Review of this power control circuit

[frankqt]

I have designed this circuit for controlling a battery operated system.

System can be turned on/off using a button (a touch button, not a mechanical on/off). Turn on is done using button and off is done using microcontroller. During on a button press would generate an interrupt for the controller and that is how controller tuns it off. (VM1, debounce in software)

The design objectives were to make it as easy and low voltage drop as possible. I like to run the battery down to 3.9V, maximum battery voltage is 6V. When USB is present power should switch to USB regardless of battery voltage.

I am a newbie to the Analog/Power Design and appreciate feedback.

The circuit below.


p.s. Switches are in there to simulate the events, button press, usb plug in and out, micro controller turn on or off etc. I can also provide a TINA simulation file if parties are interested.

Thanks

 

[BradtheRad]

You are on the right track.

Questions:

The mosfets are P-type enhancement mode. T3 and T4 turn on while the transistor T2 turns on. Is this correct?

What is direction of current flow through T4? Its orientation suggests from right to left.

Does T2 turn on when the microcontroller sends 3.3V through R5? Or does it turn off?

Does it ever happen (by intent or otherwise) that the mosfets to turn on in such as way that USB 5V charges the battery when low?

What is the voltmeter intended to measure?

 

[frankqt]

Please see my embedded comments.

You are on the right track.

Questions:

The mosfets are P-type enhancement mode. T3 and T4 turn on while the transistor T2 turns on. Is this correct?

[Frank] Yes. T2 can be turned on from uP or push button.

What is direction of current flow through T4? Its orientation suggests from right to left.

[Frank] Left to right, from battery or USB to load. I have both T3 and T4 there back to back to avoid any current coming from USB to battery if VBat is less than VUSB.

Does T2 turn on when the microcontroller sends 3.3V through R5? Or does it turn off?

[Frank] Pulling the GPIO high for the CPU will turn on the T2.

Does it ever happen (by intent or otherwise) that the mosfets to turn on in such as way that USB 5V charges the battery when low?

[Frank] I hope not. I simulated this and this is why I have 2 mosfets back to back. One of them will cut the flow in.

What is the voltmeter intended to measure?

[Frank] that is connected to an interrupt line at the uP. If user presses a button, uP will know about it and conclude it should shut down and pull the GPIO down on T2 and turn off T2 than lights out. (Except from USB but that is ok, in that case uP can face lights out.


Thanks for your support.

 

[BradtheRad]

Questions:

* It is not clear which device is supposed to control T2. Your on/off button (near R1) can turn on T1, and that turns on T2. However the controller (shown as V2) also turns T2 on-off. It appears as though the two can compete with each other. Or did you intend for one to override the other?

* I'm looking at T4, which appears to be oriented so that current goes into the source terminal. If the FDN306P is like other mosfets, then it has its terminals designated 'source' and 'drain'. The drain should get the more positive polarity of juice. It may contain an internal diode that permits current to flow the opposite way, though not under control by the gate. By putting T4 'back to back' with T3, you may find it is workable according to the simulation. However real mosfets may be damaged by reverse current.

 

[frankqt]

Questions:

* It is not clear which device is supposed to control T2. Your on/off button (near R1) can turn on T1, and that turns on T2. However the controller (shown as V2) also turns T2 on-off. It appears as though the two can compete with each other. Or did you intend for one to override the other?

* I'm looking at T4, which appears to be oriented so that current goes into the source terminal. If the FDN306P is like other mosfets, then it has its terminals designated 'source' and 'drain'. The drain should get the more positive polarity of juice. It may contain an internal diode that permits current to flow the opposite way, though not under control by the gate. By putting T4 'back to back' with T3, you may find it is workable according to the simulation. However real mosfets may be damaged by reverse current.

T2 is controlled by the uP, they can in theory compete but the reality is if user press button, this turns on T2, in return this tuns on the power, cpu boots and takes over the control of T2. I am not worried about that part.

You comment on T4 is concerning. The reason that it is there is simply when the Vbat is less than USB, I don't want current flowing to the battery. Which parameter I should look at in the datasheet to see if that will be damaged or not?

 

[BradtheRad]

You comment on T4 is concerning. The reason that it is there is simply when the Vbat is less than USB, I don't want current flowing to the battery. Which parameter I should look at in the datasheet to see if that will be damaged or not?

The diode is called the body diode, drain-to-source diode, or flyback diode. It's different from the zener diode often installed to protect the gate from static charge.

It's possible the drain-source diode can be put to use the way you propose. Link to discussion at another message board:



The link below give details about orienting mosfets to block undesired DC. It's probably better than my advice. My own knowledge is incomplete. (Specifically, what I said about the drain terminal needing to be more positive than the source. It applies to N-mosfets, and should be switched for P-mosfets because they carry 'current holes' rather than electrons.) The specs for mosfet have been getting more complicated over the years.

**broken link removed**

FDN306P datasheet is at:

**broken link removed**

If you scroll 1/3 down you'll see specs titled 'Drain–Source Diode Characteristics and Maximum Rating.'

It states .42 amps as the maximum continuous forward current. This is the reverse direction of the 'On-state Drain Current' (10A).

You may need to experiment with real devices to determine how your application works. It may or may not result in a fried mosfet. It's all part of the research-and-development process.