Forums » CircuitLab Support Forum

There seems to be a problem with inductors in CL: they do not properly conserve flux if their series resistance is set to zero.

I think this is actually a more subtle test case that has to do with the intersection of three "ideal" concepts: that of a DC solution, ideal inductors, and ideal switches:

At DC, inductors (with R_ESR=0) "look like" short circuits.

At DC (as the simulator evaluates everything for t=0), SW1 is "open", and SW2 is "closed" (shorting that middle node to ground).

I1 is forcing 1 amp into the top of L1. That current has to come out the other end. Does that current return to ground through the short circuit of L2, or does it go through the short circuit of SW2?

The answer is arbitrary -- and in this case, it picks the path through L2. (Run DC solver and ask for the inductor currents -- you'll see that both start at 1 amp, while the current into SW2 is zero.) The transient solver then starts from that DC solution, and basically you see something you didn't expect, namely both stay at 1 amp. (However, if you choose "Skip Initial", the solver doesn't start from that same assumption that L2 was a perfect short at DC, so I think you'd find the plot you were looking for.)

Adding a tiny series resistance to L2 makes it obvious to the solver that at DC all the current should go through the perfect short circuit path of SW2.

Thanks -- we'll have to think about this one.

I tried to simulate this circuit with DWS and the result seems to be correct:

https://plus.google.com/photos/118285010585894969594/albums/5734717507945270929/5743972244841283154