LTC4359

Hi Brook,
Charging an output capacitor with no inrush control certainly has potential to damage the FET, so yes, include the gate capacitor and resistor (C1 and R4) like on figure 6 to control inrush and manage stress on the FET.

You can keep the gate resistor, R4 at 10kΩ. Let’s focus on sizing the gate capacitor.

The SOA curve for the DMT3020LFDB says that it can take 200mA at 9V for 100ms. There is no Spirito region on the SOA curve (it’s a straight line, which is unrealistic), so we’ll settle for 100mA to be safe. Let’s calculate how long it will take to charge the 2.2µF output cap with the inrush current:

I_inrush=C_OUT*dv/dt
dt=C_OUT*dv/I_inrush
dt=2.2µF * 8.4V / 100mA
dt=184.8µA

It will take a brief ~185µsec to charge your output capacitor. The SOA curve shows the FET should be able to live 100mA at 9V for 1s. Even if the SOA curve omits the Spirito region, there is plenty of margin.

Now calculate the value of the gate capacitor that will give this inrush current:

C_GATE=C_OUT*I_GATE,TYP/I_INRUSH
C_GATE=2.2 µF * 10 µA / 100mA
C_GATE=220pF

Note that the tolerance of the gate pull-up current and gate capacitor may change your actual inrush current and charge time.

-Aaron

Hello Aaron,

Thank you for your response. I want to investigate furthermore on this issue.

In the input side I have a gas gauge, LTC2943. The input voltage is from battery, goes through a current sense resistor of the gas gauge, then to the MOSFET.

Both of gas gauges on the two boards are also damaged.  I took off these gas gauges and measured them, found one was pin 1 short to PGND, another one had 500ohm between pin 2 and pin 7. I don't know the gas gauge damage is caused by inrush current  related to MOSFET, or caused by overheating from the short Q1. The board was hot with the short Q1.  

Thank you.

Brooke

Hi Brooke,

2.2µF of output capacitance is relatively small, so I’d be surprised if the inrush current needed to charge it caused the FETs to fail. Using inrush control from the previous post would eliminate that potential threat though.

What kind of load is connected to V_OUT that draws 800mA? Is it a passive or active load?

-Aaron

Hi Aron,

The load of Vout  is LED driver,  maximum driving current is 700mA, and control circuit. 

Brooke

Brooke,
I don't think there is much to go wrong in the schematic you have shown me (after adding inrush control). I would look at whatever V_OUT is coupled to. Is it behaving and drawing 700mA max like you expect? Power the circuit with a current limited supply and measure input, output, current, and gate.
-Aaron

Hello Aaron,

I am also designing a switch using this ideal diode controller with back-to-back N MOSFETs as in the datasheet. On the above calculation, why did you choose 100mA (or 200mA) as inrush current while the load is 800mA? Does it mean the chip cannot turn ON while full load is connected to it? According to the SOA of the FET selected, it can handle 800mA at 9V for about 10ms.

What if the load is always connected to the output of this switch and we cannot limit its inrush? Can we design it to turn ON while full load (800mA) is connected to it on its output?

I am trying to understand this inrush current issue and how to avoid it. 

Thank you for your help

Rayy

Hi Rayy,
You are correct: Generally, the load should be kept off until the output capacitor is charged.
Otherwise, this will contribute more stress to the FET during turn-on.

-Aaron