Let me pre-empt this with saying I am a self-taught hack with a very old (and nearly useless today) computer science degree. I have never designed with a P Channel Transistor before. My project is a PIC based 60V battery charger (from 12V Solar) for a 48V LiFePO4 battery pack. (Please don't question about Li battery charging - THAT part I am clear about and a BMS is in place.)

So, in order to take readings of the battery pack voltage on a Pic A/D pin, I needed to disconnect the 60V supply. Thus, the use of a P channel MosFET.

Here is that part of the circuit.
Anything majorly wrong?



Thanks everyone.

 

What pmosfet are you going to use?

Edit: the reason I ask is you have to watch out for Vgs exceeding the max specification. If so, you'll need a voltage divider on the gate (rather than pulling it all the way down to 0V).

 

What is the PIC PWM circuit that can switch 60V?

As joeyd999 stated you need to limit the Vgs voltage (20V is a typical MOSFET maximum, so you should limit it to 10-15V)

 

What pmosfet are you going to use?

Edit: the reason I ask is you have to watch out for Vgs exceeding the max specification. If so, you'll need a voltage divider on the gate (rather than pulling it all the way down to 0V).

I haven't chosen one yet. Recommendations are welcome. Price isn't really a concern because it's just a personal project and these parts are pretty cheap compared to a more difficult design.

 

What is the PIC PWM circuit that can switch 60V?

As joeyd999 stated you need to limit the Vgs voltage (20V is a typical MOSFET maximum, so you should limit it to 10-15V)

The more I look, the more I think using an N channel to do the logic switching makes sense. Like this:

Make sense?

 

The more I look, the more I think using an N channel to do the logic switching makes sense. Like this:

Make sense?

And that is exactly the way I would do it...

 

Oh good. I'm on the right track! Like I said, I'm a hack, so I only partly know what I'm doing, but not completely ignorant.

Thanks a lot.

Chris

 

With the zener added you don't necessarily need the gate-source resistor.

 

So I just need to choose resistors for the voltage divider that will be the correct ratio and end up under the current limits of the PIC pin's sink, yes?

 

With the zener added you don't necessarily need the gate-source resistor.

Thanks. That makes sense, since the divider on the high side should limit current. Hey, maybe I'm getting the hang of this thing, huh?

 

So, I was thinking of using an IRF9540N (datasheet here: http://www.taydaelectronics.com/datasheets/A-083.pdf) for the P channel.

Max Vgs is 20V. Does this voltage affect how much it is "on," as in its saturation? In other words, as a switch, do I need to make that voltage as high as possible without going over, or is it fully on when it's at 10 or 12V or so? I've never been quite clear on that.

 

With the zener added you don't necessarily need the gate-source resistor.

It doesn't hurt. And it adds just a bit of protection in case of either the network becoming disconnected or an over-voltage situation.

Edit: More clear: either use the resistor or zener or both. If you use both, it just adds a little protection if either fails or becomes disconnected.

 

It doesn't hurt. And it adds just a bit of protection in case of either the network becoming disconnected or an over-voltage situation.

Edit: More clear: either use the resistor or zener or both. If you use both, it just adds a little protection if either fails or becomes disconnected.

Does it play in to the voltage divider equation, being on the other side?

And thanks for your responses. I really appreciate the help.

Chris

 

With the zener added you don't necessarily need the gate-source resistor.

Without the G-S resistor, the PMOS will take forever to turn off once the zener ceases conducting. You definitely need the resistor.

With TTL input, you will need a logic level NMOS. An NPN is a good alternative.

 

I must be missing something here, usually am. Why not just put the load before the drain of a Nmos and be done with it? One gate resistor and one logic level Nmos is a lot simpler.

 

Without the G-S resistor, the PMOS will take forever to turn off once the zener ceases conducting. You definitely need the resistor.

With TTL input, you will need a logic level NMOS. An NPN is a good alternative.

Your are correct. Not putting much thought in my answers....busy with a rather intense math problem at the moment. Perhaps I should shut up till I'm done.

 

Without the G-S resistor, the PMOS will take forever to turn off once the zener ceases conducting. You definitely need the resistor.

.............................

You are quit correct. Didn't think my answer through.

 

So, I was thinking of using an IRF9540N (datasheet here: http://www.taydaelectronics.com/datasheets/A-083.pdf) for the P channel.

Max Vgs is 20V. Does this voltage affect how much it is "on," as in its saturation? In other words, as a switch, do I need to make that voltage as high as possible without going over, or is it fully on when it's at 10 or 12V or so? I've never been quite clear on that.

The datasheet will tell you the Vgs required for full turn-on. That voltage is shown in the specification row that lists the ON resistance. Any voltage higher than that will have little effect on the ON resistance.

 

OK, so the resistor will stay. Does it play into the voltage divider equation?
I have some logic level N Channel Mosfet transistors here so I will use one of those to operate the P-Channel.

The datasheet says "2-4V Gate Threshold Voltage." So, it doesn't need much to saturate.

Thanks everyone.

 

I must be missing something here, usually am. Why not just put the load before the drain of a Nmos and be done with it? One gate resistor and one logic level Nmos is a lot simpler.

I don't really understand what you mean. I couldn't figure out how a 5V logic signal could turn it on in the position it is located.