Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed from November 25 through December 2. Thank you for your patience.

  • State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 65 subscribers
  • Views 1717 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.

LM5105 Mosfet Failures and Slow Turn Off High Side

Graham Sayers
Prodigy 45 points
Other Parts Discussed in Thread: LM5105

We have a solar charging circuit that uses the LM5105.  We have had a handful of cases where the HS and LS FETs have been damaged while testing the circuits in the factory before shipping.  This is light testing (maybe 3A) consisting of connecting a battery and solar panel briefly to ensure everything looks normal.  In other words we are not stressing the circuit.  The failure occurs pretty rarely (maybe 1%) so it has been tricky trying to track it down.  Here are the basics:

Solar Vin: 50Vdc max

Solar Power: 380W max

Vout: up to 30Vdc

Iout: 15A max

Duty cycle: microcontroller, limited to 30% to 99%

FETs, two HS, one LS are all: Fairchild FDB13AN06A0

HS gate has a 10ohm resistor and reverse diode for turn off.  LS gate has a 5ohm resistor and reverse diode for turn off.  I did see the other post where removing LS resistor was recommended.

Deadtime is RDT = 10k which is 100ns.

If I replace just the HS and LS FET the unit works again.  I have been trying to come up with scenarios that could cause the FETs to fail but have been struggling.  I did notice on one unit the HS Vgs was turning off much slower than on other units.  I attempted to measure the current at the source of the LS FET, but all I had was a big wirewould resistor so I am not totally confident that I wasn’t just picking up a bunch of noise.  But there did appear to be a current spike though right as the LS switched on.  Next week I will put a small smd resistor on the LS source and try to get a clean reading.

Does this look like shoot through?  Could shoot through like this destroy the FETs?  Any ideas on why the HS would be so lazy turning off on some units?  Is some sort of variation expected?

Thanks.

Normal unit, high to low transition.

Slow unit. HS turns off very slow.

Slow unit showing Vgs. Still maybe 5V on HS when LS turns on.

  • 0
    TI__Mastermind 25570 points
    Hi Graham

    If you look at the bottom screenshot you see a linear slope on Vgs - dV is about 5V, dt is about 100ns and the FDB13AN06A0 datasheet indicates Cgate of 1350pF. This would indicate a gate current of about 67mA ! - much less than I would expect. Since this is the turn-off event, the 10 Ohm resistor is bypassed by a diode - there should be nothing preventing you getting close to the peak pull down current of about 1.6A or so. Doing the same calculation on the top waveform would indicate a gate current of about 450mA which is much more reasonable.

    In the 'bad' plot., the switched node isn't moving at all in the time immediately after you attempt to turn the high side MOSFET off - this indicates that the top MOSFET hasn't turned off at the instant the low side MOSFET is turned on - shoot through current is a definite possibility here.

    It may be that the dead time is too short at 100ns - have you tried increasing it? - although I'd like to understand why the turn-off is so slow first.

    Have a look at the HB - HS supply and see if it is maintaining 12V or transiently dropping below or close to the HB UVLO threshold?

    Is there something in your layout that is adding inductance from HO to gate or HS to source - stored energy in the source inductance will cause a negative going spike at the source and can add to the gate source voltage and make turn-off more difficult.

    I have had problems with gate drive resistors going 'high' due to the pulsed nature of the current they see - 0603 resistors of unknown provenance - so I would recommend you use 1206 parts here. The diode should be carrying the gate discharge current at the turn-off event so this may not be the issue.

    Regards
    Colin
  • 0 in reply to Colin Gillmor
    Prodigy 45 points

    Thanks Colin.  Let me digest this a bit.  Will be working on this some more this afternoon.  I will check a larger sample of units and see if I can find this slow turn off in any other units.  

    Graham

  • 0 in reply to Colin Gillmor
    Prodigy 45 points
    So I spent some more time with this particular 'Bad' unit with the slow high side turn off.  I was able to take a more accurate current measurement on the low side FET source.  Currents were spiking up to 250A for up to 100ns as the high side was turning off.  Also the FETs were noticeably hot even at very low power.  Pretty clearly in this case shoot-through was a problem.
    However, while there is clearly something wrong with this one bad unit, I tested 20 other units to see if any others had shoot-through.  All of the other units looked normal.  All of these are recent units from production. 8 of the 20 units had high side and low side mosfet failures during routine production testing.  I am feeling now like the slow high-side turn off on this one unit is more likely the result of some sort of failure, rather than the root cause of the failure.   
    Does anyone have any suggestions for systematically tracking down seemingly random FET failures?  The test conditions are room temperature 5A max (design is 15A max) for maybe 60 seconds at most.  So I don't think it is a failure in normal forward operating mode.  It feels like some sort of unexpected transient event.
    Thanks!
    -Graham
  • 0 in reply to Graham Sayers
    TI__Intellectual 1875 points
    Hi Graham,
    What is the topology? Is it isolated or non-isolated?
    Isolated topology with transformer tend to see this kind of failure when there is a small imbalance in the transformer volt-second.
    Spurious volt-second imbalance can cause transformer to saturate which will result in high current through the FETs.
    Regards,
    Ritesh