How to read a timegrapher without overreading it
A timegrapher can show rate, a timing trace, beat error, amplitude, and beat rate. The order matters: decide whether the measurement is coherent before deciding what its numbers mean.
Short answer: check signal and settings, read rate with its uncertainty, inspect the trace, repeat the run, and only then interpret beat error and amplitude in the context of that movement, position, and state of wind. A clean-looking snapshot is evidence, not a service diagnosis.
The safer reading order
- Signal and settings. Is the cadence locked, the watch untouched, and the nominal beat rate plausible? Is the lift angle known or only assumed?
- Rate and uncertainty. Read the sign, the seconds per day, and any interval or stability indicator together.
- Timing trace. Look for a coherent slope or cluster, not merely one attractive number at the top of the screen.
- Repeatability. Run the same setup again. If the two rates do not support each other, stop interpreting the finer metrics.
- Beat error and amplitude. Use them with the correct lift angle, movement design, pickup method, and manufacturer guidance.
- Real-world context. Compare positions and then check the watch against a reference clock over several days.
1. Start with the measurement conditions
A timegrapher listens to the escapement during a short run. The result belongs to that watch position, wind state, temperature, pickup, and time window. Change one of those and the result may change even when nothing is wrong.
Grand Seiko's own mechanical-watch instructions make the same distinction: normal-use accuracy varies with mainspring wind, temperature, and position, and should be judged from mean daily rate over roughly a week rather than one day. That does not make a timegrapher unhelpful. It defines the question the instrument actually answers: how is this watch behaving here and now?
Primary reference: Grand Seiko, cautions for mechanical-watch accuracy.
2. Read rate as a signed estimate
Rate projects the observed timing error into seconds per day, written as s/d. A positive rate means the watch is gaining; a negative rate means it is losing. Zero means no detected rate error during that run.
Do not detach the value from its precision. “+4.2 s/d ± 0.3” is a stronger statement than “+4.2 s/d ± 7.0,” even though both have the same headline rate. If an instrument provides no uncertainty, use stability over time and a repeated run as practical checks against false precision.
There is no universal “good” rate for every watch. Compare the result with the maker's specification for that caliber and with the watch's real behavior. A short dial-up result can differ from the daily average assembled from many wrist positions.
3. Read the timing trace
A conventional timing diagram plots detected ticks and tocks as the tape moves. Its geometry provides a visual check on the rate and beat-error numbers.
- Flat: no rate error is visible over the plotted interval.
- Upward slope: the conventional display indicates a gaining watch.
- Downward slope: it indicates a losing watch.
- Two separated traces: the tick and tock timing is asymmetric; greater separation represents greater beat error.
- Scatter, waves, or breaks: the timing is varying or the pickup is not producing one clean, stable event stream. The shape alone rarely identifies one unique cause.
The official Bergeon Accuracy Motion manual describes the same conventional rules: positive and negative slopes indicate gain and loss, flat lines indicate zero rate error, and tick/tock line separation represents beat error.
Phone-chart caution: not every app draws a conventional paper tape. Hairspring plots tick and tock timing deviations as colored points against seconds and milliseconds. Read that scatter for coherence and separation; do not infer “up means fast” unless the chart's axes and documentation say so.
Primary reference: Bergeon Accuracy Motion user manual, Diagram view.
4. Read beat error after the trace is stable
Beat error, in milliseconds, describes asymmetry between the tick and tock halves of the escapement cycle. On a conventional diagram it is represented by separation between the two traces.
A stable value can be useful. An unstable value is first a reason to check pickup, noise, movement compatibility, and repetition. It is not enough by itself to name a damaged part or prescribe an adjustment. The same numerical result can have different significance across movement designs and service contexts.
5. Treat amplitude as model-dependent
Amplitude, in degrees, estimates how far the balance swings. On an acoustic timegrapher the calculation depends on which escapement events the instrument identifies and on the entered lift angle. A wrong lift angle can move the amplitude result without changing rate.
Witschi's professional acoustic instruments therefore include a manually selected lift angle. Its combined WisioScope adds optical measurement specifically to determine actual lift angle and amplitude independently of that assumption. That distinction is useful whenever two acoustic tools disagree: the displayed degree value is not a direct camera measurement of the balance wheel.
Hairspring reports a live amplitude range rather than implying that a phone microphone supports one exact degree value. Compare like with like: same watch, position, wind state, lift-angle assumption, and pickup method.
Primary references: Witschi ChronoMaster Auto and Witschi WisioScope.
6. Confirm beat rate, or BPH
Beat rate is the movement's nominal cadence in beats per hour. It is a setting or classification, not a health score. If the detected BPH is implausible for the caliber, stop: every downstream calculation may be based on the wrong cadence. Hairspring automatically detects the common beat rates it supports and visibly resets when the detected cadence changes.
What common result patterns support
| What you see | What the evidence supports | Next check |
|---|---|---|
| Stable rate, coherent trace, repeated run agrees | A repeatable rate for this position and state of wind | Test another position and compare with multi-day use |
| Rate settles, then drifts as the run continues | The behavior is changing during the measurement | Check wind state, settling time, handling, and trace |
| Scattered or broken trace | The instrument is not receiving one stable timing stream | Improve pickup and quiet; repeat before diagnosing the watch |
| Two clean, separated conventional traces | Measurable tick/tock asymmetry on that setup | Confirm the numerical beat error and repeat |
| Amplitude differs between tools | The methods or assumptions may differ | Check lift angle, pickup, position, and each tool's repeatability |
| Good snapshot but poor wrist timekeeping | One bench position does not represent daily use | Log daily offset and compare several positions |
Why universal “good reading” tables mislead
A threshold table is appealing because it converts three numbers into a verdict. It also discards the movement specification, watch position, wind state, temperature, measurement method, and uncertainty. Those omissions can turn a useful screening tool into false confidence.
Use the movement maker's specification when one is available. For everyday performance, compare the watch with a reference clock over several days. For service or regulation decisions, use controlled positions, correct settings, a repeatable pickup, and a watchmaker's movement-specific judgment.
Repeat before you interpret
- Wind the watch and document the position.
- Leave the watch and sensor untouched until the result settles.
- Save the rate and its uncertainty or stability indicator.
- Repeat the same setup without selecting the better-looking run.
- If the pair agrees, move to another position. If it does not, fix the setup before interpreting beat error or amplitude.
Use the free timegrapher repeatability calculator to compare two rates and their fit uncertainty. The companion mechanical-watch accuracy calculator turns a longer clock comparison into mean seconds per day.
Read it with Hairspring
Hairspring is an iPhone timegrapher that reports rate with statistical fit uncertainty, shows the live tick/tock timing scatter, records positions, and asks for a second measurement. It saves both runs and the agreement verdict in History and in a PDF receipt. When the signal cannot support a result, it says why.
One $9.99 purchase. No subscription, account, ads, analytics, or automatic uploads.
Frequently asked questions
What should I read first on a timegrapher?
Read the signal and stability before the headline numbers. Confirm the beat rate, let the result settle, inspect the trace, and repeat the run before interpreting rate, beat error, or amplitude.
What do the lines on a timegrapher mean?
On a conventional diagram, upward slope indicates gain, downward slope indicates loss, flat lines indicate zero rate error, and separation between the tick and tock traces represents beat error. Phone apps may draw different axes.
What are good timegrapher readings?
There is no universal range for every caliber and condition. Compare against the movement maker's specification, repeat the test, measure several positions, and check real-world accuracy over several days.
Does a good timegrapher rate prove good wrist accuracy?
No. It is a short measurement in one position and one state of wind. Daily performance combines many positions, temperatures, activity levels, and mainspring states.