Im having trouble understanding how or if it even is possible to regulate a voltage with a MOSFET. I want to regulate an output voltage of 0-24VDC 0.5A(continuosly) with the duty cycle of a 0-5V 1kHz PWM signal from a PCA9685 LED driver.

My initial solution was using a low pass filter then run the average voltage through an op-amp. But I was told that this was not a good solution. Because I want 16 analog DC outputs, using op-amps will be both costly and I will need a cooling solution. The 24VDC source is from a truck. The output should also have short-circuit protection, and shouldnt allow more than 0.5A of current.

Is this possible with a MOSFET? Will i still need cooling solutions for this? And what would such a circuit look like?


  • 1
    \$\begingroup\$ 16 analog DC outouts? Where did they come from? \$\endgroup\$ – JIm Dearden Apr 22 '16 at 11:41
  • \$\begingroup\$ The PCA9685 has 16 channels. I can change the duty cycle of each channel with that driver \$\endgroup\$ – Andy90 Apr 22 '16 at 11:48
  • \$\begingroup\$ How much current do you need on each of 16 output? What is the purpose of these outputs? \$\endgroup\$ – Master Apr 22 '16 at 11:56
  • \$\begingroup\$ Each output should handle 0.5A. It will drive e.g relays and bulbs. Maybe solenoids also. Just to be more clear, adjusting the duty cycle should let me get say 12V DC out. if thats what I need \$\endgroup\$ – Andy90 Apr 22 '16 at 12:07
  • \$\begingroup\$ If driving solenoids it's more efficient to drive them with a full-fat 24V PWM signal. \$\endgroup\$ – Andy aka Apr 22 '16 at 13:07

You could do something around this idea:

enter image description here

Then, if everything is sized appropriately (this example circuit needs tuning, don't take it for granted), it's very efficient and you don't need any cooling. But there are things to consider:

  • There is no feedback: the output voltage is poorly regulated and varies depending on the load. Because each load is not very high, it could be acceptable. Adding a feedback would require a controller per output and would be very expensive.
  • The PMOS gate (on the top) is driven with a capacitive coupling of the PWM signal. This makes the schematics very simple and avoids a level translator, but it requires that the PWM signal never stays at low level for long (there should be always a minimum duty cycle, and also note that the off state is when the PWM signal stays at high level). So you can never totally reach full-on.
  • But... Honestly, with this circuit and a 1KHz frequency, the inductor and capacitor are way too big, in size and in price. However, if you change the controller (e.g. use a general-purpose microcontroller with an appropriate number of PWM outputs) and can reach the ~50KHz range, then it becomes practical. Both the inductor and capacitor can be MUCH smaller. But, if you go too high in frequency, the switching losses in the MOS will start to be too high. Make some tries with LTSpice and see how it goes.
|improve this answer|||||

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.