Oscilloscope

A two-channel oscilloscope: clip it to any two nodes and watch the voltages over time while the circuit runs.

Electrically it is exactly what a real scope probe is (1 megohm from each channel to the ground clip), which is enough to load a high-impedance node the way a real probe does, and little enough to leave anything sane alone.

Pins

Pin	What it does
CH1	Channel 1, 1 megohm to the ground clip.
CH2	Channel 2, 1 megohm to the ground clip.
GND	The ground clip both channels are measured against.

Properties

window: how many seconds of trace the display shows, 0.02 by default. volts: the full-scale range on the vertical axis, 5 by default.

Both are also on the face, as a - and a + under the screen that step 1, 2, 5 per decade the way a real front panel is marked. Turning either of them does not restart the circuit: the trace carries on while you change range, which is the whole reason the buttons are there.

What the model gets right

The scope never creates an analog island on its own (clipping it to a digital net should not turn that net into a matrix problem), but it joins one that reaches it, so its 1 megohm input resistance is accounted for whenever it matters. A channel with nothing clipped to it reads as no signal, which is what a probe touching nothing tells you on a real bench too.

It triggers, and it measures. The screen is 256 samples and the scope keeps four screens behind it, so it can draw the newest whole screen that starts on channel 1 rising through the middle of its own swing, which is what makes a repeating waveform stand still instead of crawling. TRIG in the corner says it found an edge; with no edge to find it free-runs and draws the newest screen anyway, the way a bench scope's auto trigger does. The strip along the bottom measures channel 1 off whatever is on the screen: peak to peak, frequency, and the fraction of each cycle spent high.

What it does not model

The face triggers one way only: channel 1, rising, at the middle of its own swing. The Scope tab is where the rest lives: source, slope, level (or automatic), holdoff, and Auto, Normal or Single, plus time and voltage cursors, XY and an FFT of the screen. There is no math channel and no persistence. 256 samples a screen is the resolution, so a feature much narrower than a hundredth of the time base is not in the picture, and the FFT is of those 256 samples and nothing else.

The probe is 1 megohm and nothing else: no input capacitance, so nothing to compensate and no bandwidth roll-off, and no x10 setting. There is no analog bandwidth limit at all, no noise floor and no vertical quantization, so the trace is exactly what the solver computed. The sample rate follows the time base rather than being a fixed number, and the samples are point readings rather than a real digitizer's peak-detect, so a narrow spike between two samples is simply missing rather than showing up as a tall thin artifact.

Common mistakes

Forgetting the ground clip. Without GND wired to the same reference as the circuit under test, both channels are reading against nothing in particular, which is exactly as meaningless as it is with a real scope probe left unclipped.

See it in action

RC filter on the scope and RL filter on the scope are built around this part, and RC filter on the bench and 555 on the bench put it beside a multimeter on the same signal. See the bench instruments. Open any of them at /templates to see a real trace update live.