A scope has two jobs: draw the signal, and tell you how much to trust the drawing. de:volt's oscilloscope does both, and the second job is the one this post is about, because it is the one that turns a pretty trace into a measurement.

Channels are nodes

Open the Analysis panel on any running circuit and the channel list shows the circuit's nodes: every set of pins, holes, and wires that is electrically one point. Ticking a channel plots that node's voltage. Each node keeps one colour everywhere it appears (channel swatch, trace, and the highlight on the board when you hover), and the pencil icon renames it, so "Node 3" can become "cap" before your screenshot goes in a lab report. Renaming is free for everyone, as is a differential probe: pick nets A and B in the DIFF PROBE box and the scope plots their difference as one channel, which is how you watch a component's voltage rather than a node's.

Guests and Hobby accounts get 4 channels, Pro gets 16.

Roll for slow, trigger for fast

Roll mode, under ACQUIRE, scrolls the trace continuously. It is right for anything you experience in human time: an RC charge curve, a thermistor warming up, a 0.7 Hz blinker crawling across the screen.

Trigger mode locks repeating waveforms in place. The scope waits for the signal to cross a level in a chosen direction, then draws one screenful anchored to that crossing, so a 1 kHz square wave stands still instead of shimmering. Source, level, edge, position, and auto/normal/single behave like the bench instrument: auto free-runs when nothing crosses, normal waits politely, single grabs one event and freezes, which is the mode for catching a button bounce or a power-up transient. Persistence stacks repeated sweeps so jitter shows as a smear rather than a lie.

The TIME control sets the visible window, from milliseconds to tens of seconds. It also quietly sets your resolution, which brings us to the fine print.

The fine print is printed

Above the plot the scope states its acquisition honestly: the requested and effective sample rate, the Nyquist frequency that implies, and how many valid samples the window holds. The rule of thumb: the scope aims for about 10,000 points per window, capped at 10 kSa/s. A 1 s window at 10 kSa/s can honestly represent content up to 5 kHz; ask it to show a 20 kHz oscillation and it will warn you instead of aliasing silently. If the solver had to step coarsely (heavy circuit, busy machine), the effective rate drops and the display says so.

Gaps are explicit too. Pause and resume, or freeze the display, and the trace shows a break rather than a confident line across time nobody observed. Since July, a momentary button press no longer resets the sweep, so you can watch an event arrive in context.

Worked example: one wave, three readings

Place a Signal Generator, set it to a 100 Hz sine, 1 V amplitude, and wire it to a 1 kΩ resistor into ground. Probe the generator node:

1. In Roll with a 5 s window, the sine is a solid ribbon: 500 cycles per screen is not a drawing, it is a texture. Read the envelope, not the shape.
2. Switch ACQUIRE to Trigger, set the level to 0 V on a rising edge, and drop TIME to 20 ms. Two clean cycles stand still. This is the view for reading shape.
3. Open the display menu and enable Measurements: FREQ reads 100.0 Hz, and VPP reads about 1.90 V rather than the 2.00 V you might expect. That missing 5% is real: the Signal Generator has a 50 Ω source resistance, stated on its datasheet page, and your 1 kΩ load forms a divider with it. Swap the load for 100 kΩ and VPP climbs to 2.00 V. The simulator did not round anything off; it modelled the instrument you actually asked for.

Then change the channel's Input from DC to AC coupling and watch nothing change, because the sine has no offset. Add a 2 V offset on the generator and repeat: DC coupling shows the sine riding at 2 V, AC coupling re-centres it at zero after the declared 1.59 Hz high-pass settles. That toggle is the difference between "what is the voltage" and "what is the wiggle", and knowing which one you asked for is most of scope literacy.

The full acquisition contract, coupling model, and measurement definitions are written down at docs.devoltapp.com/user/oscilloscope-and-analysis. Tomorrow: the RP2040 board, a button, and the internal pull-up that makes it honest.