[Ktechlab-devel] feelin %)

[Alan Grimes]

I've been in an obsession mode the last few days. (see smiley in subject
line).

'been trying to get jfet working right to fix that glitchy circuit.

I've leaned down nonlinear.cpp... (The ideal for the entire project is
to simplify it to the point where it either works or it don't.)

One of the freaky things about jfet is that it uses diode models to
compute some of the PN junction stuff...

This got me to thinking, we're trying to deal with circuit elements, so
maybe the way to go is to work in ECjfet and have a basic jfet element
which controls the conductance and then implement the diodes as distinct
circuit elements. This would have the advantage of reducing the
complexity of the jfet class, if the thing worked...


I've also been trying to verify the code in inductance.

mnemonic device:

ELI the ICE man.

read as: "Voltage precedes current in an inductor" and "voltage follows
current in a capacitor". (learned that in a night school class...

But in matrix land everything is inside out... so in inductance you have
equations that take current and give voltage, and the reverse in the
capacitor...  freaky!

I'm trying to figure out how to extend it to support multiple tapings
and transformers...

The other freaky thing about the engine is that it doesn't do AC
analysis, instead it does a sequential DC analysis! =P

We do need to go to a database based parts library, there's no question
about that! but at the same time we should make an effort to design it
so that as much of this nodal analysis freakyness is hidden.

I've been going through the stuff on my HD for information about how to
model the electron tube but haven't found anything really useful...
(modified nodal analysis wasn't invented or wasn't popular when these
books were written! =P )

-- 
New president: Here we go again...
Chemistry.com: A total rip-off.
Powers are not rights.