Watch Adjustment

By Dewey Clark, President and Watchmaker, 2019

Historic Timekeepers, Inc.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
1
 “Repairing a watch is one thing. Rating it to position is something else.”

W.H. Samelius, 1944

Introduction

This article provides a description of the steps involved in returning a high

.
grade watch to precision performance. It is based entirely on modern

nc
knowledge and practice as explained by a number of sources as long ago
as 1963. Even in the US, W.H. Samelius was aware of these procedures in

,I
1944.

rs
pe
The description is not based on new knowledge created by or the opinions
of the author. It is taken from known sources that include WOSTEP, Greiner

ee
Vibrograf AG and H. Jendritzki, a Swiss watchmaking instructor.

ek
The author became aware of some of this 15 years ago and started to
m
question much of the information to which he had been subjected since
1985. After attending the WOSTEP program in Neuchatel he applied this
Ti
knowledge in his business with satisfying results.
ic

Watchmakers and collectors have been subjected to much misinformation

or

about the proper service and performance capabilities of vintage precision

ist

watches. This is especially true when it comes to American Railroad (RR)

watches, although it applies throughout the range of high grade watches.
H

This article is intended for the serious watch collector who wants to
ht

understand what a competent watchmaker will do in the course of service.

rig

It was initially written after working with several recent graduates of

py

global watchmaking schools. It became clear that after graduation, they

have a period where they are uncertain how to apply what they were
Co

taught. Having attended one of these programs, I can attest that there is a
lot of information to be absorbed and there is not much opportunity to
digest it as a complete package.

Courtesy of www.historictimekeepers.com
2
So one purpose is to summarize in one place what is taught about watch
adjusting.

Another purpose is to assist those workers without the benefit of formal

training in improving their work. While this article focuses on the
technique of dynamic poising for positional adjustment, it must be
cautioned that this technique is only useful in a correctly serviced watch.

.
Dynamic poising will not overcome faults in an improperly serviced watch.

nc
,I
To that end, the subject is presented within the context of the watch
adjustment sequence. These topics are treated briefly, as the

rs
knowledgeable watchmaker will understand and value that everything

pe
he/she does to a watch is to prepare it for positional adjustment as a
precision instrument.

ee
ek
Less knowledgeable watchmakers may come to appreciate why they get
less than desired results. This article may serve as a basis for determining
m
a plan for skill improvement.
Ti

Finally, the watch owner will have a basis for evaluating the level of
ic

expertise of a watchmaker they propose to use for the maintenance of

or

their watches.
ist

Expectations for Precision Watches

H

Watches such as the American Railroad Watch were designed to be used as

ht

precision instruments. The American RR watch (typified by the iconic

Hamilton 992) was sold, used and serviced with the expectation that it
rig

would have a maximum daily error of 5 seconds per day that was
consistent and predictable. Given RR watches recorded 432,000 seconds
py

in a day, this means the watch had a calibration error of 1 part in 86,400!
Co

That is truly an accomplishment that should be, and can be, preserved.
The original standard for the Hamilton 992 is tighter than the factory
standard for a modern Rolex.

Courtesy of www.historictimekeepers.com
3
Yet, many owners and watchmakers believe it is impossible or unrealistic
to have these 120 year old watches perform to that standard today.

There are many reasons for this. Probably the biggest is that for 20 years
or so these precision instruments were not appreciated for what they were.
They were given to curious children and used by others as practice pieces.
Secondly, and egregiously, authors popular with individuals without the

.
benefit of formal training encouraged methods that actually amount to

nc
vandalism. Finally, even trained watchmakers came to conclude that most

,I
owners were unwilling to pay for the time required to properly service
these watches and simply performed a “cleaning and lube” while telling

rs
customers they should not expect precision performance due to age.

pe
Of course, any mechanical device can be returned to original performance

ee
if it is returned to factory condition and properly serviced. Because of the

ek
necessity to correct damage done by the naive, the initial cost can be high
and may involve the use of a “donor” movement for unmolested parts.
m
Ti
But, after such work, the cost of service drops dramatically if the watch is
serviced by a knowledgeable watchmaker. And while it generally costs
ic

$750 to $1000 to have a modern watch serviced, the cost to maintain a

or

Hamilton 992 in proper condition is generally under $400. And it is a piece

of American history.
ist

Sadly, many precision watches have been altered/vandalized. There was a

H

period during which these watches were not valued and they were given to
ht

children or used for practice by untrained workers.

rig

The most common damage was to the balance assembly. For whatever
reason, many came to believe that the balance must be statically poised
py

after they replaced a staff. They then went on to remove material from
Co

screws, often then needing to add timing washers to correct for removing
too much material.

Courtesy of www.historictimekeepers.com
4
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist

Vandalized balanced screws as recommended by certain authors

H

(The bottom left is now only a hollow thin shell; all these are solid gold)
ht
rig
py

Not only was this incorrect, as you will learn from this article, it was futile
Co

from the start. Given that the methods used to remove material were
based on files, drills, counter bores or the lathe, there was no way to
accurately measure how much material was removed; let alone no way to
determine the amount of material intended to be removed.

Courtesy of www.historictimekeepers.com
5
This damage often requires the watchmaker to find a donor movement with
a good balance and balance spring. Since Railroad and other high grade
watches have the serial number of the watch scribed on the underside of
the balance, it is important the balance wheel itself be “rescued.” The
screws and balance spring from the donor watch must be transferred to the
watch under restoration. Of course this means the donor watch has now

.
been lost.

nc
While returning a watch to original factory condition is time consuming and

,I
expensive, once done and adjusted, the cost of routine service drops

rs
dramatically as long as a competent watchmaker is used.

pe
This article is primarily for young (in experience) watchmakers and owners

ee
who want to understand how a watch is maintained as a precision
instrument. To use this information, the watchmaker must first master

ek
many skills including proper cleaning technique, lubrication, critical self
m
appraisal, escapement adjustment, basic repairs such as staff replacement
Ti
and pivot/jewel fitting, etc.
ic

Watch adjustment cannot overcome a poorly serviced watch.

or
ist

What is watch adjustment?

H

It is extremely important to understand watch adjustment cannot

overcome faults present in a watch. All faults must be corrected prior to
ht

any attempt at watch adjustment. This particularly applies to

rig

magnetization.
py

A watch must be in “perfect” condition, otherwise a watchmaker is making

random changes and chasing his/her tail.
Co

There are several components to watch adjustment. All of these

components lead to the final condition in which the watch performs as a
precision timekeeper. For a Hamilton Railroad grade watch, this means
that it will keep time to within 4 seconds per day in a consistent manner.

Courtesy of www.historictimekeepers.com
6
This standard actually exceeds the factory specifications for a Rolex,
which uses an alloy balance assembly and has automatic winding to
maintain even power to the balance assembly.

From here on in it is assumed the watch has been properly cleaned and
lubricated. The details throughout this discussion highlight the need to
eliminate sources of variation due to physical differences. Banking pins

.
nc
must be straight and rigid so that the safety action of the dart and roller is
consistent even as the lever changes its position due to the required end

,I
shakes. The balance spring must be centered between the regulator pins

rs
regardless of spacing to alter isochronism. The regulator pins must be

pe
straight and rigid. Pivots must be a close fit to their hole jewels and the
pivots truly cylindrical.

ee
ek
Temperature adjustment vs. Positional adjustment. For better or worse,
m
the distribution of the mass around the bimetallic balance wheel is used to
Ti
adjust both temperature and positional rates. Both cannot be perfectly
accomplished at the same time. The way the piece will be used must
ic

determine which adjustment is most important.

or

A pocket watch can assume any position but is kept at a fairly constant
ist

temperature in the pocket. A marine chronometer is generally used in one

H

position but can be exposed to a wide range of temperature from below

freezing to equatorial.
ht

So position takes preeminence for pocket watches, temperature for marine

rig

chronometers.
py

Permanent adjustments designed into the watch. These include the major
Co

factors that define the watch’s performance in dial up, dial down and nearly
wound and nearly run down conditions. They include the pivot sizes, hole
jewels and the collet location for the balance spring.

Courtesy of www.historictimekeepers.com
7
The pivot sizes must be identical for the watch to keep the same rate
regardless of which pivot is supporting the balance assembly. Pivot size
also impacts how similar the watch rate will be when the watch is
suspended by both pivots (movement in a vertical position) as compared to
when it is resting on a pivot end (horizontal positions).

The pivots, balance spring, balance and collet location and mainspring all

.
nc
determine how equal the watch rates will be when nearly fully wound and
after 24 hours.

,I
rs
This is called isochronism. Isochronism means that the balance will

pe
complete its rotation in the same amount of time at high amplitudes (full

ee
power) and at low amplitudes (low power). Amplitude is how watchmakers
measure the extent to which the balance rotates. If it rotates 270 degrees,

ek
it is said to be at high amplitude (which is standard for a freshly serviced
m
watch) and 180 degrees is considered low amplitude.
Ti
We define amplitude as the angular displacement of the balance in one
ic

swing of the balance. The oscillation of the balance covers two swings,
one forward and one return. So a balance with an amplitude of 270
or

degrees has a total oscillation of 540 degrees. This is often a point of

ist

confusion.
H

The importance here is that there is very little a watchmaker can do to

improve isochronism. As will be seen later, a knowledgeable watchmaker
ht

can adjust the regulator pin spacing to improve isochronism. But, there
rig

has been much misinformation about how to use regulator pins for any
number of things. Their only functions are to adjust the effective length of
py

the balance spring and isochronism.

Co

The escapement. The escapement includes the escape wheel (EW), pallet
jewels (entry and exit), impulse jewel and banking pins; plus the impulse
roller and safety roller mounted on the balance.

Courtesy of www.historictimekeepers.com
8
The lever escapement is the most widely used for watches because the
balance (oscillator) is free from the force of the mainspring for all but a
small percentage of its rotation. This makes for a stable oscillator.

The escapement actually performs three independent functions. These are

to power the oscillator (balance assembly), to release the train so the
hands can advance on regular intervals, and to ensure the escapement is

.
nc
not deranged by normal shocks (safety).

,I
rs
Powering the Oscillator (Impulse). The most obvious function is to supply

pe
power to the oscillator (the balance assembly) through the escape wheel
tooth, transmitting lift to the pallet jewel which then forces the lever over.

ee
At the end of the lever is a notch (impulse slot) which transmits power to

ek
the impulse jewel mounted on the balance wheel.
m
The lever is moved from side to side by the impulse jewel on the balance.
As the impulse jewel enters the impulse slot it contacts the far side of the
Ti
impulse notch unlocking the pallet jewel, which is locking an EW tooth.
ic

The pallet jewel then receives power from the EW tooth face which
or

accelerates the lever and the impulse notch. The lever moves so fast that
ist

the other side of the notch catches up with the back side of the impulse
jewel, delivering power to the keep the balance in motion. It is true
H

impulse as defined by physicists.

ht

This is how external power is supplied to maintain the oscillator.

rig
py
Co

Courtesy of www.historictimekeepers.com
9
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic

Lock A
or
ist

The Escape Wheel tooth is locked on the EXIT pallet

H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
10
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H

Unlock
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
11
 .
nc
,I
rs
pe
ee
ek
m
Impulse
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
12
 .
nc
,I
rs
pe
ee
ek
m
Lock B
Ti

The Escape Wheel is locked on the ENTRY Pallet

ic

The above photos show the escape wheel, pallet fork assembly and roller
or

in their approximate locations. In these 4 photos, the single roller table is

ist

upside down in relation to its orientation in the watch so you can see the
H

impulse jewel.
ht
rig

The ENTRY pallet is the where the EW tooth locks on the outside face of
the jewel. The EW tooth locks on the inside face of the EXIT pallet jewel.
py
Co

In Lock A, the escape wheel tooth has just dropped on the locking corner
of the EXIT pallet jewel (drop lock). The impulse jewel has cleared the
notch on the pallet fork.

Courtesy of www.historictimekeepers.com
13
In Unlock, the balance is on its return swing. The impulse jewel is serving
its second purpose by hitting the far side of the notch and moving the lever
so the EW tooth can unlock from the pallet jewel.

In Impulse, the impulse plane of the EW tooth is delivering power to the

impulse face of the pallet jewel. If you look carefully, you can see the far

.
nc
side of the pallet notch is now delivering that power to the impulse jewel

,I
on the side opposite to the one that just enabled unlock.

rs
pe
In Lock B, you can see the ENTRY pallet jewel is now locking an EW tooth
and the sequence will start again but in the other direction.

ee
ek
These are the sounds heard by your timing machine.
m
Ti

Counting the seconds. The second function of the escapement is the

ic

release of the gear train which advances the second, hour and minute
or

hand. Each time the escape wheel provides power to the balance
ist

assembly, a second pallet jewel blocks (locks) another tooth of the escape
wheel. Each time this occurs, the second hand advances ⅕ second on an
H

American RR watch.
ht

This time interval is determined by the balance assembly (oscillator). It is

rig

a function of the balance mass, balance diameter and the strength of the
balance spring. The ⅕ second period of the American RR watch is termed
py

18,000 beats per hour (5 beats per second X 60 seconds X 60 minutes).

Co

Marine chronometers have a beat of 14,400 beats per hour (2 beats per
second) and many modern watches have beats up to 36,000 beats per hour.
As an aside, a positional error of 5 seconds in 24 hours results in a
calibration error of 1 part in 86,400.

Courtesy of www.historictimekeepers.com
14
Safety Function. There is a third, less obvious function of the escapement.
This is to ensure the lever does not cross to the wrong side of the impulse
jewel (over-banking) which would stop the watch. This is called the safety
function and there are two types.

The single roller uses the impulse roller into which the impulse jewel is
mounted and a vertical pin in the center of the pallet. If the watch is

.
nc
knocked while the lever is at rest (locked), the pin will hit the roller and
prevent the watch from over-banking.

,I
The disadvantage is that the impulse jewel roller is large and there is

rs
significant interference in timing when the safety action of the single roller

pe
is activated by a shock.

ee
The double roller minimizes this. A second smaller roller (safety) is

ek
mounted below the impulse roller. A small dart is added to the lever which
extends past the lever fork. The dart still hits the safety roller, but since
m
the roller is smaller the impact on timing is greatly reduced.
Ti
Both designs are equally effective at preventing over-banking. If a single
ic

roller escapement over-banks there is a fault which must be corrected.

or

The black dots in the next two photos represent the banking pins and the
ist

lever has been lifted off by a shock. They represent the safety action of
the two escapements.
H
ht

In the Single Roller Safety Action photo, you can see the lever lifted off the
rig

banking pin and how the vertical dart now hits against the impulse jewel
py

roller. This is why these rollers have to be clean, polished and round (not
stretched or cracked).
Co

Courtesy of www.historictimekeepers.com
15
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht

Single Roller Safety Action

rig

In the Double Roller Safety Action photo, you see the advantage of the
double roller. It is smaller in diameter and offers less of a lever arm for the
py

resistance caused when the dart contacts the roller. These small safety
Co

rollers can be cracked very easily by forcing them onto over-sized staffs
(not unusual). This deforms the roller creating unpredictable positional
adjustment challenges. Measure the original and the replacement staffs
and correct if the replacement staff is over-sized!

Courtesy of www.historictimekeepers.com
16
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py

Double Roller Safety Action

Co

Courtesy of www.historictimekeepers.com
17
The two photos below clearly show how the impulse roller in the single
roller escapement is used to provide safety while the double roller
escapement employs a second, smaller roller beneath the impulse roller.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H

Single Roller Escapement

ht
rig
py
Co

Courtesy of www.historictimekeepers.com
18
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or

Double Roller Escapement

ist
H

Adjusting locks requires a means of measurement. Many of the methods

described by American authors suggest locks can be adjusted by eye. This
ht

is patently untrue. How is the pallet stone returned to the original position
rig

if it was the wrong move? How do you reduce or increase a move by a

precise amount?
py
Co

Measurement. The Swiss for years have used a tool called an escapement
meter with a scale marked in hundredths. No one knows what the scale
really measures since the geometry and size of pallet assemblies varies
widely. But it does not matter. For THAT pallet assembly, the scale will

Courtesy of www.historictimekeepers.com
19
indicate a unit increase or decrease and after trial the pallet can be
returned to the tool, adjusted to the last scale indication and altered from
there. The math minded will know this as an interval scale (has direction
but “4” is not twice as big as “2,” like a thermometer vs. a ruler).

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
20
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Bergeon Tool

Courtesy of www.historictimekeepers.com
21
These tools are quite expensive and may be why American authors never
mentioned them.

But, there is a way to measure with a microscope equipped with a reticle. You
can measure the jewel exposure from one of the pallet frame corners. You do
need to make a clamping plate but it works. I developed it in response to a
readers who asked about an alternative.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

View through eyepiece showing reticle scale against the pallet jewel

Courtesy of www.historictimekeepers.com
22
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Plan for Precision Pallet Adjustment showing plate and heat

resistant pad under scope

Courtesy of www.historictimekeepers.com
23
While this is not the place for details on normal tasks, there is no need to buy a
special Swiss heater for adjusting locks. Simply placing the tool or clamp plate on
an adjustable hotplate does the same thing.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Adjustable hot plate

Courtesy of www.historictimekeepers.com
24
In my opinion, every watchmaker should have a good stereozoom microscope and
hot plate. Alcohol lamps are simply unneccesary. Bluing can be done on a lab
grade hot plate much more consistently.

Natural errors of the lever escapement. There are several aspects of

physics which must be understood at this point concerning forces applied
to a rotating body such as the rotating balance assembly. These forces

.
nc
can be resistive or additive.

,I
Again, the neutral point refers to when the impulse jewel is centered in the

rs
impulse notch of the pallet.

pe
A negative force applied before the neutral point will cause a loss. Such a
force is the resistance to unlocking.

ee
A positive force applied before the neutral point will be additive, but

ek
becomes negative after the neutral point is passed. Both happen during
m
impulse.
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
25
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Natural Escapement Errors from Watch Adjustment, Jendritzki, 1963

Coutesy of http://www.booksimonin.ch

Courtesy of www.historictimekeepers.com
26
The net effect of these is to produce a loss due to escapement error.

The corollaries are:

The further from the neutral point, the smaller the impact.

The smaller the amplitude, the greater the effects.

Even the resistance of the balance spring must be taken into

.
nc
account.

,I
rs
These result in several implications. First, to reduce the impact of natural

pe
errors, the watch should run with high amplitude as long as possible. Most
of the escapement error occurs while the escapement is in direct contact

ee
with the balance. In Hamiltons this is 48 degrees of the balance rotation.

ek
In Swiss watches it is 52 degrees. This is the lift angle in our charts.
m
Forty eight of 180 degrees is a much greater percentage than 48 degrees
of 270 degrees.
Ti

Secondly, it is why isochronism in the “ideal state,” where external

ic

influences are ignored, is insufficient. It must be adjusted in real

or

conditions which include the external forces required to maintain motion

ist

and those that result from the natural escapement errors. It requires trial
and error to find the best location for the pinning point of the collet and
H

even the type of terminal curve at that location for a new design.
ht

This is one of the many reasons I am skeptical about all those claims to “fit
rig

a new balance spring.” Without the factory patterns, how can one person
fit springs to a wide variety of watches, especially since it takes weeks to
py

master one spring design when supplied with the pattern!

Co

The resistance forces must be kept to a minimum. Increased depth of lock

results in increased resistance; draw too strong does the same. These are
both effects of “opening banking pins.” Banking pins (if adjustable) should

Courtesy of www.historictimekeepers.com
27
be opened only to allow minimum total lock (drop lock plus slide) required
for the safety action. Modern watches eliminate adjustable banking pins.

A less obvious impact comes from the balance spring. If the balance
spring is still winding when it should be unwinding (beat error) after the
balance has passed the neutral point, it will impact timing. The balance
spring must be fully relaxed when the balance is at its neutral point.

.
nc
This is adjusted using beat error. Beat error is the difference in the time

,I
interval it takes to deliver impulse in from one balance swing and the
impulse in the return swing from the other. It is why a beat error under 1

rs
millisecond is emphasized.

pe
The strength of force by which the pallet jewel is held against the escape

ee
wheel tooth (draw), the depth of the pallet jewel on the escape wheel tooth

ek
(lock) and the extent to which the pallet must be moved to disengage the
escape wheel tooth (slide), all have significant impact on the natural
m
errors.
Ti
ic
or

Adjusting the escapement includes beat error, lock, slide and draw . The
interfering forces applied by these conditions must be minimal.
ist

Beat error is initially adjusted with the mainspring completely let down.
H

The balance spring collet is rotated on the staff until the impulse jewel is
ht

centered in the pallet notch when the balance is at rest.

rig

However, visually setting the beat error is not precise. This is due to the
fact that the balance rotates more in one direction than the other
py

depending on whether the balance spring is winding or unwinding. Final

Co

adjustment must be made with an electronic instrument. A beat error of

under 1 millisecond has no measurable impact on timing.

The locks must be as light (shallow) as possible consistent with safe

locking on all escape wheel teeth. The draw must be just strong enough.

Courtesy of www.historictimekeepers.com
28
A watchmaker can only alter this to a small degree. The slide must be
minimal as is consistent with proper functioning of the safety roller.

Like altering balance screws, these adjustments cannot be properly

executed with guess work. Watchmakers who apply heated needles to the
pallet stone cement so they can “move it a little” are wasting time. An
accurate way to measure the alterations (and reverse them) is needed.

.
nc
The banking pins must be as closed as is safe. Wide banking pins are a

,I
sign of a mechanical fault that precludes watch adjustment.

rs
All of these things are permanent until someone consciously alters them.

pe
In a watch with a history of good and proper service, these will not need to
be touched. But vintage watches often have not been properly serviced.

ee
ek
Adjusting to Position. m
After the watch has been serviced it should be “perfect.” Pivots fit their
Ti
jewels correctly, escape wheel acts on the center of the pallet jewel faces,
pallet jewels are level and well secured, proper end shakes, ensure all
ic

parts are centered and cannot move enough to foul on another part,
or

balance spring is level and true, same with balance wheel and the watch is
ist

properly lubricated. The watch is fully demagnetized.

H

Adjusting to position ensures the watch will keep a consistent rate when
ht

worn. If a watch has a rate difference of 6 seconds among the horizontal

rig

and vertical positions, it cannot develop a rate error that exceeds 6

seconds per day when worn.
py

Watches regulated to be 15 seconds fast per day in use, but which can be
Co

made to show an error of 2 seconds after spending the night dial down, are
not adjusted! The owner is trying to compensate for a fault in the watch.

Adjusting the regulator so the watch splits its timing problems is also not
adjusting. This is exemplified by setting the regulator and /or regulator

Courtesy of www.historictimekeepers.com
29
pins so the watch is 15 seconds fast in the horizontal and 20 seconds slow
in pendant up.

An adjusted watch will show a proportional error throughout the 24 hours.

That is, if it has an error rate of plus 4 seconds per day, it will show a dial
reading error of plus 2 seconds 12 hours after being wound and plus 4
seconds 24 hours after being wound. Its dial error will be proportional to

.
nc
the time since it was last compared to a standard (like GPS time).

,I
Isochronism is checked by timing the watch at low power (180 degrees

rs
amplitude) and at high power (250 to 290 degrees amplitude) in one

pe
horizontal position (dial down).

ee
ek
A precondition is that the balance spring be centered between the
regulator pins when it is at rest and in beat (beat error under 1
m
millisecond.)
Ti
ic

There is little the watchmaker can do to alter this. However, if the watch
or

uses regulator pins, the pin spacing can be altered to bring the rates
between high and low amplitude as close together as possible while
ist

keeping the balance spring centered between them.

H
ht

Horizontal Positions. Check the dial up and dial down rates. They should
be identical. If not, find the fault now. They are a function of movement
rig

prep, including pivot condition and lubrication. Check for fouling of parts.
py

If the watch was in “perfect” condition, there should not be a problem. If

there is a difference, there is a problem.
Co

Courtesy of www.historictimekeepers.com
30
Vertical position adjustment. This is where the trouble starts. It takes
time and discipline to perform this. It requires a good electronic timing
instrument.

The traditional description of the 4 vertical positions are pendant up,

pendant right, pendant down and pendant left.

The procedure discussed here discards the pendant as a reference point in

.
nc
favor of the escape wheel. This makes it easier to determine where the

,I
adjustment should be made.

rs
This description will use a post 1920 Hamilton railroad grade pocket watch.

pe
This is because these watches have 4 adjustment screws (timing screws)
referred to as ¼ screws. The method here is extremely useful on high

ee
grade balance assemblies without timing screws; but those require the use

ek
of timing washers.
m
Ti
NOTES OF CAUTION:
ic

Never remove material from the balance assembly. Regardless of what has
or

been written, this is nothing short of vandalism. If the screws were of the
proper mass when the watch left the factory, by what mechanism did they
ist

acquire MORE mass? If adjustments are needed to a balance without

H

timing screws, add timing washers.

ht

Check the vertical rates before doing anything else. You are doing this now
rig

anyway; but this is because you will never static poise (poising tool) a
py

balance again. For an after-sales watchmaker, the poising tool amounts to

a parlor trick. Cool to do but pointless.
Co

Static poise only has value when making a completely new balance
assembly (balance and spring). It provides a starting point. But after the
watch is adjusted to positions, the balance will by definition be OUT of
static poise. It left the factory out of static poise! The old books make it

Courtesy of www.historictimekeepers.com
31
appear static poise is an end to itself. The end to be reached is
equalization of rates in the vertical positions: dynamic poise.

Dynamic poise is to watches what dynamic wheel balancing is to

automotive mechanics. It is not enough to assume a statically poised
balance will result in a poised balance assembly while in motion.

In the watch, there are various factors that influence dynamic poise. The

.
nc
easiest to understand is that the balance spring is always in motion. Static

,I
poise is undertaken without the spring mounted. Dynamic poising takes
the spring, natural escapement errors and the whole system into account.

rs
pe
Dynamic poising is in fact what was done by factory watch adjusters . It

ee
just took longer. At the factory, railroad watches were tested for 24 hours

ek
in each vertical position. The adjustment was determined and executed
m
and then another week of testing. It is known this process could take 3
months.
Ti

It is not unusual for me to need 6 trials with an electronic instrument (I

ic

very much favor the Vibrograf MU-700) and there are times it takes 15
or

trials to zero in the adjustment. Where each trial takes me 12 to 15

ist

minutes, in the early 20th century this translated into 6 to 15 weeks.

H

And yes, this means final adjustment can take up to 4 hours. But, it is like
zeroing in a rifle sight. Each successive grouping gets tighter.
ht

We are aiming for the original factory performance. For the Hamilton, this
rig

means no more than 5 seconds difference across the 2 horizontal and 4

py

vertical positions. Other companies had their own standards.

Co

Courtesy of www.historictimekeepers.com
32
Procedure for Equalizing Rates via Dynamic Poising. Research has shown
the interaction between vertical position and amplitude. For example, all
vertical rates are magnified at low amplitude (under 180 degrees) and
conversely high amplitudes minimize those positional rate differences.
Also, 220 degrees of amplitude is a “magic” value where vertical rates
have little to no differences. For these reasons, see the diagram below.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
33
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig

Interaction between Position and Amplitude on Rate

py

From Watch Adjustment, Jendritzki, 1963

Co

Courtesy http://www.booksimonin.ch

Courtesy of www.historictimekeepers.com
34
Two things to understand. Isochronism is not to be confused with testing
the vertical positions at low amplitudes. The watch must be wound to
produce 180 degrees amplitude in the vertical positions. A correctly
prepared watch will lose up to 40 degrees of amplitude between horizontal
and vertical. Also the amplitudes in the various vertical positions should

.
nc
be within 10 degrees of each other (ideally the same).

,I
Mount the watch vertically in the microphone with the balance facing you.
Set the lift angle for your watch on the instrument. Hamiltons are 48

rs
degrees; most Swiss are 52. Lift angle is the portion of the balance

pe
rotation during which the escapement is in direct contact with the balance.

ee
Wind the watch slowly until the instrument reports the amplitude is 180

ek
degrees. Now, verify the amplitude in the 4 vertical positions and the 4
positions intermediate to those. We will be using these 8 vertical positions
m
from now on.
Ti
If the amplitudes vary by more than 10 or 15 degrees, find the fault and
ic

correct. You can go no further because you will be unable to separate rate
or

differences due to dynamic poise from those due to amplitude.

ist

This next part becomes confusing for most. The neutral position (of the
balance) is the impulse jewel centered in the lever fork. The escape wheel
H

will be the positional reference, not the crown. The heavy point of the
ht

balance is located at the escape wheel position that produces the fastest
rig

rate. When the microphone is returned to the position that produces the
fastest rate, hold the balance in the neutral position. The heavy screw will
py

be located plumb (vertical line straight down) from the balance staff.
Co

This will become clear as you proceed.

You will also need plenty of paper. Make an X-Y graph. Now divide each of
those quadrants in half so you have 8 lines intersecting in the center. The

Courtesy of www.historictimekeepers.com
35
rates in each of the escape wheel (EW) positions will be recorded on this
chart.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
36
 Typical Vertical Position Rate Observation Diagram

Mount the watch on the microphone vertically so that the EW is directly

below the balance. I rotate the microphone counter clockwise so the
second position will have the EW 45 to the right of and below the balance.
The third position places the EW directly to the right of the balance. In the
fourth position the EW will be to the right and above the balance, and the

.
nc
fifth position will place the EW directly above the balance.

,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
37
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Escape wheel directly below balance

Courtesy of www.historictimekeepers.com
38
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Escape wheel rotated 45 degrees counter clockwise

Courtesy of www.historictimekeepers.com
39
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Escape wheel rotated 90 degrees counter clockwise

Courtesy of www.historictimekeepers.com
40
It likely will look “wrong” until you can get used to it. Just remember the
focus is on the EW.

.
nc
Note the position of the stopped balance in each of the above photos. It is
stopped with the impulse jewel centered in the pallet fork. THIS is the

,I
neutral position.

rs
pe
At 180 degrees amplitude, the heavy point on the balance will correspond

ee
to the microphone position at which the fastest rate was observed. Put the

ek
microphone back into that position with the balance stopped at the neutral
point. The heavy spot will be at the bottom most screw.
m
Ti
Record the rates of the watch in each position on the chart. They should
ic

form a logical progression and pattern. Slowest rates opposite fastest

or

rates; a progression from fast to slow and back to fast.

ist

Wait at least 60 seconds for the watch to settle down in each position
before recording the rate.
H
ht

If the pattern does not make sense (20 fast followed by 10 slow followed by
rig

15 fast) check the amplitudes in each position. If it varies, there is a fault.

If the watch passed to this point, verify banking pins, magnetism, draw,
py

safety action.
Co

If the amplitude does not vary, then verify the balance spring and the
regulator pins.

Assuming the watch is perfect, the heavy spot will be identified using the
fastest rate. Put the microphone mounted watch in the position that

Courtesy of www.historictimekeepers.com
41
produced the fastest rate. Hold the balance in the neutral position. The
heavy spot (screw) will be at the point on the balance directly below the
balance staff.

I mark that point with red ink so that I know where I made the alteration.

If this points directly to a timing screw, terrific. You can either move that
screw in (¼ turn per trial) or the opposite screw out. If it falls between two

.
nc
screws, move the two screws on either side ½ of the adjustment.

,I
rs
If the watch does not use timing screws, add known timing washers to the

pe
screw opposite the heavy point.

ee
Now do the testing again. The overall rate will have changed. But the
differences should be smaller. Repeat until you zero in on your accepted
performance.
ek
m
When done, test the watch again in all horizontal and vertical positions to
Ti
ensure success.
ic

Below are the charts used to zero in the rate of a Hamilton 950B.
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
42
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

First of Several Trials

Courtesy of www.historictimekeepers.com
43
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

More Trials

Courtesy of www.historictimekeepers.com
44
 .
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Final Trials and Positional Rates Achieved

Courtesy of www.historictimekeepers.com
45
It is also very useful to perform amplitude and positional tests on a watch
BEFORE service. This provides you with a way to assess the improvement
you provided as well as offer hints at what you may need to correct.

It should now be clear that even the most seemingly insignificant thing can
have major impacts. It should now be apparent why the impulse jewel

.
nc
must be straight and vertical in the roller table. Any slant in the jewel

,I
changes its effect as a new position is taken due to endshake. Likewise, it
should be clear that the impulse jewel size is not simply an issue of “what

rs
will fit.” Its size was chosen in the design to ensure unlocking and impulse

pe
were delivered at the optimum moments.

ee
The timing machine is your task master. It does not close its eyes. It is an

ek
objective witness to your work.
m
At first it is extremely frustrating and annoying. But, if you accept what
the timing instrument is telling you and troubleshoot each bad report, your
Ti
technique will improve, making you a much better watchmaker.
ic

As a historical note, dynamic poising was known to the top American

or

watchmakers as early as 1944. Samelius and Purdom highlight the

ist

importance of understanding the difference between static and dynamic

poise in their “It’s Timing that Counts.” It is clear however, their
H

understanding on how to use it is tentative and they fail to describe a

ht

procedure.
rig

Samelius (perhaps the top U.S. instructor of the period) and Purdom go into
detail about what makes a “perfect watch” ready for precision adjustment.
py

One of their quotes is “Repairing a watch is one thing; rating it is

Co

something else.”

This book also provides an example of the promotion of a theory by experts

that was subsequently discarded. Samelius and Purdom go into great

Courtesy of www.historictimekeepers.com
46
detail describing something called “needling the balance spring” that
involved stroking the balance spring with a steel needle. It sounds like
they are manipulating the magnetism of the spring.

It is interesting in their preparation of the watch for adjustment that they

ignore the importance of demagnetization.

This is not throwing rocks at two of the giants in the field. It is a

.
nc
cautionary tale for those trying to rely on older, traditional sources of

,I
information. Promoted theories that are discredited are very rarely directly
retracted; they just kind of disappear from further discussion.

rs
pe
This material presented here was drawn from the direct instruction of the
excellent teachers of WOSTEP and two books. One is the owner’s manual

ee
for the Greiner Micromat Timer published in the late 1970s. Section 2 of

ek
that manual goes into many of the factors discussed and provides a
procedure for dynamic poising. Copies of this manual are very rare
m
however.
Ti
Thanks to the generosity of Greiner Vibrograf AG, and particularly Mr.
ic

Michael Klaefiger, I have permission to provide a PDF of this manual via my

or

website documents area. You can access it at:

ist

http://www.historictimekeepers.com/documents/Micromat.pdf
H

Watch Adjustment by H. Jendritzki is seen more often. This book goes in

and out of print and can be difficult to find. It may be available through the
ht

current publisher http://www.booksimonin.ch/. Mr. Simonin is the founder of

rig

WOSTEP.

We thank Mr. Simonin for providing the PDFs of the pages from Watch
py

Adjustment for inclusion in this article.

Co

This article was completed only with the encouragement and help of a
number of people. First of whom is my wife Mary who had the unenviable
task of editing.

Courtesy of www.historictimekeepers.com
47
I want to thank the reviewers who volunteered to ensure accuracy and
readability. These include Karl Mansson, Lukas Hedman, Graham Morse,
Doug Smith, Robert Kunin, Dushan Grujich and Clint Geller. I especially
want to thank Graham for using his expertise in image
reproduction for ensuring the photos were adjusted correctly.

.
nc
,I
rs
pe
ee
ek
m
Ti
ic
or
ist
H
ht
rig
py
Co

Courtesy of www.historictimekeepers.com
48