We see by the preceding calculations that there is one definite point in the time train of a clock; the center arbor, which carries the minute hand, must revolve once in one hour; from this point we may vary the train both ways, toward the escape wheel to suit the length of pendulum which we desire to use, and toward the barrel to suit the length of time we want the clock to run. The center arbor is therefore generally used as the point at which to begin calculations, and it is also for this reason that the number of teeth in the center wheel is the starting point in train calculations toward the escape wheel, while the center pinion is the starting point in calculations of the length of time the weight or spring is to drive the clock. Most writers on horology ignore this point, because it seems self-evident, but its omission has been the cause of much mystification to so many students that it is better to state it in plain terms, so that even temporary confusion may be avoided.

Sometimes there is a second fixed point in a time train; this occurs only when there is a seconds hand to be provided for; when this is the case the seconds hand must revolve once every minute. If it is a seconds pendulum the hand is generally carried on the escape wheel and the relation of revolutions between the hour and seconds wheels must then be as one is to sixty. This might be accomplished with a single wheel having sixty times as many teeth as the pinion on the seconds arbor; but the wheel would take up so much room, on account of its large circumference, that the movement would become unwieldy because there would be no room left for the other wheels; so it is cheaper to make more wheels and pinions and thereby get a smaller clock. Now the best practical method of dividing this motion is by giving the wheels and pinions a relative velocity of seven and a half and eight, because 7.5×8=60.

Thus if the center wheel has 80 teeth, gearing into a pinion of 10, the pinion will be driven eight times for each revolution of the center wheel, while the third wheel, with 75 teeth, will drive its pinion of 10 leaves 7.5 times, so that this arbor will go 7.5 times eight, or 60 times as fast as the center wheel.

If the clock has no seconds hand this second fixed point is not present in the calculations and other considerations may then govern. These are generally the securing of an even motion, with teeth of wheels and pinions properly meshing into each other, without incurring undue expense in manufacture by making too many teeth in the pinions and consequently in the wheels. For these reasons pinions of less than seven or more than ten leaves are rarely used in the common clocks, although regulators and fine clocks, where the depthing is important, frequently have 12, 14 or 16 leaves in the pinions, as is also the case with tower clocks, where the increased size of the movement is not as important as a smoothly running train. Clocks without pendulums, carriage clocks, locomotive levers and nickel alarms, also have different trains, many of which have the six leaf pinion, with its attendant evils, in their trains.

Weights.--Weights have the great advantage of driving a train with uniform power, which a spring does not accomplish: They are therefore always used where exactness of time is of more importance than compactness or portability of the clock. In making calculations for a weight movement, the first consideration is that as the coils of the cord must be side by side upon the barrel and each takes up a definite amount of space, a thicker movement (with longer arbors) will be necessary, as the barrel must give a sufficient number of turns of the cord to run the clock the desired time and the length of the barrel, with the wheel and maintaining power all mounted upon the one arbor, will determine the thickness of the movement. If the clock is to have striking trains their barrels will generally be of more turns and consequently longer than the time barrel and in that case the distance between the plates is governed by the length of the longest barrel and its mechanism.

The center wheel, upon the arbor of which sits the canon pinion with the minute hand, must, since the hand has to accomplish its revolution in one hour, also revolve once in an hour. When, therefore, the pinion of the center arbor has 8 leaves and the barrel wheel 144, then the 8 pinion leaves, which makes one revolution per hour, would require the advancing of 8 teeth of the barrel wheel, which is equal to the eighteenth part of its circumference. But when the eighteenth part in its advancing consumes 1 hour, then the entire barrel wheel will consume 18 hours to accomplish one revolution. If, now, 10 coils of the weight cord were laid around the barrel, the clock would then run 10×18=180 hours, or 7½ days, before it is run down.

Referring to what was said in a previous chapter on wheels being merely compound levers, it will be seen that as we gain motion we lose power in the same ratio. We shall also see that by working the rule backwards we may arrive at the amount of force exerted on the pendulum by the pallets. If we multiply the circumference of the escape wheel in inches by the number of its revolutions in one hour we will get the number of inches of motion the escape wheel has in one hour. Now if we multiply the weight by the distance the barrel wheel travels in one hour and divide by the first number we shall have the force exerted on the escape wheel. It will be simpler to turn the weight into grains before starting, as the division is less cumbersome.

Another way is to find how many times the escape wheel revolves to one turn of the barrel and divide the weight by that number, which will give the proportion of weight at the escape wheel, or rather would do so if there were no power lost by friction. It is usual to estimate that three-quarters of the power is used up in frictions of teeth and pivots, so that the amount actually used for propulsion of the pendulum is very small, being merely sufficient to overcome the bending moment of the suspension spring and the resistance of the air.

It is for this reason that clocks with finely cut trains and jeweled pivots, thus having little train friction, will run with very small weights. The writer knows of a Howard regulator with jeweled pivots and pallets running a 14-pound pendulum with a five-ounce driving weight. Of course this is an extreme instance and was the result of an experiment by an expert watchmaker who wanted to see what he could do in this direction.

Usually the method adopted to determine the amount of weight that is necessary for a movement is to hang a small tin pail on the weight cord and fill it with shot sufficient to barely make the clock keep time. When this point has been determined, then weigh the pail of shot and make your driving weight from eight to sixteen ounces heavier. In doing this be sure the clock is in beat and that it is the lack of power which stops the clock; the latter point can be readily determined by adding or taking out shot from the pail until the amount of weight is determined. The extra weight is then added as a reserve power, to counteract the increase of friction produced by the thickening of the oil.

Many clock barrels have spiral grooves turned in them to assist in keeping the coils from riding on each other, as where such riding occurs the riding coils are farther from the center of the barrel than the others, which gives them a longer leverage and greater power while they are unwinding, so that the power thus becomes irregular and affects the rate of the clock, slowing it if the escapement is dead beat and making it go faster if it is a recoil escapement.

Clock cords should be attached to the barrel at the end which is the farthest from the pendulum, so that as they unwind the weight is carried away from the pendulum. This is done to avoid sympathetic vibrations of the weight as it passes the pendulum, which interfere with the timekeeping when they occur. If the weight cannot be brought far enough away to avoid vibrations a sheet of glass may be drilled at its four corners and fixed with screws to posts placed in the back of the case at the point where vibration occurs, so that the glass is between the pendulum rod and the weight, but does not interfere with either. This looks well and cures the trouble.

We have, heretofore, been speaking of weights which hang directly from the barrel, as was the case with the older clocks with long cases, so that the weight had plenty of room to fall. Where the cases are too short to allow of this method, recourse is had to hanging the weight on a pulley and fastening one end of the cord to the seat board. This involves doubling the amount of weight and also taking care that the end of the cord is fastened far enough from the slot through which it unwinds so that the cords will not twist, as they are likely to do if they are near together and the cord has been twisted too much while putting it on the barrel. Twisting weight cords are a frequent source of trouble when new cords have been put on a clock. The pulley is another source of trouble, especially if wire cords (picture cords) or cables are used. Wire cable should not be bent in a circle smaller than forty times its diameter if flexibility is to be maintained, hence pulleys which were all right for gut or silk frequently prove too small when wire is substituted and kinks, twisted and broken cables frequently result from this cause. This is especially the case with the heavy weight of striking trains of hall and chiming clocks, where double pulleys are used, and also leads to trouble by jamming and cutting the cables and dropping of the weights in tower clocks where a new cable of larger size is used to replace an old one which has become unsafe from rust, or cut by the sheaves.

Weight cords on the striking side of a clock should always be left long enough so that they will not run down and stop before the time train has stopped. This is particularly the case with the old English hall clocks, as many of them will drop or push their gathering racks free of the gathering pinion under such conditions and then when the clock is wound it will go on striking continuously until the dial is taken off and the rack replaced in mesh with the gathering pinion. As clocks are usually wound at night, the watchmaker can see the disturbance that would be caused in a house in the “wee sma’ hours” by such a clock going on a rampage and striking continuously.

Oiling Cables.--Clock cables, if of wire and small in size, should be oiled by dipping in vaseline thinned with benzine of good quality. Both benzine and vaseline must be free from acid, as if the latter is present it will attack the cable. This thinning will permit the vaseline to permeate the entire cable and when the benzine evaporates it will leave a thin film of vaseline over every wire, thus preventing rust. Tower clock cables should be oiled with a good mineral oil, well soaked into them to prevent rusting. Gut clock cords, when dry and hard, are best treated with clock oil, but olive oil or sperm oil will also be found good to soften and preserve them. New cords should always be oiled until they are soft and flexible. If the weight is under ten pounds silk cords are preferable to gut or wire as they are very soft and flexible.

In putting on a new cable or weight cord the course of the weight and cord should be closely watched at all points, to see that they remain free and do not chafe or bind anywhere and also that the coils run evenly and freely, side by side; sometimes, especially with wire, a new cable gets kinked by riding the first time of winding and is then very difficult to cure of this serious fault. Another point to watch is to see that the position of the cord when wound up will not cause an end thrust upon the barrel, which will interfere with the timekeeping if it is overwound, so that the weight is jammed against the seatboard; this frequently happens with careless winding, if there is no stop work.

To determine the lengths of clock cords or weights, we may have to approach the question from either end. If the clock be brought in without the cords, we first count the number of turns we can get on the barrel. This may be done by measuring the length of the barrel and dividing it by the thickness of the cord, if the barrel is smooth, or by counting the grooves if it be a grooved barrel. Next we caliper the diameter and add the thickness of one cord, which gives us the diameter of the barrel to the center of the cords, which is the real or working diameter. Multiply the distance so found by 3.14156, which gives the circumference of the barrel, or the length of cord for one turn of the barrel. Multiply the length of one turn by the number of turns and we have the length of cord on the barrel, when it is fully wound. If the cord is to be attached to the weight, measure the distance from the center of barrel to the bottom of the seat board and leave enough for tying. If the weight is on a pulley it will generally require about twelve inches to reach from the barrel through the slot of the seat board, through the pulley to the point of fastening.

To get the fall of the weight, stand it on the bottom of the case and measure the distance from the top of the point of attachment to the bottom of the seat board. This will generally allow the weight to fall within two inches of the bottom and thus keep the cable tight when the clock runs down; thus avoiding kinks and over-riding when we wind again after allowing the clock to run down. If the weight has a pulley and double cord, measure from the top of the pulley to the seatboard, with the weight on the bottom, and then double this measurement for the length of the cord. This measure is multiplied by as many times as there are pulleys in the case of additional sheaves. Striking trains are frequently run with two coils or layers of cord, on the barrel, time trains never have but one.

Now, having the greatest available length of cord determined according either of the above conditions, we can determine the number of turns for which we have room on our barrel and divide the length of cord by the number of turns. This will give us the length of one turn of the cord on our barrel and thus having found the circumference it is easy to find the diameter which we must give our barrel in suiting a movement to given dimensions of the case. This is frequently done where the factory may want a movement to fit a particular style and size of case which has proved popular, or when a watchmaker desires to make a movement for which he has, or will buy, a case already made.

As to tower clock cables, getting the length of cable on the barrel is, of course, the same as given above, but the rest of it is an individual problem in every case, as cables are led so differently and the length of fall varies so that only the professional tower clock men are fitted to make the measurements for new work and they require no instruction from me. It might be well to add, however, that in the tower clocks by far the greater part of the cable is always outside the clock and only the inner end coils and uncoils about the barrel. It is for this reason that the outer ends of the cables are so generally neglected by watchmakers in charge of tower clocks and allowed to cut and rust until they drop their weights. Caretakers of tower clocks should remember that the inner ends of cables are always the best ends; the parts that need watching are those in the sheaves or leading to the sheaves. Tower clocks should have the cables marked where to stop to prevent overwinding.

In chain drives for the weights of cuckoo and other clocks with exposed weights, we have generally a steel sprocket wheel with convex guiding surfaces each side of the sprocket and projecting flanges each side of the guides; one of these flanges is generally the ratchet wheel. The ratchet wheel, guide, sprocket, guide and flange, form a built-up wheel which is loose on the arbor and is pinned close to the great wheel, which is driven by a click on the wheel working into the ratchet of the drive. It must be loose on the arbor, because the clock is wound by pulling the sprocket and ratchet backward by means of the chain until the weight is raised clear up to the seat board. There are no squares on the arbors, which have ordinary pivots at both ends, and the great wheel is fast on the arbor. The diameter of the convex portion of the wheel each side of the sprocket is the diameter of the barrel, and the chain should fit so that alternate links will fit nicely in the teeth of the sprocket; where this is not the case they will miss a link occasionally and the weight will then fall until the chain catches again, when it will stop with a jerk; bent or jammed links in the chain will do the same thing. Sometimes a light chain on a heavy weight will stretch or spread the links enough to make their action faulty. If examination shows a tendency to open the links, they should be soldered; if they are stretching, a heavier chain of correct lengths of links should be substituted. Twisted chains are another characteristic fault and are usually the result of bent or jammed links. A close examination of such a chain will generally reveal several links in succession which are not quite flat and careful straightening of these links will generally cure the tendency to twist.

Mainsprings for Clocks.--There are many points of difference between mainsprings for clocks and those for watches. They differ in size, strength, number of coils and in their effect on the rates of the clock.

Watch springs are practically all for 30-hour lever escapements, with a few cylinder, duplex and chronometer escapements. If a fusee watch happens into a shop nowadays it is so rare as to be a curiosity worth stopping work to look at.

The clocks range all the way from 30 hours to 400 days in length of time between windings and include lever, cylinder, duplex, dead beat, half dead beat, recoil and other escapements. Furthermore some of these, even of the same form of escapements, will vary so in weight and the consequent influence of the spring that what will pass in one case will give a wildly erratic rate in another instance. Many of the small French clocks have such small and light pendulums that very nice management of the stop works is necessary to prevent the clock from gaining wildly when wound or stopping altogether when half run down.

Nothing will cause a clock with a cylinder escapement to vary in time more than a set or gummy mainspring, for it will gain time when first wound and lose when half run down, or when there is but little power on the train. In such a case examine the mainspring and see that it is neither gummy nor set. If it is set, put in a new spring and you can probably bring it to time.

With a clock it depends entirely on the kind of escapement that it contains, whether it runs faster or slower, with a stronger spring; if you put a stronger mainspring in a clock that contains a recoil escapement the clock will gain time, because the extra power, transmitted to the pallets will cause the pendulum to take a shorter arc, therefore gain time, where the reverse occurs in the dead-beat escapement. A stronger spring will cause the dead-beat pendulum to take a longer arc and therefore lose time.

If a pendulum is short and light these effects will be much greater than with a long and heavy pendulum.

At all clock factories they test the mainsprings for power and to see that they unwind evenly; those that do are marked No. 1, and those that do not are called “seconds.” The seconds are used only for the striking side of the clocks, while the perfect ones are used for the running, or time side. Sometimes, however, a seconds’ spring will be put on the time side and will cause the clock to vary in a most erratic way. This changing of springs is very often done by careless or ignorant workmen in cleaning and then they cannot locate the trouble.

All mainsprings for both clocks and watches should be smooth and well polished. Proper attention to this one item will save many dollars’ worth of time in examining movements to try to detect the cause of variations.

A rough mainspring (that is, an emery finished mainspring) will lose one-third of its power from coil friction, and in certain instances even one-half. The deceptive feature about this to the watchmaker is that the clock will take a good motion with a rough spring fully found, but will fall off when partly unwound, and the consequence is that he finds a good motion when the spring is put in and wound, and he afterward neglects to examine the spring when he examines the rate as faulty. The best springs are cheap enough, so that only the best quality should be used, as it is easy for a watchmaker to lose three or four dollars’ worth of time looking for faults in the escapement, train and everywhere else, except the barrel, when he has inserted a rough, thick, poorly made spring. The most that he can save on the cheaper qualities of springs is about five cents per spring and we will ask any watchmaker how long it would take to lose five cents in examination of a movement to see what is defective.

Here is something which you can try yourself at the bench. Take a rough watch mainspring; coil it small enough to be grasped in the hand and then press on the spring evenly and steadily. You will find it difficult to make the coils slide on one another as the inner coils get smaller; they will stick together and give way by jerks. Now open your hand slowly and you will feel the spring uncoiling in an abrupt, jerky way, sometimes exerting very little pressure on the hand, at other times a great deal. A dirty, gummy spring will do the same thing. Now take a clean, well polished spring and try it the same way; notice how much more even and steady is the pressure required to move the coils upon each other, either in compressing or expanding. Now oil the well polished spring and try it again. You will find you now have something that is instantly responding, evenly and smoothly, to every variation of pressure. You can also compress the spring two or three turns farther with the same force. This is what goes on in the barrel of every clock or watch; you have merely been using your hand as a barrel and feeling the action of the springs.

Now a well finished mainspring that is gummy is as irregular in its action as the worst of the springs described above, yet very few watchmakers will take out the springs of a clock if they are in a barrel. One of them once said to me, “Why, who ever takes out springs? I’ll bet I clean a hundred clocks before I take out the springs of one of them!” Yet this same man had then a clock which had come back to him and which was the cause of the conversation.

There must be in this country over 25,000 fine French clocks in expensive marble or onyx cases, which were given as wedding presents to their owners, and which have never run properly and in many instances cannot be made to run by the watchmakers to whom they were taken when they stopped. Let me give the history of one of them. It was an eight-day French marble clock which cost $25 (wholesale) in St. Louis and was given as a wedding present. Three months later it stopped and was taken to a watchmaker well known to be skillful and who had a fine run of expensive watches constantly coming to him. He cleaned the clock, took it home and it ran three hours! It came back to him three times; during these periods he went over the movement repeatedly; every wheel was tested in a depthing tool and found to be round: all the teeth were examined separately under a glass and found to be perfect; the pinions were subjected to the same careful scrutiny; the depthings were tried with each wheel and pinion separately; the pivots were tested and found to be right; the movement was put in its case and examined there; it would run all right on the watchmaker’s bench, but not in the home of its owner. It would stop every time it was moved in dusting the mantel. He became disgusted and took the clock to another watchmaker, a railroad time inspector; same results. In this way the clock moved about for three years; whenever the owner heard of a man who was accounted more than ordinarily skillful he took him the clock and watched him “fall down” on it. Finally it came into the hands of an ex-president of the American Horological Society. He made it run three weeks. When he found the clock had stopped again he refused pay for it. Three months later he called and got the clock, kept it for three weeks, brought it back without explanation and lo, the clock ran! It would even run considerably out of beat! When asked what he had done to the clock, he merely laughed and said “Wait.”

A year later the clock was still going satisfactorily and he explained. “That was the first time I ever got anything I couldn’t fix and it made me ashamed. I kept thinking it over. Finally one night in bed I got to considering why a clock wouldn’t run when there was nothing the matter with it. The only reason I could see was lack of power. Next morning I got the clock and put in new mainsprings, the best I could find. The clock was cured! None of these other men who had the clock took out the springs. They came to me all gummed up, while the rest of the clock was clean, bright and in perfect order. I cleaned the springs and returned the clock; it ran three weeks. When I took it back I put in stronger springs, because I found them a little soft on testing them. If any of your friends have French clocks that won’t go, send them to me.”

Three-quarters of the trouble with French clocks is in the spring box; mainspring too weak, gummy or set; stop works not properly adjusted, or left off by some numskull who thought he could make the clock keep time without it when the maker couldn’t; mainspring rough, so that it uncoils by jerks; spring too strong, so that the small and light pendulum cannot control it. These will account for far more cases than the “flat wheel” story that so often comes to the front to account for a failure on the part of the workman. Of course he must say something to his boss to account for his failure and the “wheels out of round” and “the faulty depthing” have been standard excuses for French clocks for a century. Of course they do occur, but not nearly as often as they are credited with, and even then such a clock may be made to perform creditably if the springs are right.

Another source of trouble is buckled springs, caused by some workman taking them out or putting them in the barrel without a mainspring winder. There are many men who will tell you that they never use a winder; they can put any spring in without it. Perhaps they can, but there comes a day when they get a soft spring that is too wide for this treatment and they stretch one side of it, or bend, or kink it, and then comes coil friction with its attendant evils. These may not show with a heavy pendulum, but they are certain to do so if it happens to be an eight-day movement with light pendulum or balance, and this is particularly true of a cylinder.

All springs should be cleaned by soaking in benzine or gasoline and rubbing with a rag until all the gum is off them before they are oiled. Heavy springs may be wiped by wrapping one or two turns of a rag around them and pushing it around the coils. The spring should be well cleaned and dried before oiling. A quick way of cleaning is to wind the springs clear up; stick a peg in the escape wheel; remove the pallet fork; plunge the whole movement into a pail of gasoline large enough to cover it; let it stand until the gasoline has soaked into the barrels; remove the peg and let the trains run down. The coils of the spring will scrub each other in unwinding; the pivots will clean the pivot holes and the teeth of wheels and pinions will clean each other. Then take the clock apart for repairs. Springs which are not in barrels should be wound up and spring clamps put on them before taking down the clock. About six sizes of these clamps (from 2½ inches to ¾ inch) are sufficient for ordinary work.

Rancid oil is also the cause of many “come-backs.” Workmen will buy a large bottle of good oil and leave it standing uncorked, or in the sun, or too near a stove in winter time, until it spoils. Used in this condition it will dry or gum in a month or two and the clock comes back, if the owner is particular; if not, he simply tells his friends that you can’t fix a clock and they had better go elsewhere with their watches.

For clock mainsprings, clock oil, such as you buy from material dealers, is recommended provided it is intended for French mainsprings. If the lubricant is needed for coarse American springs, mix some vaseline with refined benzine and put it on liberally. The benzine will dissolve the vaseline and will help to convey the lubricant all over the spring, leaving no part untouched. The liquid will then evaporate, leaving a thin coating of vaseline on the spring.

It is best to let springs down, with a key made for the purpose. It is a key with a large, round, wooden handle, which fills the hand of the watchmaker when he grasps it. Placing the key on the arbor square, with the movement held securely in a vise, wind the spring until you can release the click of the ratchet with a screwdriver, wire or other tool; hold the click free of the ratchet and let the handle of the key turn slowly round in the hand until the spring is down. Be careful not to release the pressure on the key too much, or it will get away from you if the spring is strong, and will damage the movement. This is why the handle is made so large, so that you can hold a strong spring.

It is of great importance, if we wish to avoid variable coil friction, that the spring should wind, from the very starting, concentrically; i. e., that the coils should commence to wind in regular spirals, equidistant from each other, around the arbor. In very many cases we find, when we commence to wind a spring, that the innermost coil bulges out on one side, causing, from the very beginning, a greater friction of the coils on that side, the outer ones pressing hard against it as you continue to wind, while on the outer side of the arbor they are separated from each other by quite a little space between them, and that this bulge in the first coil is overcome and becomes concentric to the arbor only after the spring is more than half way wound up. This necessarily produces greater and more variable coil friction. When a spring is put into the barrel the innermost coil should come to the center around the arbor by a gradual sweep, starting from at least one turn around away from the other coils. Instead of that, we more often find it laying close to the outer coils to the very end, and ending abruptly in the curl in the soft end that is to be next the arbor. When this is the case in a spring of uniform thickness throughout, it is mainly due to the manner of first winding it from its straight into a spiral form. To obviate it, I generally wind the first coils, say two or three, on a center in the winder, a trifle smaller than the regular one, which is to be of the same diameter of the arbor center in the barrel. You will find that the substitution of the regular center, afterwards, will not undo the extra bending thus produced on the inner coils, and that the spring will abut by a more gradual sweep at the center, and wind more concentrically.

The form of spring formerly used with a fusee in English carriage clocks and marine chronometers is a spring tapering slightly in thickness from the inner end for a distance of two full coils, the thickness increasing as we move away from the end, then continuing of uniform thickness until within about a coil and a half from the other end, when it again increases in thickness by a gradual taper. The increase in the thickness towards the outer end will cause it to cling more firmly to the wall of the barrel. The best substitute for this taper on the outside is a brace added to some of the springs immediately back of the hole. With this brace, and the core of the winding arbor cut spirally, excellent results are obtained with a spring of uniform thickness throughout its entire length. Something, too, can be done to improve the action of a spring that has no brace, by hooking it properly to the barrel. The hole in the spring on the outside should never be made close to the end; on the contrary, there should be from a half to three-quarters of an inch left beyond the hole. This end portion will act as a brace.

When the spring is down, the innermost coil of it should form a gradual spiral curve towards the center, so as to meet the arbor without forcing it to one side or the other. This curve can be improved upon, if not correct, with suitably shaped pliers; or it can be approximated by winding the innermost coils first on an arbor a little smaller in diameter than the barrel arbor itself.

Another and very important factor in the development of the force of the spring is the proper length and thickness of it. For any diameter of barrel there is but one length and one thickness of spring that will give the maximum number of turns to wind. This is conditioned by the fact that the volume which the spring occupies when it is down must not be greater nor less than the volume of the empty space around the arbor into which it is to be wound, so that the outermost coil of the spring when fully wound will occupy the same place which the innermost occupies when it is down. In a barrel, the diameter of whose arbor is one-third that of the barrel, the condition is fulfilled when the measure across the coils of the spring as it lays against the wall of the barrel, is 0.39 of the empty space, or, taking the diameter of the barrel as a comparison, 0.123 of the latter; in other words, nearly one-eighth of the diameter of the barrel. This is the width that will give the greatest number of turns to wind, whatever may be the length or thickness of any spring. If now we desire a spring to wind a given number of turns, there is but one thickness and one length of it that will permit it to do so. The thickness remaining the same, if we make the spring longer or shorter, we reduce the number of turns it will wind; more rapidly by making it shorter, less so by making it longer. It is therefore not only useless, but detrimental, to put into a barrel a greater number of coils, or turns, than are necessary, not only because it will reduce the number of turns the barrel will wind, but it will produce greater coil friction by filling up the space with more coils than are necessary.

A mainspring in the act of uncoiling in its barrel always gives a number of turns equal to the difference between the number of coils in the up and the down positions. Thus, if 17 be the number of coils when the spring is run down, and 25 the number when against the arbor, the number of turns in uncoiling will be 8, or the difference between 17 and 25.

The cause of breakage is usually, that the inner coils are put to the greatest strain, and then the slightest flaw in the steel, a speck of rust, grooves cut in the edges of the spring by allowing a screwdriver to slip over them, or an unequal effect of change of temperature, causes the fracture, and leaves the spring free to uncoil itself with very great rapidity.

Now this sudden uncoiling means that the whole energy of the spring is expended on the barrel in a very small fraction of a second. In reality the spring strikes the inner side of the rim of the barrel, a violent blow in the direction the spring is turning, that is, backwards; this is due to the mainspring’s inertia and its very high mean velocity. The velocity is nothing at the outer end, where the spring is fixed, but rises to the maximum at the point of fracture, and the kinetic energy at various points of the spring could no doubt be calculated mathematically or otherwise.

For instance, take a going barrel spring of eight and a half turns, breaking close up to the center while fully wound. A point in the spring at the fracture makes eight turns in the opposite direction to which it was wound, a point at the middle four turns, and a point at the outer end nothing, an effect similar to the whole mass of the spring making four turns backwards. At its greatest velocity it is suddenly stopped by the barrel, wheel teeth engaging its pinion; this stoppage or collision is what breaks center pinions, third pivots, wheel teeth, etc., unless their elasticity, or some interposed contrivance, can safely absorb the stored-up energy of the mainspring, the spring being, as every one knows, the heaviest moving part in an ordinary clock, except where the barrel is exceptionally massive.

Stop Works.--Stop works are devices that are but little understood by the majority of workmen in the trade. They are added to a movement for either one or both of two distinct purposes: First, as a safety device, to prevent injury to the escape wheel from over winding, or to prevent undue force coming on the pendulum by jamming the weight against the top of the seat board and causing a variation in time in a fine clock; or, second, to use as a compromise by utilizing only the middle portion of a long and powerful spring, which varies too much in the amount of its power in the up and down positions to get a good rate on the clock if all the force of the spring were utilized in driving the movement.

With weight clocks, the stop work is a safety device and should always be set so that it will stop the winding when the barrel is filled by the cord; consequently the way to set them is to wind until the barrel is barely full and set the stops with the fingers locked so as to prevent any further action of the arbor in the direction of the winding and the cord should then be long enough to permit the weight to be free. Then unwind until within half a coil of the knot in the cord where it is attached to the barrel and see that the weight is also free at the bottom of the case, when the stops again come into action. This will allow the full capacity of the barrel to be used.

When stop work is found on a spring barrel, it may be taken for granted that the barrel contains more spring than is being wound and unwound in the operation of the clock and it then becomes important to know how many coils are thus held under tension, so that we may put it back correctly after cleaning. Wind up the spring and then let it slowly down with the key until the stop work is locked, counting the number of turns, and writing it down. Then hold the spring with the letting down key and take a screw driver and remove the stop from the plate; then count the number of turns until the spring is down and also write that down. Then take out the spring and clean it. You may find such a spring will give seventeen turns in the barrel without the stop work on, while it will give but ten with the stop work; also that the arbor turned four revolutions after you removed the stop. Then the spring ran the clock from the fourth to the fourteenth turns and there were four coils unused around the arbor, ten to run the clock and three unused at the outer end around the barrel. This would indicate a short and light pendulum or balance, which is very apt to be erratic under variations of power, and if the rate was complained of by the customer you can look for trouble unless the best adjustment of the spring is secured. Put the spring back by winding the four turns and putting on the stop work in the locked position; then wind. If the clock gains when up and loses when down, shift the stop works half a turn backwards or forwards and note the result, making changes of the stop until you have found the point at which there is the least variation of power in the up and down positions. If the variation is still too great a thinner spring must be substituted.

There are several kinds of stop work, the most common being what is known as the Geneva stop, a Maltese cross and a finger such as is commonly seen on watches. For watches they have five notches, but for clocks they are made with a greater number of notches, according to the number of turns desired for the arbor. The finger piece is mounted on a square on the barrel arbor and the star wheel on the stud on the plate. In setting them see that the finger is in line with the center of the star wheel when the stop is locked, or they will not work smoothly.

There is another kind of stopwork which is used in some American clocks, and as there is no friction with it, and no fear of sticking, nor any doubt of the certainty of its action, it is perhaps the most suitable for regulators and other fine clocks which have many turns of the barrel in winding. This stop is simple and sure. It consists of a pair of wheels of any numbers with the ratio of odd numbers as 7 and 6, 9 and 10, 15 and 16, 30 and 32, 45 and 48, etc.; the smaller wheel is squared on the barrel arbor and the larger mounted on a stud on the plate. These wheels are better if made with a larger number of teeth. On each wheel a finger is planted, projecting a little beyond the outsides of the wheel teeth, so that when the fingers meet they will butt securely. The meeting of these fingers cannot take place at every revolution because of the difference in the numbers of the teeth of the wheels; they will pass without touching every time till the cycle of turns is completed, as one wheel goes round say sixteen times while the other goes fifteen, and when this occurs the fingers will engage and so stop further winding. When the clock has run down sixteen turns of the barrel the fingers will again meet on the opposite side, and so the barrel will be allowed to turn backwards and forwards for sixteen revolutions, being stopped by the fingers at each extreme. When in action the fingers may butt either at a right or an obtuse angle, only not too obtuse, as this would put a strain on, tending to force the wheels apart. If preferred the fingers may be made of steel, but this is not necessary.

Maintaining Powers.--Astronomical clocks, watchmaker’s regulators and tower clocks are, or at least should be, fitted with maintaining power. A good tower clock should not vary in its rate more than five to ten seconds a week. Many of them, when favorably situated and carefully tended, do not vary over five to ten seconds per month. It requires from five to thirty minutes to wind the time trains of these clocks and the reader can easily see where the rate would go if the power were removed from the pendulum for that length of time; hence a maintaining power that will keep nearly the same pressure on the escape wheel as the weight does, is a necessity. Astronomical clocks and fine regulators have so little train friction, especially if jeweled, that when the barrel is turned backwards in winding the friction between the barrel head and the great wheel is sufficient to stop the train, or even run it backwards, injuring the escape wheel and, of course, destroying the rate of the clock; therefore they are provided with a device that will prevent such an occurrence. Ordinary clocks do not have the maintaining power because only the barrel arbor is reversed in winding, and that reversal is never for more than half a turn at a time, as the power is thrown back on the train every time the winder lets go of the key to turn his hand over for another grip.

Figs. 83, 84 and 85 show the various forms of maintaining powers, which differ only in their mechanical details. In all of them the maintaining power consists of two ratchet wheels, two clicks and either one or two springs; the springs vary in shape according to whether the great wheel is provided with spokes or left with a web. If the great wheel has spokes the springs are attached on the outside of the large ratchet wheel so that they will press on opposite spokes of the great wheel and are either straight, curved or coiled, according to the taste of the maker of the clock and the amount of room. If made with a web a circular recess is cut in the great wheel, see Fig. 83, wide and deep enough for a single coil of spring wire which has its ends bent at right angles to the plane of the spring and one end slipped in a hole of the ratchet and the other in a similar hole in the recess of the great wheel. A circular slot is cut at some portion of the recess in the great wheel where it will not interfere with the spring and a screw in the ratchet works back and forth in this slot, limiting the action of the spring. Stops are also provided for the spokes of the great wheel in the case of straight, curved or coiled springs, Figs. 84 and 85. These stops are set so as to give an angular movement of two or three teeth of the great wheel in the case of tower clocks and from six to eight teeth in a regulator. The springs should exert a pressure on the great wheel of just a little less than the pull of the weight on the barrel; they will then be compressed all the time the weight is in action, and the stops will then transmit the power from the large ratchet to the great wheel, which drives the train. Both the great wheel and the large ratchet wheel are loose on the arbor, being pinned close to the barrel, but free to revolve. A smaller ratchet, having its teeth cut in the reverse direction from those of the larger one, is fast to the end of the barrel. A click, called the winding click, on the larger ratchet acts in the teeth of the smaller one during the winding, holding the two ratchets together at all other times. A longer click, called the detent click, is pivoted to the clock plate, and drags idly over the teeth of the larger ratchet while the clock is being driven by the weight and the maintaining springs are compressed. When the power is taken off by the reversal of the barrel in winding, the friction between the sides of the two ratchets and great wheel would cause them to also turn backward, if it were not for this detent click, with its end fast to the plate, which drops into the teeth of the large ratchet and prevents it from turning backward. We now have the large ratchet held motionless by the detent click on the clock plate and the compressed springs which are carried between the large ratchet and the great wheel will then begin to expand, driving the loose great wheel until their force has been expended, or until winding is completed, when they will again be compressed by the pull of the weight. In some tower clocks curved pins are fixed to opposite spokes of the great wheel and coiled springs are wound around the pins, Fig. 85; eyes in the large ratchet engage the outer ends of the pins and compress the springs.

Fig. 86.