• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 64 subscribers
  • Views 1122 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.

LMR12015 Failing with higher output votlage

Luke Fulton
Prodigy 90 points
Other Parts Discussed in Thread: LMR12015

I have used the LMR12015 in two designs, one with and output at 3.6V rated for 500mA and one at 7.6V rated for 500mA load.

I have had no issues with the 3.6V design, this works well.

The 7.6V option is not good. There is a high failure rate of the device (>40%) when the supply is switched on and off a number of times.The failure mode is an increase in the current draw from the supply, it ranges from 50mA to a full short of the input.

 

I have attached details of the 7.6V circuit. LMR12015 circuit details for TI.docx

I have tested this circuit with a number of modifications, the first been an increase in the switch on load to 20mA. It is not clear in the data sheet how to relate the minimum startup current to Vout and L value.This did not make any difference.

I tried increasing the inductance value to 4.7uH. This has  helped but the failure rate is still around 20%.

Below is the switching node and Boost node waveform for a working 7.6V output option with a 4.7uH inductor and 20mA startup load. In the same document is the switch node waveform were the current draw from the 12V supply has gone from 100mA to 135mA during a power up cycle.

LMR12015 waveform details for TI.docx

Is the oscillation after the NFET switching OFF correct?

How critical is the limit of the Switch Node minimum of -0.5V? On both the 3.6V and 7.6V options the Switch Node is dropping to below -1V for the soft-start period.

What is the effect of the oscillation on the switch node going above the Vsw level when NFET is ON?

Any help in trying to solve this would be great.

Many thanks

Luke

 

 

 

  • 0
    TI__Guru 59275 points

     

    Are you switching the supply on and off through the EN pin or switching the input supply??

    It could be some kind of transient on the input supply.  What is your input voltage ??

    You could try an Al cap of about 100uF on the input as an experiment.

  • 0
    TI__Genius 10095 points

    In Response to these questions:

    Is the oscillation after the NFET switching OFF correct?

    The oscillation after the FET turns of is normal when the inductor current reaches Zero.  This will happen in light loading of the part.  The Switch node will ring up to Vout with an oscillation frequency based on parasitics.

    How critical is the limit of the Switch Node minimum of -0.5V? On both the 3.6V and 7.6V options the Switch Node is dropping to below -1V for the soft-start period.

    This ringing should not be damaging the part

    What is the effect of the oscillation on the switch node going above the Vsw level when NFET is ON?

    The switch node will oscillate above the Vsw level when it is turned on.  This peak will become larger the greater the inductive parasitics in the loop between your high side capacitor and the free wheeling diode.  Your layout looked good from this respect and with 12V in you should not be damaging the part from this ringing.

    I agree with Frank that it is possible that you have ringing on Vin when you cycle the power.  Can you monitor Vin when cycling the part to see if Vin is ringing up to a high voltage.

    The waveforms look correct if there was a static load.  This makes me think that maybe the diode is getting damaged. I would replace the diode on one of the damaged boards and see if the extra load  is coming from the diode. 

     

    Marc

     

  • 0 in reply to Frank De Stasi
    Prodigy 90 points
    Hi Frank
    Thank you for responding.
    I am switching the actual input supply. I have a comparator circuit on the input that controls the EN pin. This enables the switcher when once the 12V input has settled.
    I have tried measuring the input pins with a fast scope but have never been able to capture any spikes above 12.5V. I have observed the ripple on the input pin when running normally and this was only 20mV. (This was with only a 20mA load)

    At the moment I don't have stock of the circuit for a batch test for the Al cap but will try this soon.

    Regards
    Luke
  • 0 in reply to Marc Davis-Marsh
    Prodigy 90 points
    Hi Marc
    Thanks for the reply. Glad to hear that my switch node is as expected. As I mentioned in reply to Frank I have not managed to see any ringing on Vin pin when cycling the power. I was probing directly to Vin with GND spring to CAP GND (was using best practice).
    I do have static load of 350R at startup.

    I took both of the failed units that had the increase in load current and replaced the diode. Both continued to show the same increased current draw, so don't think it was the diode.

    When I calculated the required inductor value I got two different results. The 3.3uH I took from WEBBENCH.
    I redid calculations using the formula in the data sheet and got 4.7uH.
    I then read the app note AN1197 and using a ripple current of 250mA in the formula got a value of 5.6uH.

    What is right?

    As mentioned the only time pass rate has improved was when I increased the inductor. I am going to try this next by further increasing to 6.8uH from Wurth: 74437324068

    Any other thoughts appreciated.

    Many thanks
    Luke
  • 0 in reply to Marc Davis-Marsh
    Prodigy 90 points

    Hi Marc

    I have finally made some progress on this. Thanks for your input on the diode, it has helped find the solution I think. I replaced the diode on 1 failed board that was only just a failure and it started working correctly!

    I have tested 5 boards with a larger specified freewheel diode and a 4.7uH inductor and all 5 have passed. I am now trying to prove that this is the solution. Can you please offer some advice/thoughts on this. I believe it is the didoe at fault.

    The origianl diode (NSR0620P2)  has a spec of 0.5A DC and 1A I surge and Vr = 20V. I am thinking this is a little under speced and even though the theory says it should be ok the real application is that currents are higher than expected. Can this be the case? The replacement diode I have used is PMEG4010 with a If = 1A,  I surge of 7A and Vr = 40V.

    Can you confirm if a damaged diode could result in the switcher failing?

    Thanks

    Luke

  • 0 in reply to Luke Fulton
    TI__Genius 16025 points
    Hi Luke,

    I'm happy to see your design is working a expected.

    The diode can affect the performance in multiple ways, and yes improperly sized diodes will cause device failure. Let me explain:
    The diode has two parameters of interest. The voltage rating and the dc current rating.
    Too low of a voltage rating and the diode is degraded over time causing failure. This failure will result in the SW node of the IC working overtime and will damage the internal FET's rendering the part inoperable. A good rule of thumb is to select a diode with a voltage rating twice the maximum SW node voltage.
    If the current rating is too low, the drop across the diode increases and the internal FET's again work extra hard to provide the required energy. This will also cause the part to fail due to overheating or make it very susceptible to a damaging transient event. Choose a diode with 2x the DC output of your application.

    Hope this helps,
    Thanks,
    Anston