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

[JJ_023]

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. 

I can definitely honor your request and will not thank you anymore.

The only thing I can say about QSPICE is I have gotten literally zero spam emails from them.  I think the download code that you are referring to is to validate the email.  The actual download if I remember is just a link.  Once I did that they have not once made contact with me. 

Also they never ask you to do that process again when you update the software. 

[JJ_023]

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.

Thanks for the heads up.  I never really heard of this particular simulator.  I will have to check it out.  Do you know what the price is for the full version?

What is the node limit if you know off the top of your head?

[JJ_023]

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

I understand the argument that you are presenting and all of it makes sense.  Different tools for different scenarios.

I was wondering  what type of bread boards have the least amount of limitations.  Like everything else I assume there are better quality ones.  The ones I use are generic $10 ones.  I don't mind spending money on decent things.  I have seen bread boards that are 10 times the price but I never purchased one because I never knew which ones were considered the better ones. 

[tggzzz]

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

I understand the argument that you are presenting and all of it makes sense.  Different tools for different scenarios.

I was wondering  what type of bread boards have the least amount of limitations.  Like everything else I assume there are better quality ones.  The ones I use are generic $10 ones.  I don't mind spending money on decent things.  I have seen bread boards that are 10 times the price but I never purchased one because I never knew which ones were considered the better ones.

You seem to be talking about solderless breadboards; there are many other types. For a range, see https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/

Basically there are a whole range of techniques which can be mixed and matched as appropriate.

How would you model this solderless breadboard prototype? Pay particular attention to the (missing) connection on the right (IIRC LTSpice has a neon model in its library.)

[twospoons]

 Simetrix Demo Node limit is 140 nodes.  Cant recall the price exactly and they dont list on the website, but its north of 1kUSD, depending on options. 

Demo version limits for simetrix:

Limitations
The following product limitations apply:

The command line is not present, preventing any features that require the command line
User scripts
The Verilog-A compiler is not available. (Pro and Elite full licenses only)
Verilog-HDL mixed signal simulation is not available. (Pro and Elite full licenses only)
Custom menu and key definitions
SPICE3 raw file import
Safe Operating Area testing is not available
The menu/key editor is not available
Encrypted models are not supported
User scripts can be enabled for 30 days using the unlock feature. See menu Help | Unlock Features...

SIMetrix Circuit Size Limits
There is a limit to the size of circuit that can be simulated. The exact limits for the SIMetrix simulator are:

140 analog nodes (internal and external - see note below)
The node count includes all circuit nodes and some of these may be internal to device models. Opamp models, for example, will always have a number of internal nodes depending on complexity.

@tggzzz Thats a nice demo of one of the primary gotcha's of solderless breadboard!

[JJ_023]

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

I understand the argument that you are presenting and all of it makes sense.  Different tools for different scenarios.

I was wondering  what type of bread boards have the least amount of limitations.  Like everything else I assume there are better quality ones.  The ones I use are generic $10 ones.  I don't mind spending money on decent things.  I have seen bread boards that are 10 times the price but I never purchased one because I never knew which ones were considered the better ones.

You seem to be talking about solderless breadboards; there are many other types. For a range, see https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/

Basically there are a whole range of techniques which can be mixed and matched as appropriate.

How would you model this solderless breadboard prototype? Pay particular attention to the (missing) connection on the right (IIRC LTSpice has a neon model in its library.)

I inferred that you were talking about solderless breadboard's.  That is the reason why I asked if you had a particular product in mind.

The link that you sent somebody sent me that in one of my previous posts ( it might have been you) I read that already and I am very familiar with it.

When you ask me how I would model it, what exactly do you mean by that?  In a simulator?

As far as what I see on the breadboard.

I believe that is an incandescent lightbulb.

One of the connections is not connected directly.

My guess is due to the inductance in the breadboard due to the metal clips a circuit is being completed.

My question is what would happen if you moved over an additional two rows?

[LinuxHata]

It is very funny to read about "big support team behind the ltspice" when these Gentlemans can only offer 16 color EGA palette for plot line selection and you have no control over the colors/thickness/etc. And there are certain colors, like dark blue, which are very hard to read against the black background, and you have no control over it, since software assigns it on random, so you have to change it AFTER.

[Ian.M]

Menu: Tools -> Color Preferences opens a dialog that lets you set the RGB for each LTspice color.  There's no color picker, just RGB sliders and numerics, but its certainty possible to tweak the blues and greens so dark blue and green are more visible against a black background.

[LinuxHata]

Still no control over colors - like, you can not apply rules, so say, current measurements will be always say, shades of red, voltages, shades of green and so on.
I just downloaded QSPICE. They have a bit better control over that, but still, no pre-defined color assignment rules.
By the way, while some improvements are definitely made in QSPICE, some things are even more complicated - for example, can't find how to add power source to the schematics

[RoGeorge]

That dark blue is almost invisible. 

Fortunatly, the default setting for LTspice are stored in a text file:
    ...\User\AppData\Roaming\LTspiceXVII.ini for Windows, or in
    /home/$USER/.wine/drive_c/users/$USER/Application\ Data/LTspiceXVII.ini for Linux,
and can be kept somewhere else as copy backup.  Mine looks like this file posted here:
https://www.eevblog.com/forum/eda/where-does-ltspice-keep-its-settings/msg4150582/#msg4150582
and in practice the schematic and plots have these colors:







Peltz oscillator LTspice printscreens are from https://www.eevblog.com/forum/projects/simple-sinusoidal-oscillators/, with the colors from the above LTspice settings file.

[magic]

How would you model this solderless breadboard prototype? Pay particular attention to the (missing) connection on the right (IIRC LTSpice has a neon model in its library.)

This is an easy case because row to row capacitance can be measured and included in the simulation.

It would be harder if the circuit were meaningfully affected by inductance or mutual inductance of jumper wires.
Even harder - I'm not sure how to model random and randomly time-varying resistance, when that starts to matter

[iMo]

The row capacitance is aprox 3.5pF on those solderless breadboards, afaik..

[Ian.M]

Dave measured it.  See: https://www.eevblog.com/forum/blog/eevblog-568-solderless-breadboard-capacitance/

A quality breadboard may have significantly lower capacitance due to use of plastics with a lower dielectric constant, and better contact design minimising the side area and proximity to the wall between strips.  Boss Industrial Mouldings Ltd used to quote 0.5pF between adjacent strips of their Bimboard range.

[RoGeorge]

how to model random and randomly time-varying resistance

- For pseudo random, probably a file with predefined values, then a PWL (PieceWise Linear) source to read that file.
- For time variable resistor, LTspice has a variable called 'time', which can be used inside newly defined functions or behavioral sources.

Example of using the internal LTspice variable time inside a simulation (time variable can be used in plots, too):


Hilbert Transform simulation LTspice printscreen from https://www.eevblog.com/forum/programming/extract-precise-amplitude-and-phase-from-a-frequency-sweep-(vna-from-dsoawg)/

[tggzzz]

I inferred that you were talking about solderless breadboard's.  That is the reason why I asked if you had a particular product in mind.

The link that you sent somebody sent me that in one of my previous posts ( it might have been you) I read that already and I am very familiar with it.

When you ask me how I would model it, what exactly do you mean by that?  In a simulator?

Glad you found it helpful

Yes, model it in a simulator. That question is designed to highlight the strengths and limitations of mathemetical modelling, simulations, and solderless breadboards.

As far as what I see on the breadboard.

I believe that is an incandescent lightbulb.

One of the connections is not connected directly.

My guess is due to the inductance in the breadboard due to the metal clips a circuit is being completed.

My question is what would happen if you moved over an additional two rows?

Good last question The neon bulb lights with the right hand wire in rows 1,2,3,4 but not 5 or further away. I will take a picture of that, for future reference

It wouldn't work with an incandescant bulb since that would require too much current.

What's happening is that there is capacitative coupling between the adjacent rows of the (unmodified) solderless breadboard. The internal parallel flat plates are acting as capacitors, and the 150kHz AC power supply is "passing through" that capacitance.

Now capacitive coupling occurs everywhere, e.g. it is enlightening to use the standard mathematical models to calculate the capacitance of the earth-moon pair. Hence capacitive coupling can and does occur with other breadboard techniques, but can be arranged to be less "severe": firstly the wires have smaller area and are unlikely to be parallel, and secondly a well decoupled nearby ground plane has a tendency to "short circuit" the current (to use an informal handwaving explanation).

The inductance problem (again present wih all construction techniques) is because when the current through a conductot changes (especially a ground/Vcc connection), a "back EMF" voltage opposes the change in current. That is a particular issue with digital circuits, but also with high speed or sensitive analogue applications. Keep wires short and loops small

[tggzzz]

How would you model this solderless breadboard prototype? Pay particular attention to the (missing) connection on the right (IIRC LTSpice has a neon model in its library.)

This is an easy case because row to row capacitance can be measured and included in the simulation.

It would be harder if the circuit were meaningfully affected by inductance or mutual inductance of jumper wires.
Even harder - I'm not sure how to model random and randomly time-varying resistance, when that starts to matter

Yes on all counts

The point is, of course, that the mere existence of the capacitance has to be appreciated (a problem for beginners!) and its value estimated.

As I have just noted, the question is designed to highlight the strengths and limitations of mathemetical modelling, simulations, and solderless breadboards.

Hoperfully the modelling of the consequences of time-varying resistances can best be achieved by surrounding the existing model with a large round metal object - open at the top so that it can't be simplified by assiming it is a Faraday cage 

[Ian.M]

LTspice also supports behavioral resistors, and has rand(x) and random(x) functions, which generate a new pseudorandom value as x crosses each integer value.  rand() is unsmoothed (i.e. level step level step ...) but random() fits a curve between them.  N.B. If x repeats exactly, so will the pseudorandom sequence.

You can thus create a psuedorandom resistance simply by setting the value of a resistor to:

Code: [Select]

R={Rmin}+{Rmax-Rmin}*random(time/{tscale}+{seed})tscale is the time interval between psuedorandom samples, and seed is an integer to start the sequence at different offsets.

Edit: I forgot the {} round seed.  It seems to work without them but better to do it right.

[JJ_023]

Still no control over colors - like, you can not apply rules, so say, current measurements will be always say, shades of red, voltages, shades of green and so on.
I just downloaded QSPICE. They have a bit better control over that, but still, no pre-defined color assignment rules.
By the way, while some improvements are definitely made in QSPICE, some things are even more complicated - for example, can't find how to add power source to the schematics

I am not sure if this is what you mean.  But if you hit the "V" key it will bring up a voltage source.  Then you have a command prompt to define the source.

[JJ_023]

Dave measured it.  See: https://www.eevblog.com/forum/blog/eevblog-568-solderless-breadboard-capacitance/

A quality breadboard may have significantly lower capacitance due to use of plastics with a lower dielectric constant, and better contact design minimising the side area and proximity to the wall between strips.  Boss Industrial Mouldings Ltd used to quote 0.5pF between adjacent strips of their Bimboard range.

Are the BIM boards still available?  I did a quick search and could not find anyone selling them or a website for them.

[JJ_023]

Yes, model it in a simulator. That question is designed to highlight the strengths and limitations of mathemetical modelling, simulations, and solderless breadboards.

I will be honest I personally would not know how to model this particular scenario in a simulator.  However having read some of the previous posts I would probably do it with a capacitor in series with the bulb.  Would that be the correct way?

Good last question The neon bulb lights with the right hand wire in rows 1,2,3,4 but not 5 or further away. I will take a picture of that, for future reference

It wouldn't work with an incandescant bulb since that would require too much current.

What's happening is that there is capacitative coupling between the adjacent rows of the (unmodified) solderless breadboard. The internal parallel flat plates are acting as capacitors, and the 150kHz AC power supply is "passing through" that capacitance.

Now capacitive coupling occurs everywhere, e.g. it is enlightening to use the standard mathematical models to calculate the capacitance of the earth-moon pair. Hence capacitive coupling can and does occur with other breadboard techniques, but can be arranged to be less "severe": firstly the wires have smaller area and are unlikely to be parallel, and secondly a well decoupled nearby ground plane has a tendency to "short circuit" the current (to use an informal handwaving explanation).

The inductance problem (again present wih all construction techniques) is because when the current through a conductot changes (especially a ground/Vcc connection), a "back EMF" voltage opposes the change in current. That is a particular issue with digital circuits, but also with high speed or sensitive analogue applications. Keep wires short and loops small

So I was aware of inductance and capacitance problems with wiring and bread boards especially after reading some of the linked material in this post and previous posts. 

I wrongly assumed that it was inductance that was a problem but now that you elaborated what is happening it makes sense.  I knew the metal clips would be involved somehow and it makes sense that there would be capacitance based on the air gap which in this case would be the dielectric.

As embarrassing as it is I am not familiar with the neon bulb.  I do not know the current requirements to light one versus incandescent that is why I stated incandescent. 

My follow-up question is will this work with an LED since its current requirements are very small?  My guess is it would. 

I also assume you used some sort of switch mode power supply due to the 150 kHz mentioned.

If you put a decoupling capacitor on row 2 that should make the light go out.  Is this correct?

[tggzzz]

Yes, model it in a simulator. That question is designed to highlight the strengths and limitations of mathemetical modelling, simulations, and solderless breadboards.

I will be honest I personally would not know how to model this particular scenario in a simulator.  However having read some of the previous posts I would probably do it with a capacitor in series with the bulb.  Would that be the correct way?

The key point is that the circuit works (or fails!) because of "unexpected and invisible" parasitic components. Minimising those is part of sound circuit design; almost always the intention is that the circuit's operation is determined by well-defined components. No surprise there!

There are, in general, many parasitic "components" most of which are probably unimportant. In this circuit the important parasitic component that allows the effect to be visible is a capacitance in series with the neon and voltage source. Hence the effect could be demonstrated by a model containing the AC voltage source, capacitor, neon, plus resistor representing the internal source resistance of the voltage source.

It would be non-trivial to choose the capacitance value, and given the very strange piece of test equipment I used as a voltage source, determining appropriate values for the voltage amplitude and resistance wouldn't be trivial.

I wrongly assumed that it was inductance that was a problem but now that you elaborated what is happening it makes sense.  I knew the metal clips would be involved somehow and it makes sense that there would be capacitance based on the air gap which in this case would be the dielectric.

Having a good mental model of what's present and what's happening leads to such understanding. Having a knowledge of theoretical models and behaviour is a pre-requisite for making sense when an implementation doesn't behave as expected. Simulation can sometimes be used to confirm the cause of observed behaviour.

As embarrassing as it is I am not familiar with the neon bulb.  I do not know the current requirements to light one versus incandescent that is why I stated incandescent. 

My follow-up question is will this work with an LED since its current requirements are very small?  My guess is it would. 

Don't be embarassed; neon bulbs are largely obsolete. The nearest component used in modern equipment is a gas discharge tube used to protect against high voltage transients on PSU inputs.

They were used for strange things; before I sold it I had a Fluke DVM that used a neon as the reference voltage. Since it wasn't very stable, it had a front-panel control that enabled the displayed value to be tweaked until it showed 1.019V, the same as a (stable) Weston standard cell. You can also implement digital logic with them, but I wouldn't recommend it!

Neons require 60-120V to illuminate, but the current is 10s of microamps. Thus a LED might illuminate, but not brightly.

I also assume you used some sort of switch mode power supply due to the 150 kHz mentioned.

The voltage source was just something I had to hand, a strange device used for testing bulbs, a TIS1040. It is more an oscillator than an SMPS.

[tggzzz]

Thus a LED might illuminate, but not brightly.

Confirmed. A modern high-efficiency LED illuminates dimly, on the adjacent strip only. Older LEDs don't illuminate.

[Ian.M]

Boss Industrial Mouldings Ltd used to quote 0.5pF between adjacent strips of their Bimboard range.

Are the BIM boards still available?  I did a quick search and could not find anyone selling them or a website for them.

Unfortunately not.  The company went under back in 2008, and IIRC ceased breadboard production well before that date. 

[MarkT]

And I suspect that 0.5pF number was erroneous anyway given the data from all the other breadboard!

[JJ_023]

Yes, model it in a simulator. That question is designed to highlight the strengths and limitations of mathemetical modelling, simulations, and solderless breadboards.

I will be honest I personally would not know how to model this particular scenario in a simulator.  However having read some of the previous posts I would probably do it with a capacitor in series with the bulb.  Would that be the correct way?

The key point is that the circuit works (or fails!) because of "unexpected and invisible" parasitic components. Minimising those is part of sound circuit design; almost always the intention is that the circuit's operation is determined by well-defined components. No surprise there!

There are, in general, many parasitic "components" most of which are probably unimportant. In this circuit the important parasitic component that allows the effect to be visible is a capacitance in series with the neon and voltage source. Hence the effect could be demonstrated by a model containing the AC voltage source, capacitor, neon, plus resistor representing the internal source resistance of the voltage source.

It would be non-trivial to choose the capacitance value, and given the very strange piece of test equipment I used as a voltage source, determining appropriate values for the voltage amplitude and resistance wouldn't be trivial.

I wrongly assumed that it was inductance that was a problem but now that you elaborated what is happening it makes sense.  I knew the metal clips would be involved somehow and it makes sense that there would be capacitance based on the air gap which in this case would be the dielectric.

Having a good mental model of what's present and what's happening leads to such understanding. Having a knowledge of theoretical models and behaviour is a pre-requisite for making sense when an implementation doesn't behave as expected. Simulation can sometimes be used to confirm the cause of observed behaviour.

As embarrassing as it is I am not familiar with the neon bulb.  I do not know the current requirements to light one versus incandescent that is why I stated incandescent. 

My follow-up question is will this work with an LED since its current requirements are very small?  My guess is it would. 

Don't be embarassed; neon bulbs are largely obsolete. The nearest component used in modern equipment is a gas discharge tube used to protect against high voltage transients on PSU inputs.

They were used for strange things; before I sold it I had a Fluke DVM that used a neon as the reference voltage. Since it wasn't very stable, it had a front-panel control that enabled the displayed value to be tweaked until it showed 1.019V, the same as a (stable) Weston standard cell. You can also implement digital logic with them, but I wouldn't recommend it!

Neons require 60-120V to illuminate, but the current is 10s of microamps. Thus a LED might illuminate, but not brightly.

I also assume you used some sort of switch mode power supply due to the 150 kHz mentioned.

The voltage source was just something I had to hand, a strange device used for testing bulbs, a TIS1040. It is more an oscillator than an SMPS.

Thank you for such an elaborate response.

So I googled your power supply and watched a video on it.  It's basically used for testing all sorts of lightbulbs.  It also has a continuity feature.  How does it work I didn't see a ground clip?  It talks about holding one end in hand  so I assume the person completes the circuit?