The Micrometer

A two-monthly review of the engineering of time

Issue 8   November - December 2017

A Christmas story

Some years ago I wrote about a couple of lapses by a customer which, as we approach Christmas, readers of The Micrometer may find amusing.

     On opening up a stopped mantel clock to prepare the estimate, it appeared to have been regularly serviced and was in generally good condition – but just a minute, what is all that sparkly stuff coating the insides of the movement?  I quizzed the owner:

     “Oh, that’ll be the glitter from the swags of holly we always decorate the clock with at Christmas...

More common and far more damaging is plaster dust resulting from room redecoration.  Not uncommonly, the owner sees it as unnecessary to remove their clock from the scene of the action, and equally unnecessary to wrap it up hermetically to prevent the ingress of the said dust (not that excluding plaster dust is an easy thing as anyone who has attempted so to do will attest).


     Another failure arriving with this same clock was a modern quartz movement clock accompanied with a puzzled “I changed the battery but it now won’t go”.  I turned it over to be confronted by the Ever Ready completely wrapped in kitchen foil.

     “Oh, yes, I couldn’t get it to stay in, so I thought that if I wrapped it in silver paper it would not only hold it in place but it would also conduct the electricity”.

The customer was right on both counts, the latter guaranteeing that the electricity was conducted straight from one battery terminal to the other...

A replacement battery

Staying on the theme of batteries, few engaged in the electronics industry before the year 2000 will not be familiar with the excellent AvoMeter (‘Amps-volts-ohms-Meter’), but the Avo Model 8 requires a now difficult-to-get 15 volt battery for the high resistance ohms range.  The original battery (Ever Ready B121) consisted of ten 1.5 volt zinc-carbon cells in series, which in the replacement BLR121 are replicated by stacking ten 1.5 volt alkaline cells in series.  Unfortunately the quality of construction of this replacement is not impressive, the life is not good and the cost high (£12 or so)*.   As I still use my AvoMeter where the steadiness of the analogue meter needle is preferable to a flickering digital numeric display, I decided to see what I could do for myself.

* Noting the Japanese writing on the ends of the button cells, I leave readers to judge which parts of the battery were 'Made in Britain' as advertised on the side of the yellow sleeve containing the manufacturer's roughly torn up bits of polystyrene packing.  (I am reliably informed by my brother that the writing is in katakana used for the transcription of foreign words, in this case  ボタンセル  which is pronounced bo-to-n-se-ru  (i.e. 'button cell').)


     The solution I adopted was to use five 3 volt CR2032 lithium cells in series, which are obtainable for a total of £2 or so.  With a capacity in excess of 200mAh and a perhaps 10-year shelf life, the performance should be far better than that of even a high-quality stack of 50mAh AG10s with their lower shelf life.  My casing is made from 25mm diameter acetal bar drilled and bored to take the lithium cells, and the brass terminal is a simple turning job that ideally would be plated to prevent oxidation leading to a poor contact.

A Belgian slip

Staying on a lighter theme, an 18th century longcase clock in for repair a few years ago had a neatly made slip washer ('key piece' in horologist's terminology) to the going barrel.   It is the sort of thing over which one’s imagination could run riot: noting the very crisp condition of the lettering the coin seems hardly used, so could it have been fitted by a repairer shortly after 1861 (the date on the coin)?  Was it something that was ‘liberated’ from Belgium during the First World War and fitted by Tommy Atkins on his return from the trenches to take up a career in clock repairing?  Or should one take more notice of the cheesehead BA-thread electrical screw used to lock the key piece into place, suggesting a later 20th century repair by a more economically-minded repairer?  Probably the most probable, though our more recent repairer may be a little disappointed were he to learn that Belgian 20 centime coins minted in 1861 now sell at around $25 (about £20) on the Internet in far, far poorer (worn) condition.

     I refitted the coin; after all it was a fascinating find, it was well made and fitted, and it worked very well.  All I now need to find are the owners of clocks whose repairers made their key pieces not out of centime pieces but gold sovereigns…

Another mignonette carriage clock

In the July 2017 issue of The Micrometer I mentioned the service of a mignonette carriage clock for a niece, and a few weeks ago another one came in for attention.  Originally bought as a wedding present and of great sentimental value, this one had fine champlevé enamel panels in a gilt Anglaise case.

     Checking Allix & Bonnert* (SBN 902028 25 1), mignonettes (or ‘little darlings’) made their appearance in the last quarter of the 19th Century.  The price of a plain mignonette in a plain corniche case in the 1907 Army & Navy Stores (London) catalogue listed the purchase price as £4.7s.0d (£4.35) at a time when a clock repairer might have been paid around £1.10s.0d (£1.50) a week and a solicitor might have expected to earn £500 a year.  By interpolation from the prices of other carriage clocks listed, the ornate case would perhaps result in the price of this mignonette being double this sum.


* As an aside, Allix & Bonnert ask the rhetorical question: 'What clocks did humble people have at this time?' to which they reply: 'American thirty hour ogee clocks which were so reliable and inexpensive... they would run for years on end without any attention, and when one did stop, a feather dipped in paraffin and brandished about indiscriminately in the works usually set it going again.'  Those, like me, who do not know why an ogee clock is so called, will want to check the bottom of this page.


     The issue of how carriage clocks ('pendules de voyage’ in French) were used (or carried ('carriaged'?)) is a talking point for many experts.  Clearly they could be carried, but to what extent they were used while travelling is a more vexed question.  I tend to think of them as being used like a travelling bedside alarm clock, i.e. a clock that would continue to keep time while being carried safely to and from (or between) temporary accommodation but only used on arrival or perhaps during overnight stop-overs.  But I do not know.

     Apart from the lack of a service for at least a decade, it had two primary faults; the attention of 'WD40' man (not the owner, I hasten to add) who appeared to have liberally sprayed the product inside the case so coating all surfaces with the wax residue this product is designed to leave, and an overfull barrel meaning that the mainspring could not develop the required number of turns to ensure a 7-day run time between windings.

     During the full service of the movement and platform which included removal of all wax residues and attention to the 0.5mm diameter upper balance staff end cap fixing screws, the heavily-coned mainspring (shown below both before and after removal from the barrel) was changed.  The replacement mainspring was also selected at 0.25mm thick (in a 22.1mm barrel), and the new one was cropped to 725 mm long, which is significantly shorter than the previously-fitted 950 mm long spring.  Even at 725 mm long, this is still about 5% to 10% over the theoretical optimum length but, through the 46.67 gearing ratio between barrel and centre arbor, did comfortably ensure a 7-day run time once the movement had been serviced.

     Students of horology will be able to work out the number of barrel turns a 7-day run represents, which is within the capacity of the (measured) maximum 5.1 turns available with the new spring fitted.  A pdf of the Excel spreadsheet I developed for these calculations can be viewed below.

     What these calculations show is that the theoretical maximum number of turns available from the barrel is some 24% higher than the actual number available (6.7 turns cf. 5.1 turns measured), which my initial thoughts suggest is caused by permanent set in the spring ends coupled with a reduction in the effective barrel volume by the protrusion of the barrel and arbor hooks.  The annotated image of the new spring (right) attempts to explain.

     Going back to the previous photo, the more advanced student will also notice the absent Geneva stopwork* from the barrel cover, which is not uncommonly missing from 19th Century carriage clocks.  Why it is missing is unclear; perhaps the peg on the wheel engaging the Geneva wheel had sheared off and the parts simply thrown away?  Or perhaps a past repairer had simply discarded it during an unsuccessful attempt to set it up correctly?  With 4.1 turns being needed for an 8-day run from the maximum 5.1 available and after allowing for the last one-third of a turn perhaps having insufficient strength to keep the clock going, accurate set-up of the Geneva mechanism would have been absolutely essential.


* For those who saved the download, Geneva stopwork and its set-up was discussed in the January 2017 issue of The Micrometer.


Stop press: Under Richard Price's expert narration, this carriage clock was recently featured on the BBC's Antiques Roadshow first transmitted on the 1st October 2017.  Valued by him at a retail price in excess of £5000, it was a privilege to be selected by the owner for its servicing.

Silk suspensions

The pendulum suspension sometimes found in early 19th Century French clocks is in the form of a silk thread hanging in a V-shape under the weight of the pendulum hooked onto it.  Nowadays I tend to replace the thread in synthetic pearl necklace stringing cord as being somewhat stronger than silk (the one fitted is unlikely to be either silk or original).  The problem with them is that, whatever material one uses, the suspension does not travel well when the clock is returned to its owner, which means that timing the clock in the workshop cannot be achieved with any great accuracy.  Sometimes for a short, gentle journey, leaving a light pendulum hooked in position is far from stupid, especially with synthetic 'silk' and the clock lying backwards at anything up to 45 degrees with a little bubble-wrap in the case to stop the pendulum rattling around.  Leaving the pendulum fitted is something that Laurie Penman* recommends, suggesting that a light rubber band be used to tension the suspension and so keep it V-shaped and hooked to the pendulum.  If the pendulum is removed, the silk will unravel from the rate adjusting 'windlass' leaving little choice other than to rewind the silk and re-time the clock at the owner's premises.


* The Clock Repairer's Handbook, Laurie Penman, 2000  ISBN 9780715311226

A fault with a platform escapement

An early 20th Century cylinder escapement carriage clock retailed by a jeweller from one of the former British Dominions has just come in with a decidedly 'graunchy' sounding mainspring (full of grit and lying in a decidedly non-uniform spiral on the bench out of its barrel, it was replaced as pat of the full service).  And with a balance amplitude of at best ±40 degrees it was also clearly in need of at least a clean, but another thing that I immediately noticed was a fault with the assembly of the platform escapement - see photograph.  Can you spot what is wrong with it?  (Answer at the bottom of the page.)

     As I mention above, the mainspring was replaced, and I chose a like-for-like mainspring (15 x 0.30 x 30 mm) - but on test things were not well.  Actually things went too well with an amplitude approaching ±180 degrees, the balance rim pin occasionally banking on the banking pin below the balance cock.  I decided there was no alternative but to fit a weaker mainspring, which, as a 0.28 mm thick spring was not available, I chose a 0.25 mm thick spring.  This reduced the amplitude to a comfortable ±135 degrees*, which is probably not too bad for a one hundred year old escapement.


*  For a watch which varies its position as it is worn or carried by the bearer, Gazeley suggests ±120 degrees is about right, but for a clock ideally I would have achieved ±150 degrees.

The photograph shows the clock on test, and after fifty-four 30-beat counts, the rate is just a little shy of the target 18000 beats per hour.  After the first few timing intervals to get things in the right ball-park, I always average the count over at least one full revolution of the contrate wheel to take out any fluctuation due to a less than perfectly true wheel or tooth spacing.  With the index lever close to central at the rate illustrated, no adjustment of the balance spring collet was needed, so final rate adjustment was left until the clock was re-installed in its case.


     And now a little test for the inquisitive horology student (answers at bottom of page):

(a) What error per day does 17998.83 beats per hour represent?

(b) Using the data below, what number of beats must be counted by the timing machine for one full revolution of the contrate wheel for this clock?

          Escape wheel: 15 teeth.   Escape wheel pinion: 8 leaves (teeth).  Contrate wheel (fourth wheel): 80 teeth.





USB charging point

Charging the battery in devices such as mobile phones and iPads is now almost a daily chore, and is greatly facilitated if one has access to a charging socket at table height.  Most such sockets - in the UK, 240 volt 13A 3-pin sockets - are found in the kitchen (for, for example, kettles, toasters, food processors and microwaves), but in all other rooms the sockets tend to be mounted at floor level.

     Some months ago, I noticed that 13 A wall sockets can be obtained with integral usb* charging sockets, so I decided to investigate on the basis that not only would this free up one or more of the 13A sockets for its originally intended use, but would get rid of that hideous plug adapter.  After a certain amount of research I discovered that there was some unsubstantiated concern that a wall-mounted usb charging point might not be suitable for iPads, but this was quickly set aside as probably Apple Inc. not wanting to be taking responsibility for equipment not supplied by themselves.

     Anyway, reliability and integrity are keywords for me in selecting any electrical equipment, so after studying the MK data sheet which gave me reassurance that their usb charging sockets had been tested on a number of devices, I bought and fitted a Logic Plus™ twin wall socket, part number K2743WHI.  Installation in the 35 mm deep pattress box (wall box) was a simple one-for-one exchange, and since then I have found it entirely successful in charging an iPad (see photograph) as well as charging HTC and Samsung Galaxy Smart phones.


*  The usb (Universal Serial Bus) is an amazingly user-friendly connection standard which is now commonplace as a means of interconnecting computer peripherals and power chargers.  Indeed, readers will have noticed that the mains (battery charging) plug supplied with their devices is an adapter that takes mains voltage and responsively delivers 5 volts DC to a usb socket mounted on its face.  Pull the usb plug out of the adapter and you can use the cable for downloading photos to your desktop - as well as charging from the adapter now incorporated into the MK usb wall plate.

Bryneglwys Slate Quarry

Two issues ago I published a few photos of the Abergynolwyn incline winding house, and this time perhaps a few photographs of Bryneglwys slate quarries taken by me at the same time (1968) may be of interest.  Now I am no expert of what the building were used for and cannot remember their precise location, but perhaps the map below may allow some correlation with the photos.  More detailed maps and some excellent descriptions and photographs are available in the book 'Bryneglwys Slate Quarry, Abergynolwyn, Merionethshire', by Alan Holmes with Sara Eade under ISBN-13: 978-0956565242 available at the Talyllyn Railway shop.

Source: Wikimedia Commons/public domain


     In the photos below, the first photo (this one is not mine; it is courtesy of the Talyllyn Railway Press Office) looks west and shows a general view of the lower Cantrybedd level in the 1950s.  Taken from near the top of the Beudynewydd Incline, the Cantrybedd incline winding house can be seen to the centre right.

     The second photo (1968) is a view of the Lower Mill seen in the previous photo taken looking to the south-west.  The third photo (1968) is perhaps of the Old and New Mill complex taken looking to the south.  The fourth photo (1968) is probably looking north-west towards the manager's house and, not shown on the map above, the associated buildings.

     In May 2017 I revisited the now fenced off quarry to find it heavily overgrown with little trace of any of the buildings.  The lower Alltwyllt (lit. 'wild hill') incline is accessible and a photo of the winding house taken on this visit is below (Photoshopped to remove the intrusive tourist signage).  However, the Cantrybedd (lit. 'Chantry (chapel?) grave') incline has all but disappeared under the vegetation.  Indeed, even as a teenager back in the sixties, I recall that the simple rail/sleeper type bridge across the stream at its lower end had collapsed making it something of a scramble, and today's footpath takes one well to the south of the incline.

West Country Clock and Watch Fair

The next West Country Clock and Watch Fair will be held on Sunday 19th November 2017 at the Holiday Inn, Taunton, TA1 2UA, which is close to Junction 25 on the M5 motorway.  Doors open at 9.00am and last entrance is at 2.15pm  For further details visit

     In addition to the many horological traders at the fair, both the General Manager and Membership Secretary/Short Course Officer of the British Horological Institute (BHI) will be at the BHI Stand.  Brian and Zanna will be most pleased to answer questions from enthusiasts, professionals and prospective students alike on the services the BHI can offer and, as someone who has greatly benefited (and is still benefiting) from his membership of the BHI, I can thoroughly recommend that visitors take a moment to say "hello".




Balance vibrating tool

This is the last issue in which the Balance Vibrating Tool constructional series will be available to readers, so if you have not already downloaded all parts, now is the time to do it.



Next time

The second and final part of a photographic essay in constructing a 3½ inch gauge locomotive boiler, and the design and construction of a Rotary Broach for ‘drilling’ square and hexagonal holes.



Guy Gibbons


Ogee clock

A wall or shelf clock apparently so called because of the ogee (S-shaped) surround to the case.  The one illustrated is by Seth Thomas of Connecticut.

(Photograph: JVGJ at the English language Wikipedia)


A fault with a platform escapement

The balance spring does not lie between the curb pins which, at two or three times the thickness of the spring, are also a little too widely spaced.  About 1.5 times the spring thickness is perhaps a good gap width between the curb pins (or curb pin and boot) to aim for.  The inner curb pin was also adjusted to leave the gap between it and the outer curb pin parallel; as received the gap was wider at the bottom as can be seen from the photograph.


(a)  3600 x 24 x (17998.83 − 18000)/18000  =  −5.6 seconds per day (slow).   As most platform escapements have an 18000 beats per hour escapement, a rule of thumb that is worth committing to memory is: 1 beat per second timing error represents a 5 seconds per day timekeeping error.


(b)  Number of beats for one full revolution of the contrate wheel = 15 x 2 x (80/8) = 300 beats.


Copyright (c) 2017  G E Gibbons

The Micrometer

A two-monthly review of the engineering of time

Issue 8   November - December 2017

The Micrometer

A two-monthly review of the engineering of time

Issue 8   November - December 2017

Copyright (c) 2017  G E Gibbons