Hi there,
we experienced that in our circuit the MOSFET got burned when using SHTD# to switch off.
According to datasheet and LT4363IDE-2 FET gets in short the Gate should be discharged with 1mA. But this is only the case until the Gate-voltage is down to the output voltage. After that the Gate discharging gets extremely slow. We confirmed that by using the ADI evaluation module (https://www.analog.com/en/resources/evaluation-hardware-and-software/evaluation-boards-kits/dc1935a-b.html ).
We modified the board by desoldering the 47nF CG cap, in order to accelerate Gate charging/discharging as much as possible.
We connected a 2A load and switched on and off using the SHTD# pin. The Gate-voltage was charged to 25V (13V above Vin), discharging needed 13us for 15V Gate-voltage-drop - this should be OK for 1mA discharging current and only the input capacitance of the FET. After the 13us the Gate voltage droped down way slower (see osci screenshot). It never dropped below the output voltage, which in turn resulted in bad output voltage drop.
(black: SHTD#; green: Gate; red: output
note different scaling!)
BTW: We saw the same behaviour when using UV as switching input.
I think this is an undocumented behaviour, isn't it? Our interpretation: for switching the MOSFET on/off by SHUTD# needs a MOSFET with very relaxed SOA specification - which comes with quite a big package. In *our circuit* voltage und current are higher (30V/30A). Our MOSFET would survive the discharging if it were done with 1mA - but this timing is so bad that we would need a MOSFET that does not fit to our PCB.
Is there any modification possible to ensure 1mA Gate discharging down to 0V completely ?
In case we would discharge the gate with an external NPN transistor (controlled by inverted SHTD#) - would this lead to any issues in the LT4363-2 (e.g. overload to the charge pump)? Should we use a series resistor then?
Thank you in advance!
