Post Go back to editing

LTC4364 Can the current be increased by connecting in parallel?

Hi,

I want to use LTC4364.

I have a question, please help me.

In my specification, 45A flows in the main current path.

I would like to use LTC4364 to realize the function of ideal diode (OR connection + reverse current protection).

With the LTC4364, by connecting in parallel and selecting FETs,
Is it possible to pass a current of 20A or more, which is the specification?

With similar products,
The LTC4357 achieves this by means of  "Droop Sharing" .

Best Regards,

Naohisa

  • Naohisa,

    I understand that you want a total of 45A output from multiple power supplies.
    What are the voltage levels of your power supplies (feeding into the ideal diode-ORs)?
    Do those supplies have a current limit?
    Do you intend to use current limiting on the LTC4364?

     

    In the "droop sharing" LTC4357 application, current isn't evenly shared, due to variances in the supply voltages and RDS(ON) of the FETs.

    It should be possible to do something similar using several LTC4364s. The LTC4364 has an additional FET and sense resistor in the powerpath, which adds more potential for channel mismatch/unequal current sharing.

    Also look into the LTC4370; it is a diode-OR current balancing controller. Be aware that it only runs up to 18V though.­­­

    -Aaron

  • Hi Aaron,

    Thank you for your reply.

    The voltage is 12V, so I think the specifications are OK.

    I plan to use the "current limit" .
    I plan to use it for protection from inrush current, and I haven't decided what A yet.
    45A flows continuously.

    I also understand that the LTC4364 can be used in the same way.

    Assuming that the current will be non-uniform,
    Is it possible to use it by estimating and selecting the heat generation conditions of the external FET in the worst case?

    Best Regards,

    Naohisa

  • Naohisa,

    To get an idea of temperatures during a worse case scenario, let’s assume that only one channel is supplying the entire 45A. The power dissipation across the FETs would be

    PFET=I2 * RDS(ON).

    RDS(ON) would depend on the FET you choose. If power dissipation exceeds what the FET package can handle, choose a different FET with a larger package/lower RDS(ON) or start adding FETs in parallel.

    You can then find the temperature rise of the FET using

    Trise=PFET*RϴJA.

    RϴJA also depends on the selected FET. You now have the FET's temperature rise above ambient. Keep the junction temperature under 150C.

    Also consider the power dissipation across the sense resistor. Make sure that I2*RSENSE does not exceed what the resistor package can handle. Place sense resistors in parallel to spread out the thermal stress if necessary.

    Also consider:

    • The timer capacitor on the LTC4364 needs to be sized so the SOA of the FET is not violated when the output is shorted (part will regulate at current limit).
    • The gate capacitor needs to be sized so the inrush current is under the current limit while you charge an output capacitor.
    • The PCB layout is symmetrical between channels to get better current sharing.
    • The current limit isn’t so small where a channel mismatch causes one channel to hog the current, violate the current limit, turn off, and cause a chain reaction where the other channels turn off.

    -Aaron