I've seen a lot of buck converters where the mosfet is at the +ve rail.


Would this work?
The main attraction with this is, the microcontroller can directly drive the mosfet, and I can use a n channel mosfet rather than p channel mosfet.
Most n channel mosfets has (much) lower resistances. To put this at the top would need a mosfet driver with bootstrap as it would otherwise not turn on.

I can even use an ordinary LM358 to measure currents on the low side. Measuring currents on the high side is a pain. Just about everyone resort to dedicated high side current sense op amps.

attached the netlist (from KiCad) in case anyone feel like running this in spice. and the whole kicad project. the pcb is empty.

 

Yes, it works, but with the inductor in that position, the entire output has the switching waveform as common-mode noise, which is probably not such a great idea for emissions! Move the inductor so that it is between the switching node and the capacitor.

 

This can work but you will need to connect the inductor between the anode of D1 and the -ve of C1. Link cathode of D1 and +ve of C1. Technically easier to drive the mosfet but the +ve rail then becomes the common. Basically, you have turned the circuit upside down.
Most buck converter driver ICs are designed to use an N-channel mosfet and a -ve common supply rail. The mosfet gate is driven above the input +ve rail using a technique known as "bootstrap". Not very difficult to do and most buck converter driver ICs have the circuit built in.
Have a look at this data sheet and the circuit on page 1 and see how the N-channel mosfet Q1 is driven using Cboot and D1. Basically, the pin "BOOT" is pumped up to a voltage greater than the supply pin "Vcc". This gives the gate voltage to drive Q1.

 

The main problem with a low-side drive circuit is that the load must float off ground.
Not many loads can operate that way.

 

Yep, I have used such a circuit when the load was simply an LED, but if it is anything much more complicated, the lack of ground could be a problem.

Bob

 

Done this many times. Some loads do not have to be connected to ground. fan, LED, etc

 

Yep, I have used such a circuit when the load was simply an LED, but if it is anything much more complicated, the lack of ground could be a problem.

Bob

Unless you want to convert a negative voltage to a less-negative voltage.

 

View attachment 269868

interesting