Spring, torsion or spiral. A strip of steel bent into a spiral, able to expand and contract, elastic, used by horologists in two different ways: one as a power source, and the other as a regulating force.

Springs derive all their energy from the elasticity of their material; this property is generally known and even palpable in almost all objects, but we are still profoundly ignorant of its cause. Therefore in this article I propose to discuss only its effects, and above all the use made of it by horologists to derive power and regulation. For this reason I shall avoid enumerating the various materials capable of elasticity, limiting myself to tempered steel, which horologists use to such great advantage.

We know in general that elastic force can be considered an active power that reacts in proportion to the effort pressing or compressing it; whatever shape a perfectly elastic body may have, it will resume it as soon as compression ceases. For example, when a sword blade is bent, it straightens more quickly the more force was used in bending it; this reaction lets springs take the place of weights as a motive force to power watches and clocks, and for this reason they are called main springs .

As power sources, they may be subject to various configurations that are more or less advantageous for the strength of the power, whence the question may be stated: knowing the material and the quantity, find the figure that will give the greatest elastic power. Aside from the fact that the solution is very difficult and depends on a large number of experiments still to be carried out, worthy of engaging the most skilled physicists, I must for now simply report what is known, rather than what ought to be done.

Of making springs and employing them as power sources.

To make watch springs, a steel bar is shaped in a heavy forge and then drawn as wire of a gauge depending on the springs to be made of it. Otherwise, English steel wire, the best, is cut into pieces 20 to 30 pouces long. After annealing, it is forged flat, about one-quarter ligne thick, and the faces are dressed, the file taking care of the unevenness left by the hammer; this is observed by the curve of the spring when bent here and there along its length. It is also filed to uniform width by passing its whole length through a gauge. When several springs have been prepared thus, wrap each one full length with brass wire, leaving half- pouce intervals; take one spring and bend it in a circle 7 or 8 pouces in diameter, then bend a dozen more of the same width concentrically, forming a cylindrical mass as thick as the springs are wide, and as wide as all the thicknesses together. A hole is left at the center, and all the spaces left by the brass wire; these gaps are useful because the oil or other liquid into which the springs are dipped can easily reach their surfaces. The package of twelve springs is placed on an iron ring shaped like a crown wheel with four arms; it pivots at the center on the end of an iron rod. While in the furnace the ring can be turned at its circumference by means of another rod; it will readily be seen that this mechanism is just for convenience in applying even heat to every part of the circumference.

The package is placed in an annealing furnace with a hot charcoal fire, and when the springs have reached the temperature that only experience can teach, about the red of burning charcoal, they are removed from the furnace and quickly dropped into a sufficiency of rapeseed oil; this is repeated as many times as there are dozens of springs to temper.

Remove the springs from the oil, cut the brass wire at intervals to separate them, brighten them with powdered sandstone, blue them on a hot iron, straighten them with a hammer, and file them again to even them in width as well as thickness, with this exception: the strip must become imperceptibly thinner towards what is to be the inner end of the spring. The last operation requires the fullest attention in order that the curvature may be regular and uniform throughout; when slightly curved and run through the fingers, there must be no difference or stiffness; in a word, equal flexibility throughout its length, as if one were running a ribbon through one's fingers. Experience and a sensitive touch are much better at teaching this test than everything that could be said of it.

After doing one's best with the file, in order to even them perfectly the springs must then be drawn several times between two pieces of hard wood four or five pouces square and well dressed, hinged together, and the top piece fitted with a handle a pied long for clamping. Two workers are required for putting a spring through this machine; one holds it by one end with pincers and pulls, while the other bears down on the handle. Coarse emery is put on the surfaces to start, and fine towards the end, then the strip is polished.

The last operation gives the spring the uniform flexibility so essential to it, after which it is blued again as evenly as possible with gentle heat. Each end is annealed in order to make an opening called an eye ; the end that is to wrap around the arbor is curved with round-nosed pliers, and the spiral profile is given it by wrapping it around the arbor by means of a hook in the eye of the spring; the arbor is turned with one hand while the other thumb presses against the coil, and so for the whole length of the spring. Thus coiled, the spring tends to straighten again, so it must be released gradually. It follows that the reaction is less than the action, the more so the tighter springs are compressed and the longer they are kept in that state. If spring material were perfectly elastic, far from remaining coiled in a spiral it would again become straight, as at the point of departure. On the contrary, if the material had no elasticity the spring would remain as coiled and be worthless; whence it follows that the best springs are those which react the most, or lose the least of their elasticity. Now of all materials tempered steel has the most elasticity, so horologists are right to prefer it. The elasticity of steel is mightily increased by tempering, but the tempering must be limited to avoid breaking under load; it is rightly said that the best springs are likely to break because they have kept the most elasticity. When their elasticity is decreased too much by annealing after tempering, they do not break, but they lose too much of their elasticity and consequently their strength; therefore there are extremes on each side that must be avoided. This is a point that should be possible to find and claim, but doing so is infinitely difficult, not to say impossible. Given this choice, preference goes to a spring that is closer to breaking from too much elasticity than failing from too little. Finally, to summarize what experience and reasoning have shown me concerning the various springs I have tested, I have found, all other things being equal, that a spiral spring is more elastic and retains that quality longer the thinner, wider and longer it is; when wound in a spiral around the arbor in the barrel, its radius should equal the width or height of the coils, and vice versa. That is why the springs in thin watches weaken or break more frequently than others. The spring is hooked in the barrel at the outer end, and the arbor hooks the inner end. In this condition, if the arbor is turned the spring will immediately wind itself around the arbor, as well as each coil in succession. In this condition the spring will be wound up; if it is connected to a movement to be driven by means of the teeth on the circumference of the barrel, which meshes with the first pinion, the spring will drive the movement at a speed that will decrease as the spring unwinds.

The barrel encounters equal radii in the pinion leaves it acts upon; if instead it is fitted with a chain that connects with a conical piece and wraps around it in a spiral groove, the radii decrease precisely as the force of the spring increases, and this is the fusee. See Fusee. Now since the fusee carries the wheel, the barrel will drive the first pinion at the same speed at every revolution, and consequently the power on the whole movement will be uniform.

On making the spiral spring and its use as a regulating force

The spiral spring of a conventional watch is a very fine strip of steel that may be three or four pouces in length, a ninth to a twelfth of a ligne wide, and a thirtieth to a forty-eighth [of a ligne ] thick, curved into a spiral of four or at least three coils. The coils must be more or less spaced depending on the strength of the spiral and the size of the balance wheel; the strip must grow imperceptibly thinner from the outer end to the center, so that when a small weight is hung from the inner end and the outer end is raised with tweezers, the spring assumes the shape of an inverted cone. This is the test to see if the spring flexes correctly and keeps a spacing proportional to the diameter of the spiral ; in addition, the spring coils must be exactly parallel to each other and on the same plane.

To make these little springs English steel is used, not tempered but rolled, which gives it enough body to be elastic. Several makers use it and make their spiral springs themselves; they even straighten and reshape springs already made, but only skillful craftsmen are able to make them well. Geneva is the only city I know where people work exclusively at making these springs, and they are the better at making them because habit and a delicate touch are far superior to theory. They never use English wire; they take a strip of tempered steel, annealed like a mainspring, they reduce it by filing to a certain thickness, then they cut it into small lengths. Evening, filing to width and thickness, smoothing, and bending into a spiral are all operations that are too long to describe in detail, and even then no idea could be given of their delicateness; only experience can do so.

I shall not decide which of the two steel spirals is better for being tempered or not tempered; what is certain is that I have seen good results from both. I do not think anyone knows otherwise than by conjecture which one is to be preferred, and the arguments made on either side strike me as too weak to be related here.

Of the application of the spiral spring to the balance wheel.

The balance staff has a small seat where a collet fits by friction; the circumference of the collet is drilled on a tangent to receive the inner end of the spiral , which is fixed to the collet with a taper pin. The collet is cut to make it slightly elastic in fitting the balance staff, which allows it to be turned; when the balance wheel is assembled on the plate, it rests with the stop pin at the center of the escapement. See Overbanking. The outer end of the spiral is fixed with a taper pin in a drilled stud on the plate. This way the balance wheel cannot turn in either direction without tensing the spiral spring . When the balance wheel is installed, a tooth of the contrate wheel acts on the pallet if it is a verge escapement, or on the face of the cylinder if of this type, and it tenses the spiral spring through the impulse arc. But the balance wheel does not cover its impulse arc without acquiring energy to prolong the arc, which becomes five or six times greater. See Recoil, Deadbeat, Supplement of an angle, Impulse arc, where the spiral spring plays such an important role in opposing the vibrations of the balance wheel and accelerating them. ( See Elastic regulator.) Under the balance wheel is located a mechanism called the regulating slide ; it consists of a toothed wheel meshing with a rack forming an arc three or four times the size of the wheel; the rack has teeth outside and is located concentric with the balance wheel; on a portion of its radius two pins are placed, and between them lies the outer coil of the spiral spring . The wheel carries a pointer; when it is turned the rack moves, and the two pins follow the curve of the spiral and consequently shorten or lengthen it, because its length is assumed to be measured from the fork or pins. One must therefore disregard the portion that goes from the fork to the stud where the end is fixed, because this portion must not receive any motion from the vibrations of the balance wheel; that is why the pins are placed very close together, allowing the spiral to slide between them only; because this mechanism shortens or lengthens the spiral spring , it becomes stronger or weaker, accelerating or slowing the speed of the balance wheel. It is therefore a true regulating force. I have learned from experience that all other things being equal, spiral springs that were small relative to the balance wheel allowed the greatest momentum in the balance without stopping on impulse. In a correct installation, the spiral must not bind in any direction and must leave the balance wheel free to make its greatest arc, which may readily be seen. By observing the watch as it goes, one sees if the spring works flat, if the coils flex in the right proportion, etc.

Spiral springs do not lose their elasticity through vibrations; they compress and expand in perfectly even impulses, and I have made several experiments on this point that prove it. On the machine for testing pivot friction, with the balance wheel held by the spiral , I wound it around the collet by as many as three turns and released it; not only did the spiral unwind three turns, but it made almost three turns in the opposite direction, which almost made the spiral a straight line; so it made six turns in those first vibrations, which decreased in scope until they stopped.

I have repeated this experiment several times, and I have found no alteration in the elasticity of the spiral ; therefore a fortiori it will not lose it in watches, where the greatest tension never reaches a single turn.