-1
\$\begingroup\$

Here is very interesting discussion on this topic: Proof that every circuit with diodes has exactly one solution

And this is my example:

enter image description here

Let's start. We have 3 generic diodes with cut-off voltage = 0.7 V. Voltage Vb=-0.7 V, Va = 0 V.

I assume D1 is off, others are on. Then the current across R1 is the same as across D2. Equation for the node B:

I(D2) + I(D3) = I(R3);
(5-0)/8k + I(D3) = (-0.7+5)/2k;
I(D3) = (-0.7+5)/2k - (5-0)/8k = 2.15 - 0.625 = 1.525 mA.

The current through D3 has positive value, also, the assumption is correct.

Now I assume that D2 is off, others are on.

(5-Va)/8k = (Va+10)/4k;
Va = -4.53 V.

I(D3) = (-0.7+5)/2k = 2.15 mA.

The current is positive. Assumption is correct. LTSpice confirms it.

My question is: if any given multiple diode circuit has several solutions, which one must be chosen as a correct one?

P.S. It was suggested in the answers that in case if several solutions were found, then the consistency of the whole circuit must be checked for each solution. It's obviously doable for the circuit with 2-3 diodes, but how should one proceed if the circuit contains dozens of diodes and meshes? Is there some industrial/academic method which allows to discard automatically all the inconsistent solutions?

\$\endgroup\$
3
  • \$\begingroup\$ Your question needs a large number of improvements: 1) "for more details, you can refer to a textbook of Neamen[...]" -- most people will not have access to this book, so this suggestion is not helpful 2) "We have 3 generic diodes with Vgamma=0.7 V" - what is "Vgamma"? That is not standard terminology. 3) "[...]the voltage across R1 is the same as across D2" - that's not true, I presume you meant to state that the current through R1 is the same as the current through D2. \$\endgroup\$
    – Mr. Snrub
    Commented Oct 13, 2019 at 5:58
  • \$\begingroup\$ 4) "I(D1) + I(D3) = I(R3)" - again this is not true, I presume you meant "I(D2) + I(D3) = I(R3)". 5) "[...] the assumption is correct" - just because you got an answer, that does not prove that the assumption was correct. That's just the answer you would get if the assumption was in fact true. \$\endgroup\$
    – Mr. Snrub
    Commented Oct 13, 2019 at 5:59
  • \$\begingroup\$ @ Mr. Snrub, thank you for your remarks. Corrected. \$\endgroup\$
    – tenghiz
    Commented Oct 13, 2019 at 6:20

2 Answers 2

1
\$\begingroup\$

Start with \$D_3\$. It "looks" forward-biased. So let's assume it is active (on.) This means that \$V_\text{B}=-700\:\text{mV}\$. Now, if \$D_2\$ is also active and forward biased, this implies that \$V_\text{A}=0\:\text{V}\$. But if this is the case, then it follows that \$I_{R_1}=\frac{5\:\text{V}-0\:\text{V}}{8\:\text{k}\Omega}=625\:\mu\text{A}\$. But we know that \$I_{R_3}=\frac{-700\:\text{mA}-\left(-5\:\text{V}\right)}{2\:\text{k}\Omega}=2.15\:\text{mA}\$ -- half of which must be present in \$D_2\$ and half of which must be in \$D_3\$. But if \$I_{D_2}=1.075\:\text{mA}\$, then the voltage drop across \$R_1\$ must be greater than \$8\:\text{k}\Omega\cdot 1.075\:\text{mA}= 8.6\:\text{V}\$ or \$\gt 8.6\:\text{V}\$. But that would mean that the voltage at \$V_\text{A}\$ is well below \$0\:\text{V}\$. This contradicts the assumption that \$D_2\$ is active and on. Therefore, we must conclude \$D_2\$ is off and therefore all of \$I_{R_3}=2.15\:\text{mA}\$ must flow through \$D_3\$ (which is on and confirms our earlier assumption about \$D_3\$ being on and active) and none of it through \$D_2\$. Since \$D_2\$ is off, it isolates the circuit into two parts, both of which are independent from each other and easily analyzed, now.

\$\endgroup\$
2
  • \$\begingroup\$ You imply the simplification of the circuit is the key criterion? \$\endgroup\$
    – tenghiz
    Commented Oct 13, 2019 at 10:37
  • \$\begingroup\$ @tenghiz There are more robust ways to generate a result. A spice program certainly doesn't use the logic I suggested and it certainly doesn't use ideal diodes, either. But it is an approach you might consider towards getting a result. Sometimes, circuits are bistable (not in this case), but that only means two sets of different assumptions remain consistent. Here, only one set of assumptions will survive consideration, luckily. \$\endgroup\$
    – jonk
    Commented Oct 13, 2019 at 14:50
1
\$\begingroup\$

have I made some errors in my first assumption?

Of course you did. Just because you can make one true statement about the circuit, doesn't mean that the assumption was valid. You need to be able to solve the entire circuit without contradicting that assumption.

In other words, you must find ALL the voltages and currents throughout the whole thing, and verify that, say, the voltage across D1 doesn't contradict your original assumption that it is not conducting.

In fact, it should be obvious that since V2 is the most negative potential in the entire circuit, if the anode of D1 is connected to any other point, it must be conducting!

\$\endgroup\$
6
  • \$\begingroup\$ What does it mean "solve the entire circuit"? \$\endgroup\$
    – tenghiz
    Commented Oct 13, 2019 at 4:51
  • \$\begingroup\$ Shouldn't this be a remark? I fail to read it as an answer to: if any given multiple diode circuit has several solutions, which one must be chosen as a correct one? \$\endgroup\$
    – Huisman
    Commented Oct 13, 2019 at 9:02
  • \$\begingroup\$ @Huisman: It is a direct answer to the actual question asked in the body (before editing): "have I made some errors in my first assumption?" And it answers the question in the title indirectly by negating the premise that the OP has found multiple valid solutions. \$\endgroup\$
    – Dave Tweed
    Commented Oct 13, 2019 at 11:20
  • \$\begingroup\$ @tenghiz: It means that you must find ALL the voltages and currents throughout the whole thing, and verify that, say, the voltage across D1 doesn't contradict your original assumption that it is not conducting. \$\endgroup\$
    – Dave Tweed
    Commented Oct 13, 2019 at 11:22
  • \$\begingroup\$ I would like to repeat the question which was asked in the original version: how do you deal with this kind of situations in the real life, like, during the industrial design stage or somewhere in R&D dpt? Do you literally check the consistency of all voltages node by node or there is some short way to do it? \$\endgroup\$
    – tenghiz
    Commented Oct 14, 2019 at 1:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.