• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Subscribers 63 subscribers
  • Views 2762 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.

TPS62170-Q1 Switching Node Waveform for Light Load

Luis Cordova
Intellectual 1940 points
Other Parts Discussed in Thread: TPS62170, TIDA-01002

Hello,

I wanted to verify something in the following switch node waveforms of the TPS62170. See the following three screenshots below:

It appears the switcher is going into PFM or Power save mode, based on the look of the waveform. My guess here is that for a brief moment, the load current is simply not enough to allow for continuous conduction, and causes the device to go into this mode. Once the load current picks back up, the ringing shortens and the PWM switching appears to be out of the AM band again (>1.8MHz).

Is this expected?

  • 0
    TI__Guru**** 150900 points
    Yes, this is possibly what is happening. You need to measure the inductor current to be sure.
  • 0 in reply to Chris Glaser
    TI__Intellectual 1940 points

    It appears that is the case. See the current screenshot below:

    Even though this Is relatively low current, could it potentially interfere in the AM band? Since it is only in the AM band for a brief moment and the current levels are quite low, I would not expect this to be an issue.

  • 0 in reply to Luis Cordova
    TI__Guru**** 150900 points
    It's not clear what you measured. The frequency is not at the switching frequency, but is much slower.

    I've never actually seen a real AM band interference issue, but most automotive customers just want to operate above the AM band to be sure and have no issues.
  • 0 in reply to Chris Glaser
    TI__Intellectual 1940 points
    Thanks again for the information.

    The current waveform was taken after the inductor, from the inductor-capacitor node to a 0-ohm resistor. Should it be taken right after the inductor and before the capacitor?

    Regarding EMI and AM band interference, My understanding is the voltage switching node can cause Radiated Emissions, and the current draw of load on that rail can cause Conducted Emissions.
  • 0 in reply to Luis Cordova
    TI__Guru**** 150900 points
    Ah, you need to measure in a different place. You need to measure in series with the inductor itself. This requires lifting the Vout terminal of the inductor, soldering a wire between this and the PCB, and clipping the current probe around this wire.

    Yes, the voltage transitions on the SW node are a primary cause of radiated EMI. The switching action also causes pulsed currents on the input supply, which is a primary contributor to conducted emissions.
  • 0 in reply to Chris Glaser
    TI__Intellectual 1940 points
    Aha, thank you for clarifying. I will measure in series with the inductor itself and report back.

    One more thing: since the switching action causes pulsed currents on the input supply, is it worthwhile to measure this current and the frequency content of these pulsed currents? This way, you could see where in the frequency spectrum you are causing the most noise.
  • 0 in reply to Luis Cordova
    TI__Guru**** 150900 points
    Is this the same TIDA-01002 that we looked at a few months ago? That one had an input filter which is supposed to help with conducted emissions.
  • 0 in reply to Chris Glaser
    TI__Intellectual 1940 points

    Yes, this the same one. There is a power over coax filter, which has a high impedance from around 1MHz to just over 100MHz.

  • 0 in reply to Luis Cordova
    TI__Intellectual 1940 points

    See the waveforms taken in series with the inductors on the input, 2.8V rail and 1.8V rails:

    This is the input current, through the special power over coax inductor, to the two switchers.

    This is the current through the inductor of the 1.8V rail.

    This is the current through the inductor of the 2.8V rail, but only a 2 microsecond snap of it. The screenshots below will show where it appears to be briefly entering discontinuous conduction:

    This appears to be in line with the mode change shown in the voltage waveforms at the switching node of the 2.8V switcher. Maybe this is obvious, but just to be sure: is it safe to say the switching frequency is 2.083MHz, despite brief moments of apparent discontinuous conduction? Or whatever the frequency of the switching node is?

  • 0 in reply to Luis Cordova
    TI__Guru**** 150900 points
    Everything you've shown here is ok for the IC.

    Based on the load current, it will operate in either PFM or PWM mode. In PFM mode, it stops switching when the load is light enough. This may cause a scope measurement to report a lower switching frequency but in reality the frequency between switching pulses is the same. There are just gaps between the pulses. (Actually, for your last set of measurements, I don't see any skipped pulses or PFM mode. So, the 2.083 MHz is likely the real operating frequency with your Vin and Vout conditions.)

    Finally, from the very first waveform in your first post, it seems that the input voltage is very low. Is there some concern at this corner case? Does the switching look better/different with higher Vins?
  • 0 in reply to Chris Glaser
    TI__Intellectual 1940 points
    There was initially, so we thought it would make sense to get pretty close to the corner case. However, typically the input voltage will be 6V to 14.5, and this test platform provides 5V in.

    Looking at the current waveforms on a test platform which provides 12V in, the switching looks a little better, at about 2.13MHz, with the same gaps seen. But I think the >2MHz switching seen for the low V in and low output current corner case is very good.