When does the NMOS form the channel? VGS > VTH or VGB > VTH? I see that textbook talks about VGS > VTH, channel formed, but the capacitor is formed between the gate and the body, so i think it should be VGB >VTH?

 

When a source connected to the substrate question makes no sense. Otherwise, the on threshold (Vs> Vb, Vt> Vto) increases with Vsb. The threshold is measured for the Vgs. The threshold voltage is determined by the current, and not by the presence of an inversion layer.

 

Can you explain more physically in the structure of the MOS?

 

If you have time try read this
http://www.ittc.ku.edu/~jstiles/312/handouts/Creating a Channel for Current Flow.pdf
http://www.ittc.ku.edu/~jstiles/312/handouts/Applying a Drain Voltage to an NMOS Device.pdf

 

I see that textbook talks about VGS > VTH, channel formed, but the capacitor is formed between the gate and the body, so i think it should be VGB >VTH?

In discrete (3 terminal) devices, bulk (or body) is internally connected to the source, so Vgs is essentially the same as Vgb.

 

If you have time try read this
http://www.ittc.ku.edu/~jstiles/312/handouts/Creating a Channel for Current Flow.pdf
http://www.ittc.ku.edu/~jstiles/312/handouts/Applying a Drain Voltage to an NMOS Device.pdf

From this hanout, source and bulk are tied together so Vgs and Vgb are the same. How about the case source and bulk are not connected together?

When a source connected to the substrate question makes no sense. Otherwise, the on threshold (Vs> Vb, Vt> Vto) increases with Vsb. The threshold is measured for the Vgs. The threshold voltage is determined by the current, and not by the presence of an inversion layer.

How about the state of mosfet, on or off? Is it determined by Vgs or Vgb?

 

See

 

Thank you. Your simulation proves that you are right. However, I am still confused because it seems Vgb is the voltage creates inversion layer and so it determines mosfet threshold voltage.

 

Bordodynov, could you send me the model and simulation file?

 

anhnha,
I am come file a little later (the File on my home computer). Much years back beside me to appeared such question. I came to output that channel there is, but he does not contact with source of the transistor (because of impoverished area of the spatial charge)

 

I am sending the file. I found it.

 

Thank you. I tried that file and it is OK. However, I just test the circuit below. The source terminal is left open and bulk and source are tied together. Then I plot Vs when Vg sweeps from 0 to 5V.
As you can see, Vs is equal to 5V for all values of Vg.
Does this mean that Vgb is what helps form channel and the channel conducts now?

 

Yes, LTspice all calculated correctly. Even when the channel is not present, there is a leak p-n junction diode (i_leak_D-B). This tightens the substrate to leak Drain. Load the source resistor 1 Meg, and you will see the difference.