[Ktechlab-devel] simulator

[Glen]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


There are a few decent tube equations out there.. I don't see why they'd
be to hard to inherit to use / modify. I have no idea if anyone's found
a better way to model them that doesn't involve iteration, though!

I DO want to make parts inheritable... there's no reason not to. Who
else does that?

>I partly agree here. The point why I see KTechLab in kdeedu is, because
>of
>it's really easy to use interface. It's ready to be used in (higher)
>education. Engineers do want something more practical here, don't they?
>Some
>something that plays together with all the professional tools out >there.

Easy-to-use doesn't have to mean toy or educational-use-only ;) Nothing
wrong with professionally using the same software you learned a trade
on! KTechLab has a better interface than the vast majority of apps in
the same vein. It seems that professional-quality is where we're looking
to take it and I'm all for that! Maybe one day we'll spearhead a new KDE
app category.

I suppose one thing I should do is get feature lists from all of the
favorite simulation apps and compare what we already have to it. I just
wish that publishing web pages was simpler. I'm a horrible web designer.

I'm upgrading to ubuntu 8.10 as I write this to make dealing with the
KDE stuff better and to fix a few things I've torn apart. I had KDE 4.0,
not 4.1 Classic wasn't available in 4.0 and that's a large part of why I
wasn't enjoying the experience.

- --G

(why is it I write books in response, delete them, then rewrite
different books and send?)

Alan Grimes wrote:
>> I partly agree here. The point why I see KTechLab in kdeedu is, because of 
>> it's really easy to use interface. It's ready to be used in (higher) 
>> education. Engineers do want something more practical here, don't they? Some 
>> something that plays together with all the professional tools out there.
> 
> I see the primary use case for the parts library as: you have a
> datasheet, the part isn't in the library, you create a new part, inherit
> the differential equations from a reference part, then add the
> parameters from the datasheet, then try out your part in a few test
> circuits to make sure that the curves it puts out are in the ballpark,
> then party on...
> 
> A slightly less common use case is that you have a part that nobody's
> even heard of, say the 71A. Lets say that there isn't even a DHT model
> yet, so you create a new class called "DHT" then using a combination of
> elemental properties (resistence, transconductance, capacitance) begin
> to rough in a model for the part, then you'd have to go find the old
> mathematical equations used to design the damn thing, enter those in,
> and then try it out...
> 
> 
> A more general issue in the simulator, is how nonlinear properties are
> handled... For example, the grid of a typical tube will exponentially
> begin to source a (classical) current to the cathode of the tube as its
> voltage approaches and exceeds that of the cathode... Everyone knows how
> to model linear stuff with matrices... I wonder if there is a similarly
> elegant/generalizabile technique for simulating nonlinear components so
> that the current iterative approach can be avoided...
> 
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkkWExoACgkQLstl3vProOBA8wCeIxjqPtqeGlt1oiElF0DZvtTk
xuIAn2l9vGmKY4M+7HHkmsNxsd8eeBFB
=mFAC
-----END PGP SIGNATURE-----