0

I would like to ask your opinion regarding NMOS test-bench as a varactor (voltage control capacitor). I am trying to find the properties of Vg,W for achieving a \$200fF\$ capacitor.

Here is my test-bench : enter image description here

Calculated \$C\$ using AC analysis, where

\$C = \frac{I}{(V*2*pi*f)}\$

\$I = NMOS-Gate\$ point

and

\$V = voltage\$ on gate net.

Is this correct?

question - we are looking for \$Cgs\$ here, so can I just take the value from Cadence "Results Browser"?

Aaron
  • 7,274
  • 16
  • 30
Daniel Sapir
  • 23
  • 3
  • Don't forget that you need to add DC beside the AC voltage, for a proper `.op` point and `.AC` analysis. See [this](https://electronics.stackexchange.com/a/498912/95619) for something similar. – a concerned citizen May 13 '20 at 20:36

1 Answers1

1

The following is the equivalent circuit model of a FET. Disconsidering the parasitic inductances, if the drain and source ends are short citcuited the equivalent gate capacitance is given by:

$$C_{G}=C_{GS}+C_{GD}$$

model

source

The gate capacitance can be calculated by connecting a gate resistance much bigger than \$R_G\$ and sweeping the AC gate voltage according to the following e.g.:

  1. Sweep AC gate voltage
  2. Find the frequency at which the amplitude drops by 3dB
  3. Workout the gate capacitance

In this case, the gate capacitance is given by:

$$C_{G(CALC)}=\dfrac{1}{2\pi\cdot23.4963kHz\cdot1k\Omega}\approx 6.77nF$$

It can be double checked by inspecting the model details:

$$C_{G(MODEL)}=5.2nF + 1.6nF=6.8nF$$

circuit_2

vtolentino
  • 3,529
  • 1
  • 7
  • 17
  • I did not understand few things : 1. you did not sweep the AC voltage, it is constant 1V according to your simulation, so where does Vac of the gate plays a part ? 2. why you took the 3dB point ? 3. according to your answer, it is much easier to take the model Cgs+Cgd capacitance as Cg other then calculate from simulation - is that right ? – Daniel Sapir May 14 '20 at 06:15
  • 1. It is actually being swept with the instruction `.ac dec 1000 100 1e6`. The 1V is the DC bias. 2. You are right provided that you have such information in a model and parasitics are not accounted for. Furthermore, this is a very simple model. However, in your case, if you dont have such information at hand, you have to extract them by either simulation or calculation, or a combination of both. – vtolentino May 14 '20 at 06:53
  • While the AC voltage is 1 V, what is swept is the AC Frequency – jp314 May 29 '22 at 23:56