not be necessary and the resulting transient on AVDD is poorer that I would expect. Please see the attached scope plots showing the load transient measurement results on AVDD of TPS65177 and OUTx of BUF18830; I used these  devices EVMs for the test and connected AVDD to Vs.

    Could you please measure again with a scope during the transient event AVDD on C260 and C675? I want to make sure that we optimize first this transient response because OUTx should follow then.

    Best regards,

    Diaara

  • Hello Robin,

    Do you have any update regarding this issue?

    Best regards,

    Diaara

  • Hi Diaara,

    Great thanks for your attention for this case.

    The customer reflects that adding cap close to Vs pin doesn't help a lot.

    Now it is still on a holiday of China, no more feedback from them.

    I would go to field to have a look.

    Do you have some more suggestions to figure out the problem?

    Thanks again.

    Robin

  • Hi Diaara,

    Below is the waveform at C260 and C675 node.

    Waveforms of BUF18830.docx

    There is some difference between them. After current sensing circuit, some ripple is observed on AVDD rail.

    And also, for OUT1 and OUT2, the waveforms are really different.

    I have tried to add capacitors on several Vs pins (double to 10uF*2), but no obvious improvement is observed.

    And customers themselves have also tried to add capacitors on OUTx pins, a little improvement is observed.

    Any suggestion for this case.

    Thank you.

  • Hello Robin,

    Thanks for posting the waveforms.

    On the initial schematic you posted I see that on the COMP pin there is an additional capacitor to ground C285  in parallel to C288-R110. Is this one mounted? If yes I would recommend to remove it and redo the transient measurement without it because it is much too big and this could be the reason of the bad transient behavior of AVDD converter. 

    Best regards

    Diaara

  • Hi Diaara,

    I think you refer to C289, actually I had taken it off. Before it was taken off, the transient time was almost 2ms, while now it is about tens of micro seconds.

    All the waveforms I send to you are tested in the condition C289 is taken off.

    Thank you.

    Robin

  • Hello Robin,

    Just for my understanding: You mention that this ringing is observed when a picture is displayed. Would you please give more information about the problem you are observing? Is there some visible effect on the picture?

    For the differences between OUT1 and OUT2, is it possible that the Code for these 2 outputs is set differently? What is the code for these 2 outputs?

    BTW, you might need to add some smaller capacitors to buffer the fast drop. A capacitor with lower capacitance normally has as well smaller series resistance, so for buffering it might be better to add a high quality 1uF capacitor instead of a 10uF capacitor. The dielectric should be X5R or better and please always keep the DC bias of ceramic capacitors in mind.

    You might try as well some filter on AVDD to keep in clean. The output voltage on the buffers is directly dependent on the analog supply voltage, as given in formula 1 in the datasheet on page 11. So a transient on the supply can be observed directly on the output and if the output voltage is very close to Vs, this effect will be worst.

    Best regards,
    Brigitte

  • Hi Brigitte,

    Thanks for replying. See my input.

    Just for my understanding: You mention that this ringing is observed when a picture is displayed. Would you please give more information about the problem you are observing? Is there some visible effect on the picture?

    [Robin] Actually this is a very complex problem. The visible effect on the picture (most are pure color) is the issue they call "mura". But it is not sure that this is caused by ringing of BUF18830. At the start of every frame, current sensing and maybe some other sensing should be done, a fast load transient performance and stable output of AVDD and gamma is appreciated.  

    For the differences between OUT1 and OUT2, is it possible that the Code for these 2 outputs is set differently? What is the code for these 2 outputs?

    [Robin] OUT1 to OUT9 are used for reference of Source driver to generate G1 ~G9. G1 is the highest output which is close to Vs. What is the code? output voltage?

    BTW, you might need to add some smaller capacitors to buffer the fast drop. A capacitor with lower capacitance normally has as well smaller series resistance, so for buffering it might be better to add a high quality 1uF capacitor instead of a 10uF capacitor. The dielectric should be X5R or better and please always keep the DC bias of ceramic capacitors in mind.

    [Robin] You mean add small capacitor at the output of BUF18830? Or at AVDD? At AVDD, they have added a lot of capacitors, including large one and small one. As for BUG18830, they also have do some test, they tell me a little improvement is observed, but I don't know what value has been used.

    You might try as well some filter on AVDD to keep in clean. The output voltage on the buffers is directly dependent on the analog supply voltage, as given in formula 1 in the datasheet on page 11. So a transient on the supply can be observed directly on the output and if the output voltage is very close to Vs, this effect will be worst.

    [Robin] OUT2 has small peak value but more ringing times

    Thank you, again.

    Robin

  • Hello Robin,

    Did you find some more time to work on this issue? Could you find out if the mura is dependent on the BUF18830 output voltage?

    Code means the portion of AVDD that shall be shown on the output of the gamma buffer. It is normal that the output that is closest to AVDD shows the biggest distortion by a change in AVDD.

    The buffer capacitors would be necessary on AVDD but close to the BUF18830 device to make it more immune to AVDD changes. It is as well important which type of capacitors are used. X5R or better dielectric should be used and be aware of DC bias. Ceramic caps loose a lot of their capacitance if DC voltage is applied. This needs to be checked and sometimes it is even not given in the datasheet of the capacitor and needs to be asked from the manufacturer.

    Mura means a luminance difference of parts of the panel. Is this correct? Is there a possibility to check the picture content at the moment when you observe the transient on AVDD? I could imagine that directly at the start of the frame, the content is black for some lines and then the transient effect should be gone before content is displayed.

    Best regards,
    Brigitte