Verification of John Harrison’s Claim of One Second in a Hundred Days?

At the time of writing (12th June 2015), numerous accounts of a recently claimed verification of John Harrison’s prediction of the performance of his ultimate land-based precision regulator system (as embodied in his final land-based regulator, popularly referred to as the ‘RAS Regulator’) may be found by conducting an internet search using, for example, the words: harrison one second in a hundred days

As described and explained elsewhere on this website ( Harrison’s claim (in his 1775 manuscript ‘Concerning Such Mechanism…’, often abbreviated as ‘CSM’, for which see the DOWNLOADS section) was that his precision regulator system should be capable of a mean rate to within one second in a hundred days.

Nothing would please me more than to be able to congratulate those involved in the creation of a modern embodiment of Harrison’s principles, capable of demonstrating his performance claims. Unfortunately, an adequate account of the processes by which those claims have recently been verified is, according to my investigations, not available.

May I remind those involved, together with the horological community, of the following, and ask that the principles therein be followed:

Left mouse click once on the following link for an explanation of The Scientific Method: The Scientific Method

May I draw particular attention to the paragraphs in the above link entitled: ‘Replication’, ‘External review’ and ‘Data recording and sharing’. In short, what is undoubtedly necessary is a detailed account of the dimensions, materials and processes involved in the creation of the test regulator, a detailed description of the methods of set-up and adjustment of the test regulator and a through account of the test environment, test equipment and test methods.

Regarding test equipment, particular attention must be paid to the extent to which the test regulator was exposed to variations in the ‘natural atmosphere’ (by which phrase I mean the Earth’s free atmosphere, external to the test building). Further attention must be paid to explaining the degree of sealing of the case from pressure and temperature variations in the ‘natural atmosphere’ and the extent of variation imposed upon the temperature of the ‘natural atmosphere’ (e.g. by any man-made heating system, such as hot water radiator(s) to the test area, amongst many other possibilities). I would expect that heating by sunlight would have been carefully eliminated, although confirmation would be included in any thorough explanation of the test area. I have assumed that variations in the pressure of the ‘natural atmosphere’ are transmitted with little, if any, delay to the test area, although confirmation would also be included in any thorough explanation.

May I also emphasise that, in CSM, Harrison includes a statement that two regulators are necessary for correct adjustment of his ultimate land-based regulator system. Any account of the recently conducted set-up, adjustment and test methods should therefore include an explanation of the means by which this requirement was eliminated.


The above post was published as a comment on which is a blog created by the team responsible for the test.

On 7th July, the following comment was also submitted for inclusion on that blog, in response to the publication of the test results by a team member, Mr Rory McEvoy:

Many thanks to Rory McEvoy for publishing the test results elsewhere on this site:

Unfortunately, comments regarding the test are not permitted below Rory’s post in the above link, so I’ll offer them here instead:

The results suggest an unnaturally constant test temperature (within 5 deg C), at very likely variance with that of the ‘natural atmosphere’ (by which phrase I mean the Earth’s free atmosphere, external to the test building). I doubt that the temperature of the ‘natural atmosphere’ of the test location (in the south of the UK) for an entire 100 days from April 2014 never varied by more than 5 degrees C. Given the variations in barometric pressure during the test period, the achieved performance is, therefore, at considerable odds with my understanding of Harrison’s compensation for pressure variation in the ‘natural atmosphere’ via (at least in part) an excess component of temperature compensation, as generated by exactly the same ‘natural atmosphere’ at, or close to, the same instant.

Put simply, despite Rory’s report, all of the points in my post of 8:15am, June 19th 2015 (see above) still apply, as do the questions and requests they raise.

In brief, what is therefore still required is a detailed account of the dimensions, materials and processes involved in the creation of the ‘Burgess B’ test regulator, a detailed description of the methods of set-up and adjustment of the test regulator and a through account of the test environment, test equipment and test methods. More detailed and especially essential requirements were described in my previous post. The objective must be to enable independent replication of Burgess B, the test set-up and the test conditions, followed by no less independent verification of the test results and Harrison’s claim of a mean rate to within one second in a hundred days.

Until others have confirmed the 100 day test results entirely independently (I repeat, entirely independently), Harrison’s claim and his land-based longcase regulator science will, in my opinion, remain no less shrouded in doubt, mystery and misunderstanding than they have been since the 24th of March 1776. Published results or not, a single Burgess B and a single test is, quite simply, not good enough, as any competent scientist would surely agree.

In 1775, in his final manuscript ‘Concerning Such Mechanism…’, Harrison envisaged that a copy of (something closely related to) his Final Regulator (now popularly referred to as the ‘RAS Regulator’) would sit in a dedicated building in at least every major port, all set to Greenwich Time, all no doubt achieving a mean rate to within one second in a hundred days. With that in mind, we should consider Harrison’s land-based precision regulator science and his eighteenth century vision to have been PROPERLY verified when we demonstrate that it can be consistently achieved by more than one timepiece, independently constructed and adjusted in accordance with his principles, without any manipulation of the ‘natural atmosphere’ by inappropriately far more modern means.

David Heskin (glathoppa)


All attempts to open any of the blog links mentioned in the above posts failed on 30th November 2015. All records of the claims have apparently disappeared, as have the comments and criticisms submitted by myself and Dr Paul Smith. Access may have been denied at any date before today, although I recall using the links a few months ago with success.

David Heskin (glathoppa)

John Harrison RAS Regulator ‘Replica’ – Video 1 of 13

More videos available at

John ‘Longitude’ Harrison’s ‘RAS’ Regulator ‘Replica’, researched and constructed (to the displayed state) from early 2003 to late 2005.

The original regulator, discovered after his death in 1776, was to be Harrison’s final, most perfect masterpiece of land-based precision timekeeping technology, claimed to be capable of a mean rate within one second in a hundred days (the theoretical limit of performance for a purely mechanical ‘clock’ exposed to the atmosphere). Unfinished by Harrison and tragically ignored after his death in 1776, despite vast superiority over contemporary timepieces (and apparently, if the predicted performance is correct, never bettered by any purely mechanical, single pendulum, in-atmosphere timepiece throughout history).

Optimised single pivot grasshopper escapement, thirty seconds spring remontoire, Harrison maintaining gear, anti friction arbor supports, radial wheel teeth, roller pinions, caged rolling element bearings to the great arbor and Harrison gridiron pendulum with suspension cheeks. Those devices, when optimised, constitute Harrison’s ideal combination of superior inventions.

Shown is an unpolished, uncased, ‘visual near-replica’ of the original, which is currently on display at The Royal Observatory, Greenwich, England.

Harrison Single Pivot Grasshopper Escapement

This short video presents the single pivot Harrison grasshopper escapement during construction and early testing of a ‘replica’ of his final longcase precision regulator. The date is estimated to be 2005. The pendulum and its suspension are temporary arrangements and finishing and polishing had yet to be completed.

The grasshopper escapement, in an alternative ‘twin pivot’ configuration, was originally devised by Harrison exclusively as a means of eliminating sliding friction and the consequent requirement for lubrication. Low performance oils and their rapid degradation was (and still is, albeit to a considerably lesser extent) a major source of timekeeping inconsistency. An inspection of the grasshopper action will confirm that there is no sliding friction, apart from insignificant rotation at the pallet arms pivot(s). Harrison subsequently devised two more versions of the escapement : the single pivot and the twin balance. He also identified additional, invaluable characteristics, although a complete explanation is beyond the scope of this brief description. See DOWNLOADS for a full explanation.

The cycle of operation is both mesmerizing and ingenious. The (dark coloured) pallet arms, typically carved from a suitable hardwood, are alternately captured by the clockwise-driven saw-toothed escape wheel. Static (not sliding) friction between an escape wheel tooth tip and the nib locking corner of the applicable pallet arm is all that binds them together. At an appropriate stage in the escapement cycle (shortly before the pendulum reaches the end of its swing), escape wheel recoil is induced by the second pallet nib locking corner, as it contacts, by careful design, a chosen escape wheel tooth precisely at its tip. Static friction at the first pallet nib is removed by recoil, the tail-heavy pallet arm is released and it pivots rapidly away from the escape wheel. At the commencement of recoil, the second pallet arm is captured by static friction and transmits escape wheel impulse to the pendulum. The impulse continues until the first pallet nib locking corner once again contacts an escape wheel tooth tip, induces recoil, releases the second nib, is itself captured and thereafter applies escape wheel impulse to the pendulum. A continuous cycle of the described events maintains the pendulum motion and, in combination with the train of wheels, pinions, hands and dials of the regulator movement, measures and displays the passage of time.

The nose-heavy brass ‘composers’ (identified in the video by the fitment of temporary, commercial adjusting screws) absorb the motions of captured pallet arms during recoil, act as stops for released pallet arms and position released arms correctly for future capture.

In a layout drawing, popularly referred to as ‘MS3972/3’, Harrison indicates that the mean of the start of impulse torque arms should be two thirds of the mean of the end of impulse torque arms. This is evident from three equispaced points, verified as such by Harrison’s annotations ‘1’,’ 2′ and ‘3’, point ‘1’ being at the same equal spacing from the escapement frame arbor axis. The stipulation is also repeated in Harrison’s 1775 manuscript ‘Concerning Such Mechanism’ (CSM), in which he declares that impulse increasing in a ratio of two-to-three is ‘a very important matter’, being related to his unique circular arc pendulum suspension spring cheeks, escapement recoil, degree of composer nose weighting etc. CSM adds an ‘approximate’ qualification to the two-to-three ratio, by way of an acknowledgement of the reduction in transmitted pallet arm forces, as their lines of action increasingly deviate from tangential to the escape wheel during each cycle. The manuscript also stipulates that a four-minute escape wheel should be used in combination with a ‘long pendulum’. For a seconds beating pendulum, one hundred and twenty escape wheel teeth are therefore required.

Harrison’s large ‘sea clocks’ (H1, H2 and H3) incorporate a version of the grasshopper in which each pallet arm impulses a symmetrical balance. By linking the balances such that they swing in opposition, the grasshopper cycle is accommodated (and the motions of a ship at sea are almost nullified). Again, MS3972/3 instructs that the ratio of torque arms be two-to-three.

For some reason, not yet convincingly explained, the twin pivot configuration was only used during Harrison’s early work and was not illustrated in MS3972/3.

Misinterpretations of Harrison’s grasshopper escapement stipulations are almost universal, perhaps because his poorly expressed intentions are easily misunderstood, have been completely ignored or, quite simply, because alternative constraints (or none at all!) are often considerably less demanding of the designer.