Help Choosing a Simulator - LTspice, QSPICE or TINA-TI

[watchmaker]

I really am a beginner.

But I have tried to make Spice for TI, Tina, Microcap and Kicad work.  As well as LtSPICe and Multism 17.  I forget which one, but one of those damned programs makes wiring components together waay too hard.

I do not plan on doing PCBs, so my main interest is simulation to support my learning resources and breadboarding.

I use Multisim a bit but bear in mind I am just starting capcitors and opamps via the Real Analog vids.  Once I figured LtSpce was shortcut driven, I started using it too.  Have not yet decided.

Of all the ones I tried, MultiSim live is the most intuitive.  Just sat thru the webinar today and I had already earned most of what I needed. 

Yes, it is cloud based and subscription.  The free subscription is very limited.   I am trying to see if it makes Multisim 17 more understandable. I am not the one to evaluate how limited the other MS Live subscriptions are for serious work.  But I can see being tempted to go to the first tier as I progress.

Given that we have to pay for new versions after a couple years, it may not be a bad deal. 

I still have my PC based photo programs and libraries though.  And I store nothing on the cloud except the business files I share with customers.  I never believed that the keepers of the cloud would not poach what was uploaded.

[RoGeorge]

My favorite is LTspice.  Easier to use than any other simulators I've tried, and backward compatible with most other spice versions.  There are models online for almost anything.  Very popular, large community, easy to find online answers for any problem, plenty of tutorials, books, videos, etc.  Very well maintained by top professionals (Analog Devices) and often updated, rock-solid and no bugs, works with Windows, Linux (through WineHQ), Mac, easy to share schematics (attach the .asc text file), free for personal/home use with no performance limitations, no ads, no buy-me nagging, can be called from Python to automate schematic creation/run/results postprocessing, works entirely offline, doesn't expire and doesn't need any license or account.  Probably the best SPICE out there.  Highly recommended.

Sometimes I also use QucsStudio, because it has live sliders to adjust components values, like this:

though, QucsStudio is almost unmaintaned (one developer), and the GUI style is from QUCS, so more cumbersome, even more cumbersome than TINA-TI.

Whichever you'll choose, don't expect them to be intuitive, or eye candy.  That's valid for any CAD/EDA program, in general.  They all require some learning, then practice.  It's a steep learning curve, but once one is fluent with a simulator, that will be an invaluable skill, particularly for learning, and later for design.

[LinuxHata]

Can't say anything about another two, but LTSPICE is awful - these guys have no understanding of the term ergonomics, and software appears to be designed for IBM System/360 and to be used with light pen and keyboard - absolutely not mouse friendly. Not mentioning that resistors are in zigzag shape and many common components are missing right out the box (triacs, for example) and you have to search, download and install separately.

Also, it does not considers limit of the components - say, if using some DC-DC boost converter, which can only work at 5V input and 2A current, it will simulate it working at 100V input voltage and 100A output current (as example) and won't give any errors.

[magic]

Perfectly usable and effective with a mouse and keyboard, I don't get what you are talking about.

The latter complaint is against chip models, same model will simulate the same on other engines.
SPICE simulators in general don't account for limits of discrete semiconductors either.

[RoGeorge]

LTspice is a CAD/EDA tool, don't expect it to be as easy as a videogame.  Of course it has both types of resistors, rectangle for European style drawing of R, and zig-zag for American style (default, but can be changed).

LTspice is for productivity and heavy lifting, not for glitter and eye-candyness.  Requires one to learn how to use the tool first.  Requires one to make intensive use of 20-30 keyboard shortcuts at least, to make the drawing and handling easy.  Shortkeys are programmable, make them single press and easy to remember.  Avoid keyboard alterators (not CTRL+something), use single letter shortcut keys, it's easier to touch a key than to press a CTRL+something else.  For example I use 'M' for move part and 'D' to drag a part with its wired attached, etc.

If you think LTspice is cumbersome, then name a better simulator.

[tggzzz]

You don't understand several important points.

... and software appears to be designed for IBM System/360 and to be used with light pen and keyboard - absolutely not mouse friendly. ...

Correct, but it is not limited to IBM360s. I first used a Spice on a Honeywell DDP516 minicomputer. Spice, per se, has changed little since then.

Spice input and control statements are ASCII text. Spice output is tables of numbers, and graphs drawn with ASCII characters similar to this (but normally the graph would go down the pages, not across)

Not mentioning that resistors are in zigzag shape and

That's a standard. Ditto inductors being "curly". The nice thing about standards is that there are so many to choose from.

Besides, that is merely "syntactic sugar", equivalent to obsessing about what fonts are used.

many common components are missing right out the box (triacs, for example) and you have to search, download and install separately.

Have you come across the concept of a library being separate from a language and an execution engine? That has many many significant advantages.

Also, it does not considers limit of the components - say, if using some DC-DC boost converter, which can only work at 5V input and 2A current, it will simulate it working at 100V input voltage and 100A output current (as example) and won't give any errors.

Correct; it will also allow you to pass 1A through a 1Gohm resistor

That is nothing whatsoever to do with any Spice. That happens with every single Spice component: they only model reality, and omit significant behaviour. For example the Gunnel-Poon BJT model doesn't correctly model low currents.

You always need to assess what aspects of behaviour are important for your task, and then to validate how well the models represent reality.

[tggzzz]

LTspice is for productivity and heavy lifting, not for glitter and eye-candyness.

Precisely. Should be emphasised.

If you think LTspice is cumbersome, then name a better simulator.

Oh, it is cumbersome.

But underneath all the simulators are the same. Component models.... no so much!

[MathWizard]

A lot of times using LTS I'll be unfamiliar with the computer science behind the scenes, and most guides probably expect you do know the names of stuff and equations, and the correct computer syntax.

I run into the same thing when using Octave, the math program. I need some basic computer science guide just aimed at math.

[RoGeorge]

most guides probably expect you do know the names of stuff and equations, and the correct computer syntax.

Not at all.  It's only mouse drawing and clicking the right buttons to run a simulation.  Then, after the simulation ended its calculations (sometimes instant, other times can take a few seconds or minutes at most), click with the mouse inside the schematic and the waveform at that point will appear plotted in the plots window.

The graphic schematic is stored internally as a .asc text file.  Under the hood all is text files, and there is indeed a syntax for the schematic drawing, and for the netlist if you want to tinker with that manually, but it is not needed at all during normal use.

It's all graphical drawing, click run, then click inside the graphical drawing to see the waveforms.

[JJ_023]

My favorite is LTspice.  Easier to use than any other simulators I've tried, and backward compatible with most other spice versions.  There are models online for almost anything.  Very popular, large community, easy to find online answers for any problem, plenty of tutorials, books, videos, etc.  Very well maintained by top professionals (Analog Devices) and often updated, rock-solid and no bugs, works with Windows, Linux (through WineHQ), Mac, easy to share schematics (attach the .asc text file), free for personal/home use with no performance limitations, no ads, no buy-me nagging, can be called from Python to automate schematic creation/run/results postprocessing, works entirely offline, doesn't expire and doesn't need any license or account.  Probably the best SPICE out there.  Highly recommended.

Sometimes I also use QucsStudio, because it has live sliders to adjust components values, like this:

though, QucsStudio is almost unmaintaned (one developer), and the GUI style is from QUCS, so more cumbersome, even more cumbersome than TINA-TI.

Whichever you'll choose, don't expect them to be intuitive, or eye candy.  That's valid for any CAD/EDA program, in general.  They all require some learning, then practice.  It's a steep learning curve, but once one is fluent with a simulator, that will be an invaluable skill, particularly for learning, and later for design.

Thank you for the elaborate explanation of your viewpoint.  I was aware of most of the positive attributes of LT Spice.  That is the primary reason why I am considering it. 

QSPICE seemed to me as an evolutionary product as it was developed by the gentleman that developed LTSPICE which is the reason that I am considering it.  It is supposed to be much more robust but currently lacks the user base that LTSPICE has and with it the benefits of having a large user base.

I chose TINA because as you know I have a partial bias towards Texas Instruments and its components as well as the well-written support of literature for them as well as many other generalized topics.  I figured Texas Instruments probably puts out a decent product and would be preloaded with the models of the actual chips that I would consider experimenting with.

I know that simulators aren't easy to grasp.  I am fairly proficient in several CAD packages so I know what that process entails.  This is the reason why I am trying to make the proper decision so as not to waste time on a simulator that will be not the best suited for me.

The one point where I do disagree with you is that the UI is fairly important.  For example in your video it would be nice to have sliders on resistors and capacitors.  There actually is a really neat very basic simulator that has all of that in case you are interested.
It is fairly limited in its capabilities compared to the simulators discussed here.

https://www.falstad.com/circuit/circuitjs.html

[JJ_023]

Can't say anything about another two, but LTSPICE is awful - these guys have no understanding of the term ergonomics, and software appears to be designed for IBM System/360 and to be used with light pen and keyboard - absolutely not mouse friendly. Not mentioning that resistors are in zigzag shape and many common components are missing right out the box (triacs, for example) and you have to search, download and install separately.

Also, it does not considers limit of the components - say, if using some DC-DC boost converter, which can only work at 5V input and 2A current, it will simulate it working at 100V input voltage and 100A output current (as example) and won't give any errors.

I agree with your first Statement of the simulator being terrible in terms of user interface. It is one of the reasons I am trying to explore other alternatives.  However what I am finding out is that because of its prolific user base that it is hard to discount in terms of consideration.

[JJ_023]

You don't understand several important points.

... and software appears to be designed for IBM System/360 and to be used with light pen and keyboard - absolutely not mouse friendly. ...

Correct, but it is not limited to IBM360s. I first used a Spice on a Honeywell DDP516 minicomputer. Spice, per se, has changed little since then.

Spice input and control statements are ASCII text. Spice output is tables of numbers, and graphs drawn with ASCII characters similar to this (but normally the graph would go down the pages, not across)

Not mentioning that resistors are in zigzag shape and

That's a standard. Ditto inductors being "curly". The nice thing about standards is that there are so many to choose from.

Besides, that is merely "syntactic sugar", equivalent to obsessing about what fonts are used.

many common components are missing right out the box (triacs, for example) and you have to search, download and install separately.

Have you come across the concept of a library being separate from a language and an execution engine? That has many many significant advantages.

Also, it does not considers limit of the components - say, if using some DC-DC boost converter, which can only work at 5V input and 2A current, it will simulate it working at 100V input voltage and 100A output current (as example) and won't give any errors.

Correct; it will also allow you to pass 1A through a 1Gohm resistor

That is nothing whatsoever to do with any Spice. That happens with every single Spice component: they only model reality, and omit significant behaviour. For example the Gunnel-Poon BJT model doesn't correctly model low currents.

You always need to assess what aspects of behaviour are important for your task, and then to validate how well the models represent reality.

Have you personally played around with QSPICE?  I don't have the experience nor knowledge base to personally make the statement, but from what I have researched and learned about it is it is supposed to be a much more robust simulator.  It was developed by a gentleman who developed the original LT Spice.  There is a interview video with him that I watched where he touted  some of the benefits.  Unfortunately my knowledge base and experience prevents me from fully understanding his arguments.

[Terry Bites]

All spice lies!

As ever, you get what you pay for, if you dont pay don't expect the best of the best.
LTspice is not gold standard. Its good enough for most purposes though.
Pspice say they are gold standard but you'll have to pay to find out.
You may like to check this page. https://electronicsguruji.com/best-circuit-simulation-software/
Qspice is more modern and has a more rational interface, it has a deep and easily updated model library.
Spice from TI AD etc are also there to promote their products, so they're not entirely a free gift.
I still have a (fading) preference for MC or "microcap" because of its superior schematic input. Its old and out of date like me though.
Its hard to find but I recall there is a download link on eevblog somewhere

[RoGeorge]

For example in your video it would be nice to have sliders on resistors and capacitors

Indeed, and it has.  A slider can be attached in QucsStudio for any numerical value (including the values of Rs or Cs) by simply clicking the wanted component, or the wanted parameter, and a new slider will appear.

There is another former commercial-only simulator, now freeware by the kindness of its author(s), to whom I would like to thank you very much.  It is called Micro-Cap, and has the sliders feature, too, though it was announced there will be no future versions or updates for Micro-Cap.
https://en.wikipedia.org/wiki/Micro-Cap
https://archive.org/details/mc12cd_202110 <- download link

Microcap is well polished, has tons of features, and more robust than QucsStudio, but I didn't have the time to play with it yet.  It was made for Windows, but it works in Linux, too, through WineHQ.  Microcap is based on the SPICE engine, while QucsStudio is based on QUCS. (Component models are different between SPICE and QUCS).
https://www.eevblog.com/forum/eda/qucs-qucs-s-and-qucsstudio-simulators-are-not-the-same-thing/

QUCS type of simulators also have some RF (Radio Frequency) oriented simulations that are missing from SPICE-like simulators (for example RF filters, Smith charts, etc.).

TINA-TI is based on the SPICE engine, just like LTspice, therefore the same spice models for components can be used in either of them.  However, different implementations of SPICE-like simulators are using different math solvers (for the same component models), and sometimes the same schematic might work very well in LTspice and not so well in TINA-TI, or the other way around.

QSPICE is very new, I didn't use it, but my understanding is that QSPICE has a single developer (at least for now), so it will be hard to beat a full team of programmers backed by huge companies like Texas Instrument, or Analog Devices.

Since you use mostly TI parts, TINA-TI is the best choice, indeed , though TINA-TI seemed to me really difficult to work with, by comparison with LTspice.  IIRC TINA-TI is based on a commercial version of SPICE called pspice, though I've used pspice during the university years, and the pspice version from some decades ago was a breeze to use by comparison with the current TINA-TI.  No idea why TI made a pspice-like simulator so difficult. 

[tggzzz]

Have you personally played around with QSPICE?  I don't have the experience nor knowledge base to personally make the statement, but from what I have researched and learned about it is it is supposed to be a much more robust simulator.  It was developed by a gentleman who developed the original LT Spice.  There is a interview video with him that I watched where he touted  some of the benefits.  Unfortunately my knowledge base and experience prevents me from fully understanding his arguments.

No, I haven't.

From memory (quite possibly faulty), that developer parted ways with his employer in ways that weren't perfectly amicable, and this caused disquiet in the LTSpice user community. The reasons were mentioned, but I don't remember them because they were uninteresting.

He then developed QSpice, but quite reasonably it isn't free (beer or open source). I haven't bothered to assess QSpice nor the libraries of components.

[tggzzz]

All spice lies!

It is a truism that "All models are wrong, but some are useful" https://en.wikipedia.org/wiki/All_models_are_wrong

As ever, you get what you pay for, if you dont pay don't expect the best of the best.

Paying more does not imply something better, and paying less does not imply something worse.

[alligatorblues]

One of the areas of knowledge I am looking to expand is the use of a simulator for modeling circuits.  I have been playing around with the following 3 simulators.

LTspice  https://www.analog.com/en/resources/design-tools-and-calculators/ltspice-simulator.html

QSPICE   https://www.qorvo.com/design-hub/design-tools/interactive/qspice

TINA-TI   https://www.ti.com/tool/TINA-TI

I am asking for help in steering my decision-making process.  LT Spice seems to be the gold standard but I find it cumbersome to use and the UI not as friendly.

QSPICE Was developed by the same gentleman that did LT Spice and seems to be a better evolution of LT spice.

For those of you that have read some of my posts before I am somewhat partial to Texas Instruments because I find they make quality parts but most importantly I find it interesting reading their data sheets. Not only do they give specifications but they offer a very good education on certain subject matters.  For this particular reason I decided to try TINA-TI.

TINA-TI by far to me seems the most natural and easiest to use and model basic circuits.

I am asking for people who are a lot more experienced than me to give me guidance in selecting a simulator to focus on.  If I decided to settle on TINA-TI would I be missing anything and/or would I be gaining anything over the other two?

QSPICE Seems to be superior to LT spice and it was developed by the same gentleman that developed LT spice so naturally it seems like it would be a better simulator.

Since I would like to become a lot more proficient in being able to model circuits via simulator and am going to allocate a considerable portion of my time in becoming proficient I am asking for the communities help in guiding my decision making process considering I am virtually new to utilizing a simulator.

As always thank you to all who participate in advance.

You'll find, like many other things, design the circuit comes before simulating it. You don't use a debugger to learn how to program. Command driven operating systems are so much faster than gui OSs,  but you have to take a few days to learn the commands.

I never debug a program as a trial and error method to write correct code. Nor would I use a GUI text editor. I use a command line stream editor called 'sed'. You just type a command line to edit a file the way you want, and redirect the output to a new file. That way, if you've got 1,000 '#' you have to remove and replace with '!', it takes about 2 minutes. Doing it with a mouse and keyboard would take hours.

Using a command, I can search 10,000 files for keywords, phases, or the inverse. I can let malware loose, insert my CD-ROM with known good commands, and use 'find' to find every file that has changed attributes in the last 5 minutes. Then I know which files to move into a quarantine, do a hex dump on, and analyze with other tools.

It's the same way with electronics, except your tools are mathematics. You can shave 90% of the design time by designing the circuit first, and then simulating it. The simulator doesn't tell you exactly why a circuit won't work. But, since no op amp is ideal, certain tricks must be employed to get them to settle down. And, no circuit ever lives up to the designers hopes.

There's a rule of electronic cicuits: "it's easy to get what you don't want, and difficult to get what you do want."   

[switchabl]

He then developed QSpice, but quite reasonably it isn't free (beer or open source). I haven't bothered to assess QSpice nor the libraries of components.

It's not open source but it seems to be distributed by Qorvo for free (after registration). And it looks like the license doesn't have the infamous LTspice non-compete clause. The lack of AD/LT libraries is going to be a big downside for many people though.

[JJ_023]

All spice lies!

As ever, you get what you pay for, if you dont pay don't expect the best of the best.
LTspice is not gold standard. Its good enough for most purposes though.
Pspice say they are gold standard but you'll have to pay to find out.
You may like to check this page. https://electronicsguruji.com/best-circuit-simulation-software/
Qspice is more modern and has a more rational interface, it has a deep and easily updated model library.
Spice from TI AD etc are also there to promote their products, so they're not entirely a free gift.
I still have a (fading) preference for MC or "microcap" because of its superior schematic input. Its old and out of date like me though.
Its hard to find but I recall there is a download link on eevblog somewhere

Thank you for your overview.  The link was helpful as it provided additional simulators I was not aware of.

[JJ_023]

For example in your video it would be nice to have sliders on resistors and capacitors

Indeed, and it has.  A slider can be attached in QucsStudio for any numerical value (including the values of Rs or Cs) by simply clicking the wanted component, or the wanted parameter, and a new slider will appear.

There is another former commercial-only simulator, now freeware by the kindness of its author(s), to whom I would like to thank you very much.  It is called Micro-Cap, and has the sliders feature, too, though it was announced there will be no future versions or updates for Micro-Cap.
https://en.wikipedia.org/wiki/Micro-Cap
https://archive.org/details/mc12cd_202110 <- download link

Microcap is well polished, has tons of features, and more robust than QucsStudio, but I didn't have the time to play with it yet.  It was made for Windows, but it works in Linux, too, through WineHQ.  Microcap is based on the SPICE engine, while QucsStudio is based on QUCS. (Component models are different between SPICE and QUCS).
https://www.eevblog.com/forum/eda/qucs-qucs-s-and-qucsstudio-simulators-are-not-the-same-thing/

QUCS type of simulators also have some RF (Radio Frequency) oriented simulations that are missing from SPICE-like simulators (for example RF filters, Smith charts, etc.).

TINA-TI is based on the SPICE engine, just like LTspice, therefore the same spice models for components can be used in either of them.  However, different implementations of SPICE-like simulators are using different math solvers (for the same component models), and sometimes the same schematic might work very well in LTspice and not so well in TINA-TI, or the other way around.

QSPICE is very new, I didn't use it, but my understanding is that QSPICE has a single developer (at least for now), so it will be hard to beat a full team of programmers backed by huge companies like Texas Instrument, or Analog Devices.

Since you use mostly TI parts, TINA-TI is the best choice, indeed , though TINA-TI seemed to me really difficult to work with, by comparison with LTspice.  IIRC TINA-TI is based on a commercial version of SPICE called pspice, though I've used pspice during the university years, and the pspice version from some decades ago was a breeze to use by comparison with the current TINA-TI.  No idea why TI made a pspice-like simulator so difficult. 

Again thank you for writing such a detailed response.    I was wondering what your thoughts are on this particular video on QSPICE.
I like the ability to write the parameters directly on the schematic.  I also like the fact that it provides syntax for some of the parameters.  I've done a decent amount of coding and it just seems intuitive in that regard.

[JJ_023]

Have you personally played around with QSPICE?  I don't have the experience nor knowledge base to personally make the statement, but from what I have researched and learned about it is it is supposed to be a much more robust simulator.  It was developed by a gentleman who developed the original LT Spice.  There is a interview video with him that I watched where he touted  some of the benefits.  Unfortunately my knowledge base and experience prevents me from fully understanding his arguments.

No, I haven't.

From memory (quite possibly faulty), that developer parted ways with his employer in ways that weren't perfectly amicable, and this caused disquiet in the LTSpice user community. The reasons were mentioned, but I don't remember them because they were uninteresting.

He then developed QSpice, but quite reasonably it isn't free (beer or open source). I haven't bothered to assess QSpice nor the libraries of components.

I downloaded it a few months ago and as of that time it was free.  I also updated it yesterday and that was free.

I think all you need to do is register and verify your email.

[JJ_023]

One of the areas of knowledge I am looking to expand is the use of a simulator for modeling circuits.  I have been playing around with the following 3 simulators.

LTspice  https://www.analog.com/en/resources/design-tools-and-calculators/ltspice-simulator.html

QSPICE   https://www.qorvo.com/design-hub/design-tools/interactive/qspice

TINA-TI   https://www.ti.com/tool/TINA-TI

I am asking for help in steering my decision-making process.  LT Spice seems to be the gold standard but I find it cumbersome to use and the UI not as friendly.

QSPICE Was developed by the same gentleman that did LT Spice and seems to be a better evolution of LT spice.

For those of you that have read some of my posts before I am somewhat partial to Texas Instruments because I find they make quality parts but most importantly I find it interesting reading their data sheets. Not only do they give specifications but they offer a very good education on certain subject matters.  For this particular reason I decided to try TINA-TI.

TINA-TI by far to me seems the most natural and easiest to use and model basic circuits.

I am asking for people who are a lot more experienced than me to give me guidance in selecting a simulator to focus on.  If I decided to settle on TINA-TI would I be missing anything and/or would I be gaining anything over the other two?

QSPICE Seems to be superior to LT spice and it was developed by the same gentleman that developed LT spice so naturally it seems like it would be a better simulator.

Since I would like to become a lot more proficient in being able to model circuits via simulator and am going to allocate a considerable portion of my time in becoming proficient I am asking for the communities help in guiding my decision making process considering I am virtually new to utilizing a simulator.

As always thank you to all who participate in advance.

You'll find, like many other things, design the circuit comes before simulating it. You don't use a debugger to learn how to program. Command driven operating systems are so much faster than gui OSs,  but you have to take a few days to learn the commands.

I never debug a program as a trial and error method to write correct code. Nor would I use a GUI text editor. I use a command line stream editor called 'sed'. You just type a command line to edit a file the way you want, and redirect the output to a new file. That way, if you've got 1,000 '#' you have to remove and replace with '!', it takes about 2 minutes. Doing it with a mouse and keyboard would take hours.

Using a command, I can search 10,000 files for keywords, phases, or the inverse. I can let malware loose, insert my CD-ROM with known good commands, and use 'find' to find every file that has changed attributes in the last 5 minutes. Then I know which files to move into a quarantine, do a hex dump on, and analyze with other tools.

It's the same way with electronics, except your tools are mathematics. You can shave 90% of the design time by designing the circuit first, and then simulating it. The simulator doesn't tell you exactly why a circuit won't work. But, since no op amp is ideal, certain tricks must be employed to get them to settle down. And, no circuit ever lives up to the designers hopes.

There's a rule of electronic cicuits: "it's easy to get what you don't want, and difficult to get what you do want."   

That is an insightful way of looking at it.  I have done a decent amount of coding so I understand what you are inferring.  Maybe my thought process is off but I looked at the simulator as the development environment and not a debugger.

It's also been mentioned that bread boarding is horrible and using a simulator to explore certain basic concepts seems like it would be beneficial.  All of the measuring tools are right there.  Even though I have an oscilloscope it is just so much easier if I am exploring a basic concept to do it in the simulator. 

Having said that I am also very aware that getting the circuit to work in a simulator does not guarantee that it will work once put together.  Right now I am not designing anything overly complex it's more about putting what I learned into a practical application because then what I read sinks in a lot deeper.

[RoGeorge]

I was wondering what your thoughts are on this particular video on QSPICE.

The GUI looks a little nicer, but nothing that LTspice can not do, or doesn't have.  Anyway, wanted to give Qspice a try in Linux.  He lost me at the checkbox "I agree to accept spam emails with your ads", which is a mandatory checkbox in order to download.  Then it wanted to send me an email with a download code, etc, etc.  Not for me.

Another thing, may I kindly ask you a personal favor, please?  Please stop saying how much you appreciate whatever I might have wrote, each and every time I post something in your threads.  This is distracting and awkward already.

If you noticed, in general we all here on EEVblog are making comments and statements about whatever the topic is, not about each other.  I hope you don't mind asking from you this little favor for myself, and also thanks for all your thanks. 

[twospoons]

My personal favorite is Simetrix.  I find the UI fairly intuitive, adding models is incredibly simple (drag and drop), and it seems to be quite fast. The free version is node-limited, but its enough for most simple stuff. I've done a small chip design with the free version that had 23 mosfets in it. I'm very tempted to buy the full version, but I seldom run into the node limit so its hard to justify the spend.

Tried LTSpice and hated the UI, found it a bit slow to simulate too. Tried TI-TINA, but struggled when I need to use other manufacturers parts.

[tggzzz]

That is an insightful way of looking at it.  I have done a decent amount of coding so I understand what you are inferring.  Maybe my thought process is off but I looked at the simulator as the development environment and not a debugger.

A simulator is neither a development environment nor a debugger. It is a tool.

Mathematical modelling is valuable in some ways, but has limitations.
Simulation is valuable in some ways, but has limitations.
Physical prototyoes are valuable in some ways, but have limitations.

Engineering is about choosing the right tools so that one tools strengths outweigh another tools weaknesses.

All are good. Repeat: all are bad.

It's also been mentioned that bread boarding is horrible and using a simulator to explore certain basic concepts seems like it would be beneficial.  All of the measuring tools are right there.  Even though I have an oscilloscope it is just so much easier if I am exploring a basic concept to do it in the simulator. 

Breadboards have their limitations, some more than others (I particularly dislike the unnecessary problems people find with solderless breadboards). Nonetheless, they are valuable.

Example: I recently simulated an HV capacitor multiplier. It all worked nicely in the simulator and enable me to spot and and remove a few problems. Nonetheless a quick partial prototype revealed I had forgotten to think about the electrolytic capacitor's leakage current. Whoops

Having said that I am also very aware that getting the circuit to work in a simulator does not guarantee that it will work once put together.  Right now I am not designing anything overly complex it's more about putting what I learned into a practical application because then what I read sinks in a lot deeper.

Neither does getting one prototype to work guarantee that the next will work.

Mathematical modelling is frequently based on the assumption of linearity, so that superposition can work. Without that the maths becomes intractible.

Simulation can include non-linear effects, but often obscures "the big general picture"; it is too easy to see the trees but not the wood. Many simulations can help understand sensitivity factors.