These original Kirt's Cogitations™ may be reproduced
(no more than 5, please) provided proper credit is given to me, Kirt Blattenberger.
Please click
here to return to the Table of Contents.
Cog·i·ta·tion [koj-i-tey'-shun] – noun: Concerted
thought or
reflection; meditation; contemplation.
Kirt [kert] – proper noun: RF Cafe webmaster.
You have probably seen somewhere along the line in your electronics career the resistor cube problem. The 12 edges
of the cube each contain a 1 Ω resistor, and the challenge is to calculate what the equivalent resistance is
between two opposing corners. It is a daunting problem using straight circuit analysis, since it requires writing
and solving multiple mesh equations. There are lots of opportunities for making mistakes.
One option if you
had the time and facilities would be to build the model in a circuit simulator and let it determine the result.
Usually, though, the cube is thrust upon you in a compromising situation, like in a job interview. If you are an
electrical engineer and cannot figure it out on the spot, forget that circuit design job. If you are an
electronics technician, you will be forgiven for not solving it, but you had better demonstrate an understanding
of the method once it is presented.
As it turns out, there is a relatively simple analysis based on
symmetry and a fundamental level of understanding of currents and voltages.
The
traditional method used involves recognizing sets of equipotential
points within the vertices of the cube, then shorting them together to enable calculation of parallel resistances.
Finally, those resistances are added in series to arrive at the resulting equivalent resistance. The process is
illustrated below.
After explaining the traditional method, I will present
my solution, which is a little more intuitive and direct method for arriving at the same answer. Solving via
the traditional method actually requires the same knowledge of how currents are divided at nodes.
Finally,
LTSpice is used to arrive at an answer via a Spice-based
circuit simulator.
This is the cube structure consisting of 12 resistors electrically connected between the 8 vertices. Each
resistor is 1 Ω, but any value can be used so long as they are all the same.
Here is where the intuition comes into play. Color coding is used to help keep track of the resistors and
associated nodes (below). Due to symmetry, the potential (voltage) at the three nodes labeled "
α"
are equal. Since no current flows between nodes with a potential difference of 0 V, they can be shorted together
without affecting the circuit's integrity. The same can be done for the nodes labeled "
β."
Once you short those nodes, you obtain the equivalent circuit shown below. As you can see, there are two
sets of three resistors in parallel, in series with one set of six resistors in parallel. So, you have 1/3 Ω in
series with 1/6 Ω in series with 1/3 Ω, which equals 5/6 Ω.
Now I will present my method of solving the resistor cube problem. The structure is repeated again here.
Kirchhoff's current law, which states that the sum of the currents entering and exiting a node is zero, is
essential in the analysis.
The first step is to recognize that at a node where equal resistances exist,
current entering the node will be distributed equally between the number of output branches - in this case three.
For convenience sake, I assigned an input current of 3 amperes at the corner labeled "A," so that 1 amp will flow
through each output branch. Note that 1 A flows through each branch.
On the far side of each of those
branches is another node with two output branches. Again, due to symmetry, the input current will divide evenly so
that ½ A flow into each branch. Looking at the cube's output node labeled "B," it is apparent that the same
situation exists as with "A."
Take a moment to sum the currents into and out of each node to verify that
they all add up as required.
Now that you know the current through each branch, and you know that each branch has a single 1 Ω resistor
in it, Ohms law allows you to calculate the voltage across each resistor.
The next step is to sum the voltage from input node "A" to output node "B." Any path you take travels along
three edges, and all total to 2½ volts.
Finally, apply Ohms law, which says that the resistance is equal to the voltage divided by the current. As
with the other analysis method, the resulting equivalent resistance is 5/6 Ω.
You can see that in reality, being able to make the assumptions in the traditional solution requires an
understanding of the current division principles in my method. So, IMHO it is simpler to add voltages and then
plug voltage and current into Ohm's law to arrive at the answer than to risk shorting nodes incorrectly. It's so
simple, a caveman could do it.
As a verification of the result, the resistor cube circuit was simulated using the free
LTSpice program, by Linear
Technology. Resistors are labeled in accord with the labels in the
traditional method of analysis. A 3 amp current source is placed at
input node
N001. The resulting voltage is the predicted 2.5 V. Again, Ohm's law for 3 amps and
2.5 volts yields a resistance of 5/6 Ω.
R§b
N001 N003 1 Analysis Report:
V(n001): 2.5 voltage R§d N002 N004 1
R§a
N002
N001 1
R§f N004 N003 1
R§h N005 N007 1
R§l N006 0 1
R§i N006 N005 1
R§j 0 N007 1
R§g N003 N007 1
R§k N004 0 1
R§e N002 N006 1
R§c
N001 N005 1
I1 0
N001 3
.op
.backanno
.end
Posted 6/11/2010
Thanks to RF Cafe visitor
Les Carpenter for sending me this
solution that rearranges the resistors in delta and star configurations in order to "simplify" the solution. My
head is still hurting from looking at it.
Posted 6/27/2010