spectre rf hb

Hi,

as far as I know, Spectre RF is aimed at IC design. It is indeed a simulator in Cadence. ADS is a complete environement which is more directed to system design and circuits with discrete components. A link can be done between ADS and Cadence with RFDE to use ADS simulator (HB, ...) in the Cadence environement.

That's it!
ghost

I use both tools, as standalone and with RFDE.
ADS is frequency basad simulator with a lot of stuffs to simulate complex signals
SpectreRF is the RF IC design standard and is a time based simulator.

In my opinion the two tools are complementary in some sense (and that's why Agilent and Cadence are working together into the RFDE integration).
In my experience (RF IC designs) I have seen that in general the accuracy of the simulators can be quite similar, but with different practical aspects:
-for small signal analysis (ac, noise) without frequency conversion and transients I prefer spectre, is really more and more user friendly.
-for linearity simulations on amplifiers (IIP1,IIP3) ADS is really superior in term of efficiency. This is also true if you're simulating circuits with largely spaced harmonics (such as direct conversion mixers, for example)
-for mixers with RF/IF ratio of about 10/15 the two simulators are comparable, so I prefer spectre (not for IIP3).
I hope it can help.
Bye
Mazz

Yes , Mazz's point is very correct , If the BLOCk is include inductor , The ADS is better than Spectre-RF ,But in the transient or time domain , The Spectre-RF is better than ADS , Becasue Spectre-RF is hspice like

thanks,

so, for RF IC design both ADS and Spectre RF are almost equal. Is it correct?

Cadence has very powerful tools for layout and verification

What about layout and verification in ADS?

ADS harmonic balance simulator is very powerfull compared to spectreRF HB I think...

In Transient and DC simulations, Cadence is better.

For amplifiers, mixers, partially oscillators, hard nonlinear circuits, etc...ADS is far away good.

There are some advantages in the noise and PN simualtions with ADS. Cadence has higher Noise treshold level than ADS, so it computes the noise with much error under a certain level.

Depends on application. It should be have the both...

what is the RFDE?

RFDE is Aglilent's RF Design Environment that puts ADS and Momentum inside of cadence Analog Design Environment.

ADS is a harmonic balance (HB) simulator that is good for small (<50 device) linear circuits like power amplifiers. SpectreRF is a time-domain simulator (shooting method) that is good for VERY LARGE (ie 1,000+ devices) and non-linear circuits. So use SpectreRF for large non-linear circuits (receivers, mod, demod, etc) and ADS for small linear circuits like PA.

Good paper on SpectreRF is http://home.pacbell.net/kundert/jssc99.pdf

Thanks

Who knows, does RFDE or ADS exist for Linux OS?

RFDE on linux is soon to be released, probably by the beginning of next year I think

ADS Linux is already released. But only for "remote simulations". I mean, there isn't any GUI but it is used through ADS for Windows if it's well configured that is very difficult. It should be a network between linux and windows and it's still in beta test. Next decade, i hope they will release a linux version with GUI.

Do you know when the ADS with combine with IC50
for Linux version

And what about others design tools for RF IC (chip) design?

Spectre RF is a powerful tool for time domain analysis and is generally used for CMOS anaolg ckt designs like PLL Oscillators, amplifier transient responses etc. ; it lacks a direct frequency domain analysis; In comparision ADS is a much larger package and can perform circuit & system level simulations in both frequency and time domain.; It's very useful for simulating Mixers, harmonic distortion in Amplifiers, Oscillator harmonics and a host of other circuits. moreover the GUI in ADS is a lot better than that of SpectreRF .
Still both the softwares are good in their own domains ; but i feel ADS being more powerful of the two certainly edges out SpectreRF in a broder perspective.

I have done some simulation with spectre, it is ok as long as no inductor involvoed. If there is inductor, the result really bad, and does not match the simulation result with spice.
Anyone know why?

i have experience in using both software for commercial product for Rf, below 20GHz, but above 2GHz.
my experience is that don't use spectre, it is totally useless, it is only toys.
i trust only hpads.

RFDE for Linux 2004A is just released and it may be installed under Cadence Design Environment. What it means? It means you can use Harmonic Balance Simulator Engine ( in my opinion it's much more powerful than PSS ) through Analog Artist.

It's very good news, really...
I have used them at my company and results were incredible...

For instance consider a circuit highly nonlinear and never converges with PSS or QPSS but using with HB, it converges in some minutes...

Nice collaboration...

As far as i know, the RFDE is almost the ADS for workstation under the cadance enviroment

I do not know SpectreRF, never used it but if I would compare Microwave Studio, Ansoft Designer with ADS to beat them in system level. ADS can perform time or s-param and frequency simulations I wonder why you guys have problem with it?