Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Electronic Circuit Simulator in the Browser

falstad.com

Post reply on HN
Print view View original

101–110 of 137 posts

Re: Electronic Circuit Simulator in the Browser

#101

Post by slededit »

kmm wrote: post #49
Tade0 wrote: post #14

I remember when this was still a Java applet - nice to see that it has been since rewritten in JS. I credit this app for being able to pass the first half of my sophomore year at University - just months before I failed four courses(the allowance was two) and was given a chance by the dean to redeem myself.

I was really happy to see it got rewritten in JS as well, because Java applets are a pain (or just plain impossible). Sadly, the JS version is also a lot slower, I don't know how that could have gotten prevented

Webassembly tends to help by a good margin.

Re: Electronic Circuit Simulator in the Browser

#102

Post by avip »

AYBABTME wrote: post #82
sleavey wrote: post #40

I'll throw my hat into the ring as well. I have also recently been building a small signal AC analysis circuit simulator [1]. The main idea behind it is to convert a closed source tool that has been around for years within the LIGO scientific collaboration to one that is open source and implemented in Python. It's almost functionally complete, can simulate transfer functions and (accurate) noise from op-amps, resisto…

Looks pretty cool but the input is pretty esoteric to a newbie like me: https://git.ligo.org/sean-leavey/circuit/blob/master/example... Same thing about the output format: https://git.ligo.org/sean-leavey/circuit/blob/master/example... Is this a custom file format or a well known one? Also, the references to LISO are all behind a user/password wall, which makes it hard to learn: https://wiki.projekt.uni-hannover.de/a…

It has a familiar Fortran/Academia/FEM solver vibe. This is very, very likely a format inherited from a Fortran code.

Re: Electronic Circuit Simulator in the Browser

#103

Post by ozmaverick72 »

compumike wrote: post #18
Hupriene wrote: post #5

I used and enjoyed circuitlab.io until it went for-pay only. Nice to have an easy to use free alternative for very casual hobbyists like me.

That would be https://www.circuitlab.com/ -- I built it :) Edit: if anyone wants, reply with your CircuitLab username and I'll give you a free membership upgrade when I get home this evening.

mine would be: lonewolf72. Thanks heaps

Re: Electronic Circuit Simulator in the Browser

#105

Post by illuminati1911 »

compumike wrote: post #18
Hupriene wrote: post #5

I used and enjoyed circuitlab.io until it went for-pay only. Nice to have an easy to use free alternative for very casual hobbyists like me.

That would be https://www.circuitlab.com/ -- I built it :) Edit: if anyone wants, reply with your CircuitLab username and I'll give you a free membership upgrade when I get home this evening.

Username: illuminati1911

Thank you! :)

Re: Electronic Circuit Simulator in the Browser

#106

Post by RandomException »

compumike wrote: post #18
Hupriene wrote: post #5

I used and enjoyed circuitlab.io until it went for-pay only. Nice to have an easy to use free alternative for very casual hobbyists like me.

That would be https://www.circuitlab.com/ -- I built it :) Edit: if anyone wants, reply with your CircuitLab username and I'll give you a free membership upgrade when I get home this evening.

Nice! This looks really cool :)

Username: RandomException

Re: Electronic Circuit Simulator in the Browser

#107

Post by sleavey »

AYBABTME wrote: post #82
sleavey wrote: post #40

I'll throw my hat into the ring as well. I have also recently been building a small signal AC analysis circuit simulator [1]. The main idea behind it is to convert a closed source tool that has been around for years within the LIGO scientific collaboration to one that is open source and implemented in Python. It's almost functionally complete, can simulate transfer functions and (accurate) noise from op-amps, resisto…

Looks pretty cool but the input is pretty esoteric to a newbie like me: https://git.ligo.org/sean-leavey/circuit/blob/master/example... Same thing about the output format: https://git.ligo.org/sean-leavey/circuit/blob/master/example... Is this a custom file format or a well known one? Also, the references to LISO are all behind a user/password wall, which makes it hard to learn: https://wiki.projekt.uni-hannover.de/a…

Thanks for taking a look!

> the input is pretty esoteric to a newbie like me

Yeah, those input and output files are what LISO uses. This was developed between around 1995 and 2010, and was built mostly to solve specific problems the original author was having with circuits and not as a generic tool, hence the syntax. You can avoid using this syntax at all as everything gets parsed into Python objects - see e.g. [1].

> Is this a custom file format or a well known one?

Custom. It's similar to a SPICE input file, but not identical. The author told me that he didn't know about parsing algorithms like Lex/Yacc when he wrote it, otherwise it might have been nicer.

> the references to LISO are all behind a user/password wall

Sorry about that. As I said, this was a tool used internally within the collaboration for many years. I am working on getting at least the manual made publicly available, but there are some proprietary libraries with unclear licences preventing public release of the binary and source code. Ideally, eventually my Python project will replicate all of the features of LISO and we won't even need the old binary any more.

EDIT: if you like, I can send you the manual and binary for LISO to the email in your profile - let me know.

[1] https://git.ligo.org/sean-leavey/circuit/blob/master/example...

Re: Electronic Circuit Simulator in the Browser

#108

Post by sleavey »

zokier wrote: post #74
sleavey wrote: post #40

I'll throw my hat into the ring as well. I have also recently been building a small signal AC analysis circuit simulator [1]. The main idea behind it is to convert a closed source tool that has been around for years within the LIGO scientific collaboration to one that is open source and implemented in Python. It's almost functionally complete, can simulate transfer functions and (accurate) noise from op-amps, resisto…

Is there some fundamental differences between this and the more established simulators like (ng)spice and QUCS? Trying to understand the rationale of building your own

When LISO (the tool my project is based on) was first made, SPICE simulators couldn't properly handle "realistic" op-amp parameters, like finite open loop gain, strange phase shifts due to output impedance especially of FET-based op-amps, realistic noise, stability of circuits containing feedback loops, etc. Furthermore, SPICE models provided by manufacturers typically don't represent the real behaviour of their op-amps, especially when it comes to noise, so the idea was to build a tool to which you can give arbitrary, perhaps measured, noise spectra.

LISO also contains a brilliant fitting routine which can take a measured transfer function or noise spectrum and fit circuit parameters (e.g. op-amp noise, resistor values, etc.) to the data. It first uses some standard fitting algorithm (e.g. Nelder-Mead) to get close to the possible minimum, then uses a custom optimisation in log space to get the global minimum. If I remember correctly the author published a paper on this, and I can try to find it if someone is interested.

Re: Electronic Circuit Simulator in the Browser

#109

Post by sleavey »

avip wrote: post #102
AYBABTME wrote: post #82

Earlier quoted context omitted.

Looks pretty cool but the input is pretty esoteric to a newbie like me: https://git.ligo.org/sean-leavey/circuit/blob/master/example... Same thing about the output format: https://git.ligo.org/sean-leavey/circuit/blob/master/example... Is this a custom file format or a well known one? Also, the references to LISO are all behind a user/password wall, which makes it hard to learn: https://wiki.projekt.uni-hannover.de/a…

It has a familiar Fortran/Academia/FEM solver vibe. This is very, very likely a format inherited from a Fortran code.

Ah, that might be the inspiration. The author used FORTRAN for other projects.
Post reply on HN