My brain is resistant to learning circuit analysis. Get it? Resistant? Like a resistor? Ahhh hahaha! Help.

October 31, 2012 3:11 PM

I'm struggling in Electric Circuits 1 - particularly with analyzing series-parallel circuits, thevenizing and superposition. Before I look for a tutor, can anyone suggest some resources to learn it myself?

When I'm trying to analyze a circuit, I don't know where to start. I have a myriad of formulas and rules, but I don't know which I should apply to what, and in what order. There must be a step-by-step process for breaking it down and sorting through it, right? Or maybe it's just a question of bashing my head into it from different angles until I hit on the right one? My (universally disliked) professor puts an example on the board and then quickly solves it without explaining any kind of a process. I asked him for some help during our last lab and he just took my pencil and wrote down the answer. I said, "Give me a chance to do it on my own or I won't learn," and he just kind of chuckled.

Any suggestions?

posted by SampleSize to education (5 answers total)

This page breaks the process down pretty nicely:

http://electron9.phys.utk.edu/phys136d/modules/m6/kirchhoff.htm

And you can check your answer by building the circuit with this simulator and checking the voltage drops and currents. (Or you can just blow up batteries when you get too frustrated.)

http://phet.colorado.edu/en/simulation/circuit-construction-kit-dc
posted by BrashTech at 3:31 PM on October 31, 2012

Here's another one -

Wikiversity- Elec Engrg.... and from that,
Wikiversity- Elec. Circuit Analysis

Hope that helps!
posted by drhydro at 3:45 PM on October 31, 2012

I'm a junior in Computer Engineering , and let me just say that's it's totally normal to look at unfamiliar circuits and have no idea where to start. (For me, at least. Perhaps my classmates never have trouble. :) )

All circuit analysis is basically two steps:

1. Write down equations that describe the circuit using Kirchhoff's Voltage law (KVL) and Kirchhoff's Current Law (KCL) and the I/V relationships of the circuit elements.

2. Solve the resulting system of equations. (Sometimes you're lucky and it's just a single equation, so solving it is quite simple.)

The trick with #1 is that there are a lot of different ways to solve a circuit. For example, you could write a set of equations that deal with the voltages at different nodes in the circuit, and solve it in terms of that. You could write a set of equations with deal with currents through different elements, and solve it in the terms of that, and get the same answer.

There is no trick with #2, you just have to do a lot of algebra. Sorry.

In general, I tend to pick the approach that leads to the smallest number of variables. For example, if there's some voltage source, and then 3 or 4 elements in series, and I'm being asked for, say, the voltage drop across one of the elements, I'm going to tend to solve for the current, because it's the same through all of the elements, and then use the I/V relationship (eg, for resistors, this is V=IR) of the element in question to get the voltage drop.

Alternatively, if there were a handful of elements in parallel, I might tend to solve for the voltage across them, since it would be the same for all the elements.

So, as you get more experience, you'll see why one way to solve it is easier than another and you'll develop an intuition as to what approach you should use.

But in the meantime, if you just don't know where to start on a problem, I'd say just start writing down everything you know, and hope a solution presents itself. By "everything you know" I mean, apply KCL where you can. Apply KVL where you can. Look at the resulting equations. Do you see some way to manipulate them to solve for the result that you want? What quantities are unknown? Can you use another relation that you know to solve for that, and then use that to find this other thing? etc.

Also remember that a lot of "tricks" you can use to make the analysis easier (replacing sources with a Thevenin or Norton equivalent, replacing elements of the same type that are in parallel or series with an equivalent, etc) are not always necessary to solve the problem. You could solve it without those things. It might be harder, but it can be done. In fact, I think it's instructive to, at least once, solve something "the hard way" just to convince yourself that it's the same and to see where the simplification comes from.

For example, if you're used to replacing parallel resistors with an equivalent, just once try it without that. You'll end up with more currents to solve for, and it will be more unwieldy, but you will get the same answer.

So don't worry so much about starting the problem "the right way". There are a tons of right ways to solve every circuit problem. Some of the right ways are easier than others, of course, but it's difficult to give hard and fast rules about which way is easiest.

To summarize: There is no silver bullet. Practice, practice, practice. Algebra sucks.
posted by jcreigh at 5:28 PM on October 31, 2012

Finding points of equal potential and drawing lines between them can be a useful trick to keep in the back of your mind as well. If two points are at equal potential then no current will flow through any new connection between them, which will therefore leave the circuit operation undisturbed; however, drawing those connections can sometimes let you simplify your algebra using the parallel-resistors formula.

The canonical puzzler that demonstrates this technique: consider a wireframe cube, each of whose edges has a resistance of 1 ohm. What is the resistance between any vertex and the one furthest from it?

Tackling that one the straightforward way involves a painful amount of algebra. Adding connections between points that symmetry shows must be at equal potential reduces the problem to finding the resistance of (three 1 ohm resistors in parallel) in series with (six 1 ohm resistors in parallel) in series with (three 1 ohm resistors in parallel).

There is at least some tiny chance that your universally disliked professor has not already seen this puzzle.
posted by flabdablet at 7:19 PM on October 31, 2012

How about Khan Academy?
posted by Dansaman at 10:14 PM on October 31, 2012