Load switch with FET - Is it really that simple?

[rprodrigues]

Hi guys,

I will use the load switch shown in the picture below to supply current to a 5V / 100mA load.


**broken link removed**


I made some simulations with PSPICE and the voltage drop across the MOSFET is about 50 mV at 100mA, what is fine for my requeriments.

Am I wrong or the schematic shown is really that simple and enough for my requeriments?

It seems too simple I can not believe!


Thank you

 

[smanches]

Better to put the FET after the load, not before it. To turn on the FET, you need the voltage at the gate to be +10V (+5V for logic level FETs) above the source.

Also, the cap from the gate to the source should be a resistor, not a cap. The value of the resistor will determine the FETs switching speed. 10K should be a good starting point, then modify from there.

 

[rprodrigues]

Better to put the FET after the load, not before it. To turn on the FET, you need the voltage at the gate to be +10V (+5V for logic level FETs) above the source.

Also, the cap from the gate to the source should be a resistor, not a cap. The value of the resistor will determine the FETs switching speed. 10K should be a good starting point, then modify from there.

smanches,

Thank you for your reply.

Since I'm going to interface a microcontroller to the LCD (load of the mosfet switch), I guess that putting the mosfet after the load could create a difference of potentials between the LCD ground and the Microcontroller ground. SO, I prefer to avoid it.

Note that the IRF5305 is a P-MOS, and not a N-MOS, so the CAP is from the gate to the drain.


Thank you very much!

 

[smanches]

Oh, sorry. Didn't look that close at the FET.

In that case forget everything I said.

 

[Roff]

Simulation will give you typical performance. Rds(on) is only guaranteed with Vgs=-10V. Personally, I would use a logic level PMOS, which will have guaranteed Rds(on) at -5V.

 

[rprodrigues]

Simulation will give you typical performance. Rds(on) is only guaranteed with Vgs=-10V. Personally, I would use a logic level PMOS, which will have guaranteed Rds(on) at -5V.

Do you have any model in mind?

 

[Roff]

Do you have any model in mind?

No, but you can look at **broken link removed** and select one. Other mfrs, such as Fairchild, have similar charts.

 

[nickelflippr]

I'll stick my neck out and say, with such a low load value, why wouldn't work? The data sheet looks like -3A at -4.5Vgs. Pullup the micro output to the +5V or source side of the pfet, could go 47kΩ. Others have mentioned the gate drive resistor should be in the 10-100Ω range, (1 meg, really?).

 

[Roff]

I'll stick my neck out and say, with such a low load value, why wouldn't work? The data sheet looks like -3A at -4.5Vgs. Pullup the micro output to the +5V or source side of the pfet, could go 47kΩ. Others have mentioned the gate drive resistor should be in the 10-100Ω range, (1 meg, really?).

It probably will. My engineering background steers me toward guaranteed performance, not typical. As I said, this would be my personal choice. Yours may differ.

 

[rprodrigues]

No, but you can look at **broken link removed** and select one. Other mfrs, such as Fairchild, have similar charts.

Thank you!!!! I'm going to study some of the models!

I'll stick my neck out and say, with such a low load value, why wouldn't work? The data sheet looks like -3A at -4.5Vgs. Pullup the micro output to the +5V or source side of the pfet, could go 47kΩ. Others have mentioned the gate drive resistor should be in the 10-100Ω range, (1 meg, really?).

Good suggestion!
I will implement the real circuit so I can get a better feeling of what my happen!

Thank you!

 

[audioguru]

You might need to sort through a bucket full of Mosfets to find one that turns on enough when its gate to source voltage is only 5V. Maybe somebody bought all the sensitive ones before you were there then you will not find one that works.

So it is best to select a Mosfet that is guaranteed to turn on enough when its gate to source voltage is only 5V.

 

[rprodrigues]

You might need to sort through a bucket full of Mosfets to find one that turns on enough when its gate to source voltage is only 5V. Maybe somebody bought all the sensitive ones before you were there then you will not find one that works.

So it is best to select a Mosfet that is guaranteed to turn on enough when its gate to source voltage is only 5V.

No, but you can look at **broken link removed** and select one. Other mfrs, such as Fairchild, have similar charts.

Roof and audioguru,

Do you think this Product Folder - NDP6020P - -20V P-Channel Logic Level Enhancement Mode Field Effect Transistor could be enough ?

The https://www.electro-tech-online.com/custompdfs/2009/11/FDN302P.pdf seems to be an even better choice!


Thank you!

 

[nickelflippr]

It probably will. My engineering background steers me toward guaranteed performance, not typical. As I said, this would be my personal choice. Yours may differ.

My blurt was more from a hobby perspective, and tend to use what is at hand. Its a real challenge to find reasonably spec'd and priced pfets in single quantities.

 

[rprodrigues]

My blurt was more from a hobby perspective, and tend to use what is at hand. Its a real challenge to find reasonably spec'd and priced pfets in single quantities.

For sure!

That is why for every specs I found, the second step is to search for it at the websites of my local Farnell supplier.

 

[mneary]

You might need to sort through a bucket full of Mosfets to find one that turns on enough when its gate to source voltage is only 5V. Maybe somebody bought all the sensitive ones before you were there then you will not find one that works.

So it is best to select a Mosfet that is guaranteed to turn on enough when its gate to source voltage is only 5V.

In the dark ages before modern process control methods this may have happened often.

Today, this has to be rare. It costs too much money to select parts out of a bin looking for key parameters. More likely the supplier has stabilized his process so that he rarely produces parts that are very different than 'typical'. For production designs, maybe the cost of troubleshooting drives the designer towards worst case design.

For the hobbyist, it's not unreasonable to expect close to typical performance. Fewer than 0.1% of the parts should be near the extremes.

 

[audioguru]

Manufactures have ranges for their specs because they cannot make all of the parts perfect. Some are typical, some are better and some are not very good. They throw away only the worst ones (to Radio Shack) but still sell the not-so-good ones.

Each production run has a different yield. You don't know if they have typical ones, poor ones or very good ones.

All of my designs have been "worst case" and I have never had one not work perfectly and I have never had one fail. There have been tens of thousands made and used.

 

[RCinFLA]

If you can dig up an old mobile phone or even an old mobile phone battery safety board, you can find P-ch mosfet with about 0.05 ohms Rs-ON with 3v Vgs.

 

[rprodrigues]

If you can dig up an old mobile phone or even an old mobile phone battery safety board, you can find P-ch mosfet with about 0.05 ohms Rs-ON with 3v Vgs.

The 3v Vgs do you mean the threshold value of the Vgs ?

 

[Roff]

The 3v Vgs do you mean the threshold value of the Vgs ?

He means 3V applied between gate and source (actually, -3V).

 

[rprodrigues]

He means 3V applied between gate and source (actually, -3V).

Is this the maximum, the common or the threshold value that turns on the transistor?

Thank you Roff,