I am in the need for a high-side load switch and came up with the following schematic. High-Side voltage levels and current vary between 5-12V / 2 A max. The high-side PMOS (SI5509DC) is rated for 20V V_DS and 12V V_GS and up to 4A.

A 10V Zener diode D1 and the voltage divider formed by R2 & R3 is meant to limit the PMOS Gate-Source voltage. Diode D2 at the output shunts negative going spikes due to inductive loads towards ground.

C1 controls the fall time when opening the switch. I have not yet found a way to make the turn-on behaviour a bit smoother.

The switch will be closed most of the time and switching frequency is not an issue. However, I need to have negative logic at the input to keept the switch open during bootup when the driving pin of the CPLD is tri-stated and pulled towards 3.3V by internal pull-ups.


simulate this circuit – Schematic created using CircuitLab

My Logic interface is 3.3V @ 20mA CMOS (MaxV CPLD). I need to make the thing as robust and compact as possible and would like to ask for advice and comments.


As you note, the circuit as drawn switches "on" much more quickly than it switches "off".

If you wish to make it more symmetric, simply add a single resistor (say 100K) in series with M2 gate || C1. Note that, to save space, you could use a small SMT zener and a dual MOSFET. The 1N4148 is rated at only 200mA, and you might want to use something beefier such as a small SMT 1A Schottky rectifier (for example, a CDBA120) if you're switching heavy inductive loads.

| improve this answer | |

If your pins are high impedance during boot-up, is negative logic necessary? Couldn't you just use a pull-down on the output?

What I've used in the past, and it's been robust is: Power P-channel high side switch. 10 kOhm pull-up between P-channel gate and P-channel source. Signal N-channel drain connected to P-channel gate, source going to ground. N-channel gate connected to microcontroller. 10 kOhm pull-down between N-channel gate and ground.

With a suitably low-level signal N-channel, 3.3V control is not a problem. I use BS150.

You can make the pull-ups/downs weaker, but switching times will be slower, which may actually be a problem if the load is significant compared to the rating of the switch.

| improve this answer | |
  • \$\begingroup\$ The CPLD has internal pullups that are active during startup and programming of 5k-25k which besides being quite a span - is rather strong. In my unterstanding I would need someting around 1-2k to pull those down again, which would waste a lot of current. \$\endgroup\$ – Arne Jun 18 '13 at 6:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.