Running a circuit
Shift+SUpdated v2.4.4At a glance
- Press play and the circuit runs — nets tint by voltage, current animates along the wires, and meters read live on the sheet
- Speed is simulated seconds per wall second, from 1 µs/s up to real time, with a gauge reporting what the frame budget actually delivered
- Pick the solver timestep — 1 to 100 µs — and it retunes the running circuit without losing your traces
- A timeline ruler scrubs the recording instantly: it replays what was captured, it never re-solves
- Retype a value or flip a switch while it runs and the traces bend — only adding or removing parts restarts the capture
- Stop returns to zero and clears the recording; Restart re-solves the settled operating point and runs again from the top
- The whole transport is on the keyboard: Space, Shift+Space, and the arrow keys step
Press Shift+S or the Simulate button and a floating transport appears at the top of the viewport. Press play. The circuit runs.
What you see on the sheet while it does: nets tinted by their voltage, dots moving along the wires at a speed set by the current through them, live numbers next to every test point, voltmeter and ammeter you placed, and every LED lit at the brightness its drive actually produces.
The solver behind it is a Modified Nodal Analysis engine built into JetCad3's own compute kernel — the same place the geometry work happens — so it advances thousands of timesteps per frame without leaving the app. It is interactive by design: drag a pot and watch the filter retune, without a hitch and without a rebuild.
The transport
| Control | What it does |
|---|---|
| ◀ ▶❚❚ ■ ⟲ ▶ | Step back a window · play/pause · stop · restart · step one timestep (Shift for a window) |
t = | Simulated time at the cursor — it follows a scrub, not the head |
dt | Solver timestep: 1, 5, 10, 50 or 100 µs |
SPD | Simulated seconds per wall second |
RATE | How much of the requested time the frame actually delivered |
SRC | A slider per variable supply, when the circuit has one |
CTRL | A row of controls for every interactive part on the board |
| ⌖ | The probe tool |
| ⚡ ∿ | Operating point · frequency response |
| 🎨 ⚬ | Net voltage tint · current animation |
| × | Leave simulation |
Every button's tooltip shows its current key binding, and re-reads it when you rebind, so a chip never advertises a key that doesn't work.
Speed is a rate, not a multiplier
A circuit simulator has no natural "1×" — a 555 astable and a switching supply live six orders of
magnitude apart. So the dial is simulated seconds per wall second: 1e-3 means one millisecond
of circuit per second of your time, from 1µs/s up to 1s/s, plus a halt rung that parks the sim
without leaving it (you can still scrub).
It's best effort. A heavy circuit can't always be advanced as far as you asked inside a frame, so the RATE gauge reports what actually happened. A dial that silently under-delivers is a lie, and this one doesn't.
Changing the timestep while it runs
dt retunes the live circuit, keeping topology, device state and elapsed time. Rebuilding to
change it would throw away your capture every time you turned the knob — which is exactly the
gesture you'd use to check whether a result depends on the step size.
Scrubbing
The ruler spans the recording, not a program. A circuit free-runs, so the window is whatever the capture buffer currently holds, and a cursor moves through it.
- Drag to scrub. Shift-drag is eight times finer. The wheel nudges a graduation; Shift+wheel steps a window.
- The LIVE pill is lit while the cursor follows the head. Scrub away and it dims; click it to snap forward again.
- Graduations sit on absolute simulated time, so they stay pinned to the signal and slide as the recording scrolls, the way a scope's grid does.
Scrubbing replays the recording. It never re-solves, so it is instant and cannot disturb the run — drag backwards while the circuit is still running and it keeps running.
Editing while it runs
The kind of edit decides what happens:
- A value retyped, a switch flipped, a pot dragged — patched into the live circuit. The traces continue uninterrupted, which is the whole point of turning the knob while watching the scope.
- A wire or part added or removed — the circuit is rebuilt and the capture restarts. The transport says why.
- Something the circuit can't see, like adding a probe or moving a label — nothing happens.
Stop, and Restart
Stop halts, returns to t = 0, empties the recording and clears the sheet overlay, staying in simulation mode.
Restart (⟲) is one press back to the top: stop, re-seed, run. It additionally re-solves the operating point, which a plain stop-then-play doesn't — invisible on a resistive divider, and the difference between a clean start and a phantom turn-on transient on anything with a capacitor in it.
What happens when something's wrong
Problems are named, not swallowed:
- No ground — "Place a GND symbol; every simulation needs one reference point." Ground is found
by any of the usual names, so a perfectly good circuit labelled
0Vstill simulates. - Nothing driving the circuit — said out loud, because every voltage reading exactly zero looks identical to a broken solver.
- A part with no simulation model is not an error. It's captured, it's manufacturable, and it's simply not simulated — reported by designator so you know exactly what was skipped.
- A half-finished part with a node floating doesn't stop the run. Every node carries a vanishingly small leak to ground for exactly that reason, so one unwired component can't take the whole simulation down.