Ltspice is a big dissapointment !

[Wolfgang]

Yeah, agreed. But these people are not monopolists, and EM software is quite costly. LtSpice was more or less a one man show (Mike) from a guy with a physics background, not from graphics design.
I must say that it is quite remar.kable what he achieved with a free tool that became the undisputed number one in a whole industry.

If you dont like it, dont complain, do something better.

[DimitriP]

Some side facts about LtSpice.

What you describe is complacency.
I never liked LTSpice. It's somehow simultaneously too simple and too complicated. The whole electronics simulator market feels like 1990s with bad UI,   incompatibilities , software written in archaic languages, bad scripting, bad output formats. Freeware that is operated as a walled garden.
Nobody cares how much time a simulation takes. What I care about is how much time completing a task takes, and that's a long time because the software needs serious UI improvements.

You either learn how to ride a bicycle, or you dont. You don't blame the "interface" or  gravity.

[PlainName]

You either learn how to ride a bicycle, or you dont. You don't blame the "interface" or  gravity.

Wrong analogy. It is more like the bicycle doesn't have a seat but just the bare end of the upright tube.

[shapirus]

Sorry but this is BS. Take Electro Magnetic simulators - they employ much more complex science but also have decent UI and plotting capabilities.

Well, we need to compare apples to apples. I have zero idea about the EM simulators you mentioned: are there any that are free (like LTSpice) and FOSS (like ngspice), developed pretty much by enthusiasts?

Otherwise, we can compare them with expensive commercial circuit simulators that sometimes have quite decent UI, easy learning curve and work well consistently enough. These apparently can afford (and see the necessity) to spend at least some resources to adapt the raw number crunching backend to the unprepared end user's expectations.

[Wolfgang]

You either learn how to ride a bicycle, or you dont. You don't blame the "interface" or  gravity.

Wrong analogy. It is more like the bicycle doesn't have a seat but just the bare end of the upright tube.

I have never seen a single case where such a bike became the undisputed industry standard. Maybe the comparison is flawed. 

[shapirus]

I have never seen a single case where such a bike became the undisputed industry standard. Maybe the comparison is flawed. 

Well, something becoming an industry standard doesn't automatically mean that it's good. One example is imperial units.

[DimitriP]

You either learn how to ride a bicycle, or you dont. You don't blame the "interface" or  gravity.

Wrong analogy. It is more like the bicycle doesn't have a seat but just the bare end of the upright tube.

The analogy works no matter what.
As long as there are people capable of riding it.....you either learn how to ride it , or you dont.

[tszaboo]

Some side facts about LtSpice.

What you describe is complacency.
I never liked LTSpice. It's somehow simultaneously too simple and too complicated. The whole electronics simulator market feels like 1990s with bad UI, incompatibilities, software written in archaic languages, bad scripting, bad output formats. Freeware that is operated as a walled garden.
Nobody cares how much time a simulation takes. What I care about is how much time completing a task takes, and that's a long time because the software needs serious UI improvements.

Yup. They not feel the need to change a lot for casual users. They are undisputed number one, so ... Maybe YOU can write a better LtSpice, they wont. Much luck, BTW.

This is called https://en.wikipedia.org/wiki/Appeal_to_accomplishment fallacy.
In case you are wondering, fallacies are bad way of arguing.

[Wolfgang]

Some side facts about LtSpice.

What you describe is complacency.
I never liked LTSpice. It's somehow simultaneously too simple and too complicated. The whole electronics simulator market feels like 1990s with bad UI, incompatibilities, software written in archaic languages, bad scripting, bad output formats. Freeware that is operated as a walled garden.
Nobody cares how much time a simulation takes. What I care about is how much time completing a task takes, and that's a long time because the software needs serious UI improvements.

Yup. They not feel the need to change a lot for casual users. They are undisputed number one, so ... Maybe YOU can write a better LtSpice, they wont. Much luck, BTW.

This is called https://en.wikipedia.org/wiki/Appeal_to_accomplishment fallacy.
In case you are wondering, fallacies are bad way of arguing.

Its not my program, I am just diagnosing what is going on. Wisdom is to change what needs to be changed, accept what cannot be changed and having the good judgement to keep the alternatives apart. In case of LtSpice ...

[DimitriP]

In case of LtSpice ...

Today it's LTSPICE, tommorow HPAK/Siglent/Tek/Rigol.Unit-T  will put some button or menu choice in the "wrong" place, the probe is going to be too long, too short, too expensive, the wrong color or have the wrong cable length...every so often we become entitled to every TD&H opinion  of how someone else should have gotten something done and the correct speed at which the Earth should really be orbiting the Sun.

It's a never ending story. There is even a song to go with it. Dunno if there is a T-Shirt.
  

[mikerj]

Some side facts about LtSpice.

What you describe is complacency.
I never liked LTSpice. It's somehow simultaneously too simple and too complicated. The whole electronics simulator market feels like 1990s with bad UI, incompatibilities, software written in archaic languages, bad scripting, bad output formats. Freeware that is operated as a walled garden.
Nobody cares how much time a simulation takes. What I care about is how much time completing a task takes, and that's a long time because the software needs serious UI improvements.

Yup. They not feel the need to change a lot for casual users. They are undisputed number one, so ... Maybe YOU can write a better LtSpice, they wont. Much luck, BTW.

This is called https://en.wikipedia.org/wiki/Appeal_to_accomplishment fallacy.
In case you are wondering, fallacies are bad way of arguing.

Many, many thousands of people use LTSpice daily without a problem. If you are unable to use it despite a vast wealth of information and libraries freely available, perhaps this says more about you than the application?  You could always ask for a refund I guess 

[shapirus]

Many, many thousands of people use LTSpice daily without a problem.

Another many thousands, on the other hand, use it with problems: yes it gets the job done, but it is the same as using any other tool that's awkward and inconvenient.
The fact that it gets the job done and that one can get comfortable with it despite its, uh, pecularities, does not disprove its drawbacks.

Here's a correct bike analogy:

If you are

Don't forget that we're discussing LTSpice, not its users. Especially not a particular user. Personal arguments and attacks aren't necessary.

[tszaboo]

Many, many thousands of people use LTSpice daily without a problem. If you are unable to use it despite a vast wealth of information and libraries freely available, perhaps this says more about you than the application?  You could always ask for a refund I guess 

This one is called ad hominem. https://en.wikipedia.org/wiki/Ad_hominem In case you are wondering it's an even lower tier argument style.
Oh and also https://en.wikipedia.org/wiki/Argumentum_ad_populum

[CatalinaWOW]

There is a conundrum here.  Spice of any flavor is a tool to observe circuit performance.   Just as an oscilloscope is.  You should always use the best tool for the job.  Spice was developed to observe circuit that are hard to observe with other tools.  Integrated circuits where probing ranges from difficult to impossible. 

No model is ever perfect.  And it takes experience, both success and failure to learn when a model is good enough. 

So now you have a beginner, trying to learn the modeling program.  It would be stupid to model such a simple circuit if that were the end goal.  It is the wrong tool.  Just build it and measure it. But the end goal is to model those hard to measure things.  So all the foibles of both the software interface and modeling are being fought while comparing to an oscilloscope.

Hang in there.  Learn the foibles and you will be rewarded.  The process would be frustrating even with the zoomiest new interface.

[Someone]

The process would be frustrating even with the zoomiest new interface.

Having an interface consistent with the platforms UI guidance would be nice, or at least consistent in time. Go and use the OSX version of LTspice and then explain why it doesnt match either the OS style, or the windows LTspice controls.

When the "standard" OS shortcuts aren't used, yet not mapped to anything at all. That's some pathological design.

[Wolfgang]

I absolutely agree that the GUI of LtSpice is gook and nonstandard. The problem is that LtSpice had and has so little competition so that the GUI was never a priority.
Its free, but not open source so attaching a better GUI to the otherwise quite brilliant numerics core is no option, unfortunately.
So - you may complain, even curse, whatever, but all Spicers have to live with this for the foreseeable future, sorry.

PS: To my knowledge Mike Engelhard is now writing a new Spice at Quorvo. You dont hear much of it, though, and downloads seem to be limited to the US.

[DimitriP]

........
PS: To my knowledge Mike Engelhard is now writing a new Spice at Quorvo. You dont hear much of it, though, and downloads seem to be limited to the US.

Oh good...that will delay further whining about how he did the interface wrong

[Wolfgang]

........
PS: To my knowledge Mike Engelhard is now writing a new Spice at Quorvo. You dont hear much of it, though, and downloads seem to be limited to the US.

Oh good...that will delay further whining about how he did the interface wrong

Yeah, unfortunately. We could send a full catalog of ancient greek and roman rhetorics to Lt (now AD) and see if that impresses them. 

[Gerhard_dk4xp]

PS: To my knowledge Mike Engelhard is now writing a new Spice at Quorvo. You dont hear much of it, though, and downloads seem to be limited to the US.

Hä??   I did download it to .de.

<    https://forum.qorvo.com/c/qspice/9    >  my shortcut to the forum.

regards, Gerhard

[Wolfgang]

Just tried again, but the registration mail supposed to be send by them never arrived.

[Gerhard_dk4xp]

I just did an update to state of $NOW and it works for me.

[Wolfgang]

Did not work again, maybe t-online mail is blacklisted by them. I wrote to Quorvo support and they sent me a direct download link, and this one worked.
Thanks anyway
  Wolfgang

[Nominal Animal]

I've thought about creating a Python 3 + Qt 5 interface for ngspice (open source).  That is possible only because ngspice provides a command line/file-based interface.  Many applications, including Altium Designer 19 and 20, and KiCad 7 and later, do support it already.

KiCad 7 and later have pretty good interface for ngspice, but it could be easier.  In particular, I miss the ease of creating new symbols I have in EasyEDA; and I'd prefer an easy to manage library format, preferably copying both the symbols and the NGSPICE/PSPICE/LTSPICE snippets, instead of file references, for ease of maintenance.  And being able to run more than one simulation (based on different schematics) at a time, of course.

Alas, my own user interface preferences seem weird to others, so doing that just for myself isn't worth it.  I can live with KiCad's oddities and occasional crashing; I'm just a hobbyist on the electronics side.  If I were to actually create such an application, I'd need to see and understand the actual workflows of a few different professionals and a couple of hobbyists, as I always use need that kind of information to construct proper user interfaces (except for examples and throwaway demos I've shown in other threads, of course).  Yet, professionals are hung up on LTspice, and don't seem interested in ngspice at all.

[Wolfgang]

Yup. Speed, numerical stability and accuracy as well as convergence are much more important for a pro than GUI issues. You can get along with a weird user interface,
but not with lack of results or false predictions.

[tooki]

The process would be frustrating even with the zoomiest new interface.

Having an interface consistent with the platforms UI guidance would be nice, or at least consistent in time. Go and use the OSX version of LTspice and then explain why it doesnt match either the OS style, or the windows LTspice controls.

When the "standard" OS shortcuts aren't used, yet not mapped to anything at all. That's some pathological design.

And in this case, we know for a fact why it is this way (with the Mac version being even less Mac-like than the Windows version!!): because the developer arrogantly believes that he knows the one “right” way software should be used, platform conventions be damned. He expressed contempt at users of the Windows version using the toolbar and menus, rather than right-click menus (his One True Way), so when he later made the Mac version, he didn’t provide those, to force people to use right-click menus. This is 180° contrary to the Mac way (the platform’s human interface guidelines expressly say that right-click menus are strictly a supplemental way to access commands*). As a former usability engineer this is a textbook example of a UI designed by engineer who has zero understanding of how normal users think, and basically treats them — and any usability experts who try to nudge them in the right direction — with contempt.

I’m hoping that, now that he’s left the company, LTspice will end up in the hands of developers who are user-centric.

*Yes, I am fully aware that Apple itself started ignoring many of its own guidelines, and I’m as irritated about this at least as much as when third party developers do it. (And I’m not alone in this sentiment.)