LTC4359 - Burning out blocking MOSFET

Hi,

I'm trying to create a load switch using the LTC4359 that can block 65V and handle 30A of continuous current. Here is my current setup:

The problem I'm having is that Q2 burns out and can no longer be switched off. I'v verified Q2 is the only component malfunctioning by removing Q2 and Q4 and testing them individually. I also powered the board with Q2 and Q4 removed, and verified that that Q1 and Q5 still operated as expected. Specifically, pulling EN high turns on Q1 and turns off Q5, and pulling EN low or leaving it floating turns off Q1 and turns on Q5. Also with Q2 and Q4 removed, I verified that the GATE and SOURCE pins of U3 get pulled to ground when SHDN! is pulled low, and that GATE is held around 10V above SOURCE when SHDN! is high.

The device operates nominally when I set VIN to 40V, and drive a 10 Ohm load such that my output current is about 4A. However, as soon as I pull EN low to switch to voltage-blocking mode, Q2 burns out and begins acting as a short. Here are some scope images of what happens (CH1 = Drain of Q2, CH2 = Gate of Q2, CH3 = Common source, CH4 = Drain of Q4):

Close up of initial voltage drop from first scope image

Close up of voltage rising edge from first scope image

Both Q2 and Q4 are heatsinked using the following: https://www.digikey.com/product-detail/en/aavid-thermal-division-of-boyd-corporation/7109D-TRG/7109D-TRGCT-ND/5189673

As far as I can tell thus far, none of Vgs,max, Vds,max, Id,max, or the max junction temperature of Q2 are being exceeded. Any help is appreciated, please let me know if I left out any relevant details.

Thank you,

Druck

  • 0
    •  Analog Employees 
    on Jan 28, 2019 10:59 PM over 1 year ago

    Hi Druck,

    The following is the Safe Operating Area (SOA ) curve for FDB86368 FET. As is clear from the scopeshots , there is a Vds of around 40V , for 5 ms with 4A load. In the SOA curve below, We can see that this operating region does not have any safety margin as the FET cannot survive this large Vds and current for more than few msec.  It looks like failure mode is Gate -source drain short for Q2 . 

    I would recommend to replace the Q2 FET to a higher SOA FET like the PSMN3R7100BSE or PSMN4R8100BSE. Also note you might need to put 2 such FETs in parallel to reduce the DC dissipation to around 1W per FET for 30A load current.

    You do not need to modify the Q4 FET as it never experiences stress during the voltage blocking mode and just acts as a diode. 

    Let us know if there is any other issue .

    Best Regards,

    Ritesh

  • Hi Ritesh!

    There is no current load during that 5ms interval. Channel 4 on the scope (the dark blue line) is the voltage measured directly across the resistive load. The voltage, and therefore the current, is zero during this 5ms interval. There IS significant simultaneous current and VDS if you look at the second scope image I attached, but the interval of time is on the order of 10s or microseconds, which I don't think exceeds the SOA. Thank you for your suggestion!

    Best,

    Druck

  • Hi Druck,

    Did you find what was the problem on your circuit?

    I have a similar problem where the first MOSFET ends up short on all 3 pins.

    Thanks for any feedback

    Franck.

  • Hi Frank80,

    I'm also facing the same problem. I think the paralleling of Mosfet will solve the problem. I'm going to test that. Did you find any other solution to the problem?

    Please reply ASAP as I'm doing the PCB design for the paralleling thing.

    Thanks!

  • Hi Alexander,

    I had to implement 2 changes to fix this issue.

    I use MOSFETs with better SOA (IRFH7545) and when the LTC4359 Shdn goes down, I also shut down the electronics powered by those MOSFETs.

    On my design, the LTC4359 powers a 300W DC-DC converter. When the MOSFETs are turned-off by the LTC4359, the DC-DC converter input voltage drops which result in a huge current increase as the converter tries to keep the load power constant. This current spike was killing the MOSFETS.

    Turning off the DC-DC converter at the same time as the LTC4359 fixed the problem.

    Hope it helps!