ADP1613
Production
The ADP1612/ADP1613 are step-up dc-to-dc switching con-verters with an integrated power switch capable of providing an output voltage as high as 20 V....
Datasheet
ADP1613 on Analog.com
Hi guys,
I'm reaching out for a quick design review. I'm doing a revision on my solar/power grid powered gateway that is supposed to power some external sensors for performing readings (such as temperature).
The gateway must operate without battery connected to an outlet (max 24V input) or with a 18650 battery type with solar recharging. Depending on the connected sensors, the supply voltage needs to be either 5V or 12V. It's not supposed to have both on the same gateway. The gateway should be provisioned and receive information if it should supply 5V or 12V.
I've attempted different dc-dc regulators, such as step-up/step-down but they were un-stable or with higher input voltage requirements. Since 18650 batteries can go low as 3.6V I think the ADP1613 is suitable for this revision.
As such, I've designed this schematic:

Vout_bat will be battery voltage when external power supply is off (4.2V max) or charging voltage when charging via solar (4.4V max). The 5V line is actually 5.5V (regular dc-dc with feedback resistor). I'm using 5.5V to make sure that Q1 will work as intended and use the 5V input which is directly connected to the solar panel / outlet when the value is above VOUT_BAT. Since, D4 will drop 500mV I'm playing safe.
Then Q2 pair is for implementing the true shutdown with R20 being there just to protect the MCU.
Now, the part that I've questions is the voltage selector that I've implemented with Q4. If my calculations are correct Vout should be 5.52V when VOUT_SEL is 0 logic and 12.46V when it's 1. For the compensation pin, I've used the values from the designs on the datasheet and pair Q13 should operate for the 5V and Q14 for the 12V. I know that probably the NMOS will add some capacitance between source and drain which might affect the compensation but will this work? The R31 is there to help discharging C28 in case for some reason the voltage needs to be changed at runtime.
Please not that VSEN will then be connected to an header of pins for powering the sensors like this:

So, with a 220uF capacitor connected and with the possibility of being disabled and instead using a pass through between the input line (24V). I guess that D7 is enough to protect reverse voltage to the output pin of the ADP.
Will this work? I really need that this revision is the last one, so I appreciate your input guys.
Thank you for the ongoing support,
Best regards,
Fernando Fontes