Watches and Winders: It's Not Only About Turns per Day

by Walt Arnstein

December 12, 2000


3.0 How a Watch Winder Affects a Watch

Having placed a watch on a winder, we expect it to do only one thing: Wind the watch's mainspring and keep it wound in such a way that the watch will always be ready to wear at a moment's notice and be showing the right time. But as with all things, what you get is not always wholly and uniquely what you want.

To wind a watch, the winder has to turn it, using the force of gravity to keep the rotor stationary while the watch rotates about it. Well, nothing comes without its price; so one unavoidable requirement for winding the watch in this way is that we must establish a torque differential between the rotor's gear train and the resting torque already built up in the mainspring, much as we need a level difference, however small, in order to be able to pour water from one container to another. In other words, the winding system must produce a torque greater than that to which it is adding energy and a winder does this in a steady and continuous fashion while it is turning the watch.

Naturally, the extra torque is passed on to the watch's gear train and the escapement. To be sure, the increment is not large but it is felt. Of that you can be sure. So for better or for worse, we can assert that the watch's balance wheel's amplitude of oscillation will at best increase ever so slightly while the watch is turning compared to the amplitude when the watch is resting. The isochronism adjustment built into the watch, if any, could minimize the net effect, but the effect will definitely be there. This could explain many owners' puzzlement with the fact that their watches run at a different rate (usually faster) on a winder than they do on the wrist or on the dresser top…not to mention running differently on one winder compared to another winder.

Once the mainspring reaches the slippage point, the added torque could increase significantly, particularly if the slippage threshold has moved up as a result of poor lubrication inside the mainspring barrel. Recall that this is hardly noticed on a watch that is being worn on the wrist.

But let us assume that we will not allow the watch to run on the winder beyond the point where slippage of the mainspring begins to take place. We will program our winder by suitable means to wind the watch just short of its full wind condition. Still – and this is significant – the watch's gear train will be subjected to greater-than-normal torque for any periods during which the watch is actually rotating on the winder.

Now we all know, and winder instruction manuals tell us clearly, that each watch has a required daily minimum of rotor turns (meaning also watch turns on the winder) in order to keep the watch from eventually running down. The average such requirement seems to be between 500 and 700 turns per day, but the figure 720 is a favorite setting for a winder in that it adds just enough daily turns to compensate for periods when the watch is off the winder and loses some mainspring tension as a result.

The above requirement has introduced a new variable: The time the watch spends actually rotating on the winder. And winders differ greatly in how they go about turning the watches. This will be examined in the next section.

3.1. Turns per Day versus Hours per Day.

Having established that the time the watch spends turning on a winder involves running under higher-than-normal torque, it will be interesting to compare how different winders accomplish their task and how their operating procedure affects this parameter.

At one extreme, consider a winder like Superior's Cyclomatic Due™ or its Time Mover™, both of which offer a "continuous wind" mode (intended mainly for fast winding on a watchmaker's workbench and usable only with an AC adaptor) by means of which the unit's spindle rotates at 19RPM without stopping. In this mode a watch can be given 24 hours' worth of power reserve in about 32 minutes. Following this procedure, the watch can be left on the stopped winder until its motor is restarted 23 hours and 28 minutes later.

At the other end of the spectrum, consider a winder whose spindle turns at ½ RPM, which comes out to 720 revolutions per day as well– just a bit more than needed to keep the same watch wound. This winder keeps the watch turning continuously, 24 hours per day, subjecting the watch to extra torque for the entire period.

If the same watch is placed on the two winders described above, it is quite likely that the performance of the watch will differ dramatically.

Incidentally, there is no suggestion in the above discussion that one or another of the above winders will expose the watch to unnecessary risk. Undoubtedly both winders will treat the watch safely providing the number of turns per day is not exceeded by an unreasonable margin. However, on one winder, the watch may gain or lose a matter of seconds per day while on the other, the same watch may build up a sizable error over a period of a week, say. Complaints regarding a watch's performance while on a winder have been extremely common among TZ posts over the years.

3.2. Other Environmental Considerations.

Let us return to a pair of winders such as were discussed in the previous section. This time, we will look at the winders' actions in more detail. The Cyclomatic Due™, when used in its normal operating mode, runs for 2 minutes every 40 minutes at about 19 RPM, reversing direction for each such 40-minute cycle. In this way, it executes 36 cycles of 2-minute rotations at ~19 RPM in the course of 24 hours, which for a bidirectionally winding watch comes out to 1368 revolutions per day. On a unidirectionally winding watch, this comes out to 684 winding (and 684 freewheeling) revolutions per day, an adequate number for most watches. We can run this latter watch without a timer.

The bidirectionally winding watch should be used with this winder connected via its AC adapter to the AC line through an appliance timer that limits its operation to about 12½ hours per day. It would be best if the appliance timer were programmed to operate in 2 or 4 cycles of about 6 hours or 3 hours each, respectively.

Since an IC timer controls the Cyclomatic Due's cycles, the spindle's starting and stopping position for every 2-minute rotation interval every 40 minutes is not clearly defined. As a result, the watch will stop in an arbitrary position following each period of rotation and maintain this position for the next 38 minutes or so. Further, for about ½ day, it will remain in whatever position it stopped last. The resulting performance of the watch is a variable determined more by the properties of the watch (i.e., the quality of its adjustments for position) rather than those of the winder.

The continuously operating ½ RPM winder will not face the above problem. Indeed, its slow rotation about the axis of its hands duplicates in some ways the behavior of a tourbillon cage rotating at that rate. Hence, its rate should be free of the unpredictability of rest position. The only effect the winder introduces to offset this fortuitous benefit is the extra torque it adds to the mainspring.

Again, it is difficult to predict which winder will work "better" with a particular watch, but it is reasonably certain that a difference in performance will be noticeable in the same watch using these winders as recommended by the manufacturer. The same applies to any pair of winders.

3.3. Notable Departures from Common Winder Design.

The overwhelming majority of watch winders operate on the common principle of rotating watches about the axes of their hands. That is, the winders' spindles are parallel to the axes of the watches' hands and balance wheels. However, there are a number of exceptions and each of them introduces some variations on the general rules described above – some of them favorable to the watch's well-being and better performance.

One such innovation is to be found in Orbita's original winder design, an AC-only winder which uses a "mandrel", a fat, padded, slightly tapering spindle elevated at 30 degrees from the horizontal, onto which a watch is slipped as though it might be a wrist. The mandrel rotates at 3 RPM, its axis of rotation being parallel to the line passing through the 3 o'clock and 9 o'clock points on the dial, i.e., the winding stem on most watches. If we examine what happens to the rotor of a watch mounted in this way – with the crown pointing upward, just for illustration -- we will discover that the rotor swings back and forth relative to the body of the watch between the 7 o'clock and 11 o'clock positions, one such full swing for every turn of the mandrel. The direction of rotation of the mandrel is irrelevant. The rotor will swing back and forth the same way for either direction of mandrel rotation.

Theoretically, the rotor of the watch executes 2/3 of a full turn at every turn of the mandrel, 1/3 turn in each direction. The mandrel, without an appliance timer, executes 4320 turns per day, which causes the watch's rotor to perform 2880 full turns per day on a bidirectionally winding watch and 1440 on a unidirectionally winding one – far too much if we continue to stick to geometric considerations alone.

However – and this is a big however – the alternating nature of the rotor's motion introduces a significant amount of "dead time" as the rotor direction reverses repeatedly, reducing the effective swing of the rotor to a maximum of about 210 winding degrees instead of the theoretical 240 degrees for each turn of the mandrel. Further, the rapid alternating motion of the rotor results in a gradual narrowing of the rotor's amplitude as the mainspring's tension grows, according to Chuck Agnoff, president of Orbita™ (discussed by him in a 1996 post on TZ's Public Forum). The consequence is a certain amount of self-limitation built into these winders, similar in nature to the limitation found in normal wear on the wrist. Some owners of these winders have in fact reported that their particular watches reached a power reserve well below the manufacturers' rated reserves and did not go beyond this tension even if kept on the winder all day. I have not had a chance to confirm or refute this but if it is true, the winder offers a simple method of keeping a watch wound to an optimum degree. After all, the ideal winder would keep a watch half-wound all the time, rather than fully wound.

In preference to depending on this form of limitation for all use, I would recommend putting the Orbita™ AC winder on a timer set for about 10½ hours per day for bidirectionally winding watches and about 21 hours per day for unidirectionally winding ones. This is a good, safe starting point. However, it may well be necessary to increase the time for the bidirectionally winding watch, given Mr. Agnoff's description of this winder's operation.

The Orbita™ AC winder is not the only winder that operates in the unusual manner just described. Superior's Cyclotest and the almost identical Bergeon professional winders (actually testers designed mainly for watchmakers, who use them to test automatics and manuals alike) rotate watches about the same axes, as does Superior's Windmill™. All of these winders add also a revolution of the rotating watch-holding spindles about a common "carousel" hub. The entire carousel is oriented essentially in a vertical plane, with the spindles protruding radially from the edge of its hub. The hub's rate of rotation is about ¼ that of the spindles. The complex motion of the watches on these winders tests them in all possible physical orientations over time. The rotors on the watches attached to these systems execute a long-term average of about ½ full turn for every turn of the watch-holding spindles. Again, one would expect that as the watches' mainsprings wind, the above figure would begin to shrink significantly. Still, these testers put out a very high number of turns per day – on the order of 10 times as much as needed by the watches – so an appliance timer is pretty much a must for them.

3.4. Additional Factors Affecting Winding Behavior.

We have already established the fact that excessive winding, while not immediately harmful to a watch in good state of repair, does stress the overwind protection subsystem of the watch and should be avoided. For this reason, a built-in variable adjustment for turns per day on a winder is definitely a desirable feature, most notably if the winder is to be powered primarily by its batteries. For winders that operate on household AC current, a simple programmable appliance timer, costing $10-15, will admirably perform the task of limiting the turns per day, as previously mentioned.

Several models on the market today, including some by Orbita™, TimeCube™ and Scatola del Tempo™, offer built-in adjustment. The range of turns per day should ideally include values such as 600, 700, 900, and 1200. If there is no provision for choosing a winding direction – i.e., if the winder alternates rotation direction according to a built-in program, as on Superior's Cyclomatic Due™ – the range should extend at least to about 1500 turns per day, since only half of these turns will actually be winding a unidirectional watch and many unidirectionally winding watches (e.g., the Valjoux 7750) need about 720 winding (clockwise in the case if this movement) turns per day.

One unfortunate feature of battery-operated winders is that the motor's speed of rotation keeps decreasing as the battery voltage drops with use. A fresh set of alkaline batteries will have a no-load voltage of 1.55 to 1.59 volts per battery. Over time, meaning a matter of 6 months to a year of use in a winder, the same batteries' output voltage per unit will drop to 1.4 volts or so before the owner decides it is time to replace it. Measurements I have made on winders fitted with ultra high quality Swiss Portescap or German Faulhaber motors have shown that the RPM of a spindle will drop from about 19 with a new set of batteries to about 13 with a set of batteries ready for replacement.

This inevitable loss of RPM translates into a corresponding drop in revolutions per day, since most winders operate by controlling winding cycles using absolute winding time via IC timers like the NC555 or NC556, exceedingly accurate microcircuits that are almost totally immune to operating voltage variations within their wide specified range. As a result, a watch on such a winder will always get an unvarying number of minutes per day of winding but not turns per day.

The owner of such a winder as above needs to be aware of this phenomenon and be prepared to boost the "turns per day" setting on at least one occasion during the operating life of a set of batteries and then reset it after he has placed a fresh set into the battery compartment.

3.5. Battery Life in Winders.

The question of useful battery life in a typical DC or AC/DC winder is significant not so much for the cost of the batteries, but for the inconvenience of replacing the batteries in some winders and, worse, the unpredictable nature of the individual battery's discharge curve, hence useful life. The owner of a Patek Philippe Perpetual Calendar might be quite upset if, returning from a 2-week trip out of town, he found the winder – and watch – he had left in his safe stopped, when he had expected to get at least another month's use out of the battery set.

Some winders have their batteries installed in a hard-to-reach compartment and replacing them involves partially dismantling the winder (for example, removing the watch carrier from the spindle and pulling away the face panel behind the spindle). Others require removal of the bottom of the winder box by removing special screws hidden by rubber plugs, then reaching close to or past some electrical wiring to reach the batteries. For such winders, it is desirable that the batteries' expected working life be reasonably long. Most winders do provide amazingly long operating life, even if used continuously. One year is a typical specification.

The cost factor for batteries is not a vital consideration. Four alkaline "C" size batteries cost at worst $2.50 and will last a full year in such winders as the Cyclomatic Due™. I have measured its current drain when turning 2 heavy watches to be about 8 mA, an amazingly low figure. The typical alkaline "C" cell has a claimed working life of about 3500 mAH (milliampere hours) with electronic equipment, suggesting that a set of these batteries will satisfactorily turn watches for about 450 hours. Since the spindle duty cycle of a Cyclomatic Due™ is 1:20 (2 minutes' running alternating with 38 minutes' rest), this comes out to about 375 days in calendar time, a few days past a full year.

Several other winders I have examined have similar battery life specifications. But I did note at least one – an English winder named Rapport™ -- whose four "AA" cells had a life expectancy of only a couple of weeks. In all other ways, the unit was very interesting in its method of operation and very well constructed. Still, the prospective buyer of this winder – if it makes an appearance in the US market – should be aware of this limitation. It may well be insignificant to anyone who expects to use the winder mainly with the supplied AC adapter.

Before leaving the subject of batteries and AC adapters, we should note that most or all AC/DC winders on the market today lack a vital feature: Automatic switchover from AC to batteries in case of AC power failure. A winder attached to its removable AC adapter and plugged into the household AC supply totally disconnects the batteries from access by the on-board electronics. As a consequence, the winder so connected will stop dead if the AC supply fails and will remain stopped even though there is a totally fresh and functional set of batteries sitting in its housing, just waiting to be called upon. In my opinion, this sort of oversight is inexcusable. Implementing the feature is technically very easy and straightforward – so much so that few owners are aware that it is not in fact available. Yet that is indeed the case with all present-day AC/DC winders.

4.0 Letting the Watch Instead of the Winder Set Winding Time.

Timing is not the only strategy available to the resourceful watch owner. It would, for example, be very desirable to have a means of sensing a mainspring's state of wind without actually attaching something to it or, alternatively, let the winding system somehow limit the turns of the rotor when a certain state of wind is reached.

The former item on the "wish list" is clearly not feasible for watches already in existence and probably not practical for new watches. However, it is very easy to come up with a way to stop a mainspring from winding beyond a chosen level of tension and that means any level of tension below the slippage threshold – no matter how many hours per day the winder runs. How? Simply by varying the winder spindle's angle of elevation.

I have already written on this subject before, so I will not cover it in the same detail again. The interested reader is referred to "Effect of Winder Spindle Elevation on Watch Winding" in the "Articles from the TimeZone Community" section of TZ's "Learning Zone". For purposes of this discussion, we will simply analyze the principles involved and describe how they function to meet our objectives.

A watch positioned with the axis of its rotor horizontal lets the full force of gravity act on the rotor as the watch is rotated, the rotor's unbalanced weight acting along a lever arm represented by the distance of the rotor's center of mass from its suspension point. This is the position that will give the rotor the maximum winding torque it is capable of exerting on its gear train. Even with rotors of modest mass, this position will allow the rotor to wind the mainspring up to and beyond the slippage point, the threshold of engagement of the overwind protection mechanism.

If we gradually start elevating the axis of rotation of the watch, i.e., the line of the spindle that rotates the watch holder, the amount of torque available from gravity will start to recede. A geometric analysis of the amount of available torque will disclose that it will vary with the cosine of the elevation angle. This trigonometric function, when plotted on paper, has a value of 1 when the angle of elevation is zero (i.e., the spindle is horizontal) and gradually drops to zero when the elevation is 90 degrees, corresponding to the spindle pointing straight up.

This makes sense. Consider what would happen if we tipped a winder on its back and pointed its spindle straight up. The dial of the watch mounted on it would be facing straight up and the rotor would be lying in the horizontal plane. No gravity would act on it and it would simply revolve with the rest of the watch, accomplishing no useful winding work.

But in between these two extremes lies a whole range of extremely useful attitudes. At an elevation of 60 degrees from the horizontal, for example, exactly half of the rotor's maximum torque would be available to the mainspring. At 45 degrees, the proportion would be 0.707; at 30 degrees, it would be 0.866, etc.

Consider now what might happen if you tip your winder back until its spindle's angle of elevation is 60 degrees, say. For simplicity, assume the winder turns continuously at 1 RPM, so that given free rein, it would deliver to the watch 1440 turns per day – far too much. But with this steep elevation, the situation changes. The watch's mainspring, after 24 hours of sitting on a desktop, puts up little resistance to the winder and begins accumulating energy. After 6 or 7 hours, however, the resistance of the mainspring to the winder (and gravity, which supplies the winding force) begins to be significant. The rotor has to begin climbing up from its original position near the bottom in order to keep supplying adequate torque. At around 10 hours, the resistance of the mainspring is hard enough to overcome the rotor's weakened gravitationally generated torque and the rotor begins to go around in a full circle with the watch! All winding ceases as watch and rotor continue to revolve together. The result is very entertaining with a movement that winds in only one direction, like the Valjoux 7750. As the rotor of this movement clears the top of its swing arc, it drops under gravity in its freewheeling mode, picking up a sizable angular velocity. Then, for 20-30 seconds, it spins rapidly, totally ignoring the much slower-turning spindle, until it has slowed down enough to get picked up by the spindle's rotation. Even without a transparent back on the watch, the action is amazing to hear.

The over-the-top behavior will persist until the mainspring has discharged enough energy to the main gear train and escapement for the rotor to begin taking charge again. Interestingly enough, the mainspring may not have become fully wound and probably has not even approached the threshold of engagement of the overwind protection mechanism. But it is a fact that if the watch is left on the winder indefinitely, it will never exceed the state of wind it has now reached – a very desirable condition, assuming we can set the inclination of the spindle to such an inclination that the energy delivered in 24 hours is always equal to at least 24 hours.

How does one determine the desired inclination? Experimentally, by trial and error or by formula. In the article I wrote some time ago, "Effect of Winder Spindle Elevation on Watch Winding", I described two experimental methods for determining an ideal angle of elevation, one easy one for watches with transparent windows on the back and one more complicated for watches with opaque backs. The second does require some mathematical calculations, but trial-and-error methods can work amazingly well.

It should also be obvious to the reader that if the above method is to be employed, some reasonable means of providing the extra (or reduced) tilt of the spindle must be constructed, in the form of a wedge or other type of object. This will be touched upon in the next section.

4.1 Elevation Angles of Commercially Available Winders.

Many currently available winders already come with tilted spindles. For example, the Cyclomatic Due's spindle has an elevation angle of 35 degrees. This is too low to limit the winding of any automatic watch now on the market. But, of course, the tilt can be increased by a 20-degree wedge, which would affect a number of popular watches.

The MTE AC winder, one of the most inexpensive winders on today's market, has a spindle elevated by about 52 degrees from the horizontal. This definitely limits its ability to wind fully some watches with relatively low rotor mass, for example the Swatch Automatic. This watch will reach equilibrium with the winder when it has accumulated about 22 hours of power reserve. So, removing the watch from the winder and placing it on a table will result in the watch's stopping in less than 24 hours.

But this is a favorable result. The watch, if left on the winder indefinitely, will settle down to running with about 14 hours of power reserve below its maximum of 36 hours, a good level for continuous "standby" operation. At any time, the owner could remove the watch from the winder, put it on his wrist and bring it up to full power reserve in a few hours. Who could ask for a better situation?

We have no clear indication if the MTE's elevation is a fortuitous coincidence or a result of careful planning by its designers, but it basically limits the winding turns the winder can deliver to the watch. A good thing, too, because this unit delivers 4320 turns in a 24-hour day, about 6 times as many as a typical watch needs to store 24 hours' worth of power reserve. If all these turns actually went into the mainspring barrel, they would overwork the overwind protection mechanism mercilessly.

Superior's Time Mover™, a beautifully crafted winder that uses basically the same electronics and motor as the Cyclomatic Due™, has a spindle elevation angle of about 55 degrees, even higher than the MTE AC winder. It is believed that this elevation was dictated more by considerations of compactness (short front-to-back measurement) than by those of limiting winding action. Still, by a fortunate combination of factors, this gives an ETA 2824 movement (like the one in my Fortis Pilot Auto) just enough winding turns to keep it at about 30 hours' power reserve. If left on the winder for long periods, as is the case with me, the watch can be removed from the winder and placed on the wrist with total assurance that this is the power reserve with which the watch will start the day.

At this moment, no winder manufacturer markets a winder with a variable elevation angle. This situation may change some day.

5.0. Summary.

In this article, we have found, probably to few TZers' total surprise, that a watch will behave differently while on a winder than it will on the wrist. Watch owners who feel that if a watch can be worn safely on the wrist under any conditions, it should be safe to leave it on a winder without worrying about turns per day are mistaken. The treatment a watch receives from a winder is often far more severe than that delivered by the wrist.

More significantly, a watch's behavior and performance will differ to a significant degree from one winder to another. One major reason for this is that winding minutes per day are as important as winding turns per day.

Winders that rotate watches about the axis of their winding stems (like the original Orbita™ AC winders) have a degree of winding limitation built into the design. It is fairly difficult to overwind a watch on these winders. On the other hand, the effect of the limited arc through which the rotor swings back and forth on these winders relative to the rest of the watch (+/- 120 degrees per turn of the mandrel and gradually less as the mainspring tension increases) has not been evaluated by experts.

Further, some watches will benefit from the simple process of being rotated in the vertical plane, much as a tourbillon rotates its cage, distributing the position errors in that plane evenly over time. In fact, many watchmakers use winders, or rather, testers like the Cyclotest™ or Bergeon™ to time all watches after cleaning and lubrication, even manual wind watches, simulating typical wear on the wrist.

Finally, a significant variable on winders available today is the angle of elevation of the spindle. The higher the elevation, the less winding torque from gravity is available to the watch rotor. This can be a problem in some watches with light rotors, but can also be a convenient way of automatically limiting the amount of power reserve that is allowed to accumulate in the mainspring. Basically, it allows a watch to be rotated indefinitely on the winder without ever reaching the threshold of engagement of the overwind protection mechanism.

Click here to return to Part 1


Back to top of page | Return to Time Zone Home Page
Copyright © 2001-2004 A Bid Of Time, Inc., All Rights Reserved