Reconstruction of John Harrison’s H4 for Derek Pratt. Escapement, Remontoir and timekeeping. This is the world’s first precision marine chronometer

This is the third part of a three-part series about Derek Pratt’s reconstruction of John Harrison’s Longitude Award-winning H4 (the world’s first precision marine chronometer). This article was first published in The Horological Journal (HJ) in April 2015, and we thank them for generously granting permission to republish on Quill & Pad.
To learn more about Derek Pratt, see the life and times of the legendary independent watchmaker Derek Pratt, Derek Pratt’s reconstruction of the John Harrison H4, the world The first precision marine astronomical clock (part 1 of 3), and John Harrison’s H4 for the diamond tray reconstructed by Derek Pratt, the world’s first precision marine chronometer (part 2, There are 3 parts in total).
After making the diamond tray, we move on to get the watch ticking, albeit without remontoir, and before all the jewels are finished.
The large balance wheel (50.90 mm in diameter) is made of a hardened, tempered and polished instrument panel. The wheel is clamped between two plates for hardening, which helps reduce deformation.
Derek Pratt’s H4 balance wheel hardened plate shows the balance at a later stage, with the staff and the chuck in place
The balance lever is a slender 21.41 mm mandrel with a waist circumference reduced to 0.4 mm for mounting the tray and balance chuck. The staff turns on the watchmaker’s lathe and finishes in the turn. The brass chuck used for the pallet is fixed to the worker with a split pin, and the pallet is inserted into the D-shaped hole in the chuck.
These holes are made on the brass plate using our EDM (electric discharge machine). The copper electrode according to the cross-sectional shape of the pallet is sunk into the brass, and then the hole and the outer contour of the worker are processed on the CNC milling machine.
The final finishing of the chuck is done by hand using a file and a steel polisher, and the split pin hole is made using an Archimedes drill. This is an interesting combination of high-tech and low-tech works!
The balance spring has three complete circles and a long straight tail. The spring is tapered, the end of the stud is thicker, and the center tapers toward the chuck. Anthony Randall provided us with some 0.8% carbon steel, which was drawn into a flat part and then polished into a cone to the size of the original H4 balance spring. The thinned spring is placed in a steel former for hardening.
We have good photos of the original spring, which allows us to draw the shape and CNC mill the former. With such a short spring, people would expect the balance to swing violently when the staff stands upright but is not constrained by the jewelry on the balance bridge. However, because the long tail and hairspring become thinner, if the balance wheel and hairspring are set to vibrate, only supported on the lower pivot, and the jewels above are removed, the balance shaft will be surprisingly stable.
The balance wheel and the hairspring have a large connection error point, as expected for such a short hairspring, but this effect is reduced by the tapered thickness and long tail of the hairspring.
Let the watch run, driven directly from the train, and the next stage is to make and install the remontoir. The axis of the fourth round is an interesting three-way intersection. At this time, there are three coaxial wheels: the fourth wheel, the counter wheel and the central seconds driving wheel.
The internally cut third wheel drives the fourth wheel in a normal manner, which in turn drives the remontoir system consisting of a locking wheel and a flywheel. The gyro wheel is driven by the fourth spindle through a remontoir spring, and the gyro wheel drives the escape wheel.
At the fourth round connection, the driver is provided to the remontoir, the contrate wheel and the center second wheel for Derek Pratt’s H4 reconstruction
There is a slender slender mandrel counterclockwise, passing through the hollow mandrel of the fourth wheel, and the second hand driving wheel is installed on the counterclockwise dial side.
The Remontoir spring is made from the mainspring of the watch. It is 1.45 mm high, 0.08 mm thick, and approximately 160 mm long. The spring is fixed in a brass cage mounted on the fourth axle. The spring must be placed in the cage as an open coil, not on the wall of the barrel as it is usually in a watch barrel. To achieve this, we used something similar to the former used to make balance springs in order to set the remontoir spring to the correct shape.
The Remontoir release is controlled by a pivoting pawl, a locking wheel and a flywheel used to control the remontoir rewind speed. The pawl has five arms mounted on the mandrel; one arm holds the paw, and the paw engages with the release pin on the opposite mandrel. When the top spins, one of its pins gently lifts the pawl to the position where the other arm releases the lock wheel. The locking wheel can then rotate freely for one turn to allow the spring to be rewinded.
The third arm has a pivoting roller supported on a cam mounted on a locking axle. This keeps the pawl and pawl away from the path of the release pin when rewinding occurs, and the reverse wheel keeps rotating. The remaining two arms on the pawl are counterweights that balance the pawl.
All these components are very delicate and require careful manual filing and sorting, but they work very satisfactorily. The flying leaf is 0.1 mm thick, but has a larger area; this proved to be a tricky part because the central boss is a person with the weather vane.
Remontoir is a clever mechanism that is fascinating because it rewinds every 7.5 seconds, so you don’t have to wait for a long time!
In April 1891, James U. Poole overhauled the original H4 and wrote an interesting report on his work for the Watch Magazine. When talking about the remontoir mechanism, he said: “Harrison is describing the structure of the watch. I had to grope my way through a series of troublesome experiments, and for several days I was desperate to be able to reassemble it. The remontoir train’s The action is so mysterious that even if you observe it carefully, you cannot understand it correctly. I doubt whether it is really useful.”
A miserable person! I like his relaxed honesty in the struggle, maybe we have all had similar frustrations on the bench!
The hour and minute movement is traditional, driven by a large gear mounted on the central spindle, but the central seconds hand is carried by a wheel located between the large gear and the hour wheel. The central seconds wheel rotates on the big gear and is driven by the same count wheel mounted on the dial end of the spindle.
Derek Pratt’s H4 H4 movement shows the driving of the big gear, the minute wheel and the central second wheel
The depth of the central second hand driver is as deep as possible to ensure that the second hand does not “jitter” when it is running, but it also needs to run freely. On the original H4, the diameter of the driving wheel is 0.11 mm larger than that of the driven wheel, although the number of teeth is the same. It seems that the depth is deliberately made too deep, and then the driven wheel is “topped” to provide the required degree of freedom. We followed a similar procedure to allow free running with minimal clearance.
Use the topping tool to obtain the smallest backlash when driving the central seconds hand of the Derek Pratt H4
Derek has completed three hands, but they need some sorting. Daniela worked on the hour and minute hands, polished, then hardened and tempered, and finally blued in blue salt. The central seconds hand is polished instead of blue.
Harrison originally planned to use a rack and pinion adjuster in the H4, which was common in edge watches of the time, and as shown in one of the drawings made when the Longitude Committee inspected the watch. He must have given up the rack early, even though he had used it in Jefferys watches and used a bimetallic compensator for the first time in H3.
Derek wanted to try this arrangement and made a rack and pinion and started making compensating curbs.
The original H4 still has a pinion to install the adjuster plate, but lacks a rack. Since H4 does not currently have a rack, it is decided to make a copy. Although the rack and pinion are easy to adjust, Harrison must have found it easy to move and disrupt the speed. The watch can now be wound freely and is carefully installed for the balance spring stud. The mounting method of the stud can be adjusted in any direction; this helps to position the center of the spring so that the balance bar stands upright when resting.
The temperature-compensated curb consists of brass and steel bars fixed together with 15 rivets. The curb pin at the end of the compensating curb surrounds the spring. As the temperature rises, the curb will bend to shorten the effective length of the spring.
Harrison had hoped to use the shape of the back of the tray to adjust for isochronous errors, but he found that this was not enough, and he added what he called a “cycloid” pin. This is adjusted to make contact with the tail of the balance spring and accelerate the vibration with a selected amplitude.
At this stage, the top plate is handed over to Charles Scarr for engraving. Derek had asked for the nameplate to be inscribed as the original, but his name was engraved on the edge of the skateboard adjacent to Harrison’s signature and on the third wheel bridge. The inscription reads: “Derek Pratt 2004-Chas Frodsham & Co AD2014.”
Inscription: “Derek Pratt 2004 – Chas Frodsham & Co 2014″, used for Derek Pratt’s H4 reconstruction
After bringing the balance spring close to the size of the original spring, time the watch by removing material from the bottom of the balance, making the balance a little thicker to allow this. The Witschi watch timer is very useful in this regard because it can be set to measure the frequency of the watch after each adjustment.
This is a bit unconventional, but it provides a way to balance such a large balance. As the weight slowly moved away from the bottom of the balance wheel, the frequency was approaching 18,000 times per hour, and then the timer was set to 18,000 and the error of the watch could be read.
The figure above shows the trajectory of the watch when it starts from a low amplitude and then quickly stabilizes to its operating amplitude at a steady rate. The trace also shows that the remontoir rewinds every 7.5 seconds. The watch was also tested on an old Greiner Chronographic watch timer using paper traces. This machine has the function of setting slow running. When the paper feed is ten times slower, the error is magnified ten times. This setting makes it easy to test the watch for an hour or more without sinking into the depths of the paper!
Long-term tests did show some changes in speed, and found that the center second drive is very critical, because it needs oil on the big gear, but it needs to be a very light oil, so as not to cause too much resistance and reduce the balance range . The lowest viscosity watch oil we can find is Moebius D1, which has a viscosity of 32 centistokes at 20°C; this works well.
The watch does not have the average time adjustment as it was later installed in the H5, so it is easy to make small adjustments to the cycloidal needle in order to fine-tune the speed. The cycloidal pin was tested in different positions, and sooner or later it would touch the spring during its breathing, and there were also different gaps at the curb pins.
There does not seem to be an ideal location, but it is set where the rate of change with amplitude is minimal. The change in rate with amplitude indicates that remontoir is necessary to smooth the balance pulse. Unlike James Poole, we think remontoir is really useful!
The watch was already in operation in January 2014, but some adjustments are still needed. The available power of the escapement depends on the four different springs in the watch, all of which must be balanced with each other: the mainspring, the power spring, the remontoir spring, and the balance spring. The mainspring can be set as required, and then the holding spring that provides torque when the watch is wound must be sufficient to fully re-tighten the remontoir spring.
The amplitude of the balance wheel depends on the setting of the remontoir spring. Some adjustments are needed, especially between the maintenance spring and the remontoir spring, in order to get the correct balance and get enough power in the escapement. Each adjustment of the maintenance spring means disassembling the entire watch.
In February 2014, the watch went to Greenwich to be photographed and photographed for the “Explore Longitude-Ship Clock and Stars” exhibition. The final video shown in the exhibition described the watch well and showed every part being assembled.
A period of testing and adjustments took place before the watch was delivered to Greenwich in June 2014. There was no time for a proper temperature test and it was found that the watch was over-compensated, but it ran the workshop at a fairly uniform temperature. When it operated undisturbed for 9 days, it stayed within plus or minus two seconds a day. In order to win the £20,000 prize, it needs to keep time within plus or minus 2.8 seconds per day during the six-week voyage to the West Indies.
Completing Derek Pratt’s H4 has always been an exciting project with many challenges. At Frodshams, we always give Derek the highest evaluation, whether as a watchmaker or as a pleasant collaborator. He always generously shares his knowledge and time to help others.
Derek’s craftsmanship is excellent, and despite many challenges, he has invested a lot of time and energy in advancing his H4 project. We think he will be satisfied with the final result and are happy to show the watch to everyone.
The watch was exhibited in Greenwich from July 2014 to January 2015 with all five Harrison original timers and many other interesting works. The exhibition started a world tour with Derek’s H4, starting from March to September 2015 at the Folger Shakespeare Library in Washington, D.C.; followed by Mystic Seaport, Connecticut, from November 2015 to April 2016; then From May 2016 to October 2016, travel to the Australian Maritime Museum in Sydney.
The completion of Derek’s H4 was a team effort by everyone in Frodshams. We also got valuable help from Anthony Randall, Jonathan Hird and other people in the watch industry who assisted Derek and us in completing this project. I would also like to thank Martin Dorsch for his help with the photography of these articles.
Quill & Pad would also like to thank The Horological Journal for allowing us to republish the three articles in this series here. If you missed them, you might also like: The life and times of the legendary independent watchmaker Derek Pratt (Derek Pratt) Rebuilding John Harrison (John Harrison) ) H4, the world’s first precision marine chronometer (part 1 of 3) for Derek Pratt (Derek Pratt) to reconstruct John Harrison (John Harrison) to make diamond tray H4, the world’s first A precision marine chronometer (part 2 of 3)
sorry. I am looking for my school friend Martin Dorsch, he is a German watchmaker from Regensburg. If you know him, can you tell him my contact information? Thanks! Zheng Junyu


Post time: Aug-02-2021