• Resolved

LM5161: Negative ringing at SW pin

Part Number: LM5161

Under the "Absolute Maximum Ratings", section 6.1 of LM5161 datasheet, the minimum "SW to AGND" value is -1.5. In this particular design, the customer measured a transient that peaks at -1.8V. Is this something with which we should be concerned? The datasheet is not specific whether this "SW to AGND" rating is continuous or transient only. 

Thank you!

  • Hello Cacula,

    Any voltage violation beyond abs max spec is a cause of concern. I would first ensure that the customer is indeed measuring the voltage swing accurately? Please make sue customer is measuring between SW and Agnd with a kelvin sense connection ensuring the Gnd lead of the probe is kept very short. I recommend using a spring clip wrapped around the barrel of the scope for the Gnd connection, to do this. make sure customer is measuring on the pins (very close) in question with short leads as described and verify the measurements. We need to reduce any parasitics which will cause inaccuracies in measurements. An actual scope shot would also be helpful when implementing above if its still in violation of the abs max specification.

    Hope this helps?


  • In reply to David Baba:

    Hi David,
    One common question to confirm with you is how to set the sample frequency of oscilloscope when measure the SW signal of buck/boost chip? You know, there are 20MHz(HW), 20MHz, 500MHz, 500MHz(HW), 350MHz, 350MHz(HW), etc.
  • In reply to David Baba:

    Hello David,

    As I know it appears hardly any influence if the negative overshoot exceeds around 0.3V in practical application. If this overshoot is bigger, that would be some risk. Is the -1.8V acceptable comparing -1.5V spec.?
  • In reply to Chauncey Liu:

    Not sure what the question is Chauncy?

    Can you re phrase please?

  • In reply to Keven Guan:

    Hello Keven,

    We cannot guarantee any violation of the abs max specification. I have often found that correct measurement and validation of the undershoot level is a better way to resolve this issue.

  • In reply to Chauncey Liu:

    Hi Chauncey,

    to just show the switch node is switching and measure the switching frequency the 20MHz scope setting is enough. If you want to measure switch node ringing accurately then you would need a high frequency probe at least 500MHz (1GHz is better). The switch node ringing frequency e.g. with controllers can be in the 10-100MHz range (mostly around 80-90MHz ringing frequency). With integrated FETs e.g. 1A Buck converter the switch node ringing frequency can be up in the 150-200MHz range.

    Make sure there is a -> short <- GND connection (low parasitic inductance loop) otherwise e.g. the typical alligator GND clip wire can form a large loop inductance that would change the measured switch node shape. Same is true for large parasitic capacitances of the scope probe. Lower frequency probes have usually a larger input capacitance that limits the bandwidth and can "smoothen" the switch node wave form.

    If you want to see switch node shape most accurately (without parasitic inductance, input resistance and capacitance of the scope probe) then there is a little trick by taking off the alligator GND clip and "hover" (no contact) over the switch node area with the >500MHz probe. Then you will have a E-Field / capacitive coupling into the scope probe that does not change the "shape". Of course if you want to measure the exact amplitude in Volt you will need the proper GND connection and a low capacitive High speed probe.

    Is this what you are asking for?

    Best Regards,
    Robert L
  • In reply to David Baba:

    Hi David,

    What I mean is the setting of 'Bw' on the oscilloscope as below red rectangle shows.

  • In reply to Chauncey Liu:


    When measuring high frequency ringing, its best to set the scope on the highest BW setting.  Suggest at least 100MHz.  This will capture most of the waveform with minimal filtering.

    Hope this helps,