Rdummy in spectre capacitor model

[wenadinho]

spectre capacitor model

Hi, I've been doing some simulations with passive high-pass filter configuration. With the process I use, I'm required to use MiM capacitor. Most of the results are the same with those simulated using ideal capacitor. However, I found a major discrepancy in noise result in which the input referred noise is much higher when using MiM capacitor.
The major contributions of noise come from Rdummy as described in the model.
Here is a partial copy of the Spectre model :

subckt mimcap ( top bottom )
parameters lt=15u
+ lb=lt+1u

rs ( top n1 ) resistor r=60 / (lt * 1.0e6)
cs ( n1 n2 ) capacitor c=(0.001 * ((lt * 1.0e6) ** 2)) * 1.0e-12
ls ( n2 bottom ) inductor l=(360 / (lt * 1.0e6) + 140.0) * 1.0e-12
cp ( bottom n3 ) capacitor c=(0.0036 * ((lb * 1.0e6) ** 2)) * 1.0e-15
rp ( n3 0 ) resistor r=(4.0e6) / ((lb * 1.0e6) ** 2)
rdummy1 ( top 0 ) resistor r=1e+12
rdummy2 ( bottom 0 ) resistor r=1e+12

As shown in the model, there are 2 Rdummies which seem to me that they represent a large resistance (1 TOhm) for DC path to ground. Am I correct, pls help me clarify ?
What I want to ask here is that is it safe to remove the noise coming from those Rdummies in the noise analysis result ?

Thanks

 

[sunking]

spectre resistor model

of course you can remove it.
by the way, it only effect the low freq noise figure.

 

[wenadinho]

e-6 e-12 capacitors

But I'm working on a very low frequency range, 100mHz - 1.2KHz, therefore I have to be sure that these noise won't come out after the chip has been fabricated. Anyway, any reason why would they put such dummy resistor ?
thx