The problem is models. 99% of the simulators out there come with zero models. And for 99% of the componetns out there you can't even get an accurate model !
i built a simple (4 opamps , a mosfet and some r's and c's) constant current sink. i used all LT opamps and even mosfet that was in LTspice. According to LTspice the thing would oscillate and have all kinds of unwanted behavior. on bench it worked just fine. built over 500 of those and not a single one exhibits the behaviors the simulator was showing me...
the trick ? having a ferrite bead shoved over the gate of the mosfet, correct grounding strategy keeping cross currents at bay and ohter PCb wizardry.
Good luck pluggin that into the sim ... The sim is only as good as the models you feed it. and sadly most models are inaccurate or simply wrong.
opamp in gain x 10, feed it 10 volts, and power it with 5 and -5 ... you'll be amazed how many sims show the output at 100 volts.. ( with 5 volt rails ? seriously ? )
Thanks for the insight! This is a very important point you have mentioned. Using the incorrect models and differences between grounding in the practical scenario vs in simulation causes a lots of differences between simulation vs practical results. Earlier I used to use a ORCAD student's version [it was not free but less costly]...it was good. Currently I use LTSpice a lot and TINA sometimes. I like LTSpice over TINA, but as you have mentioned, I tend not to believe just on the simulation results but also to make a quick bread board circuit and scope to check if it would work as I expect it to.
You don't always need the model. You might only care about the component tolerance. It depends upon the level of accuracy you are seeking. You can also use a spreadsheet for some calculations. Sometimes that's easier than SPICE. If you have a lot of unknows in you system, then should you spend a lot of time on the models?
Thank you for the advice. True...if I have lots of time, usually which is not the case, then a thorough investigation about the models would be possible. In general I tend to simulate to check the concept would or would not work. After ordering the prototype board and before those arrive, the time could be utilized for detailed simulation to be prepared for the mods on prototype...to avoid surprises and frustration :)
Without some good test equipment all the modelling software in the world is useless if you can't prove your designs.
For free software: An old copy of ansoft serenade RF designer whichwas bundled with a copy of an antenna text book from the ARRL, LTspice and the excellent free antenna modelling program 4NEC2.
Hardware: This along with a Bird RF power meter and slugs, na HP465 precision RF power meter/with head and suitable attenuators, an old analog TEX 465 scope, a HP8558 spectrum analyzer, a HP8640 RF generator with built in freq counter and high power option and an MFJ antenna analyzer along with the solder iron mentioned previously by another individual is all I need for anything from HIGH power RF amplifiers, large multi-element external antennas to small etched internal antenna designs.
I have tried Circuit Lab a few times. It is easy to use and for small circuits and since it includes schematic capture it seems just perfect. It runs in your web browser too which is another great feature...
Thanks DrFPGA! CircuitLab is now the de-facto circuit simulation software of choice at hundreds of universities around the world, with students and faculty telling us that we are swiftly replacing LTspice, PSpice, and Multisim for the next generation of EEs. We've also have a growing number of professional and hobbyist users on our non-free but affordable subscription plans. All in all, the CircuitLab community runs hundreds of thousands of simulations every month!
I've used LTSpice almost exclusively for analog simulation for the past several years. Including at work where a lot of us prefer LTSpice to PSpice (to the point where the company has decided to cut down on the number of PSpice seats despite the increase in the number of engineers)
I'm looking forward to learning more about the other simulation software you mentioned.
There is a glaring omission here. You assume nobody wants to use free software to check their PCB? I check things such as trace impedance, trace differential impedance, trace fusing current, trace dc resistance, resistance increase with temperature, estimate power distribution impedance, via current carrying capacity, via impedance and other things. Analog engineers look at their PCB characteristics too. I use Saturn and Ultracad freeware for this but we have an in house tool for traces impedances.
Is there a freeware equivalent of Hyperlynx, or would that be asking too much?
Of course, then we would never pay for the good stuff that works and has lots of features we can never understand because we can't attend the suppliers training program. Mentor does have online training and I have watched the videos over and over. They are still lacking.
There are many free circuit simulation software,it would be very huge to list all kinds of software.I think it maybe good to have a place for people to share a free and useful software。
I have tried EasyEDA, I think EasyEDA is the just software which I try to find .
I thought people might be interested in EasyEDA for the following reasons.
ii) work can be shared and - with a bit of agreement on who saves what, when via a 3rd party instant messaging service - it supports collaboration;
iii) it has some pretty impressive import and export options such as importing Altium, Eagle and LTspice files and exporting spice netlists and svg, png and pdf files for documentation. It also uses open standard JSON format files internally so it's easy to do some quite whizzy things just using a text editor;
iv) all use of the tool is free even to the point where the user can download PCB Gerbers and send them off to any PCB supplier;
v) it also offers a low cost PCB service: PCBs can be ordered directly from within the tool (this seems to be the way the money is made to keep the whole thing afloat);
vi) it can even be used without registering with files being saved in an Anonymous mode that works a bit like pastebin where files can be retrieved and shared just by their urls.
I use LTSpice for analog simulation and Icarus Verilog coupled with GTKWave for digital simulation. GTKWave is a top quality waveform viewer, and can compete in useability and features with the best commercial tools.
Icarus Verilog is a very complete and stable free Verilog simulator. It's a bit slow for large projects that require long time simulations, but totally adequate for many projects.
Another free simulator that I've got a lot of mileage with is the Verilator (Perl program), with turns synthesizeable Verilog into C++. If you have the need for speed, this tool will outperform the even the most expensive commercial simulator. Using the C++ model of the digital circuit that is output by the Verilator, you can model the external circuitry in C or C++. You can even make a DLL out of the model + behavior circuitry and call it with a GUI, scripting language, or whatever.
"Verilator should run any system with GCC and Perl. It is developed on 64-bit SuSE 9.3 and other users report success on Redhat Linux, HPUX, Solaris, and Windows NT under Cygwin (C++ only, no SystemC), and Microsoft Visual C++" And, I believe, OSX.
Have you considered TI's WEBENCH Designer for the analog category? This is a free on line circuit design tool which includes design synthesis, component selection and Spice simulation. It also has thermal simulation and the ability to download the design into popular CAD tools. You can get to it at http://www.ti.com/webench.
PowerESim is a new CAD tool on the Internet for Switching Power Supplies http://www.poweresim.com/ . I heard about it quite recently but it seems to be an interesting tool. Another interesting EE FEA tool is Quickfield from http://www.ocsimize.com/ because it interfaces with LTSpice so it is quite handy.... Best regards, Jgo