Help me understand Pfets a bit more

Thread Starter

robotDR

Joined Mar 17, 2020
90
My goal is to make a battery kill switch for a product so that if the user leaves the battery plugged in for too long, the battery will disconnect until charge voltage appears.

My question I have run into is, how a pfet would really work for this because the gate has to be less than the Vds right?

So I guess I should have two pfets in series so the body diode doesn't conduct.

But lets say my battery is still connected and I need it to charge. That means there will be 11-12 volts from the battery, and maybe 12.6volts on the other side of the fets, from the charging circuit. Is my Vds -0.6? Does my gate have to be lower than -0.6 then?

Of course when you turn it on, the drain and source are basically the same voltage so the gate at 0v would cause conflict with my concern above because they stay conducting. I don't understand this enough.

I have found this chip but it is likely too big: https://www.digikey.com/en/products/detail/ixys-integrated-circuits-division/CPC1907B/3908816

But it would do what I want. I can constuct it myself though right? (minus the opto isolated gate enable). It's just two pfets in series. But what does the gate need if I'm charging or discharging a battery that is connected or disconnected? Thanks!
 

Ian0

Joined Aug 7, 2020
9,847
My question I have run into is, how a pfet would really work for this because the gate has to be less than the Vds right?
No - to switch it on, the gate must be more than 4V negative of the source. The drain voltage isn’t involved.

The body diode will conduct if the drain is more than 0.6V positive of the source.
 

dl324

Joined Mar 30, 2015
16,943
My question I have run into is, how a pfet would really work for this because the gate has to be less than the Vds right?
It would be helpful if you were more precise. There are a number of different types of P channel FETs.
1674750446528.png
0V, or a sufficiently low voltage, on the gate will connect the battery.
 

Thread Starter

robotDR

Joined Mar 17, 2020
90
Ok great, thank you for the confirmation. so the only point in putting two pfets in series is so there is no current through a body diode in either direction. I suppose current into the battery via the body diode would be fine however power dissipation would need to be considered. The charger would turn the fet on very quickly once there is any voltage/current on the circuit so maybe that would be ok.
 

dl324

Joined Mar 30, 2015
16,943
I suppose current into the battery via the body diode would be fine however power dissipation would need to be considered.
Not if you put the charger on the left side of the P channel MOSFET.
The charger would turn the fet
Use the correct terminology. FET is an acronym for Field Effect Transistor, hence it's in all caps. There are dozens of types of FETs. Kindly specify the one you're referring to.
 

Thread Starter

robotDR

Joined Mar 17, 2020
90
Not if you put the charger on the left side of the P channel MOSFET.
Use the correct terminology. FET is an acronym for Field Effect Transistor, hence it's in all caps. There are dozens of types of FETs. Kindly specify the one you're referring to.
Well I really don't want the battery to have a continuous discharge path. I could live with a charge path that I had to manage. p channel mosfet
 

crutschow

Joined Mar 14, 2008
34,470
Does anyone feel that the symbol inside of the CPC1907 was wrong?
The only thing I see is no bleed resistance for the gates.
Otherwise that's a standard way to connect two MOSFETs back-to-back to block voltage in both directions.

So what do you think is wrong?
 

ScottWang

Joined Aug 23, 2012
7,409
The only thing I see is no bleed resistance for the gates.
Otherwise that's a standard way to connect two MOSFETs back-to-back to block voltage in both directions.

So what do you think is wrong?
What I felt strange is that the LED didn't face to face with the Gate, you can check other similar components that I linked in the previous post.
 
Top