• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 248 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TPS54A20: Loop Response Measurement Method

Hirotsugu Kobayashi
Mastermind 8765 points
Part Number: TPS54A20

Hi team,

I measured the Bode-plot of the TPS54A20EVM, but the waveform does not look like the data sheet unless the injection level is 50mV or more.

What is the correct injection level?

Best Regards,

  • 0
    TI__Guru 50855 points

     

    50mV signal injection into the feedback seems pretty high, especially for a converter with 300kHz of loop bandwidth.  Imposing a 25mV sine wave on an output capacitor at 300kHz can take a lot of current. 

    What are you seeing at lower injection levels?

    If you are seeing a lot of noise at low frequencies, it's possible that there is too much noise for your network analyzer to discern the received signal, or that your injection circuit is somehow attenuating the AC coupled injection so that the injected amplitude is smaller than the source amplitude.

    Sometimes, to balance the need for input signal level after loop attenuation at low frequency, and current demands to generate the required sine-wave on the output capacitor at high frequency, systems will require different injection levels at different frequencies, if your network analyzer supports such options.

  • 0 in reply to Peter James Miller
    Mastermind 8765 points

    Hi Peter,

    I measured TPS54A20EVM with R10 set to 51Ω.

    I send you the waveform for each injection level.

    If it is less than 50mV, the waveform will not be as shown in the User's Guide.

  • 0 in reply to Hirotsugu Kobayashi
    TI__Mastermind 29650 points

    Hello Hirotsugu-san,

    It looks like this is being addressed on email with Peter. I will close this, and keep working with Peter thru emails. 

    Loop stablity application note.pdf

    thanks 

    Tahar

  • 0 in reply to Hirotsugu Kobayashi
    TI__Guru 50855 points

    Hirotsugu-san,

    I do not see an e-mail or forum private message from you with the resulting measurements.

    you can check your injection level at the converter by measuring the Vout AC with an oscilloscope when injecting a low-frequency ripple with the loop has very high gain.  At low-frequency injection levels, the full injected level should appear on the output capacitors as the loop holds the feedback voltage constant.

  • 0 in reply to Peter James Miller
    Mastermind 8765 points

    Peter-san,

    Sorry, I've attached the waveform.

    TPS54A20EVM_Vinj.pdf

  • 0 in reply to Hirotsugu Kobayashi
    TI__Guru 50855 points

    Hirotsugu-san,

    Thank-you for sharing.  The BODE plots you provided are normal and are showing a well behaved system.  With the lower injection levels, light loading of the converter, combined with low DCR and low ESR capacitors are producing a very lightly damped system with a very high resonant "Q" and the phase shift of the feedback is exceeding 180 degrees, which the network analyzer is interpreting as -180 degrees.

    Comparing your measurements to the EVM reference measurements, you'll note that even with the 50mV injection level, the "Q" of the resonance is much higher, this is likely because your 5A load is connected with longer wires (more series inductance) or has higher bandwidth current regulation, both of which reduce the loading current in response to changes in output voltage.  This higher loading impedance dampens the resonance, resulting in a lower peak phase-shift so that it does not cross the 180 / -180 degree boundary.

    If you would like to to see a bode plot that more closely resembles the User's Guide at lower injection levels, I would recommend a resistive 5A load connected directly across the EVM output terminals, rather than a lab electronic load connected with long wires, to reduce the parasitic inductances, which raise the dynamic impedance of the load at resonance under the test.

    At the 100mV injection level, you can see the response of the current limit just before the loop reaches bandwidth at 300kHz as the TPS54A20 can't source enough peak current to induce a 50mV sine wave on the output capacitors at 300kHz.

  • 0 in reply to Peter James Miller
    Mastermind 8765 points

    Hi Peter-san,

    I also measured that non-inductive resistor connected EVM output terminal, but waveform only changed high frequency range. 

    I re-measured bode plot using amplitude compression function of FRA51615.

    The phase rotation is gone, is that a reasonable result?

    TPSM84A21_bode_Vout2.pdf

    Best Regards,

  • +1 in reply to Hirotsugu Kobayashi
    TI__Guru 50855 points

    Yes, that looks like what I would expect for the bode plot.