Because of the holidays, TI E2E™ design support forum responses will be delayed from Dec. 25 through Jan. 2. Thank you for your patience.

  • State
  • Locked Locked
  • Replies 11 replies
  • Subscribers 65 subscribers
  • Views 2302 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.

LM5017 noisy inductor current with higher load capacitance

Jens
Prodigy 140 points
Other Parts Discussed in Thread: LM5017

Hi.

I'm using a layout with LM5017. Buck-converting from 28...65V down to 25V. Load current approx. 500mA. I got some devices back in office with defective regulators.

Measuring shows the following:

All works fine with resisitve load only. Peak inductor current is well below 700mA. Additional capacitive load of 33µF makes the inductor current really noisy at input voltages >40V. Peak inductor current is higher than 800mA and nearly reaching inductor's saturation current. Due to I can't use a bigger inductor I have to reduce this noise. But howto?

Also the output voltages varies approx. 1V depending on input voltage. Maybe due to this noise.

External load is applied with 4"-8" of cable and a diode (wrong polarity protection) and consists of a 33µF capacitor and a switching regulators converting to lower voltages.

I added the schematic as .pdf.


Regards,

Jens

Converter.pdf
  • 0
    TI__Guru**** 191445 points

    Can you post some waveforms that demonstrate your issue?

  • 0
    TI__Genius 11490 points

    Hi Jens,

    Please share SW node voltage, inductor current, and VOUT waveforms to help identify the issue.

  • 0 in reply to Vijay86929
    Prodigy 140 points

    Hi.

    Thanks for your quick response.

    I added a picture with three waveforms. All with approx. 500mA load current. From left to right:

    35V Vin, no external capacitor, which is stable

    55V Vin, no external capacitor, more or less stable

    45V Vin, with external capacitor, unstable (it's slighlty better with higher or lower input voltages)

    I can't estimare how much the current clamp affects the switching process.

    Some of the defective devices I got had a burned Vin Pin.

    Regards,

    Jens

  • 0 in reply to Jens
    TI__Guru**** 191445 points

    I do not generally support LM5017, but reading thru your posts, I see two things.  First adding the addition additional external capacitor may be reducing the ripple on the output.  You may want to consider an external ripple rejection circuit as described in the data sheet.  Second, the switching node rise and fall times look very slow.  What is your layout like?

  • 0 in reply to JohnTucker
    Prodigy 140 points

    Hi.


    Rise and fall slopes are ~165V/µs at steepest parts after disconnecting the current clamp.

    The whole layout is very small. IC and inductor are placed on opposing sides of PCB with a very short switching node trace.

    Ripple voltage on FB-pin is ~100mVpp, no matter if additional external capacitance is connected or not.

    A smaller Cout (2,2µF) and a Rc of 2,2 Ohm makes things better, (Vripple,FB = ~200mVpp). Also the external capacitance makes no difference in ripple amplitude at feedback.

    Another thing is that I would do the following:

    At room temperature the IC has more than 100°C with a 450mA load (experimental layout, Exposed pad maybe not perfectly connected). To avoid thermal problems I tried to power Vcc from output voltage via 2,4kOhm to VCC and clamp with a 12V-Zenerdiode. But the LM5017 gets as hot as before as if it is not disabling the internal voltage regulator altough voltage at Vcc-Pin is between 10...12V.


    Regards,

    Jens

  • 0 in reply to Jens
    TI__Guru**** 191445 points

    I'm used to seeing a much more well defined rectangular waveform for the switching node, rather than the "soft" waveform you show.  For the current to change like that usually indicates actual load current variation or instability in the circuit.  That is why I suggested looking at the ripple injection.  COT devices rely on ripple on the output for proper operation of the off time comparator.  Let's see what the LM5017 support team has to say.

  • 0 in reply to JohnTucker
    Prodigy 140 points

    Hey John.


    Stupid mistake by myself - oscilloscope was filtered. Rising slope is 6kV/µs falling slope is 3kV/µs.

    Maybe I get closer to the reason for the noise. I changed some values and now it's as in added .xls sheet. With capacitance connected and slowly raising input voltage inductor current nearly exactly becomes noisy when frequency crosses the calculated "Resulted nominal switching frequency".

    Also switching frequency is very input voltage dependent. At 450mA load it changes from 310kHz @30V upt to 700kHz @60V.

    Regards,

    Jens

    LM5017QuickStartCalculator.xls
  • 0 in reply to Jens
    TI__Genius 11490 points

    Hi Jens,

    I don't think the converter is not switching right so I would not give much credance to the switching frequency value. I suggest trying th external ripple injection. Another thing to try is to reduce the nominal switching frequency to (say) 200kHz range. In some of the waveforms it seems the part is also running into minimum off time limits.

    Regards,

    Vijay

  • 0 in reply to Vijay86929
    Prodigy 140 points

    Hi.

    At very low input voltages <30V it conflicts with minimum off time. Is there a problem except output voltage will not reach desired value?

    I can't use such a low frequency because size of inductor ist strictly limited and I' m not able to place more than150µH.

    Can you please explain, why switching frequency and output voltage is so extremely varying? Regarding to datasheet it shoud be more or less constant (+-25%).

    Ripple at feedback pin is 250mVpp minimum, so I don't think problem is really caused here.

    Regards,

    Jens

  • 0 in reply to Jens
    Prodigy 140 points

    Hi.

    Just for info:

    Changed Vcc from 1µF to 2,2µF and Cbst from 10nF to 22nF. No noise any more.

    But frequency ist still extremely varying with input voltage.

    Regards,

    jens

  • 0 in reply to Jens
    TI__Genius 11490 points

    Good to hear that.

    Since you have confirmed that the part is in min TOFF at low VIN that would explain change in switching frequency. Once the converter hits minTOFF the frequency drops 1/(TON+TOFF). Normally as TON increases (VIN decreases) TOFF automatically decreases to keep the frequency nearly constant. However once TOFF reaches minTOFF then it cannot decrease anymore so frequency starts decreasing.