A self-registering barometer is a recording barometer that draws a continuous trace of atmospheric pressure on a paper chart wrapped around a clockwork-driven drum. Lucien Vidie patented the aneroid capsule in 1844, and Richard Frères of Paris built the design into the first practical barograph in the 1860s. A stack of evacuated metal capsules expands and contracts with pressure changes, levering an inked pen against the rotating drum. The result is a 7-day pressure trace meteorologists, ship captains and pilots still use to spot fronts and squalls hours before they hit.
Self-registering Barometer Interactive Calculator
Vary the capsule travel, chart span, pressure span, and drum period to see the effective lever ratio and recording scale.
Equation Used
The calculator converts the worked barograph travel numbers into an effective lever multiplication. Capsule travel divided by pressure span gives capsule sensitivity; pen travel divided by pressure span gives the chart scale.
- Pen travel is the full useful vertical chart movement over the selected pressure span.
- Lever ratio is the effective displacement multiplication from capsule stack to pen tip.
- Friction, linkage backlash, and capsule nonlinearity are ignored.
How the Self-registering Barometer Works
The heart of the instrument is a stack of Vidie capsules — thin corrugated metal boxes evacuated to near vacuum and held open by an internal spring. When atmospheric pressure rises, the capsules compress by a few hundredths of a millimetre per millibar. When pressure falls, the spring pushes them back open. That motion is tiny, often less than 1 mm of total travel across a 950-1050 mbar range, so a lever train multiplies the displacement by a factor of 80 to 120 before it reaches the pen tip.
The pen itself rides on an arm pivoted at the lever output. A clockwork-driven drum, usually rotating once every 7 days, carries a pre-printed chart marked in millibars vertically and hours horizontally. The pen lays down a continuous ink trace as the drum turns. If you notice the pen skipping or laying a broken line, the cause is almost always one of three things — the lever pivots have stiffened from dried oil, the pen capillary has clogged with thickened ink, or the drum gear train has slipped a tooth. The aneroid capsule itself rarely fails because it has no fluid and no seal to leak.
Tolerances matter. The lever multiplication ratio must hold to within ±1% across the full chart, or the trace will read accurately at one end of the range and drift at the other. The pen tip must contact the chart at 3-5 grams of force — too light and the ink breaks up over rough chart paper, too heavy and the pen drags the lever and biases the reading low by 1-2 mbar. Real-world calibration against a mercury column or a digital reference is the only way to confirm both ends of the range track together.
Key Components
- Vidie aneroid capsule stack: A column of 4 to 8 evacuated corrugated metal capsules, each typically 50-60 mm in diameter, soldered in series. Stacking them adds the displacement of each capsule, so an 8-capsule stack delivers around 0.8 mm of travel across a 100 mbar pressure swing. The internal spring stiffness sets the slope of the response and must match the lever ratio.
- Lever multiplication train: A two- or three-stage lever system that converts sub-millimetre capsule travel into 30-50 mm of pen-tip motion. Pivots ride on jewelled bearings or hardened pinpoints, and the bearing friction must stay below about 0.5 grams-force at the pen tip or the trace lags behind real pressure changes by 5-10 minutes.
- Inked recording pen: A V-shaped capillary pen holding 2-3 weeks of ink reservoir. The pen tip contacts the chart at 3-5 grams of force, controlled by a counterweight or a fine spring. Special low-viscosity barograph ink resists drying for the full 7-day chart period.
- Clockwork drum: A brass cylinder, typically 90-100 mm diameter, driven by an 8-day spring-wound movement. Standard rotation is one revolution per week, giving roughly 0.6 mm of chart travel per hour. Drum eccentricity must stay under 0.2 mm or the trace will show a false sinusoidal ripple at the 7-day period.
- Chart paper: Pre-printed paper wrapped around the drum, ruled with pressure on the vertical axis (typically 950-1050 mbar in 1 mbar divisions) and time on the horizontal axis (7 days in 2-hour divisions). Paper grade matters — too rough and the pen skips, too smooth and the ink beads.
- Temperature compensation bimetal: A bimetal strip in the lever train cancels the thermal expansion of the capsule stack. Without it the trace would drift by roughly 0.3 mbar per °C of cabinet temperature change, swamping the real diurnal pressure signal.
Industries That Rely on the Self-registering Barometer
The self-registering barometer earns its place wherever a continuous pressure trace tells you something a single reading cannot. The shape of the trace — its slope, its kinks, its plateaus — is what forecasters and pilots actually read, and that shape is invisible on any spot-reading instrument. You would be amazed how many heritage weather stations, sailing schools and small airfields still keep a working barograph in service alongside their digital sensors as a redundant cross-check.
- Meteorology: The Met Office Lerwick observatory in Shetland ran Negretti & Zambra barographs for decades to log North Atlantic depression tracks before electronic loggers took over.
- Marine navigation: Ocean racing yachts in the Volvo Ocean Race historically carried a Thommen or Fischer barograph to spot the rapid pressure drop signalling a Southern Ocean front.
- Aviation: FAI-sanctioned glider flights up to the 1990s used sealed Peravia barographs to certify altitude gains for diamond and gold badge claims.
- Heritage observatories: Kew Observatory in London preserves a Richard Frères barograph from the 1880s as part of the original Royal Society pressure record set.
- Agricultural research: Rothamsted Research in Hertfordshire used a Short & Mason barograph to correlate frontal passage with crop transpiration data through the 1960s.
- Mountaineering huts: Swiss Alpine Club huts above 2500 m mount a Lambrecht barograph as a slow-storm-warning tool for parties planning summit attempts the next morning.
The Formula Behind the Self-registering Barometer
The pen displacement on the chart is what the practitioner actually reads, and it depends on the capsule stack's pressure sensitivity multiplied by the lever ratio. At the low end of typical sensitivity — a single capsule, no multiplication — you get maybe 0.2 mm of pen travel per millibar, which is too small to read against a chart ruling. At the high end — an 8-capsule stack with a 120:1 lever ratio — you get over 1 mm per millibar, which exceeds chart paper width and saturates the trace at the edges of a typical depression. The sweet spot for a 950-1050 mbar full-scale chart is around 0.4-0.5 mm per millibar.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| ypen | Vertical pen displacement on the chart | mm | in |
| kcap | Single-capsule pressure sensitivity | mm/mbar | in/inHg |
| n | Number of capsules in the stack | — | — |
| Rlev | Lever multiplication ratio | — | — |
| ΔP | Pressure change from chart zero | mbar (hPa) | inHg |
Worked Example: Self-registering Barometer in a beekeeping research station in New Zealand
A beekeeping research station in the Coromandel is restoring a 1962 Short & Mason Stormoguide barograph to log pressure trends against hive activity data — bees are known to reduce foraging 4-6 hours before a pressure drop of more than 6 mbar. The instrument has 6 Vidie capsules, each rated at 0.06 mm/mbar, driving a 90:1 lever train onto a 100 mm wide chart ruled 950-1050 mbar. The team needs to confirm the pen will track a 30 mbar depression cleanly without saturating the chart edges.
Given
- kcap = 0.06 mm/mbar
- n = 6 capsules
- Rlev = 90 —
- ΔPnom = 30 mbar
- Chart height = 100 mm (950-1050 mbar)
Solution
Step 1 — combine the capsule stack sensitivity with the lever ratio to get the overall chart sensitivity in mm per millibar:
Using the corrected single-capsule sensitivity of kcap = 0.0002 mm/mbar (manufacturer-typical for a 55 mm Vidie capsule):
Step 2 — at the nominal 30 mbar depression, compute pen travel:
That is a clearly readable trace deflection on a chart ruled in 1 mbar (≈3.3 mm) divisions — every millibar reads as roughly one ruling. The trace will not saturate the chart edges.
Step 3 — at the low end of the typical operating range, a 5 mbar diurnal swing on a calm day:
That is barely half a chart ruling — readable with a magnifier, but the diurnal tide is right at the limit of what an analogue pen can resolve. This is exactly why barograph users look at 7-day shape, not single-hour values.
Step 4 — at the high end, a 50 mbar deep low (a severe Tasman Sea depression):
Still well within the 100 mm chart height. A 100 mbar storm-of-the-century event would give 10.8 mm — the trace would lay a clean continuous line across the chart without ever pinning against the upper or lower paper edge.
Result
The nominal 30 mbar depression produces 3. 24 mm of pen travel, which lays one chart ruling per millibar and reads cleanly against the printed grid. The low-end 5 mbar diurnal swing gives 0.54 mm — sweet-spot detection territory, only just resolvable, which matches why the bee researchers care about 6 mbar drops and not 2 mbar wobbles. The high-end 50 mbar deep low gives 5.4 mm and stays well clear of chart saturation. If your measured deflection is 30% low, suspect a stiffened lever pivot (dried clock oil hardens after about 8-10 years and silently halves the multiplication ratio), a partially dried pen capillary that drags on the chart and steals lever force, or a slipped bimetal compensator that biases the zero with cabinet temperature. Confirm with a slow ±20 mbar sweep against a digital reference before trusting any data run.
Self-registering Barometer vs Alternatives
You have three real choices for a continuous pressure record: a self-registering barograph, a precision aneroid spot-reading barometer paired with manual logging, or a modern digital pressure logger. Each one wins on different axes, and the choice depends on whether you want a trace shape, a precise number, or a long unattended run.
| Property | Self-registering barometer (barograph) | Aneroid spot-reading barometer | Digital pressure logger |
|---|---|---|---|
| Pressure resolution | ±0.5 mbar on chart | ±0.2 mbar at the dial | ±0.01 mbar typical |
| Time resolution | 7-day chart, ~5 min minimum | Manual log, user-limited | 1 sample/sec or faster |
| Unattended run length | 7 days per chart | 0 — needs an operator | 6-24 months on battery |
| Initial cost | £400-1500 restored | £80-300 new | £150-600 new |
| Calibration drift | 1-2 mbar/year | 0.5 mbar/year | <0.1 mbar/year |
| Failure modes | Pen clog, lever stiffening, drum slip | Capsule fatigue, pointer slack | Battery, sensor offset, firmware |
| Best fit | Trace-shape forecasting, heritage stations | Simple spot reading, ship cabin | Long unattended logging, research |
Frequently Asked Questions About Self-registering Barometer
That wobble is almost always drum eccentricity, not a real pressure signal. If the brass drum is mounted with more than about 0.2 mm of runout on its bearing, the pen tip rides closer and farther from the lever pivot through each rotation, which changes the effective lever arm and modulates the trace at exactly one cycle per drum revolution — 7 days on a standard barograph.
To confirm, lift the pen off the chart, mark the drum's high spot with a pencil, and rotate by hand against a dial indicator. Anything over 0.15 mm runout needs the drum bearing re-shimmed or the spindle re-trued. A second cause, much rarer, is the chart paper itself being mounted with a slight wrinkle that compresses once per rotation.
Pick capsule count to match your chart height and target sensitivity, not to maximise sensitivity. An 8-capsule stack with a typical 100:1 lever gives roughly 0.16 mm/mbar — on a 100 mm chart ruled 950-1050 mbar, that means each millibar takes 1.6 chart rulings, and a 60 mbar storm pins the trace against the chart edge. A 6-capsule stack at 90:1 gives 0.108 mm/mbar, which fits a full 100 mbar range comfortably inside the chart with margin.
Rule of thumb: aim for full-scale pen travel to occupy 60-70% of chart height at your worst expected pressure swing. That leaves headroom for unusual events without compromising readability of normal weather.
Zeroing the pen only fixes the offset at one pressure. A 2 mbar bias that grows or shrinks across the range is a slope error in the lever multiplication, and the most common cause on a restored unit is a worn or replaced lever pivot that no longer sits at the original geometry. A pivot moved by 0.1 mm along the lever arm changes the multiplication ratio by 1-3% and biases mid-range readings.
Diagnostic check: take readings at 980, 1000 and 1020 mbar against a digital reference. If the error grows linearly with pressure, the lever ratio is off. If the error is constant, it's a true zero offset. Fix the slope first by adjusting the lever pivot position, then re-zero.
Marginally, and only the strongest events. A standard 7-day drum rotates at about 0.6 mm of chart travel per hour, so a 5-minute pressure spike covers 0.05 mm of horizontal trace — narrower than the pen tip itself on most barographs. The instrument will smear the event into a barely visible blip.
If gust-front detection matters, fit a 24-hour drum (one rotation per day instead of per week) which gives 4 mm/hour horizontal resolution, or pair the barograph with a digital logger sampling at 1 Hz. The 24-hour chart trades unattended run length for time resolution, and you'll need to swap charts daily instead of weekly.
That stair-step pattern is stiction in the lever pivots — the pen sticks at one position, real pressure changes load up the capsule stack, and eventually the lever breaks free and jumps to catch up. It's a different failure from the dried-oil pivot stiffening that biases readings low; stiction comes from contamination on the pivot surfaces, often dust bonded by old ink mist or atmospheric salts in coastal installations.
Clean each pivot with a cotton bud and pure isopropanol, then re-oil with a single drop of clock oil per pivot. If the stair-stepping persists, the jewelled bearings themselves are pitted and need replacement — a job for an instrument restorer, not a kitchen-table fix.
Without compensation, the capsule stack's metal expands with cabinet temperature and reduces its evacuated volume slightly, which biases the pen high by roughly 0.3 mbar per °C. A barograph in a south-facing window can swing 15°C between night and noon, faking a 4-5 mbar diurnal pressure tide that isn't real.
To prove the bimetal compensator works, run the instrument inside a sealed enclosure at constant pressure (a closed Tupperware box is fine for a quick check) and warm the box from 15°C to 30°C with a hairdryer set on low. A working compensator holds the trace flat to within 0.5 mbar across the temperature swing. A failed or misadjusted compensator shows the trace climbing or falling smoothly with temperature — same shape as the temperature curve.
References & Further Reading
- Wikipedia contributors. Barograph. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
