Skip to navigation – Site map
Dossier

Conservation of an Early Electric Master Clock: Making a Copy as a Conservation Tool

Françoise Collanges

Abstracts

The project was to research how the making of a copy could contribute to the study and conservation of an electric master clock, c. 1855, by Constantin Detouche and Jean-Eugene Robert-Houdin. A copy of the clock was produced as well as an electric cell similar to the one used in 1855 and they were both tested, which made possible to determine how to safely run the original clock.

Top of page

Full text

Introduction

1The conservation of dynamic electric objects, especially of electric clocks, is not a frequent topic in the professional literature for conservators. As they are perceived as recent objects, electric devices still tend to be on the margins of the heritage world. In so far as they are housed in public collections, they are usually treated as static objects, made of different types of materials conserved next to one another, without investigating how they would have worked together and what consequences could have resulted from this.

2This is particularly the case for early electric clocks produced during the 19th century by inventors in small workshops or manufactories, for which little is known about how they run, to the point that sometimes it is difficult to determine if the object is complete or potentially dangerous to run. However, some of these objects are used nowadays, mostly restored or repaired by collectors, clockmakers or electricians, with much passion but little knowledge of conservation ethics and techniques.

Fig. 1 Electric Master clock, 1855, by Constantin Detouche and Jean-Eugene Robert-Houdin

Fig. 1 Electric Master clock, 1855, by Constantin Detouche and Jean-Eugene Robert-Houdin

Front and back views.

London, The Clockworks, picture FC

3The project of conserving an electric master clock, c. 1855, by Constantin Detouche and Jean-Eugene Robert-Houdin was thus a challenging exercise. The clock belongs to a private collection, exhibited on appointment at the Clockworks1 a Trust foundation supporting research and conservation of electric clocks. Part of the collection, which is one of the most complete in the world on this topic, is in running order. Moreover, historical research revealed that the clock under study was a milestone in the history of domestic electric horology.

4Hence, the project of conserving the Detouche clock aimed at understanding how it ran, to check if it was possible to run it again and, if yes, how this could be done. In that sense, this was a restoration project, as the term was defined in 2008 by ICOM2.

5Very little was known about the object. Moreover, no conservation guidelines on how to assess electric objects or conserve them could be found. So, the first action was to devise a methodology, based on three phases:

  • Researching the history of the object and gathering any technical information on its materials and how it ran;

  • Building a copy of the part of the mechanism, namely the electrified part, which needed to be fully understood in order to assess the condition of the original clock;

  • Testing the copy clock in order to address three topics: how does this mechanism run?    How can the condition of the original clock be assessed? Should the original clock run, and, if yes, what should be done to ensure safe running?  

6The present paper will present these three phases and the results achieved. In the following text, we will refer to the original object as “Clock A” and to the copy as “copy clock”.

Historical research and technical information

  • 3 See the clock by Guiseppe Zamboni, in Verona (Italy), College San Maffei. For details about his res (...)
  • 4  The first electric clock was patented in 1841 in the United Kingdom by Alexander Bain; at the same (...)

7The history of electric horology started as early as 1815, with first experiments on controlling a pendulum with electric power3.  However, this remained as laboratory experiments for twenty years, before clockmakers and pioneers of electric engineering started to think about practical applications and industrial developments for electric horology. However, the first systems considered as efficient enough for an industrial production appeared in the 1840’s.4

8The Great London Exhibition in 1851 was an important step for electric horology in general, but more particularly for the genesis of clock A. That year, Constantin Detouche (1810-1889), clockmaker in Paris, rue Saint-Martin, won a gold medal at the Exhibition for a mechanical astronomical regulator. He probably also realised, from the objects on display in the exhibition, that electric horology could be the future and would at least be a part of the horological market which needed to be explored.

  • 5  Patent taken with Brisbart-Gobert for an electric clock (Du Moncel 1880, p.127).

9After a first unsuccessful project in 1852,5 he associated with Jean-Eugène Robert-Houdin (1805-1871), who had worked on practical applications of electricity for the previous twenty years, but in a different field: Considered nowadays as the inventor of modern conjuring, Robert-Houdin had tested for his tricks all kinds of electric devices, including clocks, as his first training was as a clockmaker. Retired from show business in 1852, he decided to dedicate his life to the invention of electric systems, and especially to the making of an affordable reliable electric clock for domestic use Robert-Houdin 2006, p.427).

10This was new, as electric horology was then mainly developed to answer the needs of railways, factories and telegraphy industries, more than for domestic comfort. Robert-Houdin’s project, commercially developed by Detouche at first, and later by other makers, was developed and improved for the following fifteen years. However, the first main step was taken in 1855 during Paris’ Great exhibition, where the first electric clocks were shown on Detouche’s stand, while Robert-Houdin took out a patent for several models of electric clocks.

 Fig. 2 J.E Robert Houdin – Drawing

 Fig. 2 J.E Robert Houdin – Drawing
  • 6  3d March 1855, brevet d’invention n.22648, with addition on the 7th July 1855, 36 p, Paris, Instit (...)

Drawing associated to patent 22648 (detail)6.

Paris, INPI.

11This document describes, amongst other electric inventions, a first model of a mantel clock. He modified it some months afterwards and the drawing actually conserved with the patent shows the exact design of Clock A. The text describes briefly the principle on which the clock runs, indicates it was meant to drive a slave dial and to be powered by a Daniell’s cell. It thus provided us with major pieces of information:

  • The mechanism visible on the clock was complete; some of the materials used were described.

  • The clock was meant to drive a slave dial, which was missing;

  • The power source was a Daniell’s cell; this was one of the first efficient power sources, ancestor to today’s battery. Its voltage was about 1.1 volt.

  • The way the clock ran could be understood and the main components to check determined: the Daniel’s cell was attached to the circuit and the power was meant to compensate the inevitable loss of energy of the pendulum by giving an impulse to the pendulum every other swing. For that, the electric power was transformed into a mechanical force through the use of a coil.

12A first consequence of this information was that making a copy was a valid project, as Clock A was complete and the texts were precise enough to understand how it should run.

  • 7  One is in a private collection in the USA; two are in public collections: Oxford, Museum of Histor (...)

13Further research was performed to frame more precisely the making aspect of the project. Several clocks of the same type were found and examined.7 This indicated several choices:

  • the slave dial was an independent unit, which would not prevent the clock running if it was missing. The copy would not have to be completed by a copy of the slave dial.

    • 8  These analyses were performed at West Dean College and in Oxford’s Museum of History of Science wi (...)

    the materials to use could be determined: silk covered copper wire, platinum for the contacts; ivory for several insulating parts; soft iron for the coil; CZ108 brass for the flat springs, CZ120 brass for the other parts. XRF analysis8, performed on the original clock as well as on the Oxford’s Museum clock, enabled the precise composition of the metals and helped to identify modern equivalents.

    • 9  The Science Museum in London also brought precious information as a 19th century Daniel’s cell cou (...)

    the type of power source gave an important indication on the electric circuit: it should run on a low voltage and further research (Figuier 1856, p.143) indicated that, provided that the cell was properly maintained, the clock should run on the same cell for a year without major problem. Daniell’s cell were used, amongst other types of cells or batteries, until the end of the 19th century on an industrial scale, but few are conserved nowadays. So the choice was made to reconstruct a cell from a 19th century description (Daniell 1843, p.504) and to test and monitor its performance9. It was an important issue, as the way the clock could run was directly linked to the regular flow of current. Was a Daniel’s cell stable enough? How quickly was it decaying? These were important questions in order to understand the way Clock A could run in a reliable way.

  • the sensitive elements to test were the ones involving electric current. The motion work, a train of gears driving the hands, which was an entirely mechanical part and made in a traditional way, did not need to be fully copied. Its use would have been to check if the copy clock was keeping accurate time, which was not regarded as an essential component of this current study.

14At this stage, the materials for the copy clock could be sourced and the making planned. At the same time, a first assessment of Clock A could be attempted.

Building the copy clock and assessing Clock A

15These tasks were deeply linked, as making the copy meant taking measurements of the original pieces and manipulating them. This made it the right moment to also assess their condition. The measurements needed for the copy were both dimensions but also electric characteristics, which implied using bench testing meters and at each step, to consider how to use these devices to obtain valid data without damaging the original object.

Assessing Clock A: materials, components and condition

16Mechanical wear of the components, as well as traces of degradation linked to the electric current, were to be assessed.

Copper wire and risk of circuit breakage

17 The integrity of an electrical circuit relies greatly on the quality and condition of the wires connecting different elements. It has to be a metallic wire and be properly insulated from any contact with other metals to avoid circuit breakage and to protect persons in contact with the wires. The first wires used were made of copper insulated with silk. Degradation can occur by breakage of the copper wire and by wear of the insulation.

18In Clock A, both happened (Fig. 3).

Fig. 3 view of the back of clock A

Fig. 3 view of the back of clock A

Fragile and degraded wire

Credits: © Françoise Collange

  • 10  Heat shrink tubing is an extruded plastic which shrinks when slightly heated. It is used to protec (...)

19The main issue was probably the contact points between the flexible wire and rigid materials like the iron armature or the brass legs. This resulted in the replacement of some part of the wires at a later period (Fig.). It was probably at this moment that one of the ivory rings separating the wire from the back of the movement into the feet was broken. The wire used now was a modern plastic covered copper wire. Its diameter was similar to the original but the composition of the copper was probably slightly different. It has been soft soldered to the wire exiting of the coil and protected at the exit point of the feet with heat shrink tubing.10

20At the other end of the coil, linked to the lower contact, the wire was cotton covered with a more recent appearance. The remains of green silk were visible at the entry point to the bobbin so it can be surmised that the copper wire is original and was later covered by a cotton cloth to protect the circuit in case of contact between the wire and the other components surrounding it.

  • 11  Silverline ACDC 513121. The machine sends a low current when the probes are applied to a surface. (...)

21To detect if the circuit was broken, a multimeter11 was used. The probes were connected to the terminals at the back of the clock and proved the circuit was complete.

The coil: corrosion and risk of short circuit

22An electromagnetic coil in a clock is a device used to transform an electric force into a mechanical one: the current circulating into the coil is creating a magnetic field which will be used to move mechanical parts.

23 Depending on the mechanical effect wanted, the coil can have different forms. In Clock A, the structure is very simple, with a rectangular armature on which a coil of wire is fixed by a screw on one side. On the other side, the end of the coil is attracting or repelling a mobile part of the armature.

24From a practical point of view, the coil used in Clock A was an object made of five different materials (in the present case, copper wire, soft iron armature, brass, silk insulation, ivory bobbin and insulation rings) cut at specific dimensions.

 Fig. 4 View of the different components

 Fig. 4 View of the different components

Components of the coil in the copy clock, and coil in clock A.

Credits: © Françoise Collange

25Observation of the original coil showed several areas of degradation: the top of the external armature was corroded; some corrosion spots were visible on the mobile armature at its point of contact with the coil; under the screw head fixing the coil to the fixed armature, some black corrosion products were observed.

26None of these elements seemed to be dangerous to the mechanical running of the coil. The main area of concern remained the wire and the electric behaviour of the coil. When the coil was made, the layers of wire were apparently wound on each other without other insulation between them than their silk coverage. If the insulation of the wire was broken in one of the layers on the coil, it could generate a short circuit while the clock is running, which would end in burning the coil. Thus the coil needed to be handled and used with care.

The platinum contacts: erosion caused by sparking

27The platinum contacts showed spots of corrosion.

 Fig. 5 The corroded platinum

 Fig. 5 The corroded platinum

 Contact under the flat spring.

Credits: © Françoise Collange

28This is linked to a phenomenon called sparking. In an electric contact, the critical point is not when the contact is closing but when it is opening. At that point, the current tends to try to maintain itself, which increases the voltage at the contact point, until the mechanical force finally separates the two elements making contact. When this happens, the energy is suddenly dissipated by producing a spark or an arc. These are corroding the contact surfaces until the current stops flowing through the contacts. As it was noted that the contact on the flat spring is original, it is important to find a way to preserve it.

Conservation choices: risk assessment to run the clock

29The conservation survey pointed at two main forms of damage:

  • the copper wire was partly replaced and its insulation was fragile and already partly missing;

  • the original contact was pitted by sparking.

30However, the electric circuit was complete and the clock seemed to be in working order. So a controlled current could be applied to the object in order to obtain some electric measurements, but could it be applied safely for longer periods?

31The degradation of the copper wire will result sooner or later in breaking the circuit, and potentially in a place, like inside the coil, which would make any restoration impossible without a major modification of the clock. The wire degradation is difficult to ascertain or measure as it cannot be manipulated and observed easily. However, some electric and electro-magnetic tests may be achieved to gather more information on that point.

32Sparking has always been one of the major issues of electric horology, and what is precisely happening when the spark occurs is still not completely understood. During the 20th century, several solutions have been developed to limit the sparking, mainly based on adding electronic components. However, for Clock A, as all the mechanism is visible, it seems difficult to do so without degrading the object. If the spark could not be quenched, a solution may be found in using sacrificial contacts which could be fixed in a reversible manner to the original ones. This could be tested on the copy clock.

33The action of the current appeared clearly as a risk for the clock, but its impact was difficult to evaluate. Investigating more about the characteristics of the current used originally, and thus the power source for Clock A, and performing tests on the copy clock would prove helpful in this respect.  

Making the copy clock

34The making of the copy clock took 500 hours of intense practical work. The materials were sometimes difficult to source and/or to work. However, the experience of practical work was massively informative as it revealed the amount of practical skills and the diversity of industries which were pulled together to produce new objects such as electric clocks in the mid-19th c.

35The use of XRF analysis was very useful to determine the composition of the materials but also to find modern equivalents to them, both by reading the spectra but also by using the PDF library included in the machine.

36The most difficult component to copy was the coil and this illustrates the kind of issues met while trying to copy an historical object.

The coil

37Nowadays, soft iron used for coils is difficult to find. Traditional Swedish iron is no longer available as other magnetic iron-based alloys have been discovered, like iron-nickel ones, which are now commonly used.

38Soft iron is also difficult to machine as it is very pure and tends to have a structure mixing soft and harder crystalline parts. Clock A has a very large armature, so it was also difficult to find a piece of iron with the right dimensions. The closest iron available was in fact soft iron for forging.

39An XRF analysis was performed on a sample of this iron and compared to the original material, as well as to a material purchased on the internet advertised as being magnetic iron. The latter revealed itself very close to the original metal but with an even higher degree of purity and it was nickel plated.

Fig. 6 XRF analysis spectrum of the soft iron in Clock A

Fig. 6 XRF analysis spectrum of the soft iron in Clock A
  • 12  The material was kindly offered by a customer of the clock workshop at West Dean College.
  • 13  58x43x17x9 mm maximal dimensions for the L shaped part, which was tapered at the end holding the m (...)

40Soft iron for forging was thus selected as it had an appropriate composition and was available in larger sections.12 The outside armature was forged from a square bar by “upsetting” the metal at one point then bending it when red hot until the angle was created. It was then milled and filed to dimension.13

41The same material was turned using a Myford Super 7 lathe to a round section for the core of the coil. Both sides were drilled and tapped. On one side, a head was turned on the lathe then screwed in. In the original, the head is made in the same block as the core.

  • 14  Ivory alternative, Units 3 and 3A, Hambrook business Centre, cheesemans Lane, Hambrook PO188XP.
  • 15  Plus or minus 0.1 mm.

42Clock A bobbin was made from ivory. As a substitute, a polypropylene imitation was used. This material was made near Chichester (West Sussex) by a small production unit14. A colour and a material, which may be fibre glass, is mixed in the plastic and after polishing, shows veins similar to ivory. Polypropylene is a very stable polymer, often used in conservation, and it was fitting the purpose of providing a hard insulating surface to wind the copper wire on. It revealed to be slightly brittle and so difficult to turn in any other way than at high speed with a tungsten cutting tool. The material was drilled first to the dimension of the iron core then turned down to reproduce the two heads of the bobbin. The copy of the bobbin is still slightly larger in some dimensions but never more than 0.5 mm, and the most important dimension, in between the heads, is accurate.15

43As the bobbin and armature were made as close as possible to the original, the next step was to find a silk covered copper wire similar in dimensions to Clock A and to find out what length was necessary to wind the coil. As the original one could not be unwound, measurements of the wire were taken at its connection to the plate. It was a wire of 0.9 mm diameter including insulation. The difference between the bobbin's minimum diameter and the maximum diameter was 8.2 mm. Therefore, with a wire of 0.9 mm diameter, the number of layers of the coil was calculated as being:

448.2 / 0.9 = 9.11 layers

45However, the wire was entering and exiting the bobbin on the same side, so it must be an even number of layers, eight or ten. Eight was the most probable considering the width of the bobbin heads.

  • 16  Wires.co.uk: 125 gr enamelled and double artificial silk on copper wire (Ref: sc-0710edasc-125).

46The closest diameter of wire which was commercially available was 0.75 mm diameter including insulation.16 It was wound on the bobbin by hand using the lathe to hold the bobbin. The winding needed to be very regular in order to optimize the number of turns and eliminate the risk of wire moving under the effect of current while the clock runs. The result was to fit eight layers on the bobbin. The four first layers had 54 turns and the four last one had 56 turns. Sheets of 80 gr paper were put between the layers in order to make them more regular.

The Daniell’s cell

47This is a wet cell, generating current by putting in filtered contact a plate of copper and a rod of zinc.

 Fig. 7 Copy of a Daniell’s cell.

 Fig. 7 Copy of a Daniell’s cell.

Détails

Credits: © Françoise Collange

48The copper bathes in sulphuric acid and the zinc rod bathes in a solution of copper sulphate contained in a porous vase, which is placed in the middle of the sulphuric acid solution. As a consequence, the vase tends to be destroyed by the acid and few of them are actually conserved nowadays. From the examples found in the Science Museum in London, an unglazed porcelain vase was turned in the ceramic studio at West Dean College. Four tests were made and monitored.

49Once the cell and the copy clock were made, tests and monitoring could be organised.

The copy clock: testing and analysing an electric system

  • 17  The electric experiments were designed with the help of Kenneth Cobb, clock conservator and electr (...)

50There were no guidelines to study an electric system for conservation purposes. Now that the copy was made, it was possible to use it for preliminary tests17 and then for comparison tests with Clock A.

 Fig. 8 Experiment (1)

 Fig. 8 Experiment (1)

Experiment to monitor the current with an oscilloscope as the copy clock ran on a Daniell’s cell.

Credits: © Françoise Collange

51However, the time was limited and three objectives were defined:

  • Testing to obtain electro-magnetic values: from the usual methods to assess a modern electro-magnetic system, which ones were usable for an historic object? What information do they give?

  • Testing to understand if the power source was important for conservation: This was based on analysing the Daniell’s cell and comparing the current produced with the one provided by a modern bench power supply. Was an original source better for Clock A? Could a modern source like a bench power supply be used as a conservation device?

  • Testing a conservation method: would it be possible to control the degradation process of the contacts by sparking through the use of sacrificial contacts fixed in a reversible way onto the original ones?

Usable electro-magnetic values

52The coil is a main component of the system. Focusing on this component seemed like a sensible way to analyse the electro-magnetic behaviour of the clocks. It can be defined by several electrical and magnetic characteristics:

    • 18  The letter code for these values are following European norm denomination (as visible in Guye and (...)

    Its resistance (R)18: capacity of a material to resist the flow of electric current (I) when a voltage (U) is applied across the terminals of the coil under steady state conditions. Unit: ohm (Ω)

  • Its inductance (B): Quantity of magnetic flux created by an electrical coil and associated ferrous armature in an electromagnetic circuit. Unit : Henry (H)

  • Its pull (P): the mechanical weight the electromagnetic coil can lift at a definite intensity of electric current flowing through the coil of N turns wound around a ferrous armature. Unit: gram (gr).

  • Its “electromotive force” expressed in Ampere-turns (NI): Product of the number of amperes (I) flowing through the number of turns of wire in the coil. The ampere turns produces a magnetic flux which is enhanced by the armature material.

53These values should be determined for the new coil. However, they would have little meaning if they could not be compared to the ones of the original coil:

    • 19  An attempt to measure the inductance of both coils was made using a LCR meter (LT Putron LCR 9063) (...)
    • 20  A test could be performed by wiring in series the original and the copy coils, attaching resistors (...)

    Resistance and inductance could not be measured for the coils alone. The meters available were not precise enough to give reliable values19 on such low resistance circuits and a direct experiment could damage the original coil.20

  • Pull could be evaluated by experiment through using small weights hanging from the coil at a given current. However, the coil in Clock A is fixed to the legs, which does not permit manipulation without danger. Each time the copper wire is bent it can create cracks and hardened zones, which will make it more fragile.

  • The Ampere-turns: the current through the coil could be measured but the number of turns could only be estimated for the original coil because it could not be dismantled.  

54As a result, trying to achieve precise values for the copy coil alone seemed to be of limited use, as they could not be compared to the original coil. Looking at the whole object as a system, so looking at the circuit and not only at the coils, was revealed to be more possible and significant.  

55From that point of view, important values were:

  • Current and voltage used by the circuits

  • Resistance of the circuits

  • Inductance of the circuits

56To measure them, several tests and calculations were performed. It has to be reminded that the voltage (U), the current (I) and the resistance (R) are values which are defined by a specific relationship, described by Ohm’s Law:

57U=IR

  • 21  Using an oscilloscope Tektronix TDS 1012, 100 MHZ, 1Gs/s; bench power supply Roban Varex 30-2 Twin (...)

58As a consequence, to compare values for the two clocks, U, R or I needs to be determined first, then the others can be observed and calculated in relation to it. This value is the average current needed by the clock to run. The average current was determined by a first experiment.21

59The results were:

60For the copy clock: I=167 mA
For clock A: I=206 mA
From that, the resistance of the circuit could be calculated for a power source of 1.1 volt (A Daniell’s cell)
For the copy clock: R=V/I=1.1/0.167=6.58 ohms
For Clock A: R=1.1/0.206=5.34 ohms

61At a lower current, for example when a simple meter was connected to the clocks, the resistances of the clocks were:

62For the copy clock: 1.7 ohms
For Clock A: 3.2 ohms

63These values showed already that the two clocks are different. As they are simple and rather short circuits, it can be inferred that the resistances represent mostly characteristics of the coils. So Clock A has a higher resistance, which means it will draw more current out of its power source.

64It can also be compared to the resistance measured with the same meter on the Clock in Oxford’s Museum. This clock had a resistance of 1.7 ohms, closer to the copy clock. This could indicate a fragility of the coil in Clock A: as the wires aged, it may have faults in its insulation or micro-cracks, which are increasing the resistance of the coil.

65Measuring the inductance of the circuit completed the impression of the electro-magnetic behaviour of both clocks.

 Fig. 9 Experiment (2)

 Fig. 9 Experiment (2)

Experiment to monitor the current in order to calculate the inductance of the coil in the copy clock

Credits: © Françoise Collange

66This also required the use of an oscilloscope to observe specific points of the current variations in time. It provided more accurate information about the coils in both clocks. The graphs show a significant difference, which points to a better magnetic efficiency in Clock A.

Fig. 10 Inductance curves

Fig. 10 Inductance curves

 Top: inductance curve of the coil in the copy clock; bottom: inductance curve of the coil in Clock A.

Credits: © Françoise Collange

  • 22   Guye and Bossart 1957, p.49.

67This could be linked to a fault in the making of the copy clock coil: After being made, the soft iron armature of the coil, which was stressed during the making, should have been annealed following a special process to improve its magnetic permeability22. This was done in Clock A, improving greatly its electro-magnetic qualities.

Power sources

  • 23  This experiment required the use of an electronic kit K8055 USB Experiment interface board (P8055- (...)

68The study of the electric behaviour of the cell23 showed that, if connected to the copy clock, it was providing a regular current from the start and remained stable, whereas when attached to Clock A, the current decayed slowly from the start, with a final quick drop during the last hour of a three days period.

69The copy made of the Daniell’s cell was far from perfect. It was made from a description of 1843, more than ten years before Clock A was made, and Daniell’s cell were most probably improved during that time. However, the description of 1856 (Figuier 1856, p.143) indicated that some copper sulphate should be added to the cell every week, which fits with the present results, as the copy cell provided sufficient current for about three days. In that sense, producing a Daniell’s cell closer to the original appeared interesting from a restoration point of view, to better show how Clock A was initially running, but a less desirable solution from a conservation point of view, as it generates a slightly less regular current than a bench power supply.

70This was confirmed by a second series of tests, using the oscilloscope, which showed significant voltage peaks when the contact closed. More interferences were visible on the curves when the clocks were running on a Daniell’s cell than when they were powered by the bench power supply.

 Fig. 11 Spectrums

 Fig. 11 Spectrums

Top: spectrum of the current going through the copy clock powered by a Daniell’s cell; bottom: spectrum of the current going through Clock A powered by a bench power supply.

Credits: © Françoise Collange

71This indicated that the bench power supply was potentially a way to limit mechanical wear on the contacts by smoothing those peaks while also providing a more regular running, and so less original, than a Daniell’s cell. The bench power supply appeared as an important tool for conservation for other reasons: It is a more secure way to start the clock as it allows increasing the current little by little; It can simulate lower current than any actual batteries, which all start at a minimum of 1.5 volt; It is safe from an electric point of view being equipped with fuses and safety systems.

72However, the experiments also showed that electric clocks produce specific sounds and movement given the type and intensity of current, and providing them with a smoother source of power drastically modifies their running, a fact which needs to be considered from a conservation point of view.

Spark erosion and sacrificial contacts

73This last experiment was continued after the end of the Masters project, as it relied on the running of the copy clock for several hundreds of hours, with checks of the sacrificial contacts every hundred hours under an electronic microscope to see if the current was damaging the material under the contact. At the end of the MA, 200 hours of running had not created major damage.

 Fig. 12 Experiment

 Fig. 12 Experiment

Platinum contact on the copy clock after 200 hours of running: first discernable spots of erosion

Credits: © Françoise Collange

  • 24  The final version was clear shellac applied on the spring for 10 seconds, then the sacrificial pla (...)

74The conductive glue24, chosen after testing several materials and processes of gluing, proved to be both efficient and reversible. It was still protecting the material underneath after 600 hours of the clock running.

Conclusion

75Identified as a milestone in the history of electric clocks and their introduction to a domestic setting, the electric clock designed by Robert-Houdin was one of the first affordable electric systems to display time at home.

76Making a copy of the movement and of a contemporary electric cell as well as performing several tests on them informed three main research areas:

  • Knowledge and understanding of the original object and its condition;

  • Testing processes and electro-magnetic values which can be used in electric horology conservation.

  • Conservation processes which could be used on the original object;

77The historical research as well as the running tests, underlined the importance of running the clock to fully preserve its meaning. However, the fragility detected at several points within Clock A, and the use which can be made of an adapted power supply, led to consider not running it continuously but only for limited periods, on a current less than 0.37 Amp, preferably around 0.2 Amp, provided by a constant source.

78Using the copy clock for electric tests greatly improved the perception of which fundamental electric data could be significant in the conservation of an historic electric clock. It showed that they cannot be determined for isolated components, but have to be considered for the complete circuit. The impact of this information on fault finding was essential as it implied that several tests and measurements would be necessary to identify a weakness, like the increase in resistance due to degradation of the wire. These tests also helped to determine techniques to study an electric system which would be sound from a conservation point of view.

79The experiment of the use of sacrificial contacts as a conservation technique was also successful as it underlined how quickly sparking degrades contacts, and showed that conductive glue could be used as an alternative to soldering. In many ways, if the time dedicated to building (500 hours) and testing (a minimum of 200 hours) the copy seemed extensive, the final benefits appeared equally important.

Top of page

Bibliography

BERNER, G.A., L’horloger-electricien, Bienne and Besançon, E.Magron, 1926

DU MONCEL, Th., Exposé des applications de l’électricité, tome quatrième, applications mécaniques de l’électricité, Paris, 1880

DANIELL, J.F., Introduction to the study of chemical philosophy, London, 1843, p.504

Electrifying Time, Wadhurst, AHS, 1976

DAUMAS, M.(dir.), “Horloges électriques”, Histoire generale de techniques, 5. Les technologies de la civilisation industrielle: transformation, communication. Facteur humain. Paris, PUF, 1979, pp.81-85

“Detouche’s neue elektrische Uhr”, in Illustrirte Zeitung, n.1116, 19 nov 1864, pp.357

FIGUIER, L., Les applications nouvelles de la science à l’industrie et aux arts en 1855, Paris, 1856, p.143

FLORES, J. (2000), “Pendule électromagnétique Detouche à Paris”, Horlogerie ancienne, n.48, pp.127-140

GUYE, R.P., BOSSART, M., horlogerie électrique, Lausanne, 1957

GRANIER, J. (1935), Pendules électriques, Paris, Dunod

Horlogerie, catalogue du musée [des arts et métiers] section JB Horlogerie, Paris, 1949

HOPE-JONES, FR., Electrical Timekeeping, N.A.G. Press LTD., London, 1976,[1st published in 1940]

L’Heure électrique, La-Chaux-de-Fonds, Musée International de l’horlogerie, 2005

MERLING, A., Die Elektrischen Uhren, Braunschweig, 1884

ROBERT-HOUDIN, J.E., Comment on devient sorcier, Paris, Omnibus, 2006 (reprint of texts published by Robert-Houdin from 1858)

UNDERHILL, CH.N., Solenoids electro-magnets and electro-magnetic windings, New York, (1910, reprint 1992)

Top of page

Notes

1   London, http://theclockworks.org/, accessed 8/12/2013

2   http://www.icom-cc.org/242/, accessed 8/12/2013.

3 See the clock by Guiseppe Zamboni, in Verona (Italy), College San Maffei. For details about his research, see http://ppp.unipv.it/collana/pages/libri/saggi/nuova%20voltiana5_pdf/p__091-103.pdf (accessed 08/05/2012): article online, Tinazzi, Massimo, “The Correspondence between Alessandro Volta and Giuseppe Zamboni about the Realization of the “Dry Pile”.

4  The first electric clock was patented in 1841 in the United Kingdom by Alexander Bain; at the same period, Matthias Hipp, a Germano-Swiss telegraph engineer, invented his toggle system. Italian and French makers also produced early models (See Electrifying Time, 1976).  

5  Patent taken with Brisbart-Gobert for an electric clock (Du Moncel 1880, p.127).

6  3d March 1855, brevet d’invention n.22648, with addition on the 7th July 1855, 36 p, Paris, Institut National de la Propriety intellectuelle. He also won a gold medal for his electric inventions.

7  One is in a private collection in the USA; two are in public collections: Oxford, Museum of History of Science, and Paris, Musée des Arts et Métiers. All three have been documented. The clock in Paris was examined and had a slave dial; For the Oxford clock, materials could also be analysed with a portable XRF machine, thanks to the courtesy of the Museum Director.

8  These analyses were performed at West Dean College and in Oxford’s Museum of History of Science with a Bruker AXS Handheld Inc. S1PXRF Spectrum.

9  The Science Museum in London also brought precious information as a 19th century Daniel’s cell could be examined in their stores, as well as one of the original cells used by Daniel himself.  This supported the making of the cell, especially for the ceramic jar which was used as a main component.

10  Heat shrink tubing is an extruded plastic which shrinks when slightly heated. It is used to protect and insulate an irregular soldered part or the entry points of wires into a case.

11  Silverline ACDC 513121. The machine sends a low current when the probes are applied to a surface. If the current flows back into the meter it means the tested circuit is not broken. It gives measurements for resistance, voltage and current.

12  The material was kindly offered by a customer of the clock workshop at West Dean College.

13  58x43x17x9 mm maximal dimensions for the L shaped part, which was tapered at the end holding the mobile armature (0.2 mm wider). The mobile armature is 52x13x5.9mm.

14  Ivory alternative, Units 3 and 3A, Hambrook business Centre, cheesemans Lane, Hambrook PO188XP.

15  Plus or minus 0.1 mm.

16  Wires.co.uk: 125 gr enamelled and double artificial silk on copper wire (Ref: sc-0710edasc-125).

17  The electric experiments were designed with the help of Kenneth Cobb, clock conservator and electronic engineer, and Jonathan Butt, horologist and Science teacher.

18  The letter code for these values are following European norm denomination (as visible in Guye and Bossart 1957).

19  An attempt to measure the inductance of both coils was made using a LCR meter (LT Putron LCR 9063). The meter indicated an inductance of 0.17 mH for both coils, which was also the value indicated when the two sensors of the machine are put in contact of one another. Further tests confirmed 0.17 mH value was the meter own inductance and not the coils values.

20  A test could be performed by wiring in series the original and the copy coils, attaching resistors whose total series values would be between 5 and 10 ohms, and feeding the circuit with a current from a variable power supply. With the power applied and a recorded current, a measurement of voltage across each coil would be made while increasing the applied current.  However, this test would demand soldering the contact points on the coils and the increasing current could heat in the coils, leading to a rise in temperature. From a conservation point of view, these two points are not acceptable for the original coil.

21  Using an oscilloscope Tektronix TDS 1012, 100 MHZ, 1Gs/s; bench power supply Roban Varex 30-2 Twin;  two meters, Silverline ACDC 513121 and Mastech MS8230B and resistors, all connected in a circuit.

22   Guye and Bossart 1957, p.49.

23  This experiment required the use of an electronic kit K8055 USB Experiment interface board (P8055-1), Velleman, 2003; this gave figures measured during the running of the cell at specific voltages, stored in an excel sheet then put into graphs.

24  The final version was clear shellac applied on the spring for 10 seconds, then the sacrificial platinum contact was applied and bordered with a small amount of shellac charged with graphite powder. The under layer of shellac insulate the spring from the current, and so from the erosion, while the graphite-shellac mix provides proper electric contact between the spring and the contact.

Top of page

List of illustrations

Title Fig. 1 Electric Master clock, 1855, by Constantin Detouche and Jean-Eugene Robert-Houdin
Caption Front and back views.
Credits London, The Clockworks, picture FC
URL http://ceroart.revues.org/docannexe/image/3914/img-1.jpg
File image/jpeg, 80k
Title  Fig. 2 J.E Robert Houdin – Drawing
Caption Drawing associated to patent 22648 (detail)6.
Credits Paris, INPI.
URL http://ceroart.revues.org/docannexe/image/3914/img-2.jpg
File image/jpeg, 60k
Title Fig. 3 view of the back of clock A
Caption Fragile and degraded wire
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-3.jpg
File image/jpeg, 92k
Title  Fig. 4 View of the different components
Caption Components of the coil in the copy clock, and coil in clock A.
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-4.jpg
File image/jpeg, 68k
Title  Fig. 5 The corroded platinum
Caption  Contact under the flat spring.
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-5.jpg
File image/jpeg, 64k
Title Fig. 6 XRF analysis spectrum of the soft iron in Clock A
URL http://ceroart.revues.org/docannexe/image/3914/img-6.jpg
File image/jpeg, 64k
Title  Fig. 7 Copy of a Daniell’s cell.
Caption Détails
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-7.jpg
File image/jpeg, 92k
Title  Fig. 8 Experiment (1)
Caption Experiment to monitor the current with an oscilloscope as the copy clock ran on a Daniell’s cell.
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-8.jpg
File image/jpeg, 148k
Title  Fig. 9 Experiment (2)
Caption Experiment to monitor the current in order to calculate the inductance of the coil in the copy clock
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-9.jpg
File image/jpeg, 140k
Title Fig. 10 Inductance curves
Caption  Top: inductance curve of the coil in the copy clock; bottom: inductance curve of the coil in Clock A.
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-10.jpg
File image/jpeg, 44k
Title  Fig. 11 Spectrums
Caption Top: spectrum of the current going through the copy clock powered by a Daniell’s cell; bottom: spectrum of the current going through Clock A powered by a bench power supply.
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-11.jpg
File image/jpeg, 76k
Title  Fig. 12 Experiment
Caption Platinum contact on the copy clock after 200 hours of running: first discernable spots of erosion
Credits Credits: © Françoise Collange
URL http://ceroart.revues.org/docannexe/image/3914/img-12.jpg
File image/jpeg, 67k
Top of page

References

Electronic reference

Françoise Collanges, « Conservation of an Early Electric Master Clock: Making a Copy as a Conservation Tool », CeROArt [Online], EGG 4 | 2014, Online since 19 March 2014, connection on 17 November 2017. URL : http://ceroart.revues.org/3914

Top of page

About the author

Françoise Collanges

Après une première carrière dans les musées en tant que responsable scientifique de collections, j’ai choisi de me spécialiser d’abord en conservation préventive (Master 2, 2006) puis de suivre un cursus pratique en conservation-restauration d’horlogerie au West Dean College (Royaum-Uni, West Sussex, 2010-2012) qui m’a menée à un Master in Conservation studies de l’Université du Sussex en 2013. fcollanges@yahoo.fr or fcollanges@gmail.com

Top of page

Copyright

Licence Creative Commons
CeROArt – Conservation, exposition, restauration d'objets d'arts est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo DOAJ – Directory of Open Access Journals
  • Les cahiers de Revues.org