A timepiece consists essentially of a mechanism which keeps the time, and a mechanism which indicates it.
The timekeeping mechanism, or
timebase, implements some physical phenomenon which recurs at
The reckoning of time by humans is based on the extremely steady motion of the Earth in space,
which is perceived in the apparent motion of the Sun and other astronomical bodies.
Without getting too far into details, the average interval between successive highest elevations of the Sun in the sky
is divided into twenty–four hours, each hour into sixty minutes,
and each minute further into sixty seconds.
Apparent solar time indicates itself, without any special apparatus, although elaborate sundials have been constructed
to take into account the variation between mean and apparent time in the course of the year at a given latitude.
Since it is not always convenient to take an observation of the Sun,
other means of knowing the time have been devised.
The most familiar presentation of time to the human senses is a visual display, based on uniform circular motion :
two or three pointers or
hands are fixed on a common pivot,
of which one makes a single revolution in twelve hours, one in sixty minutes, and one in sixty seconds.
This suffices to record, on a cyclical basis, the lapse of time since the hands were set in motion.
By arranging for the hands to coincide at the moments of noon and midnight, all times of day can be indicated.
If the hands are once turned to a position corresponding to the current time, as determined by some external reference,
the time according to that reference will be continually shown.
This mechanism has even been adapted for the use of the blind, using stout reinforced hands which are locked
in place when the time is to be read.
An alternative presentation is embodied in the
repeater watch, which at the press of a stud sounds a chime to
count, first the number of hours, then the quarter hours or tens of minutes, and sometime single minutes,
since noon or midnight.
mechanical watch, the recurrent phenomenon is simple harmonic motion, and the timebase is known
balance, consisting of a rotating mass and a
hair spring, invented by the great physicist Huygens.
main spring is geared to turn the hands as it uncoils,
and is held in check by a mechanism connected with the balance, known as the
At a certain point during the rotation of the balance, the check is momentarily released,
causing the hands to move slightly, and imparting an impetus to the balance.
The balance pivots through a certain arc, and is returned to its starting point by the force of the hair spring,
to begin the cycle anew.
Ideally, the period of this oscillatory motion is determined only by the force of the hair spring and the moment
of inertia of the balance rotor, and is independent of the force imparted by the main spring.
In practice, things quickly get much more complicated than that.
As a result, it is not uncommon for a well–made watch in a good state of repair to advance (as compared with
a reliable determination of time, such as that from an astronomical observatory) ten or fifteen seconds one day,
and fall back a similar amount another day.
Until the 1950s, this was the only mechanism which had been successfully applied to personal timepieces. Today, however, it is the province of specialty luxury goods, and the “quartz watch” is ubiquitous. In this latter kind of timepiece, the timebase is an electronic oscillator circuit, built around a piezoelectric quartz crystal, which produces a regular train of electrical impulses. When hands on a dial are used to indicate the time, they are “kicked” at intervals (typically of one second) by a stepper motor, triggered by the impulses from the oscillator.
There are three types of watch which stand at intermediate points.
Seiko makes a watch called the
Spring Drive in which the force of the mainspring is held back by an
electromagnetic brake, the action of which is modulated by a quartz oscillator,
the circuitry deriving its working power from the uncoiling of the spring.
This is a noteworthy piece of engineering, but it stands outside my field of interest.
In the two other types, the hands are moved by the action of the timebase itself, which is energized electrically.
We first direct our attention to a form of timepiece with a largely conventional balance, in which an electrochemical cell effectively takes the place of the main spring. When the balance reaches a certain point in its swing, an electrical circuit is momentarily closed, supplying a pulse of current to energize an electromagnetic coil, which reacts upon a permanent magnet or a component of high–permeability metal to impart an impetus to the balance. (The magnet may be mounted on the balance, with the coil stationary, or vice versa.) A ratchet–and–pawl mechanism (which in some designs works partly by magnetism) rectifies the motion of the balance and transmits it to a wheel train and thence to the hands. Although the overall arrangement greatly resembles a classic mechanical escapement, the balance is the driving element which transmits the motion, rather than a driven element which receives the motion and holds it back. These movements appeared in the late 1950s, proliferated in the ’60s, and died out in the ’70s.
In earlier movements of this type, generally labelled
electric (but perhaps more precisely to be designated
contact–switched), the circuit is closed mechanically, but these are subject to sparking,
owing to the sudden connection and disconnection of a heavily inductive circuit,
and to getting out of order due to misadjustment, oxidation, or dirtiness of the contacts.
electronic movements use a transistor as an electronic switch to deliver the current pulse,
generally triggered by a very small current generated in a secondary winding under the influence of the magnet.
Movements using this system usually bear a marking such as LIC ATO.
(ATO is the trademark of Etablissements Léon Hatot, the French clock–maker which developed a similar
circuit for electrically–driven pendulum clocks.)
Just to muddy the waters, contact–switched types incorporating an arc–suppression diode
were sometimes marked “electronic”.
An obvious practical advantage of such an electronic watch over the ordinary mechanical,
from the user’s perspective, is that it does not require daily winding.
Rather, it typically runs for a year until the power cell requires replacement.
The significance of this feature was diminished somewhat by the appearance of the self–winding or
automatic watch in the 1950s.
One reason for the unstable timekeeping of the typical mechanical watch, however, is that the torque from the
main spring varies constantly.
In the electronic watch, by contrast, the electromotive force supplied by the power cell diminishes only slowly,
and the circuit itself can be arranged to compensate somewhat for the effects of this variation.
Furthemore, an uncoiling main spring exerts a lateral thrust, constantly varying in direction and strength,
and promoting friction and wear, effects which are absent in an electronic.
As a result of this greater inherent stability, finicky adaptations such as the “Breguet overcoil”
are not required, and the design of the balance can instead be tweaked for greater resistance to other sources of
irregularity in timekeeping, such as vibration and the variation of mechanical properties with temperature.
And with smaller loads on the mechanism, the electronic movement is slower to drift out of adjustment, and
has less tendency to break down.
The other object of our regard here is an entirely new departure in horology, built around an
electronic oscillator in which the primary energy–storage element is a tuning–fork resonator,
electromagnetically coupled to the electric circuit.
In this respect it is not conceptually too very different from the quartz type.
The time display, however, depends upon the motion of the fork, rather than the electrical impulses.
A microscopically small pawl affixed to the fork acts on a tiny
index wheel, and its rotation is communicated
through a wheel train to the hands.
The details of this intriguing system, most familiar under the name of the
Accutron made by Bulova, and
sometimes referred to as a
Mosaba (from the French phrase montre sans balancier), have been
most ably explicated by other hands.
The tuning–fork mechanism can provide superior accuracy, not (as is commonly said) because of its
higher frequency — a 1% variation is still a 1% variation — but because the
(Q) of the fork is substantially greater than that of a rotating spring–and–mass system.
In practical terms, this means that, for instance, when disturbed by a shock, the fork does not wander as far
from its normal resonant frequency, and settles back to that frequency more quickly.
Hence, the cumulative effect of disturbances is smaller.
Some watch enthusiasts like to say that the mechanical watch
has a soul which the quartz watch lacks.
If this refers to anything in particular, perhaps it is that, in the quartz watch, the stepper mechanism is
indifferent to the source of the pulses which trigger it, and the oscillator is indifferent to the destination of the
pulses it emits, whereas in the mechanical, the motion of the hands is inextricably bound up with the timekeeping
Following this distinction would place both balance and tuning–fork electronics into the
Until the beginning of the 20th century, personal timepieces were generally carried in a pocket, or worn on a chain or ribbon around the neck, occasionally attached to a bracelet. The emergence of the fashion of wearing a watch on a wrist strap is associated with the pioneer aviator Alberto Santos–Dumont, and his close friend the jeweler Louis Cartier. As such, one would expect “watch”, unqualified, to signify “pocket watch” rather than “wrist watch”, but the usage is not so firmly established. Following the First World War, the wrist watch became dominant, and the quantity and variety of the other types (aside perhaps from cheap novelty items) greatly diminished.
Watch collectors, by and large, concentrate their attention on wrist watches. The main exception is those who are especially interested in “antique” timepieces, in particular ones with mechanisms which have become exotic through disuse. They tend to ignore modern pocket watches. I, however, am not really a watch collector. I find myself intrigued by electronic watches, but not so much as to go chasing after them just to keep them in a display case ; and, while I do not like to carry a “mobile device” such as many people now use to know the time, I find that wearing a wrist watch does not suit me. Accordingly I was gratified to discover that pocket watches with electronic movements were in fact made, and that a reasonable selection of them could be had (with a little careful shopping) for under us$100 each.
I see no reason to be excessively narrow in my definitions.
If a timepiece was made to be carried upon the person, but not worn on a wrist strap,
I am willing to consider it as falling within my bounds,
even though some might term it a “pendant” rather than a “pocket” watch.
Watch hobbyists are so exclusively masculine that almost no attention is given to
among which the pendant types tend to be included.
I am, however, inclined to leave out so–called pocket clocks.
The Citizen Watch Company of Japan (which originated from the Shokosha Horological Research Institute)
produced watches using their own electronic movements of both balance and tuning–fork types.
Citizen developed two families of balance–regulated electronic movements, the
and the miniature, high–beat
Some movements of both families were chronometer certified,
although none of these seems to have been issued in pocket format.
They also manufactured Accutron movements of the 218 series under contract to Bulova, and developed their own
series of derivative movements, including a single–coil type broadly similar to the 219,
but with important design differences.
The name Cosmotron was used primarily for
watches with X8 movements, but also for some IC–12 models, while wrist watches with the single–coil
tuning fork were labeled
(As a reference to the famous Brookhaven National Laboratory particle accelerator, this is probably
coincidence — some Citizen mechanicals of the time had names such as “Cosmostar”.
More difficult to account for are wrist watches branded “Signal Cosmotron Electronic” and
“Lanco Cosmotron”, both marked as Swiss–made.)
Wrist watches with the 218–type movements were branded
Citizen can be reasonably said to have produced the largest variety of electronic pocket watches.
Some models were very traditional in appearance, while others were more innovative, although none approached the
grotesquerie of many mid–1970s wrist watches.
|3700A||1971–10||360 Hz, dual–coil, day and date|
|3701A||1972–06||As 3700A but with power–cell disconnect|
|3710A||1971–10||As 3700A but date only|
|3711A||1972–06||As 3701A but date only|
|3721A||1972–08||As 3701A but no calendar|
|3701B||1975–07||360 Hz, single–coil, day and date|
Ebauches SA, supplier of movements to the Swiss watchmakers, produced quite a variety of both balance and tuning–fork types, but did not market watches as such. Only the ESA 9154 and 9156 (similar, but with “biorhythm” dials) in the former category, and the 9162 in the latter, appear to have made it into pocket cases, although I keep hoping to come upon a pendant with the 9200 “ladies’” balance movement.
Bulova produced Accutrons in the USA and Switzerland, some of them marketed under the
Perhaps surprisingly, given the appearance of competing with the Accutron, the company also marketed
electronic–balance watches, under the
Caravelle economy brand.
Movements include the 12 OTC (0880), 12 OUC (0840), and 12 OUCD (4840) made by Citizen,
the 6 UDC (ESA 9200), 13–jewel 13 UECD (ESA 9154), and 7–jewel day–date
13 UKCB (ESA 9158) from Switzerland, and the 14–jewel 7 OT (probably the LIP RE50) made in France.
Of these, at least the 12 OTC was used in a pocket model.
Seiko, the leading Japanese watchmaker, was a late and somewhat unexpected entrant to the electronic field. During the 1960s, their attention was directed toward applying automation to produce high–quality mechanical watches more quickly at lower cost, and toward creating a practical quartz movement. In fact, their first electronic model, the EL–310 or 31EL, appeared at the same time as the famous Quartz Astron. One is left to wonder whether it was developed as a sort of insurance against unexpected obstacles in the way of commercializing the quartz technology. In short order, the two Seiko manufacturing organizations, “Daini” and “Suwa”, came out with the EL–320/32EL, EL–330/33EL, EL–370/37EL, and Elnix/07EL series. (By a strange coincidence, the 37EL movement type numbers overlap with those of the Citizen tuning–fork movements.) Production appears, judging from the catalogues, to have ceased by about 1975. I have so far observed pocket models only in the EL–370 and Elnix ranges, although an EL–320 ladies’ pendant in a solid gold case appears in catalogues.
33EL models are described in Seiko
catalogues and other reference materials as
Battery Watches, the others as
although 33–series movements were widely used in “Electronic” and “Transistorized”
watches sold under names such as LeGant
(Montgomery Ward department stores), Stellaris (Sears), and Westclox.
Investigation reveals that, notwithstanding a superficial resemblance to the LIP R148, the magnetic field produced by the long
bobbin–wound solenoidal coil does not act on the balance,
which is therefore of largely conventional design, but rather on the pallet fork.
The motion of the fork — which is markedly asymmetrical and, thanks to its attached magnet,
quite massive as well — drives both the index mechanism and the balance.
The timekeeping cycle is completed when the motion of the balance, after restoring the fork to its starting position,
closes a contact in the base circuit of the transistor, triggering the next magnetic impulse.
Certainly, in this arrangement, the contact is less prone to trouble than if it had to carry the coil current,
and its adjustment may also be less critical.
On the whole, however, it is difficult to guess what considerations shaped such a radically different design.
Even in this distinguished company, it stands out.
Bearing surfaces of extremely hard material are used in watchmaking to reduce and resist friction, the favored composition being crystalline alumina. This substance, when chromium ions lend it a red color, is known as ruby ; titanium ions give blue sapphire. The bearings, being made of gemstone, are known as “jewels”. For a century or so, these have been mostly synthetic, and could easily be made colorless, but ruby jewels are used almost universally.
A main–spring movement with 15 functional jewels is the accepted low end of “quality”, 21 being regarded as “fully jeweled”. Movements with higher jewel counts are made, but most watchmakers consider this painting the lily. When it comes to electronic movements, there is no real consensus. With no barrel, a shorter wheel train, and less friction overall, fewer jewels would seem to be called for. We see this assumption at work in the 9–jewel Seiko 33EL movements, for instance, and the 7–jewel ESA 9157. Junghans, however, used 17 jewels in the 600–series movements, and Citizen in its X8 chronometer went to the extreme of 25, although the more plebeian X8 movements have only 12.
Tuning–fork movements present us with a yet more obscure situation. The Citizen 3701A day/date movement has 15 jewels, while the 3701B, which we might expect to be identical mechanically, has 11. The ESA 9162, frequently seen with chronometer certification, boasts 12.
My investigation of electronic pocket watches has led me to several general observations. First and foremost, the movements are basically designed for wrist watches, and fitted into pocket cases. They invariably have either center (sweep) seconds hands, or none ; the seconds sub–dial so common among mechanical pocket watches is never found. The omission of the seconds hand is curiously frequent, considering that without it, it is difficult to take advantage of the superior accuracy of the movement. Similarly, many lack the index markers on the dial required for precise setting and reading at times other than 12:00. Metallic–finish dials with metallic hands, a combination difficult to read under most conditions of light, are also quite common. Open–face style cases strongly predominate, and the dials are nearly all circular even when the cases are not. Dimensions are generally modest, with dial diameters typically about 30 millimeters, whereas today the fashion seems to be for wrist pieces of 50 mm. Perhaps this is partly because, before quartz and microelectronics, the smaller a watch movement, the more difficult it was to make it keep good time, hence small size could be taken for a mark of distinction ; whereas now, when people who wear wrist watches rarely use them for telling time, simple bigness is the lowest common denominator of ostentation.
Listed below are watches which I actually possess.
I am not enough of a completionist to even want an example of every possible type, and such a thing may not be
feasible in any case, since there is no accounting for
one–offs or custom modifications.
Likewise, just as there must be watches which exist but which have not come to my attention, and never will,
there are some which I have seen mentioned somewhere, or even listed in catalogues, which may not really exist.
Case dimensions are given either as diameter (exclusive of crowns, bales, et cetera)
× thickness (crystal to caseback), or as height (dimension in the 12:00—6:00 direction) × breadth
(9:00—3:00) × thickness, or else as specifically described.
Dial dimensions are given either as diameter, or as height × breadth.
Measurements are made to the millimeter, and may include a judgement factor,
especially as regards just where the boundary of the dial is.
Mass is given to the nearest 0·5 gram, without a power cell installed.
Unless otherwise mentioned, both bale and crown are at 12:00.
Numerals are printed in black except as otherwise stated, and Arabic numerals are
(aligned in the 6:00—12:00 direction), while Roman numerals are aligned radially, bases inward.
A note about this list : Seiko watches of the period under discussion have a numerical code stamped on or inside the case back, and a closely–related alphanumeric code on the dial. Citizen watches also typically have alphanumeric codes on the dial and case–back, which are different, but clearly related. For neither company is there an obvious correspondence between these codes and those listed in the sales catalogues. Other manufacturers are not so helpful in providing unambiguous identification for their models.
stockorientation, as seen in advertisements, but the dial can be rotated 180°, if it is preferred to have crown at 12:00 and bale at 6:00
Citizen Pocket Watch (Cosmotron), EQ–58, item reference 0712–11 (4–350049)
Citizen Pocket Watch (Electronic), EQ–37, item reference 1321–12 (4–375106)
Pocket Watch, 1974 volume 1 and v.2 under item reference 07ELC 030, 1975 v.2 and 1976 v.1 under reference ENT030
These are watches which I have seen mentioned and described on a collector’s Web site, or exposed for sale, for instance on eBay or Yahoo Japan auctions, in any case with enough context and detail (first and foremost, photographs) that I can confidently identify and describe them. I am not inclined to take catalogue listings on their own without further supporting evidence ; since I have watches which do not seem to appear in any catalogue, I would not be surprised if some things appear in catalogues but do not really exist.
Electronic ƒ300 Hzwith Calibre 1250 (ESA 9162) movement, as well as Megasonic 720 Hz with the extraordinary Calibre 1220
boudoir clockwhen not being carried
Unsurprisingly, living with these watches has given me opinions about design elements and how they harmonize. My first criterion is that any watch should be easy to read.
true seconds. Some movements beat 5 times per second and index (advance the hands) only on alternate beats, which seems like poor design.
I also have some more subjective thoughts.
We have no reason to look for a revival of the electronic watch at this point. (Astonishingly enough, Citizen, which now owns Bulova, produced a small run of new Accutrons a few years ago.) It is, however, interesting to wonder how it might have developed if the quartz watch had not become dominant.
An optical switch could potentially be used in a balance–regulated movement, as an alternative to either contacts or an induction coil and transistor. Suppose a photo–Darlington, a transistor–like device which is controlled by light rather than electricity, to be positioned close to the balance rotor. Small dots of radioluminescent paint on the rotor would then trigger the electrical impulses as they passed by, in the course of an oscillation. The result would be a definite simplification of the circuit.
Using multiple oscillating elements of similar characteristics in coordination is a well–established approach to creating a high–accuracy timebase. A clock incorporating three pendulums, intercommunicating electrically, was formerly used as the time reference for the German State Railways, and is now displayed at the Verkehrszentrum of Deutsches Museum in Munich. Some early quartz watch movements incorporated balances or tuning forks, but these components thereby lost their time–keeping function, serving only to translate the pulse train from the oscillator into motion of the hands. One can certainly envision a movement incorporating, for instance, two forks mounted at right angles, so that a shock producing the largest disturbance on one would have essentially no effect on the other ; or two balances with hair springs wound in opposite directions, so that a rotation adding to the motion of one would reduce the motion of the other.
Let us for a moment consider a timepiece employing two electromagnetically–driven balances of the moving–magnet type, in which the driving coils are positioned at right angles to one another. Connecting the two balances by way of a magnetic shunt then opens the way to a further departure from conventional practice, the elimination of mechanical indexing. Somewhat as in the Megasonic 720, the alternation of magnetic polarities on the shunt could be made to rotate a kind of magnetic gear. This would even further reduce timebase inaccuracy arising from mechanical forces.
One of the obvious problems with any watch which relies on an electrochemical power cell is the periodic need to replace the cell. In the quartz field, Citizen addressed this early on with the combination of photovoltaic cells under the dial, and a supercapacitor or electrochemical storage cell, and has been very successful with this approach. Seiko has applied the rotating eccentric weight principle of the automatic mechanical watch to driving a tiny dynamo, with more limited success. Thermoelectric generators powered by body heat have been tried, but have proven unsatisfactory. Of course, many quartz watches today are so cheap that people discard them after a few months or years when the cell runs down, which is hardly a wise use of resources in a world of unmet needs.
An energy source a millionfold denser than chemical reactions is available to us within the atom. Obviously, radioisotope thermoelectric generators of very low power (the requirement for a watch movement is in microwatts) with a lifetime of decades can be built. Just such devices were used in some early heart pacemakers.
Another possibility, however, presents itself : the direct use of decay energy,
without an intermediate conversion step.
A sample of “beta–active” radioisotope continuously expels fast–moving electrons.
If some of these strike a nearby insulating surface, an electrostatic charge will build up,
and the mutual repulsion of like charges will produce a mechanical force.
Just this principle was applied by the Patek Phillipe research laboratories to the construction of what can only be
atomic clock, albeit practically the inverse of what is usually meant by that term.
As reported in the proceedings of the 1955 Geneva Conference on the Peaceful Uses of
Atomic Energy, this was a mechanical clock driven by a rotor which rocked back and forth, with a period of
about ten minutes, under the influence of the emission current of 50 millicuries of strontium–90, generating
about 1·5 microwatts.
More recently, the same basic principle has been applied to what are termed
“Micro–Electro–Mechanical Systems”, to induce vibration (with a much shorter period)
in tiny cantilevers.
This might be applied, for instance, by placing a number of these MEMS devices around the periphery of a wheel,
to act as pawls and rotate it continuously by ratchet action.
Moreover, the analogy to the Accutron mechanism should be obvious.
It is not difficult to envision a fork in a sealed capsule, coupling its motion magnetically (as in the Megasonic),
with radioisotope sources replacing the electromagnetic coils in the end cups.
Such approaches would require no circuitry : the watch would simply run as though by magic for a
generation, impelled by unseen forces, using no more radioactive material than is applied to a self–luminous dial.
In such a way, the ultimate evolution of the
electronic watch could have been the
As with practically any mechanism, an electronic watch requires periodic attention, to renew lubricants, remove dust or other abrasive materials which may have entered the case, and generally assure that everything is correctly set up. Without servicing, it will cease to keep good time, and the delicate apparatus will eventually be damaged. After such a lapse of years, and with the demise of many watch firms in the wake of the quartz revolution, replacement parts will not be easily come by. Many older mechanical watches, although they could potentially run for generations, are being mistreated today and allowed to wear out because they cost considerably more to keep running well than to purchase.
The electronic watch, because it is less subject to wear, should require less frequent service than an ordinary mechanical — considering the superior endurance of modern lubricants over those of 50 years ago, once every five power–cell changes might be a reasonable rule to adopt. Of course it should be serviced immediately if timekeeping becomes erratic or drifts beyond the range of adjustment, but the objective is to prevent these conditions from arising in the first place, as they indicate a potential for damage. Any competent watchmaker should be able to service a balance–wheel electronic, although there may be some details which call for guidance from the manufacturer’s service literature, but not every watchmaker will be comfortable taking on the job. The Accutron, although it requires only limited attention, is more of a specialty item, and finding a technician is correspondingly more difficult.
It will come as no kind of surprise to those who know me that I have acquired along the way a few duplicates. If you feel that you would like to have one of these intriguing timepieces, but do not feel like hunting for one yourself, let me know, and I may be able to assist you. At the moment, I have an open–face gold–filled 219 Accutron (which I received in what appeared to be new unused condition) and a Seiko 3702–0010 (with a small crack in the crystal) which are definitely surplus to my requirements ; I could without too much regret relinquish the non–display–back round stainless–steel Cosmotron with Roman numerals ; and for a suitable inducement I could part with my Seiko 0702–3000 (I have a second one which needs repair). All of these run and keep reasonable time, but have not been serviced, and this should be done before they are put into serious use, because they came to me without information on their service history. I also have a Seiko 37–3000 which is not running well, but should yield to servicing.