This is a circuit to calculate noise in cascode LNA. I need your help to calculate V2 as both Iin and Iout are zero.

Here is what I thought so far. Because both Iin and Iout are zero the current flowing Ls is also zero and therefore its voltage drop is zero.
The voltage V2 is the voltage drop in the two parallel current sources. As in the picture, it seems that two current sources are open => the voltage across each one is infinite.
I am wondering if that is correct.

 

T

The voltage V2 is the voltage drop in the two parallel current sources. As in the picture, it seems that two current sources are open => the voltage across each one is infinite.
I am wondering if that is correct.

Voltage across a current source is not infinite!
In this case V2 is not determinable, as there is no shunt admittance across the current source. If you assume a finite rds for the MOS device, you could find an expression for V2. Since rds has been neglected here, I guess for this case, you might need to consider the short-circuit output noise current instead of open circuit output voltage.

 

Thanks, in this case I assumed that rds is infinite. This is considered as a two-port network. I want to find the output voltage as both input and output are open. However, because V2 is infinite, it seems to me that it is meaningless.

 

Thanks, in this case I assumed that rds is infinite. This is considered as a two-port network. I want to find the output voltage as both input and output are open. However, because V2 is infinite, it seems to me that it is meaningless.

V2 is NOT infinite, it is indeterminate (which, technically, means that it could be infinite).

The two current sources are NOT open circuited -- one is connected to the other. KCL merely requires that I2=-gm*Vgs.

But I don't think you are going to get useful results unless you include r_ds into your model.

 

I am following the application note of this site: http://www.maximintegrated.com/app-notes/index.mvp/id/3169

As in the Eq.4 I need to compute Vn2(V2 in this picture) as both I1 and I2 are zero.

I have seen many calculation results for NF of cascode LNA and in their final expressions there are no rds at all.

For that reason, I decided not to include rds in my model.
Now, I think I need to include rds in my model. In the final result I need to set it to infinite and then maybe the final expression will not depend on rds.

 

I guess it makes more sense to consider short circuit output noise current as MOS primarily is a transconductor. So without including rds, you find the total short circuit output noise current and then determine the Noise Figure.