The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design Languages >> Verilog-AMS >> Switching models dynamically ?
https://designers-guide.org/forum/YaBB.pl?num=1185200550

Message started by AnalogAroma on Jul 23^rd, 2007, 7:22am

Title: Switching models dynamically ?
Post by AnalogAroma on Jul 23^rd, 2007, 7:22am

Hi Everybody,

I need to simulate a large sensor array using Cadence Spectre. One way to reduce the simulation time is to use Verilog-A model for the inactive sensor nodes.

Traditionally , I can use Cadence hierarchy editor to choose the right model for the inactive sesnor nodes. But the problem is that the status of a sensor node being inactive or active depends on the input signal. So if the position of the incident beam is random, then the right models of the sensor nodes can not be specified before-hand. And to change the views
of hundreds of sensor nodes manually in Cadence hierarchy editor for each simulation is really cumbersome.

So I am thinking about another trick. Each sensor node or amplifier will have two models --one is Verilog-A model and another is transistor level model. A wrapper model will
include the two models. Once the position of the incident beam is generated randomly from a PRBS source, the activation signals of the sensor nodes will be evaluated. During this evaluation, if any arbitrary sensor node seems to be inactive , then within the wrapper model of the corresponding sensor node, the Verilog-A model will be launched. Otherwise, the transistor-level model will be used.

Is this trick feasible ? I mean, will Cadence really allow me to switch models dynamically ?

With Regards,

Analog Aroma

Title: Re: Switching models dynamically ?
Post by Marq Kole on Jul 23^rd, 2007, 7:43am

Real dynamic switching will not be possible, as the circuit topology has to be fixed before a simulation can be run. However, if the simulation used to determine the activation is separate from the one used for actual simulation of the behavior you can do this. The trick is then to use the results of the first simulation to set module parameters for the second one.