Reading and Measuring Waveforms

Section Overview

With the waveform locked still and the probe proven honest, the trace can finally be turned into numbers — and numbers are what diagnosis compares against expectations (the-oscilloscope-as-the-diagnostic-instrument). Amplitude is read from the vertical. Peak-to-peak voltage is the full swing from lowest to highest point, read as vertical divisions times the volts-per-division setting, and it is the first number put on any signal (measuring-voltage-and-timing). Period and frequency are read from the horizontal. The period is one full cycle in horizontal divisions times the time-per-division, and frequency is simply its reciprocal — measure the time, and the speed follows. Edges get their own numbers. Rise time and fall time are the time an edge takes to cross from 10% to 90% of its swing (or 90% to 10%), and overshoot is how far the signal shoots past its settled level after a transition, read as a percentage of the swing. Duty cycle describes a pulse's shape in time. Pulse width is how long the signal sits high, and duty cycle is that width as a fraction of the period, which is the number that matters on clocks and PWM. And three methods deliver every one of them. The graticule gives a fast estimate, cursors give a deliberate reading, and an automatic measurement computes continuously — but it measures whatever is on the screen, so a clipped or truncated trace produces a confident, wrong number. Measure the size, the speed, the edges, and the shape — then set each number against what the circuit should produce, because the comparison is the diagnosis.

Why This Matters

A waveform's shape says something is wrong; its numbers say what, by how much, and against which specification — and repair decisions are made on the numbers (measuring-voltage-and-timing). This matters because amplitude is the first health check: a signal at half its expected swing points at a weak stage, a wrong rail, or a loaded node, and only a measured peak-to-peak value makes that call. This matters because frequency ties the signal to its source: a clock measured off its marked frequency indicts the oscillator, not the logic it feeds, so one timing measurement can move the whole diagnosis a stage upstream. It matters because edges carry digital health: a rise time too slow for the logic family explains erratic behaviour a voltage reading never would, and overshoot beyond spec stresses parts toward early failure. It matters because duty cycle is the message on PWM lines: a motor drive, a dimmed backlight, or a switching supply speaks in duty cycle, so measuring it reads the command the circuit is actually sending. And it matters because automatic measurements are believed too easily: the scope will happily compute the frequency of noise or the amplitude of a clipped trace, so the habit of checking the trace behind the number separates measurement from misinformation. Put numbers on the waveform and expectations beside them, and the scope stops showing pictures and starts giving answers.

Required Prerequisites

• The Oscilloscope as the Diagnostic Instrument — Section 8.1 established the graticule, the vertical and horizontal systems, and their per-division settings; this section reads numbers from exactly that grid.
• Measuring Voltage and Timing — Volume 2 introduced graticule, cursor, and automatic measurement as bench skills; this section turns them into diagnostic practice, where every number is checked against an expectation and the automatic readout is treated as a witness to cross-examine.

Recommended Consumables

• A compensated 10x probe with a short ground path — to deliver an honest waveform to measure (probing-compensation-grounding-and-loading)
• A notebook or worksheet — to record measured values beside the values the circuit should produce
• The scope's calibration output — to practise measurements on a signal whose amplitude and frequency are stated
• A signal generator — to set known amplitudes, frequencies, and duty cycles and confirm the measurements recover them
• Service documentation or datasheets for a practice board — to supply the expected values the measurements are compared against

Recommended Practice Hardware

• An oscilloscope with cursors and automatic measurements — to practise all three measurement methods on the same signal (measuring-voltage-and-timing)
• A signal generator with adjustable amplitude, frequency, and duty cycle — to check measurements against known settings
• A board with a marked crystal or clock — to measure a real frequency against its printed expectation
• A PWM source such as a motor driver or LED dimmer — to measure pulse width and duty cycle on a working signal
• A fast logic edge to measure — to practise rise time, fall time, and overshoot readings (probing-compensation-grounding-and-loading)
• A notebook of measured-versus-expected values — to build the comparison habit that turns numbers into diagnosis

Real-World Applications

Measurement is where the scope's picture becomes a repair decision. A technician checking a logic supply's ripple measures peak-to-peak voltage with cursors and compares it against the regulator's specification, turning "looks noisy" into "80 millivolts, twice the allowed ripple" (measuring-voltage-and-timing). A repairer verifying a dead-looking microcontroller measures the clock at its crystal's marked frequency and moves the suspicion from the oscillator to the reset circuit in one reading. Someone chasing erratic logic measures a rise time far too slow for the family and finds the overloaded driver that a voltage check called healthy. A technician on a motor that runs weak measures the PWM duty cycle at the driver and learns the controller is only asking for 20% — the fault is upstream of the power stage. And a repairer misled by a readout notices the automatic amplitude of a clipped trace reads exactly the screen height, increases the volts-per-division to bring the whole waveform on screen, and gets the true value the readout could not see (the-oscilloscope-as-the-diagnostic-instrument). The failures this prevents: replacing a regulator over ripple that was in spec, condemning an oscillator that was running on frequency, and acting on an automatic number computed from a trace that never fit the screen.

Common Challenges

• A number without an expectation proves nothing. A measured 3.1 V or 47 kHz means nothing in isolation — the difficulty is that the expected value lives in datasheets, schematics, and markings, so the measuring is often easier than knowing what the answer should be (measuring-voltage-and-timing).
• Automatic readouts look authoritative even when wrong. The scope computes a crisp number from whatever is on screen, clipped, truncated, or noisy — the difficulty is that a wrong number in a neat readout box is more convincing than an honest but rough graticule estimate.
• Edge measurements collide with the instrument's limits. A scope near its bandwidth adds its own rise time to the edge it measures, and a probe artifact can masquerade as overshoot — the difficulty is that the number on screen blends the signal with the measuring chain, and only setup discipline separates them (probing-compensation-grounding-and-loading).

Safety Notes

Risk Level: Medium. Measurement is performed with the probe on a live circuit, so the section carries the risk of any powered work — a live board, a probe tip that can slip while cursors are being adjusted, and a ground lead that is a hard connection to earth.

Professional Tips Before Starting

• Fill the screen with the signal. Accuracy comes from using the scope's resolution — set the vertical so the waveform spans most of the graticule before measuring amplitude, and the horizontal so at least one full cycle is cleanly visible (measuring-voltage-and-timing).
• Estimate before you read. A graticule estimate made first catches gross errors — if the cursor or automatic value disagrees wildly with your eye, something in the setup is wrong.
• Write the expectation down first. Deciding what the number should be before measuring — from the schematic, datasheet, or marking — keeps the reading from being rationalised after the fact.

Measuring Waveforms — Amplitude, Time, Edges, and Duty Cycle

Recap and Frame

Sections 8.1 through 8.3 built a truthful trace — instrument understood, waveform locked, probe honest; this section extracts the numbers that make the trace usable as evidence (the-oscilloscope-as-the-diagnostic-instrument). Four measurements carry most repair work. Amplitude for size, period and frequency for speed, rise and fall time with overshoot for the edges, and pulse width with duty cycle for shape in time — these four answer most of the questions a fault raises (measuring-voltage-and-timing). Three methods deliver them. The graticule gives a fast estimate by counting divisions, cursors give a deliberate reading between two placed markers, and automatic measurements compute continuously from the acquired record. Each method has its moment. The graticule is for orientation and sanity checks, cursors are for the careful number a decision rests on, and automatic measurements are for watching a value live — provided the trace behind them is fit to measure. The trace must be fit to measure. A waveform clipped by the vertical, truncated by the horizontal, or drowned in noise produces wrong numbers by every method, so the setup is checked before any reading is trusted (triggering-capturing-a-stable-waveform). And the number needs its expectation. A measurement becomes evidence only beside the value the circuit should produce, so the expected value is found — schematic, datasheet, marking — before the reading is made. Hold the frame — four measurements, three methods, a fit trace, and an expectation — and the scope starts answering questions.

Amplitude — Reading the Vertical

The first number put on any signal is its size, and the vertical system provides it (measuring-voltage-and-timing). Peak-to-peak is the full swing. Peak-to-peak voltage runs from the waveform's lowest point to its highest, read on the graticule as vertical divisions spanned times the volts-per-division setting — three and a half divisions at 2 V per division is 7 V peak-to-peak. Levels matter as much as swing. Where the waveform sits matters too — a logic high's voltage, a DC offset under a ripple, a baseline that has drifted — so the trace is also read against the channel's ground reference marker, not just top-to-bottom (the-oscilloscope-as-the-diagnostic-instrument). Cursors make the reading deliberate. A pair of horizontal voltage cursors placed on the two levels of interest reads the difference directly, replacing division-counting with a stated number and working just as well between any two points, not only the extremes. Resolution comes from filling the screen. A waveform two divisions tall wastes the scope's resolution and magnifies reading error, so the volts-per-division is set to make the signal span most of the graticule before the number is taken. Clipping ruins the number silently. A trace pushed past the top or bottom of the screen is clipped, and every method — divisions, cursors, and automatic — then reads the screen's limit rather than the signal, so the full waveform is brought on screen before amplitude is measured. The swing by divisions or cursors, the levels against the ground reference, the screen filled for resolution, and clipping ruled out — and the amplitude is read. Size the signal to the screen, and the vertical gives an honest number.

Period and Frequency — Reading the Horizontal

How fast a signal repeats is the second core number, and it is read from the horizontal system the same way size was read from the vertical (measuring-voltage-and-timing). The period is one cycle's width. One full cycle — from a point on the waveform to the same point one repetition later — spans some number of horizontal divisions, and that count times the time-per-division is the period. Frequency is its reciprocal. Frequency is 1 divided by the period, so a 1 ms period is 1 kHz and a 25 µs period is 40 kHz — measure the time and the speed follows by arithmetic, with no second measurement needed. Cursors read it cleanly. A pair of vertical time cursors placed at identical points on two successive cycles — rising edge to rising edge at the same level — reads the period directly, and identical points matter, because measuring from a peak to the following edge reads something other than the period. Averaging over cycles sharpens the number. Placing the cursors across ten cycles and dividing the read time by ten shrinks placement error tenfold, which is the standard trick when the period must be known closely. The measured frequency meets its expectation. Clocks and oscillators carry their frequency on the crystal's marking or the service data, so the measured value is set against the marked one — on frequency clears the oscillator, off frequency indicts it. One cycle's divisions times the time-per-division, frequency as the reciprocal, cursors between identical points, many cycles for a fine reading, and the result set against the marking — and the timing is read. Measure the period, take its reciprocal, and compare — the horizontal is that direct.

Edges — Rise Time, Fall Time, and Overshoot

Digital health lives in the edges, and edges get three numbers of their own — numbers the honest probe chain of the previous section makes it possible to trust (probing-compensation-grounding-and-loading). Rise and fall time use the 10%–90% convention. Rise time is the time an edge takes to climb from 10% to 90% of its swing, and fall time is the mirror — 90% down to 10% — with the outer tenths excluded so the number describes the transition rather than the rounded corners at its ends. The scope measures them at speed. An edge is spread across the screen with a fast time-per-division, then read by cursors at the 10% and 90% levels or by the automatic rise- and fall-time measurements, which apply the convention themselves. The instrument adds its own rise time. A scope and probe cannot draw an edge faster than their own response, so an edge near the scope's bandwidth reads slower than it truly is — a rule of thumb puts a scope's own rise time near 0.35 divided by its bandwidth, and edges approaching that figure need a faster instrument to be measured honestly (the-oscilloscope-as-the-diagnostic-instrument). Overshoot is the edge's excess. Overshoot is how far the signal shoots past its settled level after a transition before ringing down, read as a percentage of the swing — a spike to 5.5 V on a 5 V logic edge is 10% overshoot — and specifications bound it because repeated excess stresses inputs. The measuring chain must be cleared first. Probe compensation error imitates overshoot and a long ground lead imitates ringing, so the probe chain is verified and the ground path shortened before an edge's numbers are blamed on the circuit (probing-compensation-grounding-and-loading). Tenth-to-ninetieth for rise and fall, the edge spread wide to be read, the scope's own speed respected, overshoot as a percentage of swing, and the probe chain cleared first — and the edges are measured. Read the edge honestly, and it tells you whether the logic it drives can trust it.

Pulse Width and Duty Cycle — Shape in Time

On clocks, control lines, and anything switched, the fourth measurement is how the signal divides its time between high and low. Pulse width is the time at a level. Positive pulse width is how long the signal sits high in each cycle and negative width how long it sits low, read with time cursors on the two edges of the pulse or by the automatic width measurements. Duty cycle is width over period. Duty cycle is the pulse width as a fraction of the period, times 100 — a signal high for 2 ms of a 10 ms period runs at 20% duty cycle — so it is two time readings and a division. PWM speaks in duty cycle. Motor drives, LED dimming, heater control, and switching converters encode their command as duty cycle, so measuring it reads the command actually being sent — a motor driven weakly at 20% duty is obeying its controller, not failing (measuring-voltage-and-timing). Clocks expect symmetry. Most clocks are meant to run near 50% duty cycle, so a clock measured far from symmetric points at a failing oscillator or a loaded line even when its frequency is right. The measurement needs clean edges on screen. Width cursors need both edges of the pulse visible and the trigger stable, so the trace is locked and at least one full cycle displayed before width or duty is read (triggering-capturing-a-stable-waveform). Width at a level, duty as width over period, PWM read as a command, clocks checked for symmetry, and both edges cleanly on screen — and the signal's shape in time is measured. Measure how the signal spends its period, and control lines start reporting what they are being told.

Graticule, Cursors, and Automatic Measurements — Choosing and Checking

Every number above can come from any of the three methods, and choosing well — then checking — is what keeps measurement honest (measuring-voltage-and-timing). The graticule is the fast estimate. Counting divisions takes seconds, needs no menus, and anchors expectations, so the graticule reading comes first and flags any later number that disagrees wildly with the eye. Cursors are the deliberate reading. Manually placed markers read exactly the two points chosen — this edge, that level, these ten cycles — so cursors carry the careful measurement a repair decision rests on. Automatic measurements are the live readout. An automatic measurement is computed by the scope continuously from the acquired waveform — amplitude, frequency, rise time, duty cycle, and more, several at once — which makes it ideal for watching a value change while adjusting or warming a circuit. But the scope measures what it sees. An automatic measurement is computed from whatever the screen holds: a clipped trace yields the screen's height as amplitude, a fraction of a cycle yields a wrong frequency, and noise yields a confident number about nothing — the readout does not know the trace is unfit. So the trace is checked before the readout is trusted. Before believing an automatic value, the waveform behind it is confirmed unclipped, whole, and stably triggered, and the value is sanity-checked against the graticule estimate (triggering-capturing-a-stable-waveform). And every number ends at its expectation. Graticule, cursor, or automatic, a reading becomes diagnosis only when compared with what the circuit should produce, so the expected value sits beside the measured one in the notes. Estimate by graticule, decide by cursor, monitor by automatic readout, verify the trace behind every number, and compare against expectation — and the three methods become one discipline. Choose the method for the moment, and check every number against the trace and the expectation behind it.

Common Mistakes

• Measuring a clipped or truncated trace. A waveform pushed off screen reads as the screen, not the signal, by every method — bring the whole waveform on screen before taking any number (measuring-voltage-and-timing).
• Trusting an automatic readout without looking at the trace. The scope computes confidently from noise, clipping, or a fraction of a cycle — confirm the trace is whole and stable, and sanity-check against a graticule estimate.
• Measuring a period between mismatched points. Cursor from a peak to the next edge reads something other than the period — place cursors at identical points on successive cycles, and span several cycles for accuracy.
• Blaming the circuit for the chain's edge artifacts. Compensation error imitates overshoot, a long ground lead imitates ringing, and a scope at its bandwidth slows edges — verify the probe chain and respect the instrument's limits before condemning an edge (probing-compensation-grounding-and-loading).
• Measuring without an expectation. A number alone proves nothing — find the expected value from the schematic, datasheet, or marking, and write it down before reading the scope.

Troubleshooting Guidance

Measurement comes down to a fit trace, the right method, and a comparison. If the amplitude reading seems wrong: check for clipping and fill the screen with the signal before measuring again (measuring-voltage-and-timing). If the frequency readout is absurd: confirm at least one full, stably triggered cycle is on screen — the readout computes from whatever it sees (triggering-capturing-a-stable-waveform). If a fine period value is needed: place cursors across ten cycles at identical points and divide by ten. If an edge reads slow: ask whether the scope is near its bandwidth — the instrument adds its own rise time to edges near its limit (the-oscilloscope-as-the-diagnostic-instrument). If an edge shows overshoot or ringing: verify probe compensation and shorten the ground path before blaming the circuit (probing-compensation-grounding-and-loading). If a clock's frequency is right but behaviour is wrong: measure its duty cycle — a badly asymmetric clock fails logic that its frequency reading would clear. If a PWM load runs weak: measure duty cycle at the driver — a low command means the fault is upstream, not in the power stage. If the number cannot be interpreted: the missing piece is the expectation — find the specified value in the schematic, datasheet, or marking, and compare. The throughline: make the trace fit to measure, take the number deliberately, and let the comparison with expectation do the diagnosing.

Verification & Testing Methods

Confirm your numbers deserve to be trusted:

• [ ] I measured amplitude with the waveform filling the screen and unclipped, by graticule and by cursors, and read levels against the channel's ground reference.
• [ ] I measured the period between identical points on the waveform — across several cycles for accuracy — and converted it to frequency by taking the reciprocal.
• [ ] I measured rise and fall time between 10% and 90% of the swing with the edge spread across the screen, read overshoot as a percentage of the swing, and cleared the probe chain before blaming the circuit for edge artifacts.
• [ ] I measured pulse width and duty cycle with both edges cleanly on screen, and checked clocks for symmetry as well as frequency.
• [ ] I treated every automatic measurement as a witness to cross-examine — trace confirmed whole and stable behind it, value sanity-checked against a graticule estimate, and every number compared against a written expectation.

Then try the practice exercises below — hands-on measuring; scenarios differ from the quiz.

Practice Exercises

1. Measure one signal three ways (5 minutes, hands-on). On the calibration output, estimate the amplitude and period by graticule, then read both with cursors, then turn on the automatic measurements — compare all three against the stated values, then clip the trace deliberately by over-ranging the vertical and watch the automatic amplitude go wrong while the readout stays confident (measuring-voltage-and-timing).
2. Measure a real clock against its marking (5 minutes, hands-on). On a board with a marked crystal, measure the period with cursors across several cycles, convert to frequency, and compare against the marking; then measure the duty cycle and check it for symmetry (measuring-voltage-and-timing).
3. Put numbers on an edge (5 minutes, hands-on). Spread a logic edge across the screen with a fast timebase and short ground path, measure the rise time at 10%–90% with cursors, read any overshoot as a percentage of the swing, and repeat with the automatic measurement to compare; then repeat on a falling edge to read the fall time between 90% and 10% (probing-compensation-grounding-and-loading).
4. Read a PWM command (3 minutes, hands-on). On a PWM signal — a dimmer, fan, or motor driver — measure pulse width and duty cycle at two different control settings and confirm the duty cycle tracks the command.

These core steps — measuring amplitude, period and frequency, edges, and duty cycle, and cross-examining the automatic readout — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

• Amplitude is read from the vertical — divisions times volts-per-division, or cursors between levels — with the waveform filling the screen and unclipped, because a clipped trace reads as the screen rather than the signal (measuring-voltage-and-timing).
• The period is one cycle's horizontal divisions times the time-per-division — measured between identical points, across several cycles for accuracy — and frequency is its reciprocal, compared against the crystal's marking or the service data (measuring-voltage-and-timing).
• Rise time and fall time are read between 10% and 90% of the swing, and overshoot as a percentage of the swing — with the probe chain verified first, and the scope's own rise time respected on edges near its bandwidth.
• Pulse width is the time a signal spends at a level and duty cycle is that width over the period — the number PWM lines encode their command in, and the symmetry check that catches a sick clock whose frequency reads correct.
• An automatic measurement computes continuously from whatever is on screen — clipped, truncated, or noisy included — so the trace behind it is confirmed fit, the value is sanity-checked against a graticule estimate, and every number is compared against a written expectation.

Skills Learned

• You can now measure amplitude and peak-to-peak voltage by graticule and by cursors.
• You can now measure a waveform's period and convert it to frequency.
• You can now measure rise and fall time between 10% and 90% and read overshoot on an edge.
• You can now measure pulse width and duty cycle on a digital or PWM signal.
• You can now use automatic measurements critically, checking the trace behind the number and the number against expectation.

Glossary Additions

• automatic measurement — a value an oscilloscope computes continuously from the acquired waveform and displays as a live readout, such as amplitude, peak-to-peak voltage, frequency, period, rise time, or duty cycle, with several shown at once. Automatic measurements are ideal for monitoring a value while a circuit is adjusted, warmed, or exercised, but they are computed from whatever the screen holds: a clipped trace yields the screen's height as amplitude, a fraction of a cycle yields a wrong frequency, and noise yields a confident number about nothing. An automatic measurement is therefore trusted only after the trace behind it is confirmed unclipped, whole, and stably triggered, and its value is sanity-checked against a graticule estimate and compared with the value the circuit is expected to produce.
• fall time — the time a falling edge takes to cross from 90% to 10% of the waveform's swing, the mirror of rise time, with the outer tenths excluded by convention so the number describes the transition itself rather than the rounded corners where it begins and ends. Fall time is measured with the edge spread across the screen at a fast time-per-division, by cursors placed at the 90% and 10% levels or by the scope's automatic fall-time measurement, and it is judged against the logic family's or the circuit's specification. As with rise time, the instrument contributes its own response: an edge near the scope's bandwidth reads slower than it truly is, so fall times close to the instrument's limit need a faster scope to be measured honestly.
• overshoot — the amount by which a signal shoots past its settled level immediately after a transition, before ringing down to rest, expressed as a percentage of the swing — a spike to 5.5 V on an edge settling at 5 V is 10% overshoot. Specifications bound overshoot because repeated excursions past the rails stress the inputs they strike, and excessive overshoot on logic lines points at drive, termination, or layout problems. On a scope screen, overshoot is read only after the measuring chain is cleared: an over-compensated probe adds overshoot of its own and a long ground lead adds ringing, so probe compensation is verified and the ground path shortened before an edge's overshoot is charged to the circuit.

Suggested Next Sections

Must read next:

• Capturing Transients and Single-Shot Events — Section 8.5 goes after the signals that will not sit still to be measured: glitches, dropouts, and one-time events, captured with single-shot triggering, persistence, and peak-detect.

Recommended:

• Probing — Compensation, Grounding, and Loading — the honest probe chain every measurement in this section depends on.
• Measuring Voltage and Timing — the bench-skills introduction to graticule, cursor, and automatic measurement that this section builds into diagnostic practice.