I have an application where I need to drive a 24V load with PWM, and high side switching via a P Channel Mosfet.

Most mosfets appear to have a Vgs limit of +/-20V.

This feels like a well established problem, and I am trying not to re-invent the wheel.

I can see the direct solution of a pullup on the gate to 24V and an NPN BJT + Zenner to pull it down without violating the Vgs limit, but this doesn't seem like a particularly quick switching method (slow to switch off).

Is there a special gate driver chip that solves this standard seeming problem.

  • \$\begingroup\$ We use fast diodes (a 4148 maybe or a schottky) parallel to the gate stopper resistor for speeding up the switch-off of N-Ch. MOSFETs (low-side switching). I think the same approach can be applied here. \$\endgroup\$ Nov 23, 2017 at 6:55
  • \$\begingroup\$ Also consider using NMOS on the high side along with some kind of high-side driver IC. This is done fairly frequently, and may even be the standard solution you refer to. Depending on the details of what you are doing, you may end up with better performance this way, too, since PMOS still lags behind NMOS when it comes to low Rds. I do like the simplicity of PMOS when the voltage is lower and the gate doesn't require special protection. \$\endgroup\$
    – user57037
    Nov 23, 2017 at 7:29
  • \$\begingroup\$ If you use a BJT to turn on the PMOS gate, and don't let it saturate the way you normally would, it may be reasonably fast. What is your PWM frequency, and what is the PWM doing? \$\endgroup\$
    – user57037
    Nov 23, 2017 at 7:32
  • \$\begingroup\$ @mkeith Yep, this is definitely the kind of thing I was referring to, but dropped out of my searches due to filtering on drivers suitable for P-Channel. The 4440a5f from LT looks to be the kind of beast I was imagining. \$\endgroup\$
    – Hugoagogo
    Nov 23, 2017 at 9:25
  • 1
    \$\begingroup\$ @mkeith 's cct is a good one with minor mods. Scale R4:R1 = N so that Vgs_M1 is about 3/4 of Vgs max. eg if V_PWM = 5V then VR1 ~+ 0/4.3V. So make eg R4 = 3 x R1 so swing across R4 = 3 x 4V3 = 12V9. D1 strictly now not needed. | A push pull paif or emitter followers in place of Q1 (emitters common = out, bases common = in, collectors to proper supply, gives you a high current drive. R scaling as above can give correct voltages. | capacitor across R2 helps speed but here R2 = sc is OK as Q1 is a follower. \$\endgroup\$
    – Russell McMahon
    Nov 23, 2017 at 13:49

1 Answer 1


Here is the basic idea. I don't think this will switch super fast or anything, but it should be a bit faster than the most simplistic approach because the BJT will not be saturated, and so it should turn off a bit faster than a saturated switch. The current through the Zener will be around (VCC-0.6) / R1. So if the PWM is a 3.3V signal, it will be around 2.7 mA through the Zener. I have never done this and didn't even simulate it, but it seems like it could work.


simulate this circuit – Schematic created using CircuitLab

The base resistor does not have to be large because the emitter resistor limits the base current.

Best of luck.

  • \$\begingroup\$ It doesn't really solve the problem of the R4/Cgs time constant, I guess. And R4 should be smaller than what I have shown. But you can play with it in simulation. \$\endgroup\$
    – user57037
    Nov 23, 2017 at 7:55
  • 3
    \$\begingroup\$ here are some ideas on how to make switching faster \$\endgroup\$
    – bobflux
    Nov 23, 2017 at 8:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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