Spectre problem modelling ideal capacitor and current source

[Maz]

I am trying to model a very basic system however it is not performing
as I expected. I basically have an ideal current source (1 pA) feeding
into an ideal capacitor model (1uF) in Spectre. I am expecting a
linear relationship of voltage vs time based on integrating the
capacitor current equation:

Ic = c*(dVc/dt)

integrating to get

Vc = (1/c)*Ic*t = Kt, where K = (1/c)*Ic

However, I am finding that when I attempt to model this in Cadence
using Spectre, the simulation saturates follows an exponential
relationship that saturates at the constant K value (1mV). I cannot
seem to explain why it does not follow a linear function. Does the
specre ideal capacitor model include resistance? If so, is there a way
to remove that?

Any help would be appreciated. Thank you.

 

[Andrew Beckett]

On 17 Jul 2006 08:34:57 -0700, "Maz" <maz.khorasani@gmail.com> wrote:

I am trying to model a very basic system however it is not performing
as I expected. I basically have an ideal current source (1 pA) feeding
into an ideal capacitor model (1uF) in Spectre. I am expecting a
linear relationship of voltage vs time based on integrating the
capacitor current equation:

Ic = c*(dVc/dt)

integrating to get

Vc = (1/c)*Ic*t = Kt, where K = (1/c)*Ic

However, I am finding that when I attempt to model this in Cadence
using Spectre, the simulation saturates follows an exponential
relationship that saturates at the constant K value (1mV). I cannot
seem to explain why it does not follow a linear function. Does the
specre ideal capacitor model include resistance? If so, is there a way
to remove that?

Any help would be appreciated. Thank you.

Sorry for the very late reply. I expect you're seeing the effects of the 1Tohm
gmin resistor added to aid convergence, since there would be no path
to ground during DC from the current source.

Regards,

Andrew.
--
Andrew Beckett
Principal European Technology Leader
Cadence Design Systems, UK.