← All Guides

Reading a timegrapher,
from the trace up

A sixty-dollar timegrapher tells you more about a mechanical watch in thirty seconds than a week of wrist-watching — if you can read it. Most new owners can’t, panic at their first ugly trace, and either overpay a watchmaker or condemn a healthy movement. The machine is simple once you know what it is actually doing.

Last reviewed July 2026 · Independent guide · Not affiliated with timegrapher

The 20-second answer

A timegrapher listens to the escapement. Each tick has an acoustic structure, and from its timing the machine computes three numbers: rate (seconds gained or lost per day — the slope of the trace), amplitude (how far the balance swings, in degrees — the movement’s blood pressure), and beat error (the asymmetry between tick and tock, in milliseconds). Healthy ballpark figures for a serviced movement: rate within its spec, amplitude roughly 250–300 degrees dial-up, beat error under about 0.5ms. Three caveats turn beginners’ readings into nonsense: the lift angle setting must match your calibre or amplitude is fiction, one position tells you almost nothing, and some modern escapements — the Powermatic 80 family famously — confuse the machine’s assumptions entirely. Learn the three numbers, the caveats, and five trace patterns, and you can health-check any watch you own or intend to buy.

THE THREE NUMBERS, AS TRACESHEALTHYtight lines · gentle slopeRUNNING FASTsteep slope · regulationFAULT / DIRTsnow · service, not regulationSlope is rate. Line gap is beat error. Texture is health. Every trace is some combination of the three.
Fig. 1 — Every trace decomposes into slope, gap and texture. Learn to read those three properties separately and no trace is mysterious again: the slope tells you regulation, the gap tells you beat adjustment, and the texture tells you whether the escapement is clean, worn, or being lied to by its own settings.

The five-minute protocol

This sequence turns a cheap machine into a genuine diagnostic instrument.

Set the lift angle first

Amplitude is computed from it. The default 52° suits many calibres, but check yours — commonly published figures include 50° for the ETA 2824-2 and 53° for the Seiko NH35, with big pocket-watch calibres much lower. Wrong angle, fictional amplitude.

Wind it properly

Test at full wind, or after a day’s wear. A half-wound watch reads low amplitude honestly — mainspring torque falls as it unwinds, and you will diagnose a fault that is just an empty tank.

Read the slope: rate

A flat dual line is zero rate; sloping up is fast, down is slow. Judge the number against the movement’s own spec — +15 is failure for a chronometer and comfortably in-spec for a 7S26.

Read the gap: beat error

Two parallel lines close together is a small tick-tock asymmetry. Under ~0.5ms is good; up to ~1ms rarely matters in practice; large errors waste amplitude and deserve adjustment at next service.

Read the texture: amplitude and noise

Clean tight lines with 250–300° dial-up is health. Amplitude under ~200° at full wind means service. Scatter — snow — means dirt, wear, damage or magnetism.

Rotate through positions

Dial-up, dial-down, crown-up, crown-down at minimum. The spread between positions is the movement’s real character — and the thing single-position screenshots always hide.

What the machine actually does

A timegrapher is a microphone, a clock and some arithmetic. Clamped to the case, it listens to the escapement — each tick is really a cluster of tiny impacts as the escape wheel unlocks, impulses the balance and drops onto the next tooth — and it times those events against its internal reference to microsecond precision. Everything on the screen is computed from those timings plus one assumption you supply: the lift angle.

Rate is the simple one: if ticks arrive slightly early, the watch runs fast, and the dots march upward across the screen; late, and they slope down. The displayed number is that drift extrapolated to seconds per day. Beat error is the asymmetry between the two halves of the oscillation — tick versus tock — displayed in milliseconds and drawn as the vertical gap between the two dotted lines.

Amplitude is the derived one, and the one beginners misread. The machine measures the time between acoustic events within each tick and, using the lift angle — the arc across which the escapement interacts with the balance — computes how far the balance must be swinging. Feed it the wrong lift angle and the timing is real but the geometry is wrong: amplitude comes out tens of degrees high or low while rate and beat error remain trustworthy.

This acoustic method is why the machine has blind spots. A noisy room adds phantom dots. A quartz watch gives it nothing usable. And an escapement whose geometry departs from the classical Swiss lever — the Powermatic 80 family’s synthetic redesign is the famous case — produces sound signatures the arithmetic was never built for, and confidently wrong numbers. The machine is honest about what it hears; it just doesn’t know what it doesn’t know.

What the machine hearsEscapement acoustics · unlock, impulse, drop — timed to microsecondsBalance swingAmplitude — computed via lift angleTick vs tockAsymmetry — beat error in msTick intervalRate — against the beat specSound qualityNoise here becomes snow on screenThe machine assumes a classical Swiss lever escapement — exotic escapements can produce fictional numbers.
Fig. 2 — It is a microphone with opinions. A timegrapher never sees the balance swing — it times the sounds of the escapement’s events and computes everything from assumptions about their geometry. That is why the lift angle setting matters, why noisy environments ruin readings, and why movements with redesigned escapements can read as broken while keeping perfect time.

Rate: the number everyone starts with

Rate is the headline: seconds per day, positive fast, negative slow. Two principles turn it from a number into a judgement. First, judge against the movement’s own specification, not against zero. A 7S26 at +18 is inside its printed -20/+40 band — healthy. A chronometer-grade movement at +18 has failed. The reading only means something relative to what the calibre promises; our movement pages carry the specs.

Second, one position is an anecdote. Gravity acts on the balance differently in every orientation, so measure dial-up, dial-down, crown-up and crown-down at minimum, note the spread, and think in terms of the average your wrist will actually produce. A movement reading +4 dial-up and +6 crown-down is a gem regardless of grade; one reading +20 and -15 in the same positions will infuriate its owner despite an innocent-looking average — the 6R35 complaint in a nutshell.

The slope’s stability matters as much as its angle. A healthy trace holds its slope steadily; a rate that wanders over thirty seconds — the line bending, drifting, stair-stepping — indicates the escapement is not delivering consistent impulses, which is a service symptom no regulation will cure.

And the golden rule of dramatic rates: a watch that abruptly starts gaining minutes a day, with clean traces, is magnetised until proven otherwise. Phones, laptops, bag clasps. A demagnetiser costs less than lunch and resolves it in five minutes — always before deeper diagnosis, never after.

THE THREE NUMBERS, AS TRACESHEALTHYtight lines · gentle slopeRUNNING FASTsteep slope · regulationFAULT / DIRTsnow · service, not regulationSlope is rate. Line gap is beat error. Texture is health. Every trace is some combination of the three.
Fig. 3 — Every trace decomposes into slope, gap and texture. Learn to read those three properties separately and no trace is mysterious again: the slope tells you regulation, the gap tells you beat adjustment, and the texture tells you whether the escapement is clean, worn, or being lied to by its own settings.

Amplitude and beat error: the health numbers

Amplitude is the balance’s swing in degrees, and it functions as the movement’s blood pressure. A serviced movement at full wind typically swings roughly 250–300 degrees dial-up, dropping 20–40 degrees in vertical positions — normal, since the balance pivots carry weight differently on their sides. Below about 200 at full wind, something is wrong: dried oils, dirt, a tired mainspring, worn pivots. Above about 315, the overbanked condition called knocking — the balance swinging so far it strikes the other side of the escapement — which is its own fault, often heard as a galloping rate.

Amplitude is also where honest mistakes concentrate. Test at half-wind and you will see low amplitude that is merely an emptying mainspring — wind fully and re-test. Set the wrong lift angle and the number is fiction — set it per calibre. And compare like with like: dial-up figures against dial-up figures. A single low reading proves nothing; a full-wind, correct-angle, dial-up reading under 200 proves a service is due.

Beat error is the humbler number: the tick-tock asymmetry in milliseconds. Under 0.5ms is well adjusted; up to about 1ms is common and harmless in daily use; beyond that, the movement wastes energy restarting each swing asymmetrically, self-starting suffers, and adjustment is worthwhile — on most modern movements a simple job with a moveable stud carrier, on vintage fixed-stud calibres a fiddlier one best bundled into a service.

One habit ties the three numbers together: log them. A movement’s absolute figures matter less than their trend — amplitude sliding 30 degrees over two years is an oil-life story told early, and a rate that needed re-regulating twice in a year is whispering about wear. The machine’s cheapest gift is a baseline.

TWO PATTERNS WORTH KNOWING ON SIGHTBEAT ERRORwide, constant gap between linesMAGNETISEDsudden huge gain · often clean linesHEALTHY AGAINafter demagnetising — 5 minutesA dramatic overnight rate change with clean traces is magnetism until proven otherwise. Demagnetise before diagnosing.
Fig. 4 — Magnetism mimics catastrophe and costs nothing to fix. The most common dramatic finding — a watch suddenly gaining minutes a day — usually shows clean, steeply sloped traces. That is a magnetised hairspring, cured by a five-minute pass over a cheap demagnetiser, and it should always be ruled out before any conclusion about wear or damage.

Using it on watches you are about to buy

For a buyer, thirty seconds on a timegrapher converts a seller’s “runs great” into one of six rows. The first three cost little or nothing to resolve; the last three reprice the watch by the cost of a service or more. Sellers who own machines increasingly post screenshots — read them with the caveats above, especially the single-position problem and, on Powermatic-family movements, the escapement artefact.

Chronographs get one extra step: trace with the chronograph stopped, then running. A modest amplitude drop when engaged is physics — the movement is driving extra wheels. A collapse, or a rate that lurches, means the chronograph works are dragging, and chronograph services are the expensive kind. The 7750 page covers the full pre-purchase routine.

Phone apps deserve their honest sentence: they use the same acoustic principle through vastly worse microphones, and they are usable for rate on a quiet desk, marginal for beat error, and not to be trusted for amplitude. As a free first look they beat nothing; as a basis for buying decisions, the sixty-dollar machine earns itself back on the first negotiation.

And the authentication bonus: the machine reads beat rate directly, and beat rate identifies movement families. A claimed 28,800 vph Swiss automatic tracing at 21,600 has the wrong movement inside; a claimed mechanical ticking once per second is quartz. The timegrapher was built to measure health, but it incidentally audits honesty.

FROM READING TO ACTIONSteep slope, clean lines, normal amplitude?Regulation — quick and cheapHuge sudden rate, history of phones and magnets?Demagnetise first — five minutesLow amplitude at full wind, or snow on the trace?Service — regulation cannot fix thisUgly numbers but excellent wrist timekeeping?Check lift angle and calibre assumptionsConsistent throughout → most likely genuine
Fig. 5 — The machine sorts problems by price. Its real gift is triage: separating the five-minute fixes (regulation, demagnetising) from the real services (wear, dirt, damage) and the non-problems (wrong settings, exotic escapements). That sorting is worth more than the machine costs, on the first watch.
ReadingLikely meaningMoney consequence
In-spec rate, 250–300°, tight linesHealthy, serviced or luckyPay the asking price happily
Steep clean slope, good amplitudeNeeds regulation onlyMinor — negotiate a token
Sudden huge gain, clean linesMagnetisedTrivial — demagnetise and re-test
Low amplitude at full windService due — oils, dirt, springPrice a full service in
Snow, wandering rateWear, damage or deep dirtService plus risk — or walk
Chrono on: amplitude collapsesChronograph works draggingChronograph service — the dear kind

The four beginner errors

Each produces a confident wrong conclusion.

Wrong lift angle

Amplitude is computed from it. Default 52° on a calibre that needs 50° or 53° shifts the reading enough to fake or hide a fault. Set it per movement, every time.

Testing at half wind

Mainspring torque falls as it unwinds; low amplitude at half wind is normal physics. Full wind or a day’s wear before any conclusion.

One position, one screenshot

Positional spread is half the story and single-position readings hide it. Four positions minimum before judging — or believing — any watch.

Diagnosing exotic escapements as broken

Powermatic 80-family and other redesigned escapements confuse the machine’s acoustic assumptions. Days-on-wrist timekeeping outranks the trace on these calibres.

Check what your numbers mean

Photograph your trace and the watch. An instant read will sanity-check the numbers against the calibre’s specification.

Scan Your Watch

For high-value purchases, we recommend pairing your AI scan with an in-person inspection by a certified watchmaker — in-person inspection remains the gold standard.

Frequently asked questions

What do the numbers on a timegrapher mean?

Three things. Rate: seconds gained or lost per day, shown as the slope of the dotted trace — judge it against the movement's own specification, not zero. Amplitude: how far the balance swings in degrees, roughly 250–300 dial-up on a healthy movement at full wind — the movement's blood pressure. Beat error: the asymmetry between tick and tock in milliseconds — under about 0.5ms is well adjusted.

What is a good amplitude for a watch?

Roughly 250–300 degrees dial-up at full wind for most serviced movements, dropping 20–40 degrees in vertical positions, which is normal. Below about 200 at genuine full wind indicates a service is due — dried oils, dirt or a tired mainspring. Above about 315 risks knocking, where the balance overswings and strikes the escapement. Always confirm the lift angle is set for your calibre first, or the number is fiction.

What is lift angle and why does it matter?

The arc of balance swing across which the escapement interacts with it — the geometric assumption the machine needs to convert acoustic timings into amplitude. Calibres differ: commonly published figures include 50° for the ETA 2824-2 and 53° for the Seiko NH35, with large pocket-watch movements much lower, and 52° as the usual machine default. A wrong lift angle skews amplitude substantially while leaving rate and beat error readings valid.

What does snow on a timegrapher trace mean?

Scattered dots instead of clean lines mean the escapement's sound events are irregular or the machine cannot lock onto them: dirt, wear, damage, deep magnetism — or simply a noisy room or loose clamp. Re-seat the watch, quiet the environment and re-test; persistent snow at full wind is a service finding that no regulation will fix.

Why does my watch suddenly gain minutes per day?

Magnetism, almost always — a magnetised hairspring's coils stick together, shortening the effective spring and speeding the watch dramatically, usually with clean traces. Phones, laptop lids and bag clasps are the usual culprits. A demagnetiser costs very little and fixes it in five minutes. Rule this out before any other diagnosis of a sudden dramatic rate change.

Can I trust a timegrapher reading on a Powermatic 80?

Only partially. The machine's arithmetic assumes a classical Swiss lever escapement, and the Powermatic 80 family's redesigned, partly synthetic escapement produces sound signatures those assumptions misread — beat error especially can display alarmingly on a movement keeping excellent time. On these calibres, days of wrist timekeeping outrank the trace. Snow is still snow, though: genuine fault patterns remain meaningful.

Are phone timegrapher apps any good?

As a free first look at rate on a quiet desk, yes — the acoustic principle is the same. For beat error they are marginal, and for amplitude they should not be trusted: phone microphones and processing cannot resolve the intra-tick events the calculation needs. For buying decisions or service judgements, a dedicated machine is cheap enough that the first negotiation it wins pays for it.

Keep reading

Other brands