by Walt Odets

Mark XII on timer microphoneI have already discussed the concepts of watch timing (“Tweaking the Mark XII: Part 2.1“). In this part of the series, I will discuss the principles used in actually timing a watch.
Understanding these principles is predicated on a certain familiarity with the escapement, and a previous Horologium article, “The
Anchor Escapement,” is recommended for those unfamiliar.
While this exposition is not intended as a pragmatic manual for
timing procedures, it will allow the watch owner to understand
the issues involved, and to appreciate the complexity of a well-adjusted
watch. Inside the watch that “keeps excellent time”
one finds both beautifully adjusted watches, and those whose
good time-keeping can be attributed to little more than a fortuitous
combination of technical short comings and the owner’s personal


In daily running, the regulation of a watch
is, of course, accomplished by the balance wheel and balance
spring. But there is an old and correct watchmaker’s adage that,
“Good Timing Starts at the Mainspring.” The timing
of a watch can be no better than the mainspring, center wheel,
third wheel, fourth wheel, escape wheel, pallets and pallet lever,
and all their associated pivots, pinions, teeth, and leaves.
If the mainspring–through age, mishandling, or incorrect lubrication–delivers
power unevenly, that will be reflected in the timing. The power
delivery all the way from the mainspring to the balance impulse
jewel must be adequate, smooth, and relatively flawless for a
mechanical watch to provide consistent, excellent function.

HEIGHT=”136″ ALIGN=”RIGHT” BORDER=”2″ NATURALSIZEFLAG=”3″ ALT=”Timing tape, amplitude”>The best single measure of the condition
of a watch is the amplitude
of the balance. Amplitude is expressed as the number of degrees
of swing of the balance from rest (in the centered position)
to the end of its excursion in either direction. In a healthy,
freshly-serviced watch, amplitude should be between 275 to 310
degrees in the horizontal (dial up or dial down) position. In
a vertical position (crown down, left, up, or right), amplitude
will drop about 45 degrees due to increased friction in the balance
pivots. Thus, an acceptable minimum for vertical positions
is about 230 degrees. Because correct amplitude–good motion
in the parlance of watchmaking–depends on every component of
the movement, most faults will be reflected in inadequate (or,
occasionally, excessive) amplitude. Without good motion, a watch
cannot be properly adjusted. No amount of fiddling with the escapement
can correct for poor motion.


For purposes of this discussion, it is
assumed that the entire train of the watch is in good condition,
and that both horizontal and vertical amplitude are within correct
range. It is assumed that pallet jewels and escape wheel provide
consistent and correct geometry. And, finally, it is assumed
that the balance wheel, itself, is properly poised. An unintentionally
out of poise balance will, obviously, create unwanted effects
in the vertical positions of the WIDTH=”280″ HEIGHT=”285″ ALIGN=”LEFT” BORDER=”2″ NATURALSIZEFLAG=”3″
ALT=”Balance cock schematic”>watch. (I have discussed balance
wheels and their poising in a previous Horologium article, “The
Balance Wheel of a Watch.”)

Those basic and necessary requirements
met, the adjustment of the watch is accomplished with adjustments
to five critical regulating components. As shown in the schematic
of a balance cock (in blue, at left) seen from
below, these include the balance spring (1);
the stud carrier (2), to which the outer end of
the spring, with its stud, is anchored (at 2A); the regulator
(3), with its adjustment mechanism (3A); and
the collet (4), which rides on the balance wheel
shaft, and to which the inner end of the balance spring is attached.

The same components, identically numbered,
are also shown in profile, below right. In this
view, it can be seen that the balance spring (1) surrounds
the collet (4) to which it is attached. The spring spirals
outwards, eventually passing between the regulator index (3B)
and regulator boot (3C). The spring ends at the
stud (2A), to which the spring is attached
with a pin or adhesive. The stud is, in turn, fastened into the
stud carrier (2) by means of a se WIDTH=”350″ HEIGHT=”194″ ALIGN=”RIGHT” BORDER=”2″ NATURALSIZEFLAG=”3″
ALT=”Schematic profile”>t screw (2B). The regulator
on top of the balance cock (3D) carries the
upper balance jewel set (hole and cap jewels) and shock absorber.
In the case of the Triovis regulator on the caliber 887,
the fine regulator mechanism is also part of the regulator housing.
The balance wheel (5) and balance shaft (5A) are
also shown in the profile view.

The ring of the regulator and Triovis
ride below the regulator housing. All are shown below
left. The regulator ring (3) lies on top of the
balance cock with the regulator housing (3D) and Triovis
ring (3F) on top of the regulator ring, in that order.
In other words, the Triovis ring is between the regulator ring
and regulator HEIGHT=”242″ ALIGN=”LEFT” BORDER=”2″ NATURALSIZEFLAG=”3″ ALT=”Regulator parts”>housing.
The tab of the Triovis ring (3G) is moved by the regulator
screw (3A) and the Triovis ring moves the regulator ring.
It is by this means that the effective length of the balance
spring is adjusted. In watches with regulators, adjusting the
effective length of the balance spring is how regulation
(adjustment for daily rate) is accomplished. (Watches without
regulators must be regulated by means of changes to the balance
wheel itself.)


The two horizontal positions should measure
very close to each other, and this is the first measurement taken
in evaluating the adjustment of a watch. In good quality or better
watches, rate and amplitude should be within a few seconds per
day and about 10 degrees, respectively. If these two positions
are significantly different, this must be corrected before any
further adjustment.

When differences are found, they can usually
be traced to problems with balance pivots or their lubrication,
or to the adjustment of the regulator index (discussed below).
Other parts of the gear train and escapement may also, occasionally,
be at fault. Differences in tolerances or end shake on the pallet
lever and other components can shift alignment with changes in
position. Large differences in amplitude often suggest that the
balance spring may be rubbing on the underside of the bridge
(dial up) or on the balance itself (dial down).


ALIGN=”RIGHT” BORDER=”2″ ALT=”Collet point of attachment” WIDTH=”79″
HEIGHT=”88″ NATURALSIZEFLAG=”3″>The collet, as previously mentioned, attaches
the inner end of the hairspring to the balance shaft. Unlike
the attachment of the outer end of the spring (at the stud holder),
the inner attachment (and collet) rotates with the balance. The
collet is simply a cylinder that fits on the balance shaft, with
a channel to hold the end of the spring. At right the
collet is indicated in pink, the cross-section of the
balance staff in yellow. It can be see that the spring
attaches to the collet at the red arrow. (The green
indicates the slot that allows the collet to be spread
and placed over the balance staff.) The four-sided (rather than
round) collet of the IWC caliber 887 is shown below left,
the spring attached to the collet at the red ALIGN=”LEFT” BORDER=”2″ ALT=”Caliber 887 collet” WIDTH=”136″
HEIGHT=”118″ NATURALSIZEFLAG=”3″>arrow. The balance shaft and
upper balance pivot are also visible, in the center of the collet.
This collet shape simply makes the collet easier to work with.

In the late nineteenth century Jules Grossman,
a watchmaker from LeLocle, discovered perhaps the most critical
principle of modern watch adjusting: the point of attachment.
The point at which the balance spring attaches to the collet
(relative to the case of the watch, and thus to the position
of the crown) affects the rate differently in each of the four
vertical positions. Although traditionally used to adjust watches
with overcoil balance springs, the principle of point-of-attachment
is also applicable to flat balance springs and is now used in
all finely adjusted watches.

ALIGN=”LEFT” BORDER=”2″ ALT=”Pivot, horizontal positions” WIDTH=”210″
HEIGHT=”210″ NATURALSIZEFLAG=”3″>For purposes of adjusting a wristwatch
(as opposed to pocket watch), the vertical positions in order
of importance (starting with the most important) are crown down,
crown left, and crown up. Crown right is not normally adjusted
in a wristwatch, but for an owner who wears the watch on the
inside of the right wrist, crown right is substituted for crown
left. The ordering of these
positions is determined simply from their frequency in average
wear, and is universally accepted among watchmakers and manufacturers.

One of the most difficult problems in adjusting
a watch is the faster rate of vertical positions (compared to
the horizontal positions). In vertical positions, the friction
of the balance pivots is much higher, which reduces amplitude.
The reduced amplitude produces a faster rate (because shorter
swings complete more quickly). To the extent that the innate
characteristics of the balance spring provide good isochronism
(causing shorter swings to complete more slowly, longer swings
more quickly), the effect of vertical positions is reduced. But
it is never eliminated entirely
through good isochronism alone.

ALIGN=”RIGHT” BORDER=”2″ ALT=”Pivot, vertical positions” WIDTH=”280″
HEIGHT=”256″ NATURALSIZEFLAG=”3″>Each attachment point of the balance spring
to the collet (at 12, 3, 6, or 9 o’clock) provides predictable
corrections of rate: little error in two positions, slow rate
in the third, and fast rate in the fourth. Thus, in wristwatches,
the point of attachment is usually chosen to lower the rate
in the crown down position
, the most common vertical position
for a wristwatch. Theoretically this is the position illustrated
below right. This provides minimal error in the crown
right and left positions, fast rate in the crown up, and slow
rate in the crown down. While this is clearly a compromise, other
points of attachment often provide less satisfactory adjustments.

In a watch adjusted to five (versus three)
positions, this theoretically correct attachment may make the
crown up position difficult to keep within acceptable parameters.
Friction effects on the sides of the balance pivots (particularly
with a flat balance spring, with which shifts in the spring’s
center of gravity are larger than with overcoil springs) also
affect adjustment. Thus, the best pragmatic compromise on point
of attachment varies from caliber ALIGN=”RIGHT” BORDER=”2″ ALT=”Geomoetry of POA” WIDTH=”205″
HEIGHT=”178″ NATURALSIZEFLAG=”3″> to caliber. One generally finds
Jaeger Le Coultre movements (and IWC’s, like the 887, derived
from them) adjusted as illustrated below left. This point
of attachment tends to provide moderate error of similar magnitude
in all four vertical positions: crown down is slightly slow,
crown left slightly fast, crown up slightly fast, and crown right
slightly slow. In most high quality watches, correct selection
of point of attachment alone should bring vertical positions
to within eight to 15 seconds per day of the horizontal positions.
Additional correction must be provided by the inate characteristics
(mostly metallurgical) of a balance spring that improves isochronism
and thus reduces rate in vertical positions, despite the reduction
in BORDER=”2″ ALT=”JLC POA” WIDTH=”143″ HEIGHT=”143″ NATURALSIZEFLAG=”3″>amplitude.
Obviously, any other faults in the watch that reduce amplitude
in vertical positions must also be corrected to minimize the
increase in rate. Occasionally, a balance wheel may intentionally
be thrown out of poise to improve performance in one or two vertical
positions. But this approach is likely to worsen performance
in other vertical positions, and is usually considered bad technique.
Clearly it is an approach most applicable to watches adjusted
in three (or possibly four) positions.