I've been using LTSpice and Micro Cap 11 for circuit simulations for a short time, I'm still learning some things, mostly on simulation settings. My focus is photovoltaic cells, which are simple circuits (but it took me a while to simulate them).

My professor told be about ADS (Advanced Design System), a Keysight software that is (very) expensive and, by his words, the best to simulate circuits, yet better at simulating at high frequency. I knew ADS as a T-CAD software, which should model and simulate semiconductor devices, and not as a circuit simulator. It's quite hard to learn how to use this software, because I don't have ways to use it at home, and coming to university everytime just to do tutorials is quite... unpractical.

Ok, when I (and if I ever) start modeling semiconductor devices, ADS is a no brainer. But for circuit simulation, do you think that LTSpice and/or Micro Cap 11 are enough or I should really start messing with ADS? I choose these two because they're free (I'm from Brazil, so spending even US$60 on softwares is quite expensive) and have a good fame on electronics community, but I'm afraid that I'm losing something ignoring ADS for now.

Thanks in advance,

Henrique.

 

What is your experience level?

Simulators have their place, but you should understand the basics well before using one. I recently started learning LTSpice because many people post .asc files; and I wanted to find out what all of the fuss was about. What I learned is that LTSpice doesn't do things the way you might think it would. Without some effort on your part, you won't get a realistic simulation. If you understand the basics, you can sanity check the simulation. If you assume a "good" simulation run means the circuit will work, you're in for a rude surprise.

I liken depending on simulators too heavily to a child using a calculator for arithmetic before they've learned the basics. Take away the calculator and the child will be helpless...

 

LTSpice is not unlike a programming language, it can be used for almost anything, not just circuits. You can model many non-circuit physical systems. It is more like Excel or MatLab.

For example, here is a simple model of a solar cell showing how you find the MPPT:

 

Please, you both have different opinions, but let's keep the education around, okay?

For now, I'd only use simulators for passive circuits, like filters. The solar cell was something I looked up because it's something that I have some interest. Of course, in Electrical Engineering I'll see more complex circuits with active devices and the likes, but probable the next thing I'll see is diodes and transistors.

Neither of you even cited ADS, so I'll assume that LTSpice and Microcap are enough. Thanks.

 

I'm not familiar with ADS but it appears to be an advanced tool that will do RF and microwave simulation including X parameters and EM fields, which Spice can't do.
So for lower frequency and DC simulation LTspice and Microcap are fine.
For the more advanced simulation you may need to use ADS.

But as long as you understand the limitations of Spice (doesn't consider component tolerances or power and voltage ratings, for example) and always breadboard your circuits to verify the simulation before you build them, you'll be fine.
You can ignore the negativity of those who don't like or are not really familiar with simulators (or those newfangled horseless carriages).
I would never build a circuit without simulating it first. I have found many design errors that way and have very rarely found a circuit that does not work if it simulates properly.

 

Different simulators may or may not handle certain problem classes correctly. It all depends on the available models. MikeML from AAC, and Helmut Sennewald from the LTSpice Yahoo Group are two of the most talented innovators when it comes to creating new ways of doing things. I am constantly amazed at what they have each come up with. A special shoutout goes to Mike Englehardt, the maintainer of LTSpice for doing a bang-up job on making a stable and reliable piece of software over many years.
http://www.theamphour.com/196-an-interview-with-mike-englehardt-spice-simulator-synteresis/

I don't doubt that someone could come up with an RF transistor model that would model the s-parameters of that transistor from DC to Blue Light. It would not surprise me if someone had already done that. Other CAD/Simulations work in their area of expertise.

Free is good when you are learning or for personal projects. Spending money on the right tool is essential to avoid costly mistakes with other peoples money. I'll stick with free until I have a reason to do otherwise.

 

do you think that LTSpice and/or Micro Cap 11 are enough or I should really start messing with ADS?

As a user of Genesys, I would say that your Professor is mostly wrong, and likely clueless.

Circuit simulators like ADS/Genesys/LTSpice use a set of models to describe the behavior of a circuit, or even a non-circuit if the right models can be used to describe the same set of behaviors. The "simulation" part of the tools is largely well understood and a commodity. What distringuish one simulator from another is largely the models, as well as the focus: Genesys / ADS is RF focused. Proteus is more of a mixed signal tool. LTSpice is a generic tool.

So if you could provide the same ADS models to LTSpice, you could perform the same analysis.

But that's a minor point. The key with any simulator is to understand its limitations (aka what it cannot do). Simulators are useful only to the extent that you understand what it cannot do, and then figure out a way to mitigate it, in simulation or in a real circuit.

For most of what I do, LTSpice is a quick and dirty starting point, for analog stuff. I would use OrCAD for more analog analysis, or Proteus for mixed signal / digital stuff.

 

All simulators are only as good as what you put into them. good models in - good simulations out. bad models in - the best you'd get is polished bad out.
Mod edit: language