LTC 3705 auxiliary supply - lockup on startup

To stay in normal mode, the LTC3705 is looking for approximately >= 1.3V from NDRV to Vcc (meaning NDRV - Vcc >= 1.3V) and 50uA into NDRV.

Thanks for the reply. I tried lovering the resistance of R48 to 27k. I measure 12.2V at NDRV when I do this (VCC is 12.02V). I think this is caused by driver for the linear regulator. It is trying to turn off the external NMOS because the voltage is higher than 8V. The current into NDRV is (24-12)/27k=444uA. This problem seems to be biting its own tail. The chip is waiting for the voltage to rise to 13.4V while at the same time it is driving NDRV low as long as the voltage is >8V? Any suggestions are welcome.

The majority of the applications appear to use a "standard 3V threshold type (i.e. not logic level)" MOSFET. Page 18 of the LTC3725 Data Sheet shows another option (which may or may not have the same problem as you already have - I am not sure.) Where is your 12V coming from (and when with respect to the input voltage)? What is the input range? Depending on this information, I would think a variation of the standard MOSFET solution is the best bet.

The majority of the applications appear to use a "standard 3V threshold type (i.e. not logic level)" MOSFET. Page 18 of the LTC3725 Data Sheet shows another option (which may or may not have the same problem as you already have - I am not sure.) Where is your 12V coming from (and when with respect to the input voltage)? What is the input range? Depending on this information, I would think a variation of the standard MOSFET solution is the best bet.

Thanks. The 12V is coming from an LM317 regulator. The answer to "when" depends on the rise time of the input 24V feeding both the lm317 and the LTC3705. I agree that it was a dumb decision to not use the standard MOSFET solution on this board. The problem is that we have a number of boards in the field and some of them are failing when the 24V feeding the board is connected to another board with a heavy capacitor bank. When the 24V rises very slowly, the LTC3705 locks up. I have sort of solved the problem by fitting a 3.3k resistor from 24V to 12V_SR that lifts the 12V_SR to 14V when the LTC3705 is locked up. When the LTC3705 then starts switching it consumes more power and the 12V_SR drops to 12V again. It's a small hack but it gets around the problem. If there is a more elegant solution that I could retrofit to these boards please let me know.