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.

TPS23758EVM-080: Acoustic noise issue with LDT1018 on a circuit similar to EVM

Part Number: TPS23758EVM-080
Other Parts Discussed in Thread: TPS23758

Hi,

I am having issues with acoustic noise on a LDT1018 which is used in a circuit virtually identical to your EVM. I am reaching out in search of help for troubleshooting this and also to hear whether you have any experience dealing with acousting noise on this design?

Any pointers would be very much helpful! We see a clear spike at 11kHz. Some of our designs are louder than others, though no differences are seen when probing the pins of the transformer.

Thanks!

  • Hello Magnus,

    What is the load of the PD?

    Also, how did you implement FRS and DTHR? 

    Lastly, can you provide waveforms of the Drain?

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • Hi,

    We've tested with close to no load and up to 11W. I'll PM you the waveform and a cut out of the schematic including the DTHR and FRS.

    -Magnus

  • Magnus,

    Please include the full schematic when you send it. Thank you.

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • So we realized that the acoustic noise is directly related to the dithering and by pulling DTHR to ground with a 273K and FRS down to ground with 60K3, we fixed the acoustic noise. The C33 cap to ground from DTHR was changed from 2n2 to 1n.

  • Magnus,

    This sounds right. Typically dithering is set to 11K to dampen that frequency, so that is why I want to check it. Glad to hear the noise is fixed.

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • The following changes were performed. The noise is not gone, but changed to ~23KHz. We wish to use a microphone in our design and so the frequency component of the noise ought to be as high as possible. Any advice on this?

    R32 is changed from 237K to 300K R36 is changed from 0 to 1K

    C33 changed from 2n2 to 1n

    Thanks,

    Magnus

  • Hello Magnus,

    Where exactly is the 23KHz noise?

    Can you specify which pins on the transformer, or is it somewhere else?

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • The noise is what we measure acoustically, but the frequency components are visible on the switching signal (pin1). I'm sending you two screenprints from my scope where you can clearly see the 11kHz and 23-24kHz noise components with an FFT.

  • Hello Magnus,

    Thank you for sharing the screen shots. I think we can either: alter the DCR snubber on the drain, add another snubber to ground on the drain to help filter, or add a capacitor on CS to ground to help stabilize the sense pin. 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • Thanks for the suggestions and excuse the late feedback.

    Could you give me some recommendations on designing the snubber circuit? Is that from drain to power-ground/return or RSNS/source? Looking at the resource below, they suggest a DCR snubber across the primary winding, or a CR snubber from drain to source.

    http://www.ridleyengineering.com/images/phocadownload/12_%20flyback_snubber_design.pdf


    I did attempt to set a 100uF and 100nF from CS to ground, but to no vail. At 100nF a screetching noise appeared at 1Hz intervals and the 100uF cap just reduced the output voltage to ~1/5th of the initial level. 

  • Hello,

    Adding the snubber to RSNS would be better since that would block the sense resistor from the capacitor discharge during startup. This design already has a snubber across the primary winding so that's why I think to block the additional frequencies we should add a new one. 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • I am finding it challenging to dimension this as most guidelines are for designs where the noise is of a higher frequency. Here we have:

    Fsw=250kHz

    Fno=11kHz (dithering carrier frequency is 11kHz and the dithering amount is 13% of 250kHz)

    So finally, an RCD snubber between the DRAIN pin 24 and the RSNS pin 1. Is the diode essential?

    To be honest, I can't really see how the snubber will fix the frequency component at 11kHz. Additionally, we have harmonics of course. Any clarifications are very much appreciated!

    Thanks.

  • Hello Magnus,

    I do not think the diode is essential since this is across the FET. The diode would act more as a clamp, whereas the RC snubber would be more like a filter. I think the best step forward is to use RC. 

    What I have found with snubber design is that we have plenty of calculations, but parasitic capacitances/inductances can be difficult to factor in. Therefore adjusting the R and C values on board usually needs to take place. 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • Hi,

    It's just that snubbers are basically meant to remove and attenuate high frequency components, right? While in this very case we are struggling with components lower than the switching frequency. And since we are in fact dithering at this frequency, it would be counter-productive to attempt to go ahead and remove it.

    -Magnus

  • Hello Magnus,

    Yes the fact that the noise is below the switching frequency makes this particularly difficult. Most noise canceling techniques are used for high frequency or EMI noise. 

    That is why we consider spread spectrum frequency dithering a competitive feature for the TPS23758 and our other products that have this feature. 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments