Draw the small signal equivalent circuit and make sure to account for all of the conductances and capacitances.  Note this analysis is only valid around the DC operating point of your circuit.  Note that the capacitance may be quite different at high versus low frequency due to the inversion capacitance for minority carriers.   I'd look in a graduate text like "Fundamentals of Modern VLSI Devices" for a discussion of all of the charge storage mechanisms.

If you don't know the capacitances, how did you figure out Cdb?  Cgd and Cds would have a significant effect as well.  Don't forget the Miller effect as well.

By having the small-signal equations you can then find the model in which to fit your data.  Using some form of optimization, you can then get the capacitance values.  The conductances in your model should come right out from the DC i-v characteristic or a low frequency AC analysis.  The remaining unknowns should be your capacitances.  Don't forget to include your bias circuit in the calculations.  By changing these bias resistors, you can even choose which capacitance would be dominant in your measurement.

This is much like what your trying to do with your v=i/(sC) equation, but you would be accounting for more small-signal elements.

If you use a digital spectrum analyzer with a random noise source, you would see the frequency response all at once.

You could also  switch the device on and off and look at the resultant output waveform in the time domain.  Looking at the time delay, you may then figure out the switching Resistance and the output capacitance.

Hi all,

in my opionion the fastest way to get the numerical value of the total drain capacitance is scroll at the DC operationg point of the device looking for Cdd.
Cdd is defined as: Cdd=d(Qdrain)/d(Vdrain): in other words it is the variation of the overall drain charge with respect to the drain(-to-ground) voltage. It should include all the capacitive effect from the drain to the ground!

Regards,
Cri

You should be able to put an ac magnitude on the voltage source you're using for the drain, and then measure the imaginary drain current. You are correct that the drain (and source and gate) resistances will affect the results.  The cdd reported by the simulator is a partial derivative with respect to the internal drain voltage, assuming the other voltages are held fixed.  So, not only is the variation in the external drain voltage different from the internal drain, the internal source voltage is not fixed if there is a source resistor.