Author Topic: Do you start in a simulator?  (Read 5161 times)

0 Members and 1 Guest are viewing this topic.

Offline pilotchupTopic starter

  • Supporter
  • ****
  • Posts: 27
  • Country: us
Do you start in a simulator?
« on: July 29, 2018, 05:30:57 am »
Hi all,

Hope this isn't a ridiculously overdone question.

Do you guys start in a software package that allows simulation of a circuit if you are brainstorming or designing a circuit? Or do you just skip the simulator and start drawing a schematic in something like KiCad/Altium/Eagle/etc?

I feel like, as a someone still learning, a simulator is the natural place to do circuit design - and those other tools are for when your schematic is practically complete and/or you are going to do board layout. However, I hear a lot of people just skip simulators all together or hardly use them.

If you use a simulator, do you "export" your schematic to one of those schematic and layout tools later? What is the industry standard for simulation packages? SPICE based programs seem to be the most popular - what's your favorite?

I find that there really doesn't seem to be a "all-in-one" tool, where you can do schematics, simulations, and board layout. Maybe it's better that way, to prevent a mess of a software package.

Thanks for your time
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 11228
  • Country: us
    • Personal site
Re: Do you start in a simulator?
« Reply #1 on: July 29, 2018, 06:06:21 am »
LTSpice for specific small analog sub-circuits. The rest is straight up in EDA.
Alex
 

Offline ikrase

  • Regular Contributor
  • *
  • Posts: 151
  • Country: us
Re: Do you start in a simulator?
« Reply #2 on: July 29, 2018, 06:09:15 am »
For me? 

NO SIMULATION

NO PLANS

NO CALCULATION

NO PCB DESIGN

ONLY SOLDER.


Ok, that's slightly exaggeration but I don't do much simulation at all.
 

Offline pilotchupTopic starter

  • Supporter
  • ****
  • Posts: 27
  • Country: us
Re: Do you start in a simulator?
« Reply #3 on: July 29, 2018, 06:14:05 am »
@ikrase - haha, a 'kick the tires and light the fires' type huh?


@ataradov - Thanks for the response. That makes sense, I seem to only be simulating small analog circuits too. Any reason why LTSpice? I don't mean to make this a 'which one is best' type question, just curious if you tried the others.

Thanks!
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 11228
  • Country: us
    • Personal site
Re: Do you start in a simulator?
« Reply #4 on: July 29, 2018, 06:15:20 am »
Any reason why LTSpice?
It is there, it is free and it works. I really don't care much past that. It serves the purpose.
Alex
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19279
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Do you start in a simulator?
« Reply #5 on: July 29, 2018, 06:28:39 am »
Do you guys start in a software package that allows simulation of a circuit if you are brainstorming or designing a circuit? Or do you just skip the simulator and start drawing a schematic in something like KiCad/Altium/Eagle/etc?

If you ask the right questions, a simulator will tell you how an idealised abstraction/simplification of your circuit will work. It is good when thinking about "what if" questions.

KiCad and similar define the physical components in your circuit, but don't tell you anything about how they will behave.

In other words, a simulator helps you decide what to build, whereas KiCad helps you build it.
If you know what to build you don't need a simulator.
If you aren't going to build anything, you don't need KiCad.

Those two categories are merged in microwave design tools, where conductor geometry defines behaviour.

Quote
I find that there really doesn't seem to be a "all-in-one" tool, where you can do schematics, simulations, and board layout. Maybe it's better that way, to prevent a mess of a software package.

Yes. It is better to have one thing done well rather than two things done poorly.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: pilotchup

Offline pilotchupTopic starter

  • Supporter
  • ****
  • Posts: 27
  • Country: us
Re: Do you start in a simulator?
« Reply #6 on: July 29, 2018, 08:11:53 am »
Thanks for the response tggzzz, great simplistic explanation. Weird, but you made me breathe a sigh of relief. So easy, at least for me, to get caught up in all these tools when trying to research them in depth
What you said about microwave design makes a ton of sense, and gives me a better understanding why some of these tools are such big complex systems (Keysight ADS for example).

 

Offline ucanel

  • Regular Contributor
  • *
  • Posts: 134
  • Country: tr
Re: Do you start in a simulator?
« Reply #7 on: July 29, 2018, 09:29:36 am »
It depends on the project,
if it is about just algorithm and simple input output thing
simulatin is enough,
if it consist pwm, analog input output, screen other than 2x16 type lcd
breadbord is enough,
if it needs high frequency sort of stuff
dead bug and perfboard is good start.
These works for me because my projects are
commonly microcontroller based projects.
 
The following users thanked this post: pilotchup

Offline Berni

  • Super Contributor
  • ***
  • Posts: 4922
  • Country: si
Re: Do you start in a simulator?
« Reply #8 on: July 29, 2018, 09:39:30 am »
Oh hell yes.

If its analog and its not a known working design first thing to do for me is LT Spice. That way i can get things tuned in and the kinks worked out before i even start doing the schematic.

The simulation sometimes gives me a lot of extra insight in the working of the circuit to find potential problems with it and address them at this stage to possibly save me a PCB revision later. But you do need some experience to know what can and can't be trusted in the simulator. As good as LT Spice is, it is still not perfect, you need to be aware of potential parasitics of PCB traces and components, tolerances on parts, overly ideal components, gorse overload conditions, uneven temperature across your circuit etc. The fact that it works in LT Spice does not completely grantee it will work in real life, but if you can't get it working in Spice no matter how much you tweak it, then it will also likely not work in real life.

Also you don't usually simulate the entire end product in spice. Rather you simulate self contained modules of the product, like just an output amplifier and yet you feed it from just a signal generator rather than build all the signal generation circuity too.
 
The following users thanked this post: pilotchup

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19279
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Do you start in a simulator?
« Reply #9 on: July 29, 2018, 11:01:45 am »
Thanks for the response tggzzz, great simplistic explanation. Weird, but you made me breathe a sigh of relief. So easy, at least for me, to get caught up in all these tools when trying to research them in depth
What you said about microwave design makes a ton of sense, and gives me a better understanding why some of these tools are such big complex systems (Keysight ADS for example).

All hardware and software tools have a learning curve - and therefore cost in your time. As a rule of thumb, the more expensive the tool, the longer the learning curve. That's one reason why it can be counterproductive for a beginner to buy the "best/expensive" tool.

Once you know how to use one tool in a class (e.g. analogue simulation, or digital simulation, or oscilloscope, or spectrum analyser, or PCB CAD), it is relatively easy to learn another in the same class. Even so, there will be a cost in learning exactly which button to push on a new tool.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: pilotchup

Offline ikrase

  • Regular Contributor
  • *
  • Posts: 151
  • Country: us
Re: Do you start in a simulator?
« Reply #10 on: July 29, 2018, 11:42:17 pm »
Quote
@ikrase - haha, a 'kick the tires and light the fires' type huh?
 

On hobbyist applications, yes. Getting something sorta working is what is satisfying in the short term, and between low budget, low quality tools, etc extensive simulation is often a waste of time because it won't conform anyway.  Then, when actually building, quasi-Manhattan construction is extremely agile as is wire wrapping.

For professional stuff, the concerns are obviously very different (I don't do much professional electronics as much as deal with other peoples')
 
The following users thanked this post: pilotchup

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Do you start in a simulator?
« Reply #11 on: July 30, 2018, 01:31:14 am »
Use whatever is suitable.  There is no one best method, only all of them!

Examples:



I breadboarded this circuit first, then built up the soldered prototype pictured.  A very noncritical circuit, not bothered by the limitations of the solderless breadboard.


This one I simulated:





As it's battery powered, I was very concerned about capacitance.  The active logic transistors have 1pF capacitance, so the 4pF row-to-row capacitance of the breadboard is intolerable.  I started on the breadboard, as a demonstration.  I quickly reached a limit of 400ns switching speed, at a cost of almost 1mA current consumption.  After tweaking in the simulator, and only making a few resistor value changes in the final build, I got it down to 200ns and 0.3mA.  The improvement in switching speed allowed a reduction in the inductor value, convenient as I can't buy an SMT dual inductor any larger than that!

This one, straight to PCB:



Based on an earlier project, the analog part is already verified and the logic is just that, logic, not a big deal to review.  Most of the logic is already working, too. :)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: pilotchup

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: Do you start in a simulator?
« Reply #12 on: July 30, 2018, 02:54:57 pm »
LTSpice for specific small analog sub-circuits. The rest is straight up in EDA.

Exactly the same for me, seems to work well.

Some projects - no simulations needed at all.

Others - countless of hours brainstorming in LTSpice.

Really depends if an IC is available that performs the required "analog" tasks easily and sanely. Sometimes I need to do everything from scratch; in these cases, simulations are very valuable.
« Last Edit: July 30, 2018, 02:58:18 pm by Siwastaja »
 
The following users thanked this post: pilotchup

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Do you start in a simulator?
« Reply #13 on: July 30, 2018, 11:43:17 pm »
For me, a decent RF simulator has always been a very powerful tool to have at my disposal. When I started doing serious RF design at work about 30 years ago, the things I wanted most of all (to play with at home) were the powerful RF CAD tools we had at work. The $$$ RF test gear from HP and Marconi and R&S were desirable too but were a very distant second to these old DOS based CAD tools. It's possible to learn an awful lot from skilled use of a simulator and the learning curve can be really fast.

Fast forward to today and the gap is even wider because modern RF CAD tools are very versatile and very powerful and accurate. At work we often design RF stuff (up into the GHz region) entirely on a simulator including RF filters, couplers amps etc and go straight to PCB and (low qty) production without any hardware prototyping. The first batch of 'finished' boards are sometimes the first chance we get to play with real hardware.

To try and do it the other way around and design the RF hardware via numerous PCB prototypes (without any initial simulation work) would be very time consuming and expensive.
« Last Edit: July 30, 2018, 11:45:25 pm by G0HZU »
 
The following users thanked this post: pilotchup

Offline MT

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: aq
Re: Do you start in a simulator?
« Reply #14 on: August 01, 2018, 02:09:15 pm »
Use whatever is suitable.  There is no one best method, only all of them!

Examples:



I breadboarded this circuit first, then built up the soldered prototype pictured.  A very noncritical circuit, not bothered by the limitations of the solderless breadboard.
Tim

I agree not critical att all with GND and +V tied to same copper plane! :)...Nice boards anyway!



 

Offline MT

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: aq
Re: Do you start in a simulator?
« Reply #15 on: August 01, 2018, 02:18:43 pm »
Do you guys start in a software package that allows simulation of a circuit if you are brainstorming or designing a circuit? Or do you just skip the simulator and start drawing a schematic in something like KiCad/Altium/Eagle/etc?
Anything goes, if you stick to one and only your doomed!

Quote
I feel like, as a someone still learning, a simulator is the natural place to do circuit design - and those other tools are for when your schematic is practically complete and/or you are going to do board layout. However, I hear
a lot of people just skip simulators all together or hardly use them.

The more you know the less you simulate! :)

Quote
If you use a simulator, do you "export" your schematic to one of those schematic and layout tools later?
Hell no, im not insane, rather lazy so i poke around in schematics first!

Quote
What is the industry standard for simulation packages? SPICE based programs seem to be the most popular - what's your favorite?

Its not about favorites , its about which one works seemingly accurate. You have to remember Spice component
models are "approximations" models not complete, not exact, not realtime, dont have a sample rate, etc. When i
have to i use LT Spice and Tina TI. Chip designers and RF people seams to simulate all the time these days
but their tools and models cost a fortune im told. LT and TI is free!

Often you can read in IC sheets "by design not tested" which suggest its entirely simulated yet multi national elephants screw their designs up like ST with F4 ADC signal integrity which make on wonder which part of the simulation did they screw up? Floorplaning? subcirquit placement? or the entire sub design? Who knows, but it cost them dearly.
« Last Edit: August 01, 2018, 02:42:20 pm by MT »
 
The following users thanked this post: pilotchup

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: Do you start in a simulator?
« Reply #16 on: August 01, 2018, 02:40:36 pm »
if i have to make something like a filter : simulation
for anything io , digital , microcontroller and simple analog : straight to board.

most of the design work can be done using a 4 function calculator or excel.
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 
The following users thanked this post: pilotchup

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19279
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Do you start in a simulator?
« Reply #17 on: August 01, 2018, 03:38:46 pm »
for anything io , digital , microcontroller and simple analog : straight to board.

Only for simple stuff.

It is highly educational to simulate a multi-output digital system driving a moderate load. To pick a few figures:
  • a digital output, say 3V, with a series 100ohm resistor, 1ns risetime
  • one side of that digital output is connected to an inductor to ground. That inductor represents the inductance of the ic's ground lead, 1nH/mm so say 5nH
  • that resistor is connected to a transmission line, 100ohms, 1ns, and the other end of the transmission line is connected to a 5pF capacitor representing another ic's input capacitance

Do a transient analysis and look at both ends of the transmission line.

Then simulate eight such outputs switching simultaneously, by reducing the resistor to 12.5ohms.

If you think that is academic, I suggest you look at and understand the IBIS models published for digital ics - and why they need to exist.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: pilotchup

Offline rx8pilot

  • Super Contributor
  • ***
  • Posts: 3634
  • Country: us
  • If you want more money, be more valuable.
Re: Do you start in a simulator?
« Reply #18 on: August 01, 2018, 04:11:32 pm »
I have been using simulation more and more lately. I have been limited by my modest skills in LT Spice so I had to make an effort to learn the tool a bit better. In power electronics, I learned that I can save quite a bit of time and burned MOSFETS if I simulate and iterate the unkown areas of the circuit. I can also work through SOA issues which is important because its hard to troubleshoot a circuit after the smoke has escaped. Seeing what to expect and gaining an understanding of tolerance for various parasitics is worth the effort.

As I have become more skilled in simulation, it encourages me to learn more about circuits I was previously 'scared' of. My design time has gone down and my first prototype successes are way up.

I just wish LT Spice was not so crude in terms of UI. The solver engine is great.... the rest of it is a trip back to 1990.

Short and misplld from my mobile......

Factory400 - the worlds smallest factory. https://www.youtube.com/c/Factory400
 
The following users thanked this post: pilotchup

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Do you start in a simulator?
« Reply #19 on: August 01, 2018, 04:24:16 pm »
I typically simulate portions of the circuit that are less familiar to me, so I can play around with them and experiment with the effects of component variations. If it's something like a linear regulator or simple op-amp circuit then I usually don't bother to simulate it, I know enough about these blocks that I know how they will work.
 
The following users thanked this post: pilotchup

Offline blackfin76

  • Regular Contributor
  • *
  • Posts: 79
  • Country: nl
Re: Do you start in a simulator?
« Reply #20 on: August 10, 2018, 08:01:29 pm »
Calculations on paper for the simple stuff, sometimes a spreadsheet for the repetative stuff and a spice simulator for more advanced circuits like filters. Measurements to verify.

In the time of OTP MCU's I did simulation of source code before 'burning' them in the MCU. Today with the flash devices I have become a trial and error person  :horse:
 
The following users thanked this post: pilotchup

Offline pilotchupTopic starter

  • Supporter
  • ****
  • Posts: 27
  • Country: us
Re: Do you start in a simulator?
« Reply #21 on: August 12, 2018, 09:52:44 pm »
Thanks everyone for you insights! Invaluable!!


@ucanel   Thank you, good breakdown

@Berni   Thanks for the very detailed response, I've been looking into the differences between simulation and real life and you are right!

@tggzzz   Oh my so true, I've been going with some online tools and now LTSpice to get the hang of it first

@ikrase   Yeah, easy to waste time with this stuff rather than just experimenting :)

@T3sl4co1l   Wow, nice pictures! I was trying out what you said about capacitance with the breadboard traces. It's amazing how much capacitance and inductance there is and how much it changes vs simulation

@Siwastaja   Thank you, makes sense

@G0HZU   Wow, thank you for the valuable insight working with RF. That is precisely what I was wondering about, how you start in simulator and are even able to skip the protoype step. That is also an interesting point about sim tools vs the high dollar gear when dealing with RF as a hobby

@MT   I would think T3sl4co1l just used 2 different color wires and both are ground? I see a bare lead sticking up by the top left inductor that is probably for the power :)
   Yeah, its tricky to understand the differences between SPICE and real life circuits, as others have pointed out here too. Guess best way is just through experience to know whats going to not translate well. I hadn't heard about the ST and their F4 ADC thing before!

@free_electron   Thank you, that seems to be the consensus here

@tggzzz   I have never heard of IBIS. That is a really interesting breakdown you described, its pratically RF even though the circuit is "digital"? I am going to try this in LTSpice

@rx8pilot   Thank you for sharing your experience! I know what you mean. Also yeah, it's interesting how UIs of all these tools are circa 1990s.. not that I'm bothered by it, but it just shows where the priorities for the creators are (functionality over beauty or ease of use, which I understand)

@james_s   Right, makes sense. I suppose one day that will be the same for me, but for now - I'm simulating everything :)

@blackfin76   Gotcha. What things would you use Excel for? I guess I haven't had to work on a circuit with a lot of repetative calculations. Yeah, that's the popular software way of debugging.. which is basically just changing 'random' things and try again until you get the right solution lol
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19279
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Do you start in a simulator?
« Reply #22 on: August 13, 2018, 06:52:06 am »
@tggzzz   Oh my so true, I've been going with some online tools and now LTSpice to get the hang of it first

You're welcome; thank you for listening and thinking (which can't be guaranteed!). It is always useful to understand the capabilities of the tools in your box, even if you leave them there most of the time.

Quote
@tggzzz   I have never heard of IBIS. That is a really interesting breakdown you described, its pratically RF even though the circuit is "digital"? I am going to try this in LTSpice

It isn't much of an exaggeration to say everything is analogue, with the exception of femtoamp circuits (counting individual electrons) and avalanche photodiodes (counting individual photons). "Digital" circuits are merely analogue circuits where the voltages/currents are interpreted to represent (usually) binary signals.

Whole careers are based on that understanding. As a quick introduction, see https://www.edn.com/collections/4435129/Bogatin-s-Rules-of-Thumb
I don't recommend reading them all in detail (yet), but it is worht skimming the earlier rules of thumb.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: pilotchup

Offline Sudo_apt-get_install_yum

  • Regular Contributor
  • *
  • Posts: 161
  • Country: se
Re: Do you start in a simulator?
« Reply #23 on: August 14, 2018, 02:07:52 pm »
I mostly use LTSpice for the analog parts of my circuits but most of my projects evolve around digital electronics like µC so no simulator is needed.
LTSpice is really good when you’re testing a completely new analog design and want to know if you’re in the right ball park!
 
The following users thanked this post: pilotchup

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: Do you start in a simulator?
« Reply #24 on: August 14, 2018, 03:24:51 pm »
for anything io , digital , microcontroller and simple analog : straight to board.

Only for simple stuff.

It is highly educational to simulate a multi-output digital system driving a moderate load. To pick a few figures:
  • a digital output, say 3V, with a series 100ohm resistor, 1ns risetime
  • one side of that digital output is connected to an inductor to ground. That inductor represents the inductance of the ic's ground lead, 1nH/mm so say 5nH
  • that resistor is connected to a transmission line, 100ohms, 1ns, and the other end of the transmission line is connected to a 5pF capacitor representing another ic's input capacitance

Do a transient analysis and look at both ends of the transmission line.

Then simulate eight such outputs switching simultaneously, by reducing the resistor to 12.5ohms.

If you think that is academic, I suggest you look at and understand the IBIS models published for digital ics - and why they need to exist.
Why would you want to do that ? That is bad design practice to begin with.
The driving IC output is already in the order of 20 to 25 ohms. So you are  mismatching your transmission line.
1nS rise time is NOT a GPIO of a microcontroller anymore. Your run of the mill micrcontroller can't even go there. Besides those are current limited and slope controlled these days.

If i need to make a controller for let's say a reflow oven or a plant watering system : there is no need to grab a simulator. that's straight to board.

Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 
The following users thanked this post: pilotchup


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf