Voltage of ideal diodes in parallel?

[Carultch]

So I guess my main point is that with non-linear elements, we can't guarantee network analysis will give us a correct answer.

We can, if we have a network analysis that accounts for the exact nature of the IV curve, as Junkhound's circuit model did.

 

[junkhound]

We can, if we have a network analysis that accounts for the exact nature of the IV curve, as Junkhound's circuit model did.

Interesting how a quote within a quote comes out attributing the "we can't guarantee network analysis will give us a correct answer". to me vs. to Steve. AND, how Carultch reply to Steve appears to be attributed to Steve !


FWIW, agree with Carultch.

Here is a more complex type model as an example, usually refine a model until it perfectly matches test data, then one knows the model is good and then can tweak it to find out why failures occur over time and temperature and tolerance changes - or due to design errors due to packaging or board layout.

Phase shift resonant full bridge converter model example - will produce pulse by pulse circuit responses. A very nice tool to find design problems.

The waveforms below are an example of FEA capability. The Schematic reproduced the full bridge circuit, and stray coupling paths could be added in.
The blue curve was the initial model response for the full bridge waveforms (upper and lower should be symmetrical in a good full bridge circuit) , and the red curve the actual test data. The model then is revised for the blue curve to match the red curve, and the circuit changes due to poor wire layout in a VFD package can be found and corrected.

 

[steve66]

Interesting how a quote within a quote comes out attributing the "we can't guarantee network analysis will give us a correct answer". to me vs. to Steve. AND, how Carultch reply to Steve appears to be attributed to Steve !

Well, if you come to an internet forum, someone is going to try and put words in your mouth sooner or later.

Actually, I think that started with me and my post - it has an extra quote tag at the top, and it seems like it got copied through the next two posts.

As a moderator, I can edit anyone's post, and for some reason they put the "edit post" button right next to the "reply with quote" button. Sometimes I hit the wrong one by mistake, and the next screen looks very similar in either case. I wonder how often I have edited a post instead of replying to one without even realizing it.
So it could be worse

 

[steve66]

We can, if we have a network analysis that accounts for the exact nature of the IV curve, as Junkhound's circuit model did.

Yes, but that wouldn't be the standard network theory taught in school. My point was that Sahib's analysis failed because he tried to apply standard network theory to a non-linear circuit.

I'm sure you can massage the rules and add to them with some caveats and such, but again, that's not standard network theory.

Standard network theory is a set group of analysis methods guaranteed to give correct results on circuits that meet the limits, one being that all components are linear.

With the diode example, the superposition theory won't work. That theory is;

"In a linear circuit with several sources, the current and voltage for any element in the circuit is the sum of the currents and voltages produced by each source acting independently."

And in that simple context, there is no way to account for the non-linear diode.

 

[steve66]

I get it now: Junkhound and Carultch are using a slightly different definition for "network analysis" than I am.

I assumed Sahib was talking about network analysis as engineers are taught in School. And it is generally limited to linear circuits.

Now I believe you are talking about network analysis as a solution method computers use. Yes, I agree, that type of network analysis totally works with non-linear components.

Do I dare mention that that type of analysis only gives an approximate answer, and not an exact answer? Sure, you can make it as accurate as you want (even to the millionth decimal), but its still not the exact answer network theory gives. Not an issue for any real life problems, but its the type of thing theoretical mathematicians get worked up about

 

[Besoeker]

I get it now: Junkhound and Carultch are using a slightly different definition for "network analysis" than I am.

But we have no information about the network from the OP.