Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

CMOS PA + Output Matching Netowork Simulation

Status
Not open for further replies.

dshoter13

Member level 4
Joined
Oct 20, 2014
Messages
72
Helped
4
Reputation
8
Reaction score
4
Trophy points
8
Activity points
619
CMOS PA + Output Matching Network Simulation

Greetings,

I am currently simulating a CMOS Power Amplifier. To complete the schematic level simulations, I need to capture all the effects of bondwire + external inductors parasitics. This way, I decided to mount the external MNT in ADS. However, I am facing difficulties whenever trying to simulate the MNT inside Virtuoso (both w/ S-parameters and spectre model generated by ADS). The simulations sometimes convergence, others not, mostly whenever I change the PA output level. Any advises on how to proceed with this verification, or other methodologies to perform the described simulation?

Thank you for your time.

Regards.
 

Re: CMOS PA + Output Matching Network Simulation

Hi,

Generally I used only RLC to mimic the passive devices like inductor or transmission line, so that the transient simulation is converged. Sometimes the broadband spice generated by ADS can still fail in PA or oscillator design in Spectre. You may fit the MNT S parameter by only RLC components if possible.In addition, better to check if the voltage swing of the PA has exceed the model's capability.
 

Is MNT a representing black-box to model package and other parasitic ( external and internal) component ?? If it's so,
Cadence Virtuoso Spectre simulator does not like floating nodes,very large parallel capacitances and inductances,weird loops and branches etc.Therefore you should try add Cmin option,connect huge valued resistances to floating nodes, use different curve fitting alghoritms for s-parameters etc. to make your circuit to converge.In fact there are so many things to try but since I don't know your circuit, I cannot say something concrete.Essentially s-parameters must be defined with reasonable small steps and they should cover the operating frequency that is interested in and beyond.For instance, if your bandwidth is between 2.4-2.5GHz, s-parameters have to be at least starting from 2GHz and 3GHz with let's say 5MHz steps otherwise while spectre is extrapolating for out of band frequencies, it may face to face convergence problems.There are many many such troubles in simulation environment but you have to learn and solve your specific problems so there is no unique solution.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top