Dual supply load control using MOSFET

[Kian]

Hi all,

In my application, I want to drive a load using either 3.3V (from a LDO) or from a 5V USB supply. I also want to be able to enable or disable this 3.3V by using a PMOS with the gate connected to a micrcontroller. The circuit I have in mind is shown below.

If the 3.3V supply is connected, and the 5V supply is disconnected, the microcontroller will be able to enable or disable the 3.3V to flow into the load. Diode D1 will prevent 3.3V from flowing into the 5V supply. In another case, if the 5V supply is connected and the 3.3V is disconnected (microcontroller will be shut off too), 5V will flow directly into the load through the diode.

But what if both 3.3V and 5V is connected at the same time? It seems like this could cause a problem, especially if the PMOS is turn on, in that case, both 3.3V and 5V will flow into the load. How can I avoid this problem? Hope to get some advise here. Thanks in advance!

 

[Nigel Goodwin]

Stick another BAT20J between D1 and Q1, to stop the 5V feeding back through Q1 - this will also prevent the 3.3V feeding the load if the 5V is present. Not that the 3.3V would feed the load anyway, as the the 5V would feed the load AND feed back to the 3.3V (potentially damaging things).

 

[Kian]

Stick another BAT20J between D1 and Q1, to stop the 5V feeding back through Q1 - this will also prevent the 3.3V feeding the load if the 5V is present. Not that the 3.3V would feed the load anyway, as the the 5V would feed the load AND feed back to the 3.3V (potentially damaging things).

Thanks Nigel. Is this what you suggest? Have another diode D2 as in the image below. Would adding Diode D2 reduce the 3.3V voltage/current going into my load?

 

[Nigel Goodwin]

Thanks Nigel. Is this what you suggest? Have another diode D2 as in the image below. Would adding Diode D2 reduce the 3.3V voltage/current going into my load?

View attachment 142305

Yes, that's right - it will reduce the 3.3V by the forward voltage drop across the diode (hence the use of a schottky one).

 

[Diver300]

As Nigel says, the problem is the 5 V will flow into the 3.3 V line.

Look up LM5050-1. It is an IC designed to control a MOSFET for that application.

 

[crutschow]

what if both 3.3V and 5V is connected at the same time? It seems like this could cause a problem, especially if the PMOS is turn on, in that case, both 3.3V and 5V will flow into the load. How can I avoid this problem?

Yes, you cannot have both the 3.3V and the 5V on at the same time.

Here's a circuit that uses two P-MOSFETs in series back-to-back to block the 5V from feeding back through the transistors, without any diode drop when they are ON.
That works because MOSFETs conduct equally well in both directions when ON.

This configuration requires an additional transistor to control the MOSFETs, since the MOSFET gate-source voltage must be near zero to turn them off (which here is provided by R1 when the control transistor is OFF).

The connection of the Q1 transistor emitter to the +5V prevents the 3.3V from being turned on when the 5V is present.

LTspice sim of example circuit below:
The output (yellow trace) is 3.3V when the Ctl input (green trace) is high, and the 5V is not present.
Otherwise the output is 5V minus the D1 diode drop when the 5V is supplied (red trace), independent of the 3.3V Ctl input voltage.

The MOSFETs and diode can be the ones you chose, of course.

 

[Kian]

Yes, you cannot have both the 3.3V and the 5V on at the same time.

Here's a circuit that uses two P-MOSFETs in series back-to-back to block the 5V from feeding back through the transistors, without any diode drop when they are ON.
That works because MOSFETs conduct equally well in both directions when ON.

This configuration requires an additional transistor to control the MOSFETs, since the MOSFET gate-source voltage must be near zero to turn them off (which here is provided by R1 when the control transistor is OFF).

The connection of the Q1 transistor emitter to the +5V prevents the 3.3V from being turned on when the 5V is present.

LTspice sim of example circuit below:
The output (yellow trace) is 3.3V when the Ctl input (green trace) is high, and the 5V is not present.
Otherwise the output is 5V minus the D1 diode drop when the 5V is supplied (red trace), independent of the 3.3V Ctl input voltage.

The MOSFETs and diode can be the ones you chose, of course.

Thanks crutschow! This should work for me! Is there anything to take note of when choosing Q1? Or I can just use a 2N3904?

 

[Nigel Goodwin]

You might perhaps have a read of this:

https://www.ti.com/lit/pdf/slyy168

Which shows a number of options which might help you.

However, EXACTLY what are you trying to do? (not how you think you should be doing it), there 'may' be a simpler way?.

 

[crutschow]

Is there anything to take note of when choosing Q1? Or I can just use a 2N3904?

It can be just about any small NPN BJT (there are about 20 with models in LTspice that should work).
I used a 2N3904, since it is a common small transistor and only about $0.35 US each in a TO-92 leaded package.