• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 62 subscribers
  • Views 110 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.

TLV62585: Noise due to feedforward capacitor value connected in FB pin

Rahul Suresh
Prodigy 30 points
Part Number: TLV62585


Tool/software:

Hi Team, 

Above is our design using the TLV62585 buck converter. We had used this Buck to convert 4.6V to 3.8V. We are using this buck to power the LTE module with a maximum current consumption of 2A.

This LTE module is on a different board and we connected the LTE board to the buck regulator board through a B2B connector.  Our issue is that we hear an audible noise near the buck circuit. probably from the inductor. When we remove the LTE board, the noise will vanish. This means LTE module load only causes the noise. We tried with different values of feed-forward capacitor(C320 in schematics), and below are the results

With different capacitor values, the intensity of noise and output voltage ripple values are noted down in the table below. As of now, it seems keeping C320 NM is the best. But do you have any other way to avoid this noise issue and to get an optimum transient response? LTE module load will require a sudden burst of load currents, so the transient response should not be compromised.

Sl No
Test 1
Observations
 Ripple test 1
Max
1 C320 NM Slight sound 110mV
2 C320 mounted with 22pF capacitor high hissing sound 131mV
3 C320 mounted with one 10pF capacitor. sound greater than C320 NM scenario 113mV
  • 0
    TI__Intellectual 2770 points

    Hi Rahul,

    Thanks for posting your query. Since the device is operating in Auto Mode (lighter load->PFM and higher load->PWM) therefore the switching frequency is varying as shown in one of the plots in datasheet:

    Currently we don't know much info about the load transients like what is low load level, high load level and the slew rates but i am assuming that if the low load for device is below ~75mA then the switching frequency in Pulse Frequency Modulation would be in the audible frequency range. The only way we can get rid of this is by using the resistor divider in such a way that it bleeds >=75mA of current. Currently with this combination of resistor divider, if there is no load then my calculation shows bleeding current of 62.8mA. 

    Secondly, if you see that the LTE board is withdrawing >=13mA which we consider as low load (in total current withdrawn from device is >=75mA), then the problem could be due to the LC combination of the B2B connector and the capacitors on the LTE board that can resonate due to the LC tank formed on the other side as we don't know how long the B2B connector is and is there any capacitor on LTE board or not.

    Moreover, with 3.8V (higher Vout) you are currently using a Cout of 2x22uF which is <47uF therefore feedforward capacitor is not needed as mentioned in datasheet.

    Finally, it would be nice to have the scope shots of the SW node of the device, VOUT at buck side, load transient profile of LTE Board (to know about the LOW, HIGH current and slew rate of the load transient) and also VOUT at the LTE side to see that there is not resonance happening.

    Looking forward to hearing from you! Slight smile

    Best Regards,
    Zakir

  • 0
    Prodigy 30 points

    HI Zakir,

    Please see the attachment for the details you asked. If you need more clarification or scope shots at a different time/slot, then let me know.

    Also, as I said earlier, we found less ripple and less audible noise when we made the feed forward cap not mounted. ( We tested feed forward cap 10pf, 22pf and the result was worse. not tested more than 22pf) And we don't have a mic in our application. So we are planning to proceed with C320, not mounted. Let me know your thoughts.

    LRLC_ECAM __ LTE power experiments.pdf

  • 0 in reply to Rahul Suresh
    TI__Intellectual 2770 points

    Hi Rahul,

    Thanks for providing the details. Yes you can go without C320 and it would be fine doing that in your case.

    From scope shots, we can see the frequency content lies in the audio range which means that you have to modify the resistor divider network in such a way that the switching frequency is away from the audio range. Can you confirm this by not connecting the LTE module and observe the VOUT signal. If there switching frequency is less than 20kHz then i would suggest to reduce the resistor divider network with, R1 = 38.3kOhm and R2 = 7.15kOhm (Vout = 3.813V). Try this and let us know what changes you observe in VOUT signal's frequency when no load.

    The next thing which was interesting to see was low frequency content even at full load. Is the current withdrawn by LTE module constant once or it pulsates depending on when it needs energy? If it pulsates, shall we know how quickly it withdraws current from no load to full load in A/us. From last image i assume that it is a constant load but still if we see some low frequency content then it is most probably due to the second LC tank formed by the B2B connector and Capacitors on LTE module. When you use a constant resistance load withdrawing 2A lets say, do you see low frequency content there? Have you checked that? 

    Best Regards,
    Zakir