LT4256-2 output short test problem

Hi Andy,

The LTC4256 has variable over current response time and will react much faster during severe over current events. In order to limit the peak current overshoot , you can increase the gate resistor R1068. This essentially decouples the gate capacitor from the MOSFET Cgs and allows the part to pull the gate low much faster. 
Also make sure that the current is being measured at the input and not the output , as the current at output will also include current being dumped into the short and is uncontrolled. That does not give the true short circuit current flowing through the MOSFET.

Best Regards,
Ritesh

Dear Ritesh,

Thanks for your support.

I have other questions about LT4256. I am using LT4256-1, not LT4256-2.
R1068 I used 3 resistance values for testing. The test conditions are a bit different from previous tests. The previous test output short-circuit point was at the output of the MOSFET. The test output short-circuit point is now on the PCB's Connector.
Here are the results of the tests. The increased resistance of R1068 does speed up the MOSFET turn-off. But the peak current is about the same.
Ch1: short current
Ch2: Vds
Ch3: Vg to GND
Ch4: Output voltage
Math1:Ch3-Ch4

Gate R=100ohm

Gate R=402ohm

Gate R=620ohm

Question 1: The gate voltage will not be pulled to 0V after the MOSFET is turned off. Will this have a bad effect?
Question 2: Which resistance value is better for R1068.
Question 3: Is there any way to reduce the peak current of the output short circuit.

The following is the waveform of LT4256-1 doing Output short before power on.

Ch1: short current
Ch2: Timer to GND
Ch3: Input voltage
Ch4: Output voltage

Question 4: Is the short-circuit current and Timer waveform normal? What causes that current waveform?

HI Andy,

From the scope shot the current spike earlier was 550A , whereas now it seems more like 100A or so.  As explained earlier the output short current will include current dumped from load capacitors as well into the sort and does not flow through the pass MOSFET. As for the actual current peak, it is governed by the short circuit impedance (inductance + resistance  mainly ) and the voltage supply capacity. LTC4256 will detect the short circuit current and and respond as per the plot shown below. 

1. The gate will be pulled to GND and the MOSFET will be turned off  irrespective of Rgate. After the  cooldown duration the gate will again be allowed to turn back on in the LTC4256-2.

2. For the resistor value , you can use 620  ohm as it seems to reduce the gate turn off duration. Also not increasing the resistor will reduce the immunity to ride through input voltage steps. If you expect large input voltage steps you may verify at that test  condition as well with 620 ohm Rgate.

3. Already answered. This peak is decided by the impedance of the short and power supply capacity. Some of our more advanced controllers such as LTC4260, ADM1272  have fast pull down feature and current limit fold back to  reduce the peak currents. 

4. This does not seem correct. Ideally the timer should ramp linearly and current should be flat. Check if the power supply current limiting?

Best Regards,
Ritesh

HI Andy,

From the scope shot the current spike earlier was 550A , whereas now it seems more like 100A or so.  As explained earlier the output short current will include current dumped from load capacitors as well into the sort and does not flow through the pass MOSFET. As for the actual current peak, it is governed by the short circuit impedance (inductance + resistance  mainly ) and the voltage supply capacity. LTC4256 will detect the short circuit current and and respond as per the plot shown below. 

1. The gate will be pulled to GND and the MOSFET will be turned off  irrespective of Rgate. After the  cooldown duration the gate will again be allowed to turn back on in the LTC4256-2.

2. For the resistor value , you can use 620  ohm as it seems to reduce the gate turn off duration. Also not increasing the resistor will reduce the immunity to ride through input voltage steps. If you expect large input voltage steps you may verify at that test  condition as well with 620 ohm Rgate.

3. Already answered. This peak is decided by the impedance of the short and power supply capacity. Some of our more advanced controllers such as LTC4260, ADM1272  have fast pull down feature and current limit fold back to  reduce the peak currents. 

4. This does not seem correct. Ideally the timer should ramp linearly and current should be flat. Check if the power supply current limiting?

Best Regards,
Ritesh