user 37518
Well-known member
I get that. But to me, is not about the derivation or re-doing what has already been done before. For me is about understanding where it came from, and I am not talking about the algebra steps, but rather the physical meaning of where this or that equation came from the circuit analysis. Just knowing the formula does nothing for me.
Calculators and sims only work on numerical values, they tell you nothing about the circuit if you do not input numerical values. That is why it is important to know how to derivate the results analytically from the circuit rather than numerically. Anecdotally, during the pandemic, I gave my students circuits without any values in them, they had to tell me how the circuits worked and/or the expected voltages/currents in terms of the input voltage/current as expressions rather than numbers; many failed. I had noticed earlier that many had been cheating by simulating the circuits and just reading the voltages/currents with a probe.
As for the sims, I agree that they are not equal to the real deal, but almost. It also depends on the Sim and the circuit/part models you are using. If you use simple and rough models, you won't get accurate results. I've regularly used really high-end sims like ADS, which make use of full wave simulators using the Method of Moments (MoM), you can draw the PCB with the parts in it and it analyses the EM fields in it, and it is almost the same thing as the lab. Granted, those simulators are used for high-frequency circuits mostly and are highly specialized. But I've also experienced that when I simulate low-frequency circuits (like audio circuits), using good and accurate circuit/part models, I experimentally get 99.9% the same of what I simulated. These days, if you are designing something at low frequencies, if your simulation results are not almost the same as the real deal, you are doing it wrong.
I don't know which simulator are you using, but even the free simulators as LTSpice or TiNA are amazing. It depends of course on the models you use, if your transistor or op-amp model are just 10 lines of SPICE code, it is most likely that you won't achieve good results.
Most people who complain about simulators, in my experience, are not very experienced with simulators. The days of Bob Pease throwing away his PC from the rooftop of National Semiconductor because the thing "lied to him" are long gone. There is no commercial circuit done today that hasn't gone through countless hours of simulation.
A colleague of mine based his entire PhD on high-power high-frequency transistor modelling, he used extremely complicated models and neural networks to achieve the same behavior as the real thing. Models these days are amazing, albeit not perfect, but close.
Last edited:
