Linear CAD software

Updated March 28, 2007

Click here to go to our main CAD page

New for September 2005! Below are some links to get free linear analysis software!

What do we mean by "linear" CAD software? This is simulation stuff that obeys "normal" transmission line theory with closed-form equations rather than finite-element analysis of Maxwell's equations, does not produce results that are have multiple frequency solutions (like mixers) or does not produce rsults that are functions of power (non-linear simulations). We'll also stay away from the topic of filter synthesis, and anything to do with SPICE... Granted, almost all CAD programs cross these boundaries today, but we will deal with these topcis separately to make the discussion more focused.

Here's an index to our imformation on linear CAD software so far:

Synthesis of microwave circuits

Analysis of microwave circuits

Optimization of microwave circuits (now a separate page!)

Physical versus ideal elements

Schematic capture versus netlist interface

Linear CAD software vendors, big and small

Free linear analysis softeware!

Synthesis of microwave circuits

Synthesis is the "art" of selecting a network that will provide a useful circuit response to a given design problem. For example, selecting an L-C network in a lumped-element filter. Most synthesis is done "by hand", meaning the designer looks for a similar circuit built by a previous engineer, and reverse-engineers his matching networks. There is one exception: there are a lot of software packages that do filter synthesis quite well (which we will one day cover on a separate page).

Analysis of microwave circuits

Analysis is where the software takes your network and all of the component values and crunches through a daunting amount of math to provide you with a description of the response of your network, for example, the calculating the S-parameters of an amplifier over frequency.

Other typical outputs that you can get from linear simulations include gain and return loss (in decibels or in magnitude), group delay, and various stability factors for amplifier circuits.

Optimization of microwave circuits

This has been moved to a separate page.

Netlist versus schematic capture interface

In software driven from a netlist, the user defines the nodes of the network and assigns them unique numbers. A lot of old farts (like the UE) like this interface, it works great so long as you are aware that mistakes can be made if you accidentally duplicate a node number.

Using a schematic capture interface, the user grabs and moves itty-bitty pictures of the overall circuit, such as capacitors and transistors. Most people prefer this interface. The errors that are most prevalent are due to mistakes due to misreading tiny fonts. Note to software vendors: no one's eyesight ever got better doing computer-aided design!

Ideal versus physical elements

Ideal transmission lines are analyzed as lossless structures of fixed characteristic impedance (no dispersion) with lengths measured in electrical degrees at one frequency. Ideal lumped elements have no real resistive loss, and no parasitic components that cause unwanted resonances. Physical elements are "more real", such as transmission lines realized in microstrip or stripline with finite conductor heights and losses associated with the skin effect and surface roughness.

Quite often a design will start out using ideal transmission lines and lumped elements, because the designer can better apply microwave theory this way. Then later when he/she improves the design fidelity using inductors with finite Q, capacitors with series resonant frequencies, and microstrip transmission lines with loss and dispersion, the design takes one step backward as the ideal response is degraded. Getting the design to provide nearly the ideal response by optimizing all of the physical elements is why you get the big bucks!

Linear analysis versus nonlinear analysis

Linear simulation implies that power and voltage levels have no effect on the outcome of an analysis. For filters and small-signal amplifiers like LNAs, this is good enough. For power amplifiers and frequency conversion devices such as mixers, you can't finish a design without nonlinear analysis.

One common method of non-linear analysis is harmonic balance. More to come!

Time domain versus frequency domain

Everyone knows that using a little advanced math you can describe a linear network response in either the frequency domain or the time domain. Some linear analysis software can convert back and forth, which is great if you are trying to analyze filter reponses to pulsed RF.

Linear CAD software vendors big and small

Today, it is almost impossible to split off the linear simulation tools from all of the other stuff, since all of the vendors want to sell you an entire "design suite". There is merit in having a design suite from one vendor, especially if the interface is seamless between each part of the package.

Here are the "Big Four" for linear simulation:

Applied Wave Research Microwave Office. AWR has done the unthinkable and provided a "socket" interface that allows the user to plug in the competitors' EM analysis software (i.e. Momentum), this shows a remarkable pair of cashews!

Agilent ADS (Advanced Design System). The URL for this software changes on a daily basis, so go to Agilent.com and start clicking stuff. Bet you can't find ADS info in ten clicks! Speaking of ADS, when we asked one of their staff why it runs slower than a one-legged dog, we were advised to run it on some type of super-computer with parallel processors. Too bad we can only afford a bottom end Dell, Dude! But Agilent has made some improvements, the 2003 ADS version seems to run fine on a bottom-end 2003 Compac computer with 600 MB of RAM. One of theses days we will devote a web page to suggections on how to improve ADS.

Eagleware Genesis (note: you have to register to use this web site, so in protest we don't offer a link, but you can figure it out we'd bet!) Eagleware is less money than its competitors, but it is just as accurate all the way through millimeterwave frequencies, it runs quickly and is easy to use.

Ansoft Designer rounds out the big four.

Microwave Software is a small business that provides some very low-cost linear analysis software that uses a netlist interface. President Jim Lev tell us the story of his small company:

"What a GREAT site (Microwaves101), but you've TOTALLY overlooked one of the (almost) "founding fathers" of RF/Microwave Circuit Design Software. A fellow (me) who "rubbed shoulders" with "GREATS" like Les Besser, Ulrich Rhode, George Vendelin, Wenzel the filter god, and more. I got Besser his first sale to put the original COMPACT on an HP-3000 at Hughes Aircraft, back when (almost) many of the cars in the parking lot had vacuum tube radios in them. I first used COMPACT on a TTY using an acoustic dial-up modem.

I was there (almost) before "the walls went up," and at a time when Eagleware was known as "Circuit Busters" (ho ho ho - who ya gonna call?) I was ham radio guy K6DGX way back then."

Jim Lev
President - Microwave Software
(949)492-9715
mwsoft@cox.net

Free analysis software!

Here we will list some free downloads, none of which we have tried out. If anyone can rate these tools for us, we have a free tool for you!

Vipec is free:

http://vipec.sourceforge.net/

Puff is a program that is available from CalTech:

http://www.its.caltech.edu/%7Emmic/puff.html

Ansoft has a "student version" of their software that you can download for free:

http://ansoft.com/products/hf/ansoft_designer/