A Compound Compensating Pendulum is a temperature-compensated pendulum that uses two or more metals with different thermal expansion coefficients, arranged so their length changes cancel and the effective distance from pivot to centre of oscillation stays constant. A well-built one holds rate within 0.1 seconds per day across a 10°C swing, compared to roughly 1 s/day per °C for an uncompensated steel rod. The purpose is to remove temperature as a variable in precision timekeeping. You see them in 19th-century astronomical regulators like the Riefler and Shortt clocks before quartz arrived.
Compound Compensating Pendulum Interactive Calculator
Vary the metal expansion coefficients and temperature swing to size the compensating bar length and see the cancellation diagram update.
Equation Used
The compensation condition makes the downward steel growth equal the upward compensating-bar growth. With steel at 11 um/m/C and brass at 19 um/m/C, the required total brass length is 11/19 of the steel length, about 0.58.
- Steel down-bars and compensating up-bars move the bob in opposite directions.
- Expansion coefficients are treated as constant over the temperature swing.
- Lengths are total effective bar lengths in the compensation train.
- Joint friction, humidity, and suspension changes are ignored.
Operating Principle of the Compound Compensating Pendulum
The Compound Compensating Pendulum, also called the Compound bar compensation pendulum in older horology texts, works on a simple principle that gets fiddly fast in execution. Two metals — typically steel and brass, or steel and zinc — are arranged in opposing parallel bars so that when the steel bars try to lengthen the pendulum downward with a temperature rise, the brass or zinc bars push the bob back up by an equal amount. The effective length stays put. Period stays put. Rate stays put.
The ratio of bar lengths is what does the work. Steel expands at roughly 11 × 10-6 /°C and brass at about 19 × 10-6 /°C. To make the downward steel growth cancel the upward brass growth you need the brass bars to total around 11/19 of the steel bar total length — call it 0.58. Get the ratio wrong by 5% and you reintroduce thermal drift you were trying to eliminate. Get it wrong by 20% and you may as well have built a plain steel rod.
What goes wrong in real builds: bar friction at the cross-pieces. If the brass bars cannot slide freely against their cross-bars as they expand, they bind, store strain, then release it suddenly when the load tips over a stiction threshold. You see this on a rate trace as a sudden 0.5 s jump rather than the smooth drift of an uncompensated pendulum. The fix is loose-fit pivot pins at every cross-bar joint and never paint the bars. The other classic failure mode is moisture absorption in wooden cases — wood swells, the suspension spring's effective mounting height shifts, and now the compensation is fighting a humidity error rather than a temperature error.
Key Components
- Steel down-bars: The primary length-defining bars running from the suspension spring down toward the bob. Coefficient of expansion around 11 × 10-6 /°C. Typically 4 or 6 bars in a gridiron arrangement, each 700-900 mm long for a 1-second pendulum.
- Brass up-bars: Counter-expansion bars carrying the load upward from the cross-frames. Coefficient around 19 × 10-6 /°C. Total length sized at roughly 0.58 × steel-bar total length to balance expansion.
- Cross-frames: Light brass plates joining alternating bars at top and bottom. Pin joints must be slip-fit — clearance of 0.05 to 0.10 mm — so bars slide freely as they expand. A press fit here will bind and produce step errors in rate.
- Suspension spring: A flat steel ribbon, typically 0.10 to 0.15 mm thick, that flexes as the pendulum swings. Defines the effective pivot point. Must be clamped, not pinched, since clamp height drift directly enters the period equation.
- Bob: The mass at the bottom, usually 5 to 8 kg of cast lead or brass for a seconds pendulum. Heavy enough to dominate the moment of inertia, with a fine-thread rating nut underneath to trim period in the 0.01 s/day range.
Real-World Applications of the Compound Compensating Pendulum
Where you see a Compound Compensating Pendulum is anywhere a steady rate matters more than portability. The Compound bar compensation pendulum dominated precision timekeeping from roughly 1730 to the introduction of invar around 1900, and it still appears in restoration work, demonstration clocks, and reference instruments where the conservator wants the original mechanism rather than a modern substitute.
- Astronomical observatories: The Greenwich Observatory's Dent regulator clocks used gridiron Compound Compensating Pendulums to hold sidereal time within 0.1 s/day before they were retired in favour of Shortt free-pendulum clocks in the 1920s.
- Public tower clocks: The Westminster Great Clock — Big Ben — uses a Compound bar compensation pendulum with a stack of pre-decimal pennies on the bob platform for trim, holding rate within 2 s/week despite the bell tower's 15°C seasonal swing.
- Precision regulator clocks: Vienna regulators by Lenzkirch and Gustav Becker fitted Compound Compensating Pendulums in their top-tier models from the 1860s onward, marketed against Black Forest clocks specifically on rate stability.
- Museum and university physics demonstrations: The Deutsches Museum in Munich runs a 2-second Compound Compensating Pendulum as a working exhibit showing the gridiron arrangement, with a temperature display next to it so visitors see the rate hold steady through the building's HVAC cycle.
- Horological restoration: Conservators restoring 19th-century English long-case regulators routinely rebuild damaged compound pendulums, sourcing matched steel and brass bar stock and re-cutting cross-frames to original drawings rather than substituting an invar rod.
The Formula Behind the Compound Compensating Pendulum
The compensation condition tells you the bar-length ratio that makes total thermal length change equal zero. At the low end of the typical operating range — say a stable observatory basement at 18°C — even a poorly compensated pendulum looks fine because the temperature barely moves. At a nominal 10°C annual swing in a heated room, the ratio matters and a 1% error shows up as roughly 0.5 s/day drift. At the high end, an unheated tower or a workshop swinging 25°C between summer and winter, the ratio must be within 0.5% or you'll see seconds-per-day errors. The sweet spot for a careful build is matching the ratio to within 0.2% — past that, suspension spring temperature effects and air density variation dominate and further bar-tuning is wasted effort.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Lbrass | Total length of brass (or zinc) up-bars contributing to upward expansion | m | in |
| Lsteel | Total length of steel down-bars contributing to downward expansion | m | in |
| αsteel | Linear thermal expansion coefficient of steel, ≈ 11 × 10-6 | /°C | /°F |
| αbrass | Linear thermal expansion coefficient of brass, ≈ 19 × 10-6 | /°C | /°F |
Worked Example: Compound Compensating Pendulum in a watchmakers' college teaching gridiron pendulum
A horology programme at a technical college is building a 1-second gridiron Compound Compensating Pendulum for student rate-measurement labs. The classroom sits in an old brick building with no climate control — temperature swings 14°C between a January morning and a July afternoon. The brief is to size the steel and brass bar lengths so the pendulum holds rate within 0.05 s/day across that swing. Effective steel down-bar total length Lsteel = 994 mm (the standard length for a 1-second pendulum). Coefficients αsteel = 11 × 10-6 /°C, αbrass = 19 × 10-6 /°C.
Given
- Lsteel = 994 mm
- αsteel = 11 × 10⁻⁶ /°C
- αbrass = 19 × 10⁻⁶ /°C
- ΔT (annual swing) = 14 °C
Solution
Step 1 — apply the compensation ratio to find the brass bar total length needed:
Step 2 — at nominal conditions (a 14°C swing, perfect ratio), the residual length change is essentially zero. Period change scales with √(L), so rate drift is below 0.01 s/day. The pendulum holds within the brief.
Step 3 — at the low end of build error, suppose the student cuts the brass to 580 mm instead of 575.4 (a 0.8% over-cut). Net length change across the 14°C swing is:
That tiny residual translates to about 0.05 s/day rate drift — right at the brief's tolerance. Acceptable but no margin.
Step 4 — at the high end of build error, suppose the brass bars come in at 590 mm (a 2.5% over-cut, easy to do if the student forgets to subtract cross-frame thickness):
Now the pendulum drifts roughly 0.17 s/day — over three times the brief — and worse, it drifts in the opposite direction to a plain steel rod, which makes diagnosis confusing for students who have not seen over-compensation before.
Result
Nominal answer: brass bar total length Lbrass = 575. 4 mm against a steel total of 994 mm. In practice a student cutting to ±0.5 mm holds rate within 0.02 s/day across the seasonal swing — the clock visibly stays in step with a GPS time reference for weeks at a stretch. At the low-error build (580 mm) drift just meets the 0.05 s/day brief; at the high-error build (590 mm) the pendulum over-compensates and runs faster in summer than winter, the opposite of an uncompensated rod and a useful teaching moment. If a measured rate drift exceeds the predicted value, check three things in order: (1) cross-frame pin joints — if any are press-fit instead of slip-fit the bars bind and inject step errors that look like compensation failure, (2) suspension spring clamp — a loose clamp lets the effective pivot height shift by tens of microns with temperature, dwarfing the bar effect, (3) bob trim nut backlash — if the rating nut threads have play, the bob settles to a different height each time the case is moved, masquerading as thermal drift.
Choosing the Compound Compensating Pendulum: Pros and Cons
The Compound Compensating Pendulum sits between a plain steel rod (cheap, drifts badly) and a modern invar rod (expensive material, almost no drift to begin with). The Compound bar compensation pendulum was the dominant precision solution for nearly two centuries because it gave observatory-grade rate stability using ordinary workshop materials. Today you choose it for authenticity in restoration work or for educational demonstrations of how compensation actually works.
| Property | Compound Compensating Pendulum | Plain steel rod pendulum | Invar rod pendulum |
|---|---|---|---|
| Rate stability across 10°C swing | ≤ 0.1 s/day if ratio within 0.5% | ≈ 10 s/day | ≈ 0.05 s/day |
| Material cost (1 s pendulum) | £150-300 for matched steel and brass stock | £20-40 for a single steel rod | £200-500 for invar bar |
| Build complexity | High — 5-9 bars, slip-fit pin joints, careful ratio sizing | Trivial — single rod with rating nut | Moderate — single rod but invar machines poorly |
| Failure mode | Bar binding at cross-frames produces step errors | Steady seasonal drift, no surprise behaviour | Slow long-term creep from invar's secular dimensional change |
| Application fit | 19th-century regulator restoration, teaching, demonstration | Domestic wall and mantel clocks | Modern precision regulators, scientific reference |
| Service interval before re-rating | 5-10 years if joints stay free | 6-12 months — re-trim each season | 10-30 years — set and forget |
Frequently Asked Questions About Compound Compensating Pendulum
You are over-compensated. The brass bars are too long relative to the steel, so as temperature rises the brass pushes the bob up by more than the steel lengthens it downward, the effective length shrinks, and the period shortens. The clock runs fast.
Quickest check: measure the brass total length and the steel total length, then compare the ratio to αsteel/αbrass ≈ 0.579. If your ratio is above 0.585 you are over-compensated. The fix is to shorten the brass bars or, if you cannot disassemble, fit a small steel slug to add downward expansion. Under-compensation produces the opposite symptom — runs slow in summer.
The bar count does not change the compensation principle — it changes the symmetry and the manufacturing tolerance budget. A 5-bar gridiron (3 steel, 2 brass) is simpler to cut and assemble but the bob hangs slightly off the geometric centre, so any side-loading at the suspension spring shows up as small period asymmetry. A 9-bar gridiron (5 steel, 4 brass) is fully symmetric, which matters if you are chasing 0.05 s/day or better.
For a tower clock or domestic regulator the 5-bar is fine. For an observatory-grade build or a teaching pendulum where students will measure and diagnose errors, go to 9 bars so the geometry does not muddy the diagnostic signal.
Two causes, both real. First, the wooden case absorbs moisture, swells vertically, and shifts the suspension spring clamp upward by tens of microns. Effective length increases, period lengthens, clock runs slow. Second, air density changes with humidity — denser air gives more buoyancy on the bob (reducing effective weight) and more aerodynamic drag (changing the swing amplitude, which couples weakly into period through circular error).
For a regulator-grade build, mount the suspension cock to a metal back-plate not directly to wood, and seal the case if you can. The Riefler clocks of 1900 went one further and ran the pendulum in a partial vacuum specifically to kill humidity and density effects.
Yes, and it was common in late-19th-century practice. Zinc has a thermal expansion coefficient of about 30 × 10-6 /°C, much higher than brass. That means you need shorter zinc bars (ratio Lzinc/Lsteel ≈ 0.37) for the same compensation, which makes the pendulum more compact and lighter.
The trade is mechanical: zinc creeps under sustained load. Over years the bars slowly elongate plastically, the compensation ratio drifts, and the clock needs re-rating. Brass does not creep meaningfully at room temperature. Choose zinc for compactness in a museum demonstration; choose brass for a clock you want to leave running for decades.
A flat steel ribbon 0.10 to 0.15 mm thick by about 8 mm wide is standard for a 5-7 kg bob on a 1-second pendulum. Thinner than 0.08 mm and the spring fatigues at the clamp edges within a few years. Thicker than 0.20 mm and the bending stiffness becomes a meaningful fraction of the gravity restoring torque, which adds a temperature-dependent term you have not compensated for — steel's modulus drops about 0.025% per °C, and that bleeds straight into rate.
Rule of thumb: keep the spring thin enough that its restoring torque is under 1% of the gravity restoring torque at full amplitude. Above that, you have introduced a second thermal error your gridiron is not correcting.
Yes — same mechanism, two names. Compound Compensating Pendulum is the modern horological term you will see in restoration manuals and Wikipedia. Compound bar compensation pendulum is the older descriptive name, more common in 19th-century English texts and German-translated horology references where the bar arrangement is being emphasised over the function.
Both refer to the same family of pendulums where multiple parallel bars of different metals cancel thermal length change. The gridiron pendulum is the most famous specific example.
References & Further Reading
- Wikipedia contributors. Gridiron pendulum. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
