rfdesigner
Newbie level 6
Here's my experience:
Started graduate studies using switcap for switched capacitor simulations. It is text-based and not user-friendly. Moved to hspice and Cadence Edge (yes this dates me a bit) for simple circuits (e.g. op-amps). Did my Ph.D. "on-site" at a large semiconductor company - used Cadence with an in-house interface and hspice as the simulator. Was the only one working on RF simulations (in CMOS) which was a problem. Often did IP2 and IP3 simulations using long, long, long transient simulations and FFTs. Then I found spectre and PSS analysis, which was a blessing. Convergence was a problem for the first while, but now I know the tricks. Then I said to myself: "why run spectre on a 500MHz sparc when I can run ADS on a 1.5GHz Athlon?". Did a few comparisons and switched to ADS (still use Cadence Virtuoso for layout). The HB in ADS is wonderful, and the integrated data display is ok, although recently I have been using the Cadence calculator tools and appreciate it's "hidden" functionality through ocean scripting. Lately I've been playing with switched capacitor circuits using ADS (back to square one!) and HB. I've also used Pspice on a few projects - a useful tool for quick simulations (good transient simulations, decent waveform viewer, and now has the ability to continue transients after interruptions).
Important point: Do you have device models for the target simulator? Converting models is not trivial, and I generally hesitate to use converted models (after seeing a colleague burned badly by using incorrectly converted models on a taped-out design - ouch!).
<font size=-1>[ This Message was edited by: rfdesigner on 2002-04-14 21:55 ]</font>
Started graduate studies using switcap for switched capacitor simulations. It is text-based and not user-friendly. Moved to hspice and Cadence Edge (yes this dates me a bit) for simple circuits (e.g. op-amps). Did my Ph.D. "on-site" at a large semiconductor company - used Cadence with an in-house interface and hspice as the simulator. Was the only one working on RF simulations (in CMOS) which was a problem. Often did IP2 and IP3 simulations using long, long, long transient simulations and FFTs. Then I found spectre and PSS analysis, which was a blessing. Convergence was a problem for the first while, but now I know the tricks. Then I said to myself: "why run spectre on a 500MHz sparc when I can run ADS on a 1.5GHz Athlon?". Did a few comparisons and switched to ADS (still use Cadence Virtuoso for layout). The HB in ADS is wonderful, and the integrated data display is ok, although recently I have been using the Cadence calculator tools and appreciate it's "hidden" functionality through ocean scripting. Lately I've been playing with switched capacitor circuits using ADS (back to square one!) and HB. I've also used Pspice on a few projects - a useful tool for quick simulations (good transient simulations, decent waveform viewer, and now has the ability to continue transients after interruptions).
Important point: Do you have device models for the target simulator? Converting models is not trivial, and I generally hesitate to use converted models (after seeing a colleague burned badly by using incorrectly converted models on a taped-out design - ouch!).
<font size=-1>[ This Message was edited by: rfdesigner on 2002-04-14 21:55 ]</font>
