1
\$\begingroup\$

I have a differential amplifier. Problem of "gain" of the op-amp or "how it is wired."

  • I have simulated this as a voltage source controlled voltage (VofV.)
  • As a generic amplifer.
  • As a good old LM741C.

Very strange behavior when there is an error of "wiring." But quasi good results.

What is happening?

Precision : it is a "Dynamics-DC" analysis (so called) by microp12 Spectrum-Software ... and the results are, I think so, "numerically" correct.

EDIT : for obtaining what are "real" and correct results, one must make some other analysis on these examples, as TRANsient analysis. So, when you have a circuitry to be analysed, always try different kind of simulation. And, "in fine", ... Realize in a 'real world' by assembling it on a breadboard ... and viewing on a scope (or other toosl) the behavior. When you use "simulators" or "mathematical" tools, don't forget to verify some "answers" given. Be critical ...

enter image description here enter image description here enter image description here

\$\endgroup\$
5
  • \$\begingroup\$ It’s an interesting question that leads to more questions. What conditions or assumptions differ between simulators and real ccts? What is the peaking factor for self-compensated OA’s in these config’s for each cct. In Differential mode. The 5 uV difference is just the error due to gain. 1/Aol*Vin. The low f phase is inverted in both cases because the current loop does not change. \$\endgroup\$ Jul 3 at 14:22
  • \$\begingroup\$ How can it work , makes it weird. Examine the current loop as output phase leads input instead of lagging . \$\endgroup\$ Jul 3 at 14:53
  • \$\begingroup\$ There is an error in computing the difference voltage across the input source of 10 uV yet defined as 1V. In the 1st plot. Why? \$\endgroup\$ Jul 3 at 15:36
  • \$\begingroup\$ @JRE Thanks for "correcting" my poor english. Simulation are made in DC-dynamic mode with microcap12 spectrum-soft.com/index.shtm. It is an error often found and made ... even for me. :) \$\endgroup\$
    – Antonio51
    Jul 3 at 16:08
  • \$\begingroup\$ @Tony Stewart EE75 Some errors ... because, i think, simulators add one high resistor ( 1 Gohm) to ground for nodes when generators are found "floating". \$\endgroup\$
    – Antonio51
    Jul 3 at 16:19
1
\$\begingroup\$

It appears that you have fallen into the trap of using a simulator with positive feedback on an op-amp and gotten what appears to be a stable result. Those circuits drawn incorrectly as differential amplifiers i.e. those circuits that use positive feedback will give conditionally stable outputs that appear valid - they are not AND, whenever using a sim, you need to be sensible and not create this situation.

\$\endgroup\$
13
  • \$\begingroup\$ There are some interesting gain/ phase effects near unity gain on each version ; normal and wrong. \$\endgroup\$ Jul 3 at 13:54
  • 1
    \$\begingroup\$ Yet the current loop is the same in both cases below breakpoint, so it is negative feedback in both cases as it senses differential voltage in the current loop. Thus inverting voltage in both cases except near GBW limits. That’s when it goes unstable with a phase shift >180 deg \$\endgroup\$ Jul 3 at 14:30
  • \$\begingroup\$ @Tony Stewart EE75 OK for stability of the circuits at right circuits. It is here a DC-dynamics analysis. Just to say that we have to be carefull with simulators. :) \$\endgroup\$
    – Antonio51
    Jul 3 at 16:22
  • 1
    \$\begingroup\$ @Antonio51 just performing a dynamic-DC analysis in micro-cap tells you the average voltages that the nodes settled at in a very short (and predicted time). In reality, it tells you nothing about the real DC voltages that a real circuit eventually settles upon. The voltages shown in dynamic-DC analysis are the very first voltages the simulator calculates before beginning the much more computation-extensive transient analysis hence, user beware and don't plug-in circuits that are known not to be comparable to a standard linear circuit with negative feedback. \$\endgroup\$
    – Andy aka
    Jul 3 at 18:57
  • 1
    \$\begingroup\$ @Andy aka Yes. Closed. Sorry. \$\endgroup\$
    – Antonio51
    Sep 9 at 12:01
1
\$\begingroup\$

I like the fact that Falstad’s simulator gives real results with real Op Amp’s like the 741 (saturated) yet it works with infinite GBW ideal Op Amps by default. Here I toggled the 741 inputs (normal, inverted) using select IC> edit as the plot shows normal & saturated outputs alternating. The differential input is the 120 Hz square wave with a 60 Hz sine included for Vcm.

enter image description here I guess that makes it better than some other simulators.

\$\endgroup\$
1
  • \$\begingroup\$ I don't know this simulator and what it can do. But it is interresting for animation. I don't know if this can use Laplace functions or other things such defining PWL datas. I have used microcap for 40 years (in fact, since it begins). The more complicated schematic I have simulated was a digital circuit (gates, D-FF 7474, counters 7490) on 8 pages. (it was a 8-storey elevator complete system, with all buttons/memory at each stage and intelligent up/down). \$\endgroup\$
    – Antonio51
    Sep 9 at 12:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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