MOSFET Source follower

[crutschow]

So meaning it doesnt matter what type of configuration that my NMOS is in (Source grounded or Source follower config). All i need to see is Vgs ya? For example the output of buck converter needed is 6V, so meaning that wil be my Vs? and from there if Vg is somewhere around 14V~18V it can drive my nmos to on properly? because Vgs more than or equal to around 10V?

to off it on the other hand my Vgs needs to be 0 right? meaning, Vg can be 6V or less right?

Wrong. During the switching period the inductor voltage must be equal to the input voltage for the transistor to be fully on (think of the transistor as simulating a zero-ohm switch). Thus Vg must be 10V more than the input voltage (which is also the transistor output source voltage when the transistor is fully turned on).

To turn off, Vgs should be below the minimum threshold voltage for the transistor, not 6V, since in a switching buck regulator, the MOSFET switch output voltage goes to ground (or slightly below) due to inductor current flowing through the flyback diode to ground during the transistor off period.

 

[simayonis]

Ok, so from the datasheet, i found that Vgs(th) min 2V, max 4V...
sorry chow, i now understand the on time... but could you explain more about the off time? Vgs should be below minimum threshold voltage... but during that time i wouldnt know what Vs is right? I mean i wouldnt know what value is needed for my Vg to turn off the mosfet since Vs is unknown? i only know Vd?

 

[crutschow]

.. but could you explain more about the off time? Vgs should be below minimum threshold voltage... but during that time i wouldnt know what Vs is right? I mean i wouldnt know what value is needed for my Vg to turn off the mosfet since Vs is unknown? i only know Vd?

We do know what Vs is. Re-read my last post.

The transistor source will be one diode drop below ground (about -0.7V) due to the inductor current flowing through the flyback diode, D38.

 

[simayonis]

ok so can you let me explain it again just to make sure im on the right path...

so basically, during on time, in order for the mossfet to switch on properly Vd must equal Vs... and if thats the case my Vg should be around 26V?

and during off time Vs must be 0, so Vg should be around 10?

What are the drawbacks of not letting the gate to have enough voltage pulse? meaning right now instead of sending a pulse of +26V +10V, im sending a pulse of +16V (same as drain) and +0V as the PWM?

 

[RCinFLA]

The 15.4 volt gate voltage minus the 12 volt source voltage is the 3.4 vdc gate drive of the NMOS. The voltage drop on source will match what the current through the NMOS yields to develop the Vgs voltage. It is a simultaneous solution between Vgs and NMOS current.

Operating in linear mode the NMOS dissipation will be high. As you increase the output load current it will reach a point that the source voltage cannot drop any more to develop the needed Vgs to support the higher Ids so the output voltage will have to drop below regulation point.

 

[RCinFLA]

Ok, so from the datasheet, i found that Vgs(th) min 2V, max 4V...
sorry chow, i now understand the on time... but could you explain more about the off time? Vgs should be below minimum threshold voltage... but during that time i wouldnt know what Vs is right? I mean i wouldnt know what value is needed for my Vg to turn off the mosfet since Vs is unknown? i only know Vd?

That is gate threshold voltage. That is the point where MOSFET just starts to conduct. Not relavant to discussion here.

Look up Rds-ON versus gate voltage. That is drain to source resistance for a given Vgs. For switching you want low Rds-ON to minimize the power dissipated in MOSFET.

 

[simayonis]

ok Rds(on) Vgs is 10V on the datasheet.....

ok i am abit confused now. RC so basically, for a source follower configuration. I have to see Vgs for Rds(on) am i right? But when i started designing the buck converter. I would neeed to know the high pulse and low pulse needed to trigger the mosfet to on/off for proper operation of the buck converter. What im confused now is about what they said about the source of the mosfet. If the source of the NMOS is connected to ground i would understand beccos Vs = 0V. So i can easily calculate Vgs for proper operation. But for a source follower, i wouldnt know what my source voltage is you know what i mean? thanks

 

[MikeMl]

Yes, as I tried to tell you about 10 posts ago, the gate voltage needs to be 10V more positive than the instantaneous source voltage. That is what Vgs means! You can do this with a pulsed floating charge-pump type power supply, the negative terminal of which is tied to the source, and the positive terminal is tied to the gate. As you were told before, a much better way to do what you are trying to do is to use a PFet, where the gate can be switched to ground to turn it on.

 

[ronv]

ok Rds(on) Vgs is 10V on the datasheet.....

ok i am abit confused now. RC so basically, for a source follower configuration. I have to see Vgs for Rds(on) am i right? But when i started designing the buck converter. I would neeed to know the high pulse and low pulse needed to trigger the mosfet to on/off for proper operation of the buck converter. What im confused now is about what they said about the source of the mosfet. If the source of the NMOS is connected to ground i would understand beccos Vs = 0V. So i can easily calculate Vgs for proper operation. But for a source follower, i wouldnt know what my source voltage is you know what i mean? thanks

Here are a couple of pictures that may help.

 

[audioguru]

This is the third person who is posting this same circuit. Are they all in the same school class?
It should use a P-channel switching Mosfet, not an N-channel source-follower.

 

[crutschow]

and during off time Vs must be 0, so Vg should be around 10?

Why would you put 10V on the gate, if you want the transistor off? Why is so hard to understand that Vgs=10V turns the transistor on, and Vgs=0V turns it off?

 

[RCinFLA]

Last chance.

Disconnect the R3/Q3 16v supply and make it a separate 26 vdc supply. Rerun sim and look at source of MOSFET.

 

[CafeLogic]

Why do I feel like I've woken in 1995. This is the second thread I have seen insisting that mosfets must be driven to 10V. Anyway, TIs new Nexfet has 2 milliohm RDS ON at 3V. You can buy fets all day long that are characterized at 1.8V. I have a Core 2 Duo, its core voltage is 1.3V that means it has 500 million fets flipping on and off 3 billion times a second at 1.3v. It is mostly true though, that once you start pushing 100V devices you need a strong gate voltage. Although, if you want to pay a few bucks more you can get some of those gallium arsenide fets rated for 200V and they work at 3V - at a ridiculous 1.9Qg total charge at that.

 

[koolguy]

how to find MOSFET Enchanment or Depleted type & What the different in there working...............!

 

[crutschow]

Why do I feel like I've woken in 1995. This is the second thread I have seen insisting that mosfets must be driven to 10V. Anyway, TIs new Nexfet has 2 milliohm RDS ON at 3V. You can buy fets all day long that are characterized at 1.8V. I have a Core 2 Duo, its core voltage is 1.3V that means it has 500 million fets flipping on and off 3 billion times a second at 1.3v. It is mostly true though, that once you start pushing 100V devices you need a strong gate voltage. Although, if you want to pay a few bucks more you can get some of those gallium arsenide fets rated for 200V and they work at 3V - at a ridiculous 1.9Qg total charge at that.

You didn't awaken in 1995, you just woke in the wrong discussion.

The actual gate voltage required to completely turn on is not the focus of this thread. 10V was used because that's typical of many available power MOSFETS. But whether it's 3V or 10V, the point is, that with a source follower configuration, the gate voltage must be at least that amount of voltage above the drain voltage to completely turn the transistor on as a switch.

And the voltage that microscopic 1.3V logic transistors require to switch in a Core 2 Duo, and the voltage required to switch a quarter-inch square, high voltage, power MOSFET is rather comparing apples to oranges.

 

[RCinFLA]

Why do I feel like I've woken in 1995. This is the second thread I have seen insisting that mosfets must be driven to 10V. Anyway, TIs new Nexfet has 2 milliohm RDS ON at 3V. You can buy fets all day long that are characterized at 1.8V. I have a Core 2 Duo, its core voltage is 1.3V that means it has 500 million fets flipping on and off 3 billion times a second at 1.3v. It is mostly true though, that once you start pushing 100V devices you need a strong gate voltage. Although, if you want to pay a few bucks more you can get some of those gallium arsenide fets rated for 200V and they work at 3V - at a ridiculous 1.9Qg total charge at that.

And the gate of every one of the N ch MOSFET's is taken near full supply above source to switch them on. For high speed >90 nm geometry, they also have significant leakage with zero Vgs.

 

[koolguy]

Plz reply & give answer.......!

 

[CafeLogic]

OK, calm down. A depletion mode mosfet is like a normally open switch. It conducts until you apply a negative voltage to the gate and then it stops. They are rarely used because they are delicate and have bad frequency response. Enhancement mode is the typical mosfet that conducts when you apply positive bias to the gate. Both examples refer to n-type.

 

[koolguy]

PLZ tell some eg of it as code no. & why there are less no of p-type mosfet....!

 

[audioguru]

PLZ tell some eg of it as code no. & why there are less no of p-type mosfet....!

We speak English here. Which language are you speaking?

What is PLZ?
What is eg?
What is code no.?
What is less no?