Harrison Maintaining Power: Winding and Torque Scenarios | FIRGELLI

← Back to Engineering Library

Harrison maintaining power keeps a clock’s going train driven while its main winding drive is temporarily unavailable. A loaded auxiliary spring supplies the train during winding. This calculator compares entered torque percentages and hold-time scenarios; it does not calculate spring dimensions, stored energy or how long a real clock will run.

Harrison Winding Device Interactive Calculator

Vary the normalized train torque, maintaining-spring torque window, and hold-time window to see the torque bridge during winding.

Min Torque
--
Max Torque
--
Min Hold
--
Max Hold
--

Equation Used

Tspring_min = Ttrain * pmin/100; Tspring_max = Ttrain * pmax/100; hold = tmin to tmax

This scenario calculator multiplies entered train torque by the entered spring percentages. The reference torque is normalized; it is not a measured or certified rating. Hold times are entered assumptions, not predictions of spring endurance.

  • Train torque is normalized to a chosen reference; no physical torque unit is calculated.
  • The spring percentage window is entered, not derived from a torque-versus-angle curve.
  • The hold-time window is an entered assumption, not a predicted run time.
Harrison maintaining-power mechanism during winding
Blue — Great wheel. Grey — Intermediate ratchet. Gold — Maintaining spring and frame click. The three wheels share an arbor. The frame click holds the intermediate ratchet while the barrel reverses for winding. The curved spring connects that ratchet to the great wheel, which continues forward. The smaller click is carried by the intermediate wheel. This illustrative cycle releases the spring from 90 to 72 degrees before reloading it; it does not determine an actual clock’s torque or running time. Tooth and spring shapes are explanatory rather than manufacturing profiles. Use the interactive animation above to follow the motion.

Drive continues during winding

In the Howard clock arrangement documented by Princeton’s Time Team, an intermediate ratchet wheel drives the great wheel through a circular spring. During normal running, the main drive keeps that spring loaded. During winding, a frame-mounted click prevents the intermediate wheel from reversing, while the spring continues driving the great wheel.

The sequence is the important distinction: the main drive loads the spring, the restrained spring supplies the train during winding, and the main drive restores the spring’s working deflection afterward. Actual arrangements vary between movements.

The illustration shows three coaxial wheels and two clicks, following the arrangement in Hiscox figure 1168. The smaller click travels with the intermediate ratchet; the other pivots on the clock frame. The curved maintaining spring remains attached to the intermediate ratchet and great wheel. During winding, the intermediate wheel holds while the blue great wheel advances and the spring relaxes. Its drawn length stays constant as it changes shape. The ten-second running, winding and reloading sequence uses an illustrative 90 to 72 degree spring deflection and is independent of the torque and hold-time inputs.

Examples in clock collections

The British Museum dates Harrison’s form of maintaining power to around the mid-1720s and describes its use by later makers. Its collection includes a marine chronometer with a six-turn fusee fitted with Harrison maintaining power.

Royal Museums Greenwich also holds a fragment of H1’s original maintaining-power spring. These records establish particular examples; they do not imply every precision clock or marine chronometer used the same arrangement.

What the calculator actually computes

p_low = min(p₁, p₂); p_high = max(p₁, p₂)
T_low = T_n p_low/100
T_high = T_n p_high/100
t_low = min(t₁, t₂); t_high = max(t₁, t₂)

T_n is an entered train torque expressed as a percentage of a reference torque. The spring percentages are relative to that entered train torque. Their products are percentages of the same reference torque; they are not N·m or lbf·in.

The two hold-time values are entered assumptions. The calculator sorts them into ascending order and displays them unchanged. It does not derive hold time from spring energy or test whether the entered torque is sufficient to sustain an escapement.

A physical hold-time estimate would need usable energy over the permitted spring travel and the train’s power demand, referred to consistent shafts and including losses. For a linear torsion spring, released energy between two angular deflections is ½k(θ₁² − θ₂²). Those variables are not inputs to this tool. The preset percentages and times are illustrative, not universal horological specifications.

An illustrative torque window

At the preset T_n = 100% of reference torque and spring inputs of 70% and 90% of train torque, the displayed spring window is 70–90% of reference. Entered times of 30 and 60 seconds appear as a 30–60-second assumed interval.

If T_n changes to 120% while the spring percentages remain 70% and 90%, the displayed window becomes 84–108% of reference torque. The hold interval stays 30–60 seconds because no energy or duration model connects it to torque.

Entering the larger percentage or time first does not produce a reversed range: the calculator sorts both pairs. Changing a hold input changes that displayed assumption, not the duration of the animation.

From scenario values to a real movement

Question What is needed
How much torque reaches the train? A measured spring torque curve, transmission ratios and losses
How long does maintaining power last? Usable spring travel and train power demand
Does winding disturb the rate? Timing and amplitude measurements on the particular movement
What spring or ratchet should be fitted? Movement-specific geometry, materials and competent horological assessment

This scenario tool does not specify preload angles, tooth counts, clearances, hardness, retrofit changes or service procedures for a historic clock.

Frequently asked questions

Are 70–90% torque and 30–60 seconds recommended settings?

No. They are illustrative presets. The required torque and duration depend on the actual movement and winding operation.

Does changing torque calculate a new hold time?

No. Hold times are separate entered assumptions. The tool only sorts and displays them.

What does a torque result of 84% mean?

It is 84% of the chosen reference torque. For example, 70% of an entered train torque of 120% gives 84% of reference.

Why does the animation use a fixed cycle?

It uses a ten-second sequence to make the running, winding and spring-reloading phases visible. The hold-time inputs remain separate scenario assumptions.

Does the diagram show the exact wheel arrangement?

It shows the coaxial arrangement, moving clicks and spring attachments. Tooth counts, profiles, spring shape and preload are illustrative; they are not dimensions or adjustment settings for a particular clock.

Can this tool specify a replacement spring?

No. It does not include spring material, geometry, torque versus angle or available travel.

Where can I see documented examples?

The British Museum describes a marine chronometer fitted with Harrison maintaining power, and Royal Museums Greenwich records a surviving part of H1’s original maintaining-power spring.

References and further reading

Building or designing a mechanism like this?

Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.

← Back to Mechanisms Index
Share This Article
Tags: