0
\$\begingroup\$

I need to design a circuit able to enabling and disabling a voltage source and its respective reference (GND). The power supply has a adjustable voltage output and can supply 12V, 5V, 3.3V and up to 5A at 12V. It will be a source for powering external circuits that can be any, as long as they are within the capabilities of the source.

To accomplish this task, I thought about using the following circuit:

enter image description here

Does anyone have any experience or observation about this type of circuit?

Corrections: The mosfets were mirrored

enter image description here

\$\endgroup\$
9
  • 4
    \$\begingroup\$ Why does it have to disconnect the GND? That's usually a bad idea. Have you tried a relay? \$\endgroup\$
    – Justme
    Mar 24, 2021 at 13:45
  • 2
    \$\begingroup\$ Be aware that switching power or ground can cause unforeseen issues. Sneak paths were one of my biggest heartaches in my career. Be really careful that signal lines can't partially power up your switched circuit. \$\endgroup\$
    – Mattman944
    Mar 24, 2021 at 14:15
  • 2
    \$\begingroup\$ Your top MOSFET will never switch the supply off due to its internal body diode always conducting. \$\endgroup\$
    – brhans
    Mar 24, 2021 at 14:36
  • 1
    \$\begingroup\$ How much current are we talking about here? Might be easier just to use a relay. \$\endgroup\$
    – vir
    Mar 24, 2021 at 15:36
  • 1
    \$\begingroup\$ So you don't have to manage driving your MOSFETs with a range of voltages. \$\endgroup\$
    – vir
    Mar 24, 2021 at 16:27

3 Answers 3

2
\$\begingroup\$

Your circuit needs a small modification: connect the low-side n-FET gate directly to the control input, as follows (simulate it here):

enter image description here

Choose appropriate pass FETs, or even parallel multiples of them to get the desired current handling you seek. (Yes, it's ok to connect MOSFETs in parallel.)

Here's an improved version that counteracts the effect of the FET body diodes (simulate it here):

enter image description here

The back-to-back FETs prevent backdrive from the load.

\$\endgroup\$
2
  • \$\begingroup\$ I liked the idea of to connect the low-side n-FET gate directly to the control input. In the second circuit, the improved version, you said that the MOSFETs are in parallel, but they wouldn't be in series ? About backdrive from the load, I did not understand how the load can backdrive the power suplly \$\endgroup\$
    – Daniel
    Mar 25, 2021 at 0:58
  • 1
    \$\begingroup\$ You're confusing a few things. You can parallel FETs to increase current handling, but I don't show that in the diagram. It's just something to know about FETs. The second drawing uses back-to-back FETs to block reverse current, if needed. You may need that if there is a possibility of more than one power connection at a time to your target system (this is called 'backdrive'). You can combine the approaches (series and parallel) to get both protection and increased current. \$\endgroup\$ Mar 25, 2021 at 1:03
1
\$\begingroup\$

I believe that the circuit below can meet your need:

I didn’t carry out a sizing with so much attention. The components are just a suggestion, so you should check the electrical characteristics, and see if they meet your needs.

The OpAmp used is of the rail-to-rail type, so I can get the voltage levels to polarize the MOSFET Gate.

The selected MOSFETs can conduct a current of about 7A, and have a voltage VGS(th) of at most 1 V, therefore, these values end up being suitable since you want to use for voltages of 3.3V to 12V.

Below, I leave images of some simulations, working with 3.3V and 12V as supply voltages:

- 3.3V:

Output switched off Output switched off

Output switched on Output switched on

- 12V:

Output switched off Output switched off

Output switched on Output switched on

It is observed that the voltage drops are very small in the transistors. When the circuit is turned off, the impedance of the MOSFETs is so high that they end up being like voltage dividers, so that when turned off, the voltage over them ends up being half of VCC.

You just have to be careful only with the type of load you will connect to the output, because if not, there will be voltage in it anyway (MOSFETs do not isolate the circuit, for this, consider using a relay):

Observe the multimeter of the output, the VCC voltage is still there even with the output turned off, due to the impedance of the multimeter in relation to the impedance of the MOSFETs.

I believe that the circuit meets your needs.

\$\endgroup\$
2
  • \$\begingroup\$ Very, good! But, I didn't understand why you used OPAMPs to trigger the mosfets \$\endgroup\$
    – Daniel
    Mar 26, 2021 at 14:38
  • \$\begingroup\$ In fact, I used an OpAmp as a driver for the MOSFET, so that at least you would be isolating the ENABLE signal, which I don't know where it will come from, and sometimes the current levels at the moment of triggering the MOSFET can be harmful to some components such as microcontrollers. \$\endgroup\$
    – Cleisson
    Mar 27, 2021 at 22:33
0
\$\begingroup\$

There's a potentially flawed assumption here. Your enable/disable signal will more than likely come from a micro-controller (hence the pull-down resistor). But unfortunately you'll need the gate of device U15 to be pulled all the way to supply or ground to be conduction or in cutoff.

enter image description here

Sometimes the simple solution is the most elegant one. And this comes up on hardware interviews and it's something like this:

I have an MCU that outputs voltages from 0 to 5V but need to control an ADC that only takes logic levels from 0 to 10V, what can I do?

You can be as fancy as you'd like an make circuits up that behave like level-shifters (level translators). But why reinvent the wheel? the same circuit you have but instead of U15 and the resistors associated with them, you can just use a level shifter that takes inputs from your MCU and level shifts them to your PSU output.

enter image description here

\$\endgroup\$

Your Answer

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

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