A trace answers "what does it look like". Measurements answer "what is it", and the difference matters the day your lab report needs a number. de:volt's scope grew a proper measurements strip in July; this post is how to read it critically, because each value has a definition, and knowing them is what separates measuring from transcribing.

Turning it on

On any running circuit, open the Analysis panel, open the scope's display-settings menu (the sliders icon), and enable Measurements. A strip appears under the plot with a channel selector, a Cursors toggle, and four readouts: VPP, RMS, FREQ, DUTY. They are computed from the same coupled, quantised samples the plot draws, before any display noise, so what you measure is what you see.

A signal with texture: PWM

Square waves make every definition visible, so build one worth measuring. Place a Pulse Generator (in the Power group; it is the ideal-source cousin of the Signal Generator) and wire it across a 1 kΩ Resistor to ground. Its Inspector speaks in levels and times rather than frequency: leave Low level at 0 V and High level at 5 V, set Period to 1 ms, and set Pulse width to 0.25 ms. That is a 1 kHz square wave at 25% duty. Probe the node and trigger the scope on a rising edge, 5 ms window.

Now read the strip like an examiner:

• VPP 5.00 V: maximum minus minimum over every valid sample in the window. Peak-to-peak is insensitive to duty, so this one matches the amplitude knob exactly.
• DUTY 25.0%: fraction of each period spent high, judged by midpoint-crossing detection with hysteresis rather than by eyeballing pixels.
• FREQ 1.000 kHz: from the same crossing analysis, needing several full cycles in the window before it will speak.
• RMS 2.50 V: the interesting one. For a 0-to-5 V square wave at duty D, RMS is 5 × √D = 5 × 0.5 = 2.50 V. Not the average (that is 1.25 V), the root-mean-square: the value that predicts heating in the resistor.

Widen Pulse width to 0.5 ms (50% duty) and watch RMS climb to 3.54 V while VPP stays put. If you have ever needed to explain why PWM at half duty does not deliver half the power of DC, this pair of numbers is the whole argument, live.

Cursors, for the things strips cannot know

The Cursors toggle adds two draggable time markers with Δt, 1/Δt, and ΔV readouts. Automated measurements describe the whole window; cursors answer local questions: how long is this glitch, how far apart are these two edges, how much did the capacitor sag between refreshes. Freeze the display first (the Freeze control next to the time window) so your quarry stops moving; the simulation keeps running underneath and the trace resumes with an explicit gap when you unfreeze. Cursor positions and the measurement basis both respect gaps, so a paused stretch cannot fake a period.

When it says Unavailable, believe it

The strip refuses to guess. FREQ and DUTY need enough clean cycles in the longest gap-free stretch of the window; a 0.7 Hz blinker on a 1 s window gets a shrug, and the fix is a longer time window, not a complaint. Amplitude statistics need valid samples at all. If you want to see the evidence quality behind a number, switch on Show engine diagnostics in Settings → Editor and the strip grows a basis field with the valid-sample count and coverage assessment; the same setting reveals acquisition detail everywhere the instruments report anything.

One more honesty note: the strip measures the instrument-domain signal, after your chosen coupling and the declared 12-bit quantisation. Flip a channel to AC coupling and RMS now describes the ripple, not the rail. That is a feature, and also the classic way to fool yourself; the coupling badge next to the channel is there so you cannot.

Everything in this post works for guests and Hobby accounts. The definitions, estimator rules, and coverage semantics are specified at docs.devoltapp.com/user/oscilloscope-and-analysis, and tomorrow the 555 returns in monostable form: one button, one timed pulse, one formula to check.