Circuit Simulation (Sticky?)

[tautech]

Unfortunately a subject I know little about. 

There appears to not be a sticky on this subject, shouldn't this be addressed ?

Where might be best for any sticky to reside, Beginners ? Projects, Designs, and Technical Stuff?

Coming from a time when one could breadboard circuits and tweak  component values without to much difficulty but today with most new designs and SMD componentry and now traditional breadboarding is somewhat trickier needing SMD adapter daughterboards and so on. Dead bug prototyping has to some degree offered a breadboarding alternative however I see circuit simulation as the way of the future. 

We see many forum users using some of the circuit simulation software and I wonder if this thread can provide links to them and even provide guidance to simulation newbies in it's correct use.


I now hand over the baton to those more knowledgable than I. ........ 

[T3sl4co1l]

Alas, simulation is almost as useless (or futile) as breadboarding, nowadays.

Manufacturers that do provide models for their ICs, typically lock them within PSPICE or HSPICE or LTSpice encrypted blocks, and when they aren't doing that, they're putting custom functions in them.

Most simulators today work with e.g. PSPICE behavioral sources,

Code: [Select]

Ename node1 node2 TABLE {V(in1, in2)}((x1,y1)(x2,y2)...)
(or instead of TABLE, VALUE and an expression), and other originally-proprietary PSPICE syntax.

But it doesn't help that most simulators have their own proprietary digital logic functions.  PSPICE has (I think) event-driven primitives, and an "analog" library built from SWITCH parts.  LTSpice has a general-purpose digital primitive with almost a dozen parameters (most of them tied 0/1 to create a standard gate or flop unit), which curiously, LT doesn't want you to know about (implementation hints used to be talked about on the forums, then they were censored..).  XSPICE is the original, but it's horrible; Altium uses SimCode; Multisim uses some proprietary event-driven code; they're all different.

Engineers have long since resigned themselves to build-and-test, so it doesn't much matter.

Meanwhile, AFAIK, simulations have continued advancing, but only in six-digit software for ASIC design.  There's so little interest in board-level sim, I haven't seen new tech (i.e., since SPICE3) offered in any free or board-level packages, anywhere.

Tim

[rs20]

I think simulation is much-maligned. The times I've seen demonstrations of simulation "failures" have actually been PEBKACs (most commonly the random selection of a transistor/part that is totally inappropriate for a task; the simulator demonstrates exactly why the circuit doesn't work with part precisely correctly; and then rather than debug the problem like with a real world circuit, the criticizer simply gives up and exclaims that the simulator is broken.)

Having said all that, simulation is indeed of limited utility to the hobbyist. I do think simulation has its place for playing with little control loops and gaining familiarity with concepts.

What information are you proposing would actually be kept in this sticky? Other stickies have a very clear purpose (e.g. list of currently available oscilloscopes).

[StillTrying]

I'd be happy with just an LT Help Thread somewhere, - I don't care what other simulators people are using.

[Ian.M]

Dave says his favourite SPICE simulator is  LTSPICE and has done a number of tutorials:

EEVblog #516 - LTSPICE Tutorial - DC Operating Point Analysis

Part 1 in a series of LTSPICE tutorial videos.
In this introduction Dave explains what LTSPICE is and how to do the simplest of the SPICE analysis options - basic DC operating point analysis.
(comment thread)

EEVblog #301 - LTspice Temperature Sweep Tutorial

Dave shows how to do temperature and parameter sweeping of your circuit using LTspice
(comment thread)

EEVblog #260 – Tracking Pre-Regulator Simulation in LTspice – PSU Part 13

Some tracking pre-regulator simulation fun in LTspice.
Part 13 of the PSU design series.
(comment thread)

There have been a *LOT* of 'which simulator' discussions on the forum. Doing a site specific Google search for best circuit simulator finds most of them.

The general consensus seems to be that  Analog Devices' LTSPICE generally comes out ahead, especially if you want to share simulations with others.   Its a totally free PRO-grade simulator with a large device library of  Linear Technology^* and 'jellybean' parts + a lot more 3rd party devices can be found on its Yahoo group.  Its UI is somewhat clunky, with a lot of stuff being driven by text SPICE dot commands, but its well worth the learning curve.   You can easily pay a significant amount of money for a competing commercial product that isn't as powerful or is node-limited, and the only benefit is often a glitzier GUI.

* Analog Devices have just acquired Linear Technology - deal closing early 2017 - and no one knows what's actually going to happen yet, so if you have *ANY* interest in simulation, grab LTSPICE and archive the installer, just in case.

[tautech]

What information are you proposing would actually be kept in this sticky? Other stickies have a very clear purpose (e.g. list of currently available oscilloscopes).

Links
Tutorials
Simulator usage advice/guidance
FAQ's

I'm just throwing this out there with little idea if it will fly or not, maybe partly of my ignorance on the topic but partly spurred by something I read in a post earlier today AND there being no existing sticky.
Checked Daves Wikki too but nothing there.

I understand there is no substitute for good understanding of core electronic principles and the maths involved however not all have these strengths, so is circuit simulation a valid alternative?

[Ian.M]

There should be sticky topics on several subject not already covered. However the current sticky topics fall into disrepair because people post conversations and miscellaneous posts that may be of some value in and of themselves, but over time they serve to make finding valuable info cumbersome.

Sticky topics, and I know I am repeating myself here, should be carefully moderated to organise the data to minimise repetition, filter obsolete or outdated info or links, and to remove conversational posts. (was that an Oxford comma?)

Maybe the current sticky topic situation is worthwhile but I think they are in need of some active maintenance by knowlegable members. For example for each sticky topic have a contributions topic where members can make suggestions and the appointed maintainer can edit the sticky posts to organise the changes.

Of course that needs volunteers to put in the effort. Maybe it won't be too much. Maybe it will become a giant bun-fight.

+1 
"Which is the 'best' SPICE simulator?" is an incredibly divisive question, possibly more so in a community of EEEs than politics in an election year!   It would be a nightmare trying to keep a single sticky thread relevant and on-topic.  One locked? list (only) of simulators used by forum members and a separate page or topic for each of them would probably be required to keep on top of it.

Probably the best option would be to reclaim the wasteland that is the EEVblog Electronics Resource Wiki.  If the Wiki could be locked to registered editors only, and the registration criteria to be: a forum member in good standing for at least one month, and >100 posts, it would go a lot easier as currently the vandalism level is quite high and we don't have enough active manpower to revert it in a timely fashion. 

[tautech]

Another developing thread:
https://www.eevblog.com/forum/beginners/confusing-choice-of-circuit-simulators-is-multisim-good/

[b_force]

I don't agree with saying that simulations are almost useless.
Simulating an entire circuit is completely useless.

The trick with simulating is to break things down in little bits and investigate certain behaviors.
From there on you can make assumptions which can be translated to the actual board.
In some cases it is good enough to find a starting point or to find out if a certain idea works.
It can also be used as an helping hand to pin point certain specs in components you're looking for.

[VEGETA]

Try PartSim: http://www.partsim.com/

totally free and online... also supports Digikey. Do you need anything bigger?

[R005T3r]

Simulations of complex circuits require a LOT of computing power and time... In some cases even more than video editing/making especially if you want everything in real time.
Also another aspect about simulation is that you have to program your circuit and then debug it. Not to mention that in the top of it you have to:
1. have a hell of a computer to have results in real time (realtime matters when you want to compare designs)
2. pay the software license on it (OS + software)
3. someone who projects the circuits
4. it's a simulation, in an ideal world. You also need to make your products in a real environment and test them before releasing them on the market

So overall it may give you an advantage in the product design ambient but then you need your lab to test your prototypes(because in the 90% of the cases things won't work as expected!). Another advantage of simulators is that complex calculus and circuit analysis lead to errors if made by humans, but if you draw and then let the calculus part to the pc, things change quite drastically: you won't have any error, but if you make a mistake by drawing it, well, you need to find it out and it's better than having a bugged design in addition with possible calculus errors...

Also, since you can't visualize heat or RF a simulation may be a good choice to have an idea on how to deal with them... But nothing avoids the fact you need to test the prototype.

[Ian.M]

I think you are loosing something in the translation.
Realtime simulation rarely matters.  All that is required is that a simulation run for the time period you are interested in is short enough not to significantly disrupt your workflow.
You may waste some time capturing/modifying the schematic for simulation, but it is often possible to export schematics to or import from your PCB layout package, avoiding duplication of effort. (E.g. LTSPICe <=> EagleCAD (ver>= 6.4, or an all in one package like Proteus)

[StillTrying]

Unfortunately a subject I know little about. 

Have you tried LTspice yet? We need to recruit more LT soldiers.
http://techsology.net/2016/09/08/ltspice-tutorial

[VEGETA]

For LTSpice lovers, if you ever want to design a chip, refrain from LTSpice. Its license file stated clearly that it explicitly prohibits the use by semiconductor companies.
For this reason, I use ADE at work, and NGSPICE at home. When designing boards without custom IC, I also use LTSpice.

it is the first time I know this info, they always said it is free. Do people really obey these rules?


I guess you wanna try this: http://www.partsim.com/

tell me if it is good or not, I tried it on a small circuit.

[Ian.M]

There are two applicable clauses in the LTspice licence

...
You are granted a non-exclusive, non-transferable, non-sublicenseable, royalty-free right to evaluate LTC products and also to perform general circuit simulation.
...
This program is specifically not licensed for use by semiconductor manufacturers in the design, promotion, demonstration or sale of their products. Specific permission must be obtained from Linear Technology for the use of LTspice for these applications.

That's not too arduous at all.   *IF* you are a semiconductor manufacturer, you'd need to contact Linear Technology for a licence and they'd probably ask you for a list of your products.  If you are a competitor, they'll tell you to 'go pound sand', but if there is synergy between your product range and the L.T. product range, and you are going to be recommending your customers use L.T. chips with your parts, they're likely to be more accommodating.

[tautech]

A couple more so to attempt to keep them all together and in one place...from this thread:
https://www.eevblog.com/forum/beginners/easy-to-learn-circuit-simulator/

http://everycircuit.com/
http://www.falstad.com/circuit/
https://www.systemvision.com/

Thanks to the members that offered them. 

[Zero999]

For LTSpice lovers, if you ever want to design a chip, refrain from LTSpice. Its license file stated clearly that it explicitly prohibits the use by semiconductor companies.
For this reason, I use ADE at work, and NGSPICE at home. When designing boards without custom IC, I also use LTSpice.

it is the first time I know this info, they always said it is free.

Please don't confuse free of charge with completely free to do whatever you like with it. Most proprietary software has restrictions which limit what you're allowed to do with it. Even some open source software has restrictions on what you can do with the source code, i.e. if you modify it, you have to make the source code public.

...
You are granted a non-exclusive, non-transferable, non-sublicenseable, royalty-free right to evaluate LTC products and also to perform general circuit simulation.
...
This program is specifically not licensed for use by semiconductor manufacturers in the design, promotion, demonstration or sale of their products. Specific permission must be obtained from Linear Technology for the use of LTspice for these applications.

Interesting wording. So you're not allowed to use LTSPICE if you're a semiconductor manufacturer, but what if you're fabless? You only design the chips and get them made externally. The licence doesn't seem to prohibit that!

[b_force]

Can people really only think in licenses/suing each other nowadays?  

If people wanna be SO picky, I would respond with, first prove that this program was used on a significant way.

[b_force]

If people wanna be SO picky, I would respond with, first prove that this program was used on a significant way.

It's more or less like being prudent. Spending time and effort on learning an unlicensable or unaffordable software may end up wasting time.
It reminds me how software market works in China. Simple, piracy everywhere, so not the most affordable, but the most powerful tools get learned and used.
No copyright holders will sue a student nor an individual -- they won't get a red cent, and they will get bitched over the Internet.
Then, copyright holders simply sue those companies who hires employees using pirate software, and get a big fat chunk of money.

Ehm I don't see a difference to be very honest???

These licenses are ALL about people with the biggest wallet.
Ones you start with disclaimers, you have to hunt them and sue people who 'infringe' them (whatever that means)
In my opinion, if you start a free tool, it's free. As simple as that.
Otherwise, you WILL end up in a big ship container of worms, where everything is grey (besides the fact it is EXTREMELY easy to bypass all security keys, even for professional companies)

[Zero999]

If people wanna be SO picky, I would respond with, first prove that this program was used on a significant way.

It's more or less like being prudent. Spending time and effort on learning an unlicensable or unaffordable software may end up wasting time.
It reminds me how software market works in China. Simple, piracy everywhere, so not the most affordable, but the most powerful tools get learned and used.
No copyright holders will sue a student nor an individual -- they won't get a red cent, and they will get bitched over the Internet.
Then, copyright holders simply sue those companies who hires employees using pirate software, and get a big fat chunk of money.

Ehm I don't see a difference to be very honest???

These licenses are ALL about people with the biggest wallet.
Ones you start with disclaimers, you have to hunt them and sue people who 'infringe' them (whatever that means)
In my opinion, if you start a free tool, it's free. As simple as that.
Otherwise, you WILL end up in a big ship container of worms, where everything is grey (besides the fact it is EXTREMELY easy to bypass all security keys, even for professional companies)

The moral of the story is, always read the small print. If a company is offering you their software for free and you want to use it in a commercial setting, then check the licence permits you to use it.

As far as wasting time is concerned, I wouldn't say that learning LTSpice is a waste of time, even if the licence doesn't let you do what you want, it's SPICE and the end of the day and you can transfer the knowledge gained to other simulators.

[Ian.M]

That makes sense.

As LTspice contains encrypted models of many of their chips, why should anyone expect them to hand a free tool to their competitors for the competitors to check their clone chips against?  If you think you may fall into the category of commercial semiconductor manufacturer, all you have to do is check with them before starting commercial product development or marketing to find out if you need to specially licence LTspice or whether as a competitor they will require you to go elsewhere for your SPICE simulation needs.   

[rstofer]

There is a very active LTspice forum on Groups.Yahoo.Com and it has a couple of true experts to provide guidance.  It's a great place to get started.  Yes, I know, Yahoo...

I use LTspice for little things.  Before I got Matlab, I used LTspice to model analog computing.  It's kind of fun to be able to plot 2d order differential equations.  I also use if for trivial things like demonstrating a capacitor charging.  RLC circuits are easily modeled.

Now that I have Matlab and a couple of add-ons, I'm not using LTspice as often.

[VEGETA]

There is a very active LTspice forum on Groups.Yahoo.Com and it has a couple of true experts to provide guidance.  It's a great place to get started.  Yes, I know, Yahoo...

I use LTspice for little things.  Before I got Matlab, I used LTspice to model analog computing.  It's kind of fun to be able to plot 2d order differential equations.  I also use if for trivial things like demonstrating a capacitor charging.  RLC circuits are easily modeled.

Now that I have Matlab and a couple of add-ons, I'm not using LTspice as often.

MATLAB is very expensive if you want Simulink and other stuff. However, is it really practical to simulate circuits? last time I used it was back at university when we simulate electrical drive systems.

[T3sl4co1l]

MATLAB, and the free alternative Octave (and I think there's another one nowadays too), are good for anything you do with matrices, vectors and differential equations.

AC steady state, and some transient analyses (preferably of linear systems) are easy enough to perform.

Tim

[VEGETA]

How can you simulate electronics circuits in Octave? I don't know if it is possible or not. As for Matlab, it is done in Simulink which is why we used it in our university labs.

[Rerouter]

http://www.falstad.com/circuit/

Another vote for this one, its my go to for go / no go, on bashing out bits of circuits,

Is it 100% accurate, almost certainly not,
does it let me work through for a solution in 15 minutes vs 5 hours of hunting down the correct models, very much so.

[T3sl4co1l]

How can you simulate electronics circuits in Octave? I don't know if it is possible or not.

Linear networks (i.e., RLC circuits) are just systems of equations, which MATLAB is made for.  Nodal analysis, it's EE101 stuff...

Tim

[G0HZU]

My background is mainly RF and if I go back to the start of my design career the thing I most wanted (at home) was not the exotic RF test gear we had. It was the RF simulation package(s) we used. I still value a decent simulator as the most powerful tool available to me in a typical RF design lab. It's a great tool to use for learning, for research, for troubleshooting and for processing real measured data from test gear.

[Nerull]

If people wanna be SO picky, I would respond with, first prove that this program was used on a significant way.

It's more or less like being prudent. Spending time and effort on learning an unlicensable or unaffordable software may end up wasting time.
It reminds me how software market works in China. Simple, piracy everywhere, so not the most affordable, but the most powerful tools get learned and used.
No copyright holders will sue a student nor an individual -- they won't get a red cent, and they will get bitched over the Internet.
Then, copyright holders simply sue those companies who hires employees using pirate software, and get a big fat chunk of money.

Ehm I don't see a difference to be very honest???

These licenses are ALL about people with the biggest wallet.
Ones you start with disclaimers, you have to hunt them and sue people who 'infringe' them (whatever that means)
In my opinion, if you start a free tool, it's free. As simple as that.
Otherwise, you WILL end up in a big ship container of worms, where everything is grey (besides the fact it is EXTREMELY easy to bypass all security keys, even for professional companies)

Are you willing to lose everything when you get caught? A lot of companies aren't.

There are companies who will pay your employees thousands of dollars to rat you out for using software you haven't paid for. Then you will be on the hook for a lot of money and a possible loss of your business. It's easy for an individual to say "Who cares what the license says", it's a lot harder for a company.

And if you're using ltspice as an individual, you almost certainly aren't designing semiconductors.