Hello again,

Hey that's pretty interesting. I'll have to look into this more myself too.

I see already that the reviewer chose the most extreme circuit to illustrate his objection to the theory by Leach, but it does show that it will not work all the time and that seems to be what his whole goal was. I cant say i agree with his view however because i've used this idea so many times in the past without realizing there was a theory out there on this method. Therefore i must think that it works in so many cases and there would be less cases where it does not work. There will be cases where it does not work, that's without doubt, but as you already implied there is a criterion that must be met and it's not that hard to figure out, so it is worthwhile knowing i believe.

I seem to remember coming up with the rule that if the dependent source is shorted out and all the node voltages go to zero, then it can not be shorted out, but i am not sure how far this can be applied at present. I do like the idea about the 'sense' voltage or current not being able to go to zero though, that's probably the better idea.

I think we also have the oscillator circuit, where if we short the dependent source (the amplifier) there's no output and thus no input and thus no response when clearly there should be.

Perhaps we can look into this more by considering some more circuits.

LATER:
I can see something is up with this topic here. I found an example in a pretty good text where there is one independent voltage source, one independent current source, and one current controlled voltage source, and the text reads:
"Superposition may be used to analyze this circuit by first replacing the 3 amp source by an open circuit nd then replacing the 10 volt source by a short circuit. The dependent voltage source is always active (unless ix=0)".

However, using straight up superposition with all three sources also leads to the CORRECT solution as Leach would have found. This tells me that authors tend to lay back on older theories that had already discounted certain ideas and will find ways to hold on to those theories regardless of any reasonable contradictory argument. It is unfortunate that this kind of social behavior is found in many branches of science, and i believe in itself it is contradictory to the scientific method (ie social behavior outweighing scientific behavior), and i believe it also holds back progress.

A while back, i found an interesting method to analyze a circuit where there are (at least) two voltage sources. You'll find this hard to believe, but the two ideal voltage sources can be combined in parallel to form a SINGLE voltage source. I know it sounds hard to believe, but it works as long as we follow a simple rule or two, and as we all know one voltage source is simpler than two voltage sources. That reviewer would literally gawk at this concept
It just goes to show as we see all the time in physics, theories can be very very strange to the common experience, but as long as the calculations work out we have tools to use that allow us to solve real life problems.

 

Marshall Leach's paper is certainly enlightening and filled with examples.

The example from the peer reviewer does have problems simulating ... until you have a shunted resistance across the two current sources and the sources are not equal.

 

Hello Joe,

Did you use a dependent current or voltage source for g*v1 in the reviewers example? I used a dependent voltage source but i see he has an arrow drawn next to it. I think it's too demanding to use a dependent current source, and a dependent voltage source seems to illustrate ok.

 

I used a dependent current source. The two independent current sources were shunted with a 1 M resistor internally. When using the same value for the independent sources, there is no voltage drop across the sensing resistor.

I got the irregular circuit error without the shunted resistors, but that is normal when having series or parallel perfect sources.

 

Hello again,

Well if you look at the circuit you'll see what i mean. If we have two constant current sources like that in series with another current source, then the third current source MUST equal the sum of the other two currents. So if one is 1amp and the other is 2 amps, then the one on the bottom must be 3 amps exactly. That's too much of a constraint i believe, but that's not the end of it. Because that bottom controlled current sourse with the control being a voltage in anther part of the circuit, that voltage must be exactly equal to this 3 amps divided by the 'gain' of the source, And if that wasnt enough, these currents MUST be also able to create the right control voltage in the 'sense' resistor. Because of these facts, i believe the reviewers circuit should be considered invalid because there are too many constraints which i dont believe exist in real circuits.
For example, a current source in series with a current source. That makes no sense either, even if both current sources are both the same value.
What this suggests is if the reviewer wanted to do such an extreme circuit then there are simpler examples out there.
The problem still persists i think though, because if we change that lower source to a controlled voltage source we then have a valid circuit to argue about.
Also, the arrow he drew could have been a voltage arrow not a current arrow, but because he indicated the control voltage with + and - it seems he might have really meant a current arrow. But also some authors use both because it is quite common to use + and - symbols when showing the controlling voltage polarity.

Using shunt resistors is invalid i believe because then we change the circuit too much and then do not properly illustrate the problem anymore. The shunt resistors take away the extremely strict criteria explained above for one thing, and they also change the currents quite a bit. In addition they cause other currents to flow which are not supposed to be there.
In fact, without the resistors we might end up with a singular matrix, which would prove that his circuit is invalid to begin with.

In fact, if I2 is greater than I1 and both are positive, then the sense voltage becomes negative, and with a positive gain for the controlled current source the current through that source would have to be negative, which just doesnt work because then we would have something like:
3=-3
which can not ever work.

Changing the lower source to a dependent voltage source does create a valid circuit though.