Every simulator builds the equations. This one still has them.
LTspice, PSpice, Falstad and CircuitLab all assemble a nodal system, solve it, and throw it away. Orchard keeps it.
Solve this one opens the editor with it loadedRuns in the browser. Nothing to install, no account, and the solver is on your machine rather than someone's server.
The equations it solved
Everything below is that circuit, solved, with the working left in. Not a description of the method: the equations themselves, in the notation the course teaches. Orchard writes Kirchhoff's current law at each plain node, and every term is one component's contribution.
Rendered as MathML, not as an image or a canvas. A screen reader reads it as mathematics, because a derivation nobody can hear is not an explanation.
The check
Run the sweep and Orchard measures where the response turns over. Then it works the design equation out from the parts on the board. Two numbers, reached two ways.
This is the corner frequency — where the output has fallen 3.01 dB below its passband. The first figure was measured on the sweep by bisection; the second is the design equation f_c = 1/(2πRC) worked out from the resistor and capacitor you placed. Neither was derived from the other.
When it goes wrong
This is where the other tools hand you their internal state. Three separate conditions, and none of them is the circuit above, which solves cleanly. The left column is what those tools give you; the right is what Orchard says about the same condition.
singular matrix- Node 2 reaches ground only through a capacitor, which is an open circuit at DC. The wiring is fine; there is no DC operating point for this node.
Timestep too small- This circuit changes too fast for the transient solver to follow. It kept shortening the timestep and never reached an accurate one, so it stopped rather than draw a waveform it cannot stand behind. Try a shorter run, a smaller largest-step, or a looser tolerance.
add a 1e12-ohm resistor to ground- No ground reference — every voltage needs something to be measured against. Place a ground symbol on any node.
The first two on the left are printed by the tools themselves; the third is the standard forum remedy rather than a message, and is set here as one because it is what you are actually handed. All three are quoted from vendor documentation and support forums, and described from published material rather than from a benchmark run here.
And when it cannot
This circuit produced a result we cannot describe yet. The numbers above still stand; the explanation for this particular condition is missing.
Orchard ships that sentence. An explanation engine that always has something to say is an engine that will eventually make something up, which is exactly the failure of the AI tutors, whose steps are fluent prose that need not correspond to any computation. Orchard's explanations are derived from the solve, so where the solve has no account, neither does it.
Take it from here
The same circuit, on a real board you can edit. Change the capacitor and watch the corner move: both figures still arrived at separately, still agreeing.
Solve this one RC low-pass, loaded and ready or start from a blank sheet