.99,
SPICE Simulators don't lie, they produce results based on what you give them. If your source data is bad, then the results will be bad. If your source data is good, the results will be good. Source data being the models, and the circuit and sim settings you have input.
We should not forget the capabilities endowed and methods used in the application by the software designer
That is bandwagon science.
As my hair continues to fall out and loose color, I'm learning much of science is bandwagon science.
All,
I’m going to paraphrase myself from several years ago (Why not? Over the past couple of weeks I’ve found modified or strangely similar versions of my original 1996 ideas dealing with design procedural flaws in power distribution design).
The good thing… modified KVL in SPICE (Simulation Program with Integrated Circuit Emphasis) allows less rigorous analog or digital folks quicker and easier circuit analysis. The problem is, any flavor of KVL does not explain or allow us to understand what really happens in the circuit (especially confusing for those well familiar with Faraday’s works). The reason is that physics doesn’t allow an equivalent circuit where the incorrectly applied KVL works. It can't work, because moving the meter physically, without changing the connection points, changes the measured voltage. Digital and poorly informed analog circuit theories do not allow this effect. Faraday does.
This can be demonstrated with any simple loop circuit where two or more loop sections act as secondaries of a transformer with resistors between. Attach a single meter to what we have been calling A & D using connections which allow rotation of that meter and meter circuit around the horizontal axis of the circuit loop. Even if SPICE could handle spatial movement it would fail. As far as I know, Faraday’s laws are not included as an algorithm. Only Faraday's Laws (applied with the meter & meter loop left, right and above or below) can answer that question. The meter will read differently in all four of those positions and every position between without changing the meter’s connection points.
All, without swapping connections or meter polarities!
KVL/KCL only pertain to static situations. They are extremely important and valuable but only fit as a subset of the Laws of Faraday when working with non conservative fields. They are only correct when looking for electric field values (NOTICE: I did not say ‘EMF’. If you don’t know the difference that is the first problem you MUST address.
)
In the end, the functions of the ‘Real Circuit’ will obey Faraday’s laws because EMF falls in the realm of Faraday’s work, not Kirchhoff, Faraday trumps Kirchhoff, and the Laws of Physics demand obedience. SPICE, as I know it, will fail to provide accurate results. It was not designed to handle the more wild issues possible. Hell, the first time I used it, the voltage drops across inductors was correct. They were handled as shorts.
KCL/KVL was only attributed to Kirchhoff during his eulogy and thereafter . You’ll have to research his history to find this info.
I’m not going to put more time into this as the result is always the same. The statement will be repeated “KVL HOLDS!” when it doesn’t even apply. When I see that, it registers in my mind as ‘Faraday’s laws are invalid for this experiment’ . Let’s leave the issue where Faraday’s laws fall apart to a much later time.
When I see statements about ‘voltage drop across inductors’, I know Kirchhoff didn’t consider time so the speaker is confusing the issue. Then, I await the use of the ‘inductor/spring analogy’ contradicting the use of ‘inductor voltage drop of the electric field’.
.99,
May I ask what flavor of SPICE you are using?
If it is open-source I would be interested in seeing the actual algorithm used to simulate and inductor.
This was an easy one:
http://www.compliance-club.com/pdf/Issue82.pdfI would think the folks at Agilent Technologies know a thing or two about network analysis.
Attached is a clipping from that document.