Kirchoff rule violation in Multisim circuit

[GszaboLuther]

Hi,

I have a Multisim circuit (which is a reconstructed circuit of the LT1083 Linear voltage regulator SPICE model).

The reconstructed circuit works well as same as the component itself from the component library, e.g. the output voltage is as expected (in the example attached it is 13.25V).

The only contradiction is that the 55uA current at the Adjust pin goes INTO the pin and not FROM the pin, though this pin is the most negative point of the component. (the 55uA value is conform with the specification of the component but it should be -55uA).

I reconstructed the circuit from the SPCE model of the library component (the result is attached) and started to study it.

Intriguingly I found that at the A and B point near the Adjust pin the Kirchoff nodal rule was violated.

I have some questions:

1. How can it be that the current flows INTO the Adjust pin and not FROM the pin though it is the most negative pin of the component?

2. How can it be that there is a 110uA current which flows from the B point to the Adj On-page connector though there is no any wire connected to the On-page connector?

3. How can it be that the sum of currents going into the A nodal is not zero (namely -110uA)?

[tipa]

Привет! Пробникам не стоит доверять, они "кривые" и неточны. Проведите анализ рабочей точки, он поставит всё на место. Для извлечения токов из нужных цепей поставим нулевые резисторы (можно поставить нулевые источники напряжения). Результат должен Вас успокоить, как и мистера Кирхгофа (

[GGoodwin]

Hi GszaboLuther,

I don't have Multisim (Desktop) application, I am only a Free Subscription user of Multisim Live. Thus, I can't open the circuit that you attached and LT1083 is not available in Multisim Live Free Subscription. I hope you can post screenshots of your circuit with the probes/meters showing the results of simulation.

I have observed this gross violation of very important fundamental law of electric circuits in a short period after I started contributing circuits to Multisim Live. I noticed this when I created voltage regulators, this blunder manifests both in discrete and in three-terminal (LM317H) voltage regulator circuits. I thought of posting this problem in this forum but before I did that I tried to search if there are existing discussion(s) about this problem. I found some discussions relating to problems about Kirchhoff's Current Law but I noticed that the posts were not given attention although the grievances are valid. I then decided to just ignore this defect and deferred the completion of some of my voltage regulator circuits.

The LM317H is the only available three-terminal voltage regulator IC in Multisim Live Free Subscription. Just as you have observed in LT1083, the output voltage is also correct in my LM317 circuits. The error is the same, violation of KCL but what is somewhat ridiculous

 is no current flows into or out of all three terminals. See the screenshot below

This circuit can be reconstructed from my circuit LM317 Adjustable Voltage Regulator. However, the probes must be reinserted because I did not save the modifications to preserve the content of my circuit.

For your first question:

This only happens in (erroneous) simulation, in real-world circuit this is not possible. We can apply the Generalized Kirchhoff's Current Law to the three-terminal voltage regulator to discern this. For a positive-output voltage regulator (input) current flows into the input terminal, (output) current flows out of the output terminal. Since the magnitude of the input current is always greater than the output current, current flow in the adjustment terminal must be outward.

For your second and third questions:

Currently I can't answer these, refer to the first paragraph of this reply.

I commend you for paying attention to details and using your knowledge of foundational tools in circuit analysis to detect errors. It's more than six years since you registered in this forum, I give you your first kudos 😀.

Best regards,

G. Goodwin

P. S.: I may not be able to reply immediately.