Assumption justification

Hi all.
In many papers I read, I noted that authors neglect the cds capacitance of their transistors and considers only cgs and cgd.
How to justify that ?
Furthemore they assume that cgs=(2/3)×W×L×Cox and cgd=(1/2)×W×L×Cox.
Is that related to a specific process technology or it can be also used to simplify life for other processes like 65nm and beyonds ?

How these figures are derived are based on a simplistic assumption of carrier distribution in the channel, and probably cannot be used for any practical use. The capacitance varies in the operating mode of the transistor, and a whole lot of other parameters.

All you have to know is that Cds is usually too small compared to the other capacitances and hence has the weakest effect, often ignored unless you are working on RF frequencies. Cgs is the dominant capacitance. Cgd in saturation is way smaller than Cgs due to pinch-off, and not 3/4 of Cgs as you have wrongly assumed. However, Cgd is important for transistors in an inverting configuration, where the Miller effect amplifies it by the gain, often such that it is more likely to form the dominant pole rather than transistor Cgs.

Thanks checkmate for your detailed explanation. Actually, by spectre simulation I noted that Cds is too small compared to Cgs and Cgd and I'm very curious to know why that is true depending on the operation's regions of the mos.

I found the asumption Cgd=3/4 Cgs in a paper talking about the design of a ring VCO in 65nm. However in other papers I found the assumption Cgd=1/2 Cgs. Do you find this assumption reallistic ?