[SOLVED] Threshold voltage of MOSFET

[ashR]

I was working on 32nm NMOS. However am finding it difficult to find the approximate threshold voltage of MOSFET. Can anyone suggest me a circuit to find the threshold voltage?

 

[crutschow]

Do you have a physical transistor that you want to measure, or just a simulation?

 

[ashR]

Simulation.
This is what i have tried :-
Connected the source of NMOS to ground. Drain to 1V. Using dc analysis varied voltage at the gate from 0 to 1v.I measured the current from drain to source .Ideally the conduction should have started after the threshold voltage, however in my case it begins quite early.
Another problem is if I reduce voltage at the drain from 1v to some value lesser say 0.5 or 0.4. I get a slight difference in charactersics.

I need a method that can give me an accurate value of threshold voltage.

 

[ThisIsNotSam]

I need a method that can give me an accurate value of threshold voltage.

transistors are not ideal switches. so there's that.

 

[ashR]

ok. But then is there a way to determine the approximate value of threshold Voltage for the device, considering that it is behaving differently for different values of Vds? What value of Vds should i consider?

 

[Audioguru]

The threshold voltage does not change much when the Vds is changed. Why do you care about the threshold voltage anyway since the Mosfet is barely conducting or is almost cutoff at the threshold voltage anyway? Usually a Mosfet is used at a current that is either much higher than the threshold current or at no current.

 

[Dominik Przyborowski]

The threshold voltage does not change much when the Vds is changed. Why do you care about the threshold voltage anyway since the Mosfet is barely conducting or is almost cutoff at the threshold voltage anyway? Usually a Mosfet is used at a current that is either much higher than the threshold current or at no current.

It is not true for decananometer CMOS processes. The DIBL (Drain Induced Barrier Lowering) effect results with tens and even hundred mV threshold voltage shift with Vds changed from 50mV to 1V.

@ashR: Usually the Vth is calculated through DC simulation as operation point measurement.
If You want to obtain Vth from transfer curve, there is a few methods (check the literature for mosfet threshold voltage extraction). One of those is to plot gm/id curve and use crosspoint for 2/3 of max value (some papers recommend 1/2) - but You have to check it in appropriate papers.
The most commonly people use Vds equal to 50mV for linear region (to achieve Vth without DIBL) and 1V or max VDD.

 

[CataM]

The MOSFET has current even below the threshold voltage which is called sub-threshold region I believe. Usually threshold voltage is the one that gives 250 uA drain current or sometimes is measuered to 1 mA. So it is up to you.