I recently created a high-side switch in order to secure a servo channel for my team's avionics, obviously without any knowledge about MOSFETs switching and alike.
We have an airbrake systems powered by a decently sized servo motor. We feared the servo activating before we had the chance to set the intended position. This, with an open pwm channel, might mean a sudden movement of the servo and the connected airbrake system.
Jerking the airbrake could mean impacting it against either the external launch rail, or internally against the main body structure. Depending on the streght of the impact, this could mean permanent damage to the airbrake or to the servo itself.
But, other than such sudden impulses, a disaligned servo could mean that the zero value is placed in an unreachable position, forcing the servo against a wall and wasting lots of valuable energy. So, we determined, that a way to control if the servo recives power or not could do the difference in multiple scenarions.
+BATT
===
|
+--------+
| |
--- 100k |
| | |
| | |
--- ||-+
| ||->
+-----||-+ PMOS
| |
||-+ \------| OUTPUT
||<-
INPUT |----||-+ NMOS
|
|
|
===
GND
So, we would like to use the NMOS capacity to be turned on only with an active signal on its gate, but cannot sustain the losses due to the Rdson present on nearly all the NMOS MOSFTEs. So, what we can do, is mix the positive on-signal of the NMOS with the low Rdson of the PMOS.
In this way, until we don't supply enough voltage on the "INPUT" label, we can assure ourself that any power that passes trough the PMOS will work to close the PMOS itself. This will render it nearly impossible to power the servo without powering the NMOS from our MCU.
Obviously the 100k resistance is there only to prevent the activation of the NMOS to suck dry the battery voltage.
As usual though, we can make sure of it by using a digital simulator like falstad