diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Software Tools (http://www.diyaudio.com/forums/software-tools/)
-   -   Standard, platform independent simulations for DIYAudio (http://www.diyaudio.com/forums/software-tools/251710-standard-platform-independent-simulations-diyaudio.html)

spookydd 4th February 2014 06:26 PM

Standard, platform independent simulations for DIYAudio
 
I don't know of a repository of good tested models anywhere. If there is such a thing, I'd like to know about it, but I suspect that doesn't exist.

Perhaps this could be done on diyaudio with a wiki?

Maybe users could rate the models posted, so when they are checked out by many, they could be considered accurate and reliable.

And that could be the best place to post for model requests.

keantoken 4th February 2014 07:20 PM

1 Attachment(s)
I've tried to do that. Here it is. Open it in Notepad and read the comments to understand it. It's even got my 2SC4883A/A1859 models.

I asked people years ago whether anyone wanted to do a collaborative model collection thing. I don't remember there being any replies at all.

It would be nice if we would all add our custom and fully tested models to a single database accessible to anyone.

spookydd 4th February 2014 07:31 PM

Quote:

Originally Posted by keantoken (Post 3805991)
I asked people years ago whether anyone wanted to do a collaborative model collection thing. I don't remember there being any replies at all.

It would be nice if we would all add our custom and fully tested models to a single database accessible to anyone.

I'm all for that!

And I don't understand why nobody bothered replying to that.

This is needed, and I'd surely be a user of it.

Can this be done? Is a wiki on this site usable for this?

keantoken 4th February 2014 07:48 PM

I think it would be neat to try and set up a version control system and treat the model file as if it were the source code for some open-source software. This would streamline the collaboration. BUT, only people who understand version control systems would be willing to try that.

Trying to do this on a wiki will be a bit clunky, with everyone following their own rules, but it sure beats nothing.

spookydd 4th February 2014 08:00 PM

Quote:

Originally Posted by keantoken (Post 3806029)
I think it would be neat to try and set up a version control system and treat the model file as if it were the source code for some open-source software. This would streamline the collaboration. BUT, only people who understand version control systems would be willing to try that.

That was one of my reflections today. I was thinking about something similar to the wikimedia commons from wikipedia.

Perhaps something like this may be easier to use than a full version control system like cvs or whatever. But version control I think is a must.

Quote:

Originally Posted by keantoken (Post 3806029)
Trying to do this on a wiki will be a bit clunky, with everyone following their own rules, but it sure beats nothing.

There may be some other way that we haven't yet thought of. Maybe some people will propose other usable methods. If nothing is found, then a wiki would still be a plus for sure, and I second that thought that it beats nothing!!!

I think that it would be good to have feedback from users of those models, so enough people could ***** them and with a large number of people thinking a model is good, then perhaps it may be so.

In any case, there is nothing available out there (that I know of), so this would be tremendous!

keantoken 4th February 2014 08:37 PM

Quote:

Originally Posted by spookydd (Post 3806041)
***** them

You forgot an S. ;)

Here's another idea. What if we can streamline the process of triaging models or creating them from the ground up? There are several tools already available for modeling semiconductors - my big problem is I don't know how to use Excel/spreadsheets, or X program is command line, or archaic, or requires special configuration before it can even work and I don't know how to do that.

If we could get people who know how to use those tools and then make a how-to to streamline the process, maybe even use scripts to automate it a bit and do the configuration, many more people would be able to create and refine their own models.

If so, we would then have the input and cooperation of a number of users, maybe enough that we could create a standardized modeling process which was not any more complex than necessary.

spookydd 4th February 2014 09:07 PM

Quote:

Originally Posted by keantoken (Post 3806080)
You forgot an S. ;)

OOps!! Typo. This can happen when we type too fast and don't re-read what we typed.

Quote:

Originally Posted by keantoken (Post 3806080)
Here's another idea. What if we can streamline the process of triaging models or creating them from the ground up?

That I wouldn't know how to do.

Quote:

Originally Posted by keantoken (Post 3806080)
There are several tools already available for modeling semiconductors - my big problem is I don't know how to use Excel/spreadsheets, or X program is command line, or archaic, or requires special configuration before it can even work and I don't know how to do that.

But this I may be able to help some.

I'm on mac, I don't do windoze, so I have a unix below the hood, with a command line that I don't have any hesitation in using. OS-X also has X built-in, so this may be usable, and I use excel and I may be able to help with that too.

Quote:

Originally Posted by keantoken (Post 3806080)
If we could get people who know how to use those tools and then make a how-to to streamline the process, maybe even use scripts to automate it a bit and do the configuration, many more people would be able to create and refine their own models.

I'm all for that and I'm ready to learn how to do this.

A wiki can be a good place as a knowledge base for this.

Quote:

Originally Posted by keantoken (Post 3806080)
If so, we would then have the input and cooperation of a number of users, maybe enough that we could create a standardized modeling process which was not any more complex than necessary.

Count me in!

keantoken 4th February 2014 10:33 PM

Personally I'm using Linux with Wine for running Windows programs.

I think we should start simple with a format that we could expand to different electronics modeling contexts. For now, why not just post a model collection to start with and invite people to work on it?

I think this project should be started under the wider aim of standardizing simulations for DIYAudio. Bad models are a subset of this - ultimately what we want are standards, and a platform that will streamline sharing of accurate simulations so that we don't constantly have the problem of "are you using the models I'm using? Will the simulation get the same results on your computer as it got on mine? Will it even work on another computer?".

Kgrlee, Cordell, Edmond and the like are constantly sharing simulations so I think they would welcome a platform for collaboration and sharing that would eliminate the simulator wrangling so they could get collaborate transparently. If the system was made ready for them and then they were asked to participate, they may give it a try. It sounds like fun.

I don't really understand Git very well but here's my idea. We create a Git repository designed to be downloaded alongside an LTSpice installation that will integrate with the simulator, provide standardized model sets, and a directory structure containing public simulation projects which everyone would share. The system would be designed so that the path of least resistance would be to create a simulation that complied with all the standards, which would be the default setting.

I guess the public shared simulations would need to be separated from the project repository, so we would need to find another solution for that, but something similar to Git so that public projects were accessible by everyone and could be imported and updated by a user-friendly script.

Since Git is platform-independent IIRC, it would be accessible to Windows, Mac and Linux users.

Essentially what I'm describing is a hub for collaborative open-source simulation and electronics design, and it's basic form wouldn't be much of a step up from a version-controlled standardized simulation kit.

As far as I'm concerned, there is a catch. I don't want to spend much time on a project like this unless I know that I can walk away at some point and it will not fizzle out. If I am the only one that really wants it, then I would rather spend time on my other pursuits.

spookydd 4th February 2014 10:58 PM

Quote:

Originally Posted by keantoken (Post 3806232)
I think this project should be started under the wider aim of standardizing simulations for DIYAudio. Bad models are a subset of this - ultimately what we want are standards, and a platform that will streamline sharing of accurate simulations so that we don't constantly have the problem of "are you using the models I'm using? Will the simulation get the same results on your computer as it got on mine? Will it even work on another computer?".

Oh boy, this is even more than I imagined. But I'm all for it.

Quote:

Originally Posted by keantoken (Post 3806232)
Kgrlee, Cordell, Edmond and the like are constantly sharing simulations so I think they would welcome a platform for collaboration and sharing that would eliminate the simulator wrangling so they could get collaborate transparently. If the system was made ready for them and then they were asked to participate, they may give it a try. It sounds like fun.

Fun and very useful.

Quote:

Originally Posted by keantoken (Post 3806232)
I don't really understand Git very well but here's my idea. We create a Git

Unfortunately I don't know anything about it either. I only heard the name before, so I couldn't be much help in setting that up. But I'm sure there are other things I can contribute.

Quote:

Originally Posted by keantoken (Post 3806232)
Since Git is platform-independent IIRC, it would be accessible to Windows, Mac and Linux users.

That's a must! Platform independent.

Quote:

Originally Posted by keantoken (Post 3806232)
Essentially what I'm describing is a hub for collaborative open-source simulation and electronics design, and it's basic form wouldn't be much of a

I think the key point is "open source", and with enough collaborators in it, it doesn't need to rely on a single person or small group to stay alive.

If it's well put together, it should find enough interest and perpetuate.

AndrewT 5th February 2014 09:28 AM

What about handing this "project" to the Spice users' Forum?


All times are GMT. The time now is 04:17 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2