The Art of Being

Drunk
Second Edition

Richard Watkins

© Copyright 2017, 2018

 The Art of Being Drunk

On Planetary Gears
Planetary gears consist of three wheels and a carrier, Figure 11:
(a) Sun gear s with Ns teeth (blue).
(b) Planet gear p with Np teeth (green).
(c) Annulus gear a with Na teeth (red).
(d) Carrier c (yellow).
The sun, annulus and carrier rotate on three concentric arbors,
and the planet gear rotates on an arbor mounted on the carrier.
To derive the formula for planetary gear motion we shall look at
the rotation Tp of the planet gear p in terms of the rotation of the
Figure 1
other three components Ts, Tc and Ta.2
First consider how p rotates relative to the rotation of s and c. Obviously p is rotated by s and c and
so Tp must be the sum of these influences. Considering the rotation of s by itself:
Tp = - (Ns/Np) Ts
which is negative because s is rotating p in the opposite direction.
Now, ignoring the motion of s, p has to rotate in the same direction as c rotates around s, because
of its teeth meshing with s (which it can do by turning a). A single turn of c causes p to pirouette
once around s and so Tp = (Ns/Np) Tc
Therefore, the actual rotation of p must be the sum of these two influences and:
Tp = (Ns/Np) Tc - (Ns/Np) Ts
Or:
NpTp = NsTc - NsTs
Note that the annulus a is irrelevant because it is being driven by, and so following, the other
movements.
Now consider the rotation of p relative to c and a ignoring s.
Considering the rotation of a:
Tp = (Na/Np) Ta or NpTp = NaTa
because the internal teeth of a rotate p in the same direction.
Also, p has to rotate in the opposite direction to c as c rotates around a, because of its teeth meshing
with a. A single turn of c causes p to pirouette once around a and so:
Tp = - (Na/Np) Tc or NpTp = - NaTc
Again adding the two together, the actual rotation P is:
NpTp = NaTa - NaTc
Now p can be eliminated by combining the two motions of p:
NsTc - NsTs = NaTa - NaTc
In other words:
NaTa + NsTs = (Na + Ns) Tc
1 This and some other diagrams were created using http://geargenerator.com
2 This proof is taken from Meditations on Breguet and Mathematics (available from http://www.watkinsr.id.au/
Breguet.html).
2
 The Art of Being Drunk
Note that the planet gear does not appear in this formula.
If all three gears are in the same plane then, Figure 2, the radii
are Ra = Rs + 2Rp. As their meshing teeth must be the same size,
the numbers of their teeth are related by: Na = Ns + 2Np and the
formula for the planetary gears becomes:
(2Np + Ns)Ta + NsTs = (2Np + 2Ns) Tc
The unconstrained motion of planetary gears is not interesting,
because two of s, c and a can be rotated arbitrarily and the third
component will rotate to suit. However, imposing a constraint
produces two useful results:
Figure 2
(a) Lock one of s, c, and a so that it cannot turn. For example,
if the annulus is locked then Ta = 0 and NsTs = (Na + Ns) Tc. The most interesting case is
when the carrier is locked. Then Tc = 0 and NaTa = - NsTs. That is, the sun gear and the
annulus rotate in opposite directions, as indicated by the minus sign.
(b) Force two of s, c and a to turn at the same rate. For example, if a and s rotate together then
Ta = Ts and NaTs + NsTs = (Na + Ns) Tc. That is, Ta = Ts = Tc and all three gears rotate
together.
Another feature that might be useful is that locking the planet gear so that, in the two formulae for
Tp above, Tp = 0, and this also results in Ta = Ts = Tc.

On Clicks, Drivers and Switches

In a normal watch with a fusee:
(a) When the watch runs, the mainspring rotates
the fusee and the first wheel 1 clockwise,
Figure 3. (This view is from underneath the
dial, and so the center (second) wheel and
the hands rotate anti-clockwise from this
perspective.)
(b) To wind the mainspring the fusee has to be
turned anti-clockwise (using the square on
its arbor) without rotating the first wheel.
This requires the addition of a ratchet and Figure 3
click, Figure 4.
From left to right, the two views are:
(a) The ratchet R attached to the fusee cone, viewed from underneath.
(b) The first wheel 1 with the click 1' attached to it viewed from above.

Figure 4
3
 The Art of Being Drunk
Figure 5 is a view of the click, ratchet and first wheel from the top
through a transparent fusee. All the diagrams in this article are from
this perspective.
When the fusee cone rotates anti-clockwise to wind the mainspring,
the ratchet R turns with it and the click 1' rides over the ratchet
teeth; the first wheel 1 does not move. When the mainspring rotates
the fusee clockwise to run the watch, the ratchet R rotates clockwise
with it and the click 1' forces the first wheel to rotate with it, driving
the watch train.
Figure 5
In this design there are two concentric wheels R and I connected by a
click, and four different movements are possible.
First, either of the wheels can be the driver; that is, power is provided to that wheel. In the above
example the fusee is always the driver.
Second, the driver can rotate clockwise or anti-clockwise. In Figure 5, the red arrow represents the
movement of the first wheel 1 when it is the driver, and the yellow arrow is the movement of the
other wheel R when it is the driver. The four possible movements of the wheels are:
(a) R is the driver rotating clockwise and it drives 1 clockwise through the click. (This is normal
running in a watch).
(a) R is the driver rotating anti-clockwise, so the click does not act in the ratchet teeth and 1
is motionless. (This is normal winding in a watch).
(c) 1 is the driver rotating clockwise, so the click does not act in the ratchet teeth and R is
motionless.
(d) 1 is the driver rotating anti-clockwise and it drives R anti-clockwise through the click.
This fusee does not provide maintaining power. That is, while the fusee cone and R are turning
anti-clockwise no force is provided to the first wheel and the watch train. Without maintaining
power the watch could stop while it is being wound. This is not much of a problem with old
key-wound watches because winding only takes a few seconds and old watches were usually not
very accurate; so losing a few seconds did not matter. But it could be a serious problem with very
accurate timepieces, like clocks with fusees.
It is important to note the asymmetry, and the choice of the driver
and the direction of the click are very important.
Figure 6 is the same as Figure 5, but the ratchet and click have
been reversed. As a result, the rotations have been reversed; R drives
1 when it rotates anti-clockwise and 1 drives R when it rotates
clockwise.
A useful way to look at clicks is that they act as switches, turning
rotation on and off.
Indeed, often it is important that a wheel does not rotate, as when Figure 6
winding a watch, and clicks can be used for this purpose.
In the context of planetary gears, clicks as switches can be used to lock a wheel’s rotation or to made
two wheels rotate together.

4
 The Art of Being Drunk

On Being Drunk
A drunken fusee is a fusee designed so that the mainspring is wound when the fusee arbor is rotated both
clockwise and anti-clockwise.
Presumably it was so named because a drunk person could try to wind his watch by turning the key
the wrong way, clockwise, whereas a sober person would not do this. So an alcoholic could buy a
drunken watch with a drunken fusee in it, and not risk damaging it by forcing the key the wrong
way.
A drunken watch is impossible with an ordinary fusee where the fusee arbor and the fusee cone are
rigidly attached to each other. In this situation, Figure 5, R cannot wind the mainspring when it
rotates clockwise because the watch train 1 and the click 1' stop R from rotating. (Actually, during
running R rotates clockwise very slowly to provide power to the watch train.)
So the first point is that:
The fusee arbor cannot be rigidly attached to the fusee cone; the arbor must be free.
Separating the arbor and the cone means that there are now three concentric “wheels” (the arbor, the
fusee cone and the first wheel) and the problem of drunkenness changes significantly:
(a) When winding the mainspring the fusee arbor drives the fusee cone and the fusee cone does
not drive the first wheel.
(b) When the watch is running the fusee cone drives the first wheel.
In (b), with a key-wound watch, it does not matter if the arbor is or is not rotated by the cone,
because there is nothing to impede its movement when the watch is running.
Again we must separate the clockwise and anti-clockwise cases:
(a) During normal (anti-clockwise) winding, both the arbor and the cone rotate in the same
direction and a concentric wheel and click system should work provided it can connect the
three components.
(b) When the arbor rotates clockwise the cone must
rotate anti-clockwise and a concentric wheel and
click cannot achieve the reversal of motion.
Two wheels in a train do reverse motion, Figure
7. But then the arbor and the cone cannot be
concentric and two simple wheels like this cannot
be used! Figure 7

There is only one mechanism that makes it possible for the arbor to rotate the cone in the opposite
direction, planetary gears.
There are several requirements for these gears:
First, the planetary gears in the fusee must be arranged so that the fusee arbor is linked to the sun gear
and the fusee cone is linked to the annulus. Then if the arbor rotates clockwise the annulus rotates
anti-clockwise, providing the drunken winding. As a result, the carrier must be linked to the first wheel
and, during winding, Tc = 0, so NaTa = – NsTs.
(In principle, the fusee arbor could be linked to the annulus and the fusee cone to the sun gear, but
Figure 1 suggests that this would be very difficult because of the separation between the arbor and
the annulus.)
5
 The Art of Being Drunk
Second, the fusee arbor must be linked directly to the fusee cone to provide sober (normal) winding
when the arbor rotates anti-clockwise.
Planetary gears can be used by making two of the three components s, c and a rotate at the same
rate, and consequently the planetary gears are locked and rotate together. Any two gears will do, and
setting Ts = Tc, Ts = Ta or Tc = Ta will achieve the same result of Ts = Tc = Ta.
Forcing two gears to rotate together can be achieved by a click acting between them. The link must
be a click to allow the two pieces to rotate separately or together, depending on whether the watch is
being wound drunken or normally. That is, the click acts as a switch between the two modes. So this
switch must be a click acting between the sun s and the carrier c, or the annulus a and the carrier
c, or the sun s and the annulus a.
There is a fourth option of placing a click between the carrier c and the planet p. In Figure 1, as I
have noted, if the planet cannot rotate relative to the carrier then Ts = Tc = Ta.
However, when the sun rotates anti-clockwise Tc = Ts and the carrier will also rotate anti-clockwise,
which seems impossible because it is attached to the first wheel 1.
Third, during winding the fusee cone is being driven by the fusee arbor. To run the watch, the fusee
cone becomes the driver turning the first wheel through the carrier. Now the carrier attached to the
first wheel must turn clockwise.
Consequently, because the carrier must be able to rotate both clockwise and anti-clockwise it
cannot be mounted on the first wheel; that wheel can only rotate clockwise. So the carrier must be
a separate “wheel” loose on the fusee arbor.
However, although counter-intuitive, there is one fusee where the carrier and first wheel are a single
piece. So designs with the carrier and first wheel joined must be considered.
If the carrier is loose then there is no link between the planetary gears and the first wheel! So there
must be another click somewhere to link the first wheel and the fusee cone so that the watch will
run. And this click might make it possible for the carrier and the first wheel to be a single piece.
If the planetary gears are used for all three functions (two winding and one unwinding) then:
(a) Drunken winding: The planetary gears act normally.
(b) Sober winding: There is a switch that makes the planetary gears rotate anti-clockwise
together.
(c) Running: The fusee cone and annulus rotate clockwise, rotating the first wheel clockwise
That is, in principle a drunken fusee can be constructed using planetary gears and a switch.
From the above, there are eight possible designs:
Carrier and first wheel Switch Designer
Separate Sun and carrier Massotéau, ca 1740
Separate Annulus and carrier Virgo, ca 1740
Separate Sun and annulus Moore, ca 1729
Separate Planet and carrier Watkins
Joined Sun and annulus Tappy, ca 1750
Joined Sun and carrier
Joined Annulus and carrier
Joined Planet and carrier
The design attributed to me is a design I have not seen described elsewhere, but it may well have
been created by someone else before me.

6
 The Art of Being Drunk

On Switching the Sun and the Carrier,

Massotéau
In 1742 Massotéau de Saint Vincent
published the drunken fusee illustrated
in Figures 8 and 9.3 His design uses a
switch to make the sun and carrier
rotate together. This fusee has three
layers; from the bottom up they are:
layer 1: the first wheel 1; layer 2: the
carrier c; and layer 3: the sun s, planet p
and annulus a wheels.
As required, the annulus a is attached to
the fusee cone; in Figure 9 the annulus
is transparent to show the click 1'
underneath it. In this design the three Figure 8
functions are:
(a) Drunken, clockwise winding: When the fusee arbor and sun gear s rotate clockwise the
switch sw has no effect and the planetary gears behave normally. The sun gear causes the
planet gear p to rotate anti-clockwise and, because the teeth of p mesh with the annulus, p
attempts to pirouette around the sun and rotate the planet carrier c clockwise. But the click
1' locks the carrier, Tc = 0 and so NaTa = – NsTs
That is, the annular gear and the fusee cone rotate anti-clockwise winding the chain onto
the fusee.
In Figures 8 and 9, Na = 28 and Ns = 12, and so Ts = - 28/12 Ta. That is, it takes more than
two turns of the key to turn the fusee once.
(b) Sober, anti-clockwise winding: The
sun s rotates anti-clockwise and the
switch sw forces the carrier to rotate
with it. So Ts = Tc = Ta, and a rotates
anti-clockwise (winding the chain
onto the fusee) and the planet carrier
c rotates anti-clockwise. Now the click
1' has no effect and the first wheel is
free.
(c) Running: During winding s is
the driver and a is driven. Now a
becomes the driver, rotating clockwise
as the chain is pulled off the fusee. It
attempts to rotate p clockwise and
p in turn attempts to rotate s anti-
clockwise. But the switch sw locks the
carrier c to the sun s so that Ts = Tc =
Ta and the clockwise rotation of the
fusee cone is transmitted to the first Figure 9
3 Paris, 1742, Memoires pour l’Histoire des Sciences & des Beaux Arts, Volume XLII, Paris: Chaubert. (See also Origins
of Self-Winding Watches, http://www.watkinsr.id.au/Origins.html)
7
 The Art of Being Drunk
wheel through the click 1'. As the planetary gears are locked and rotate together the fusee
arbor s rotates with them and the planet carrier c rotates clockwise.
During drunken winding, for the planet gear p to turn anti-clockwise and rotate the annulus anti-
clockwise against the force of the mainspring, its teeth necessarily creates a clockwise force on the
carrier c against the click 1' and provides maintaining power. However, during normal winding the
click 1' slides over the teeth of 1 and there is no maintaining power.

On Switching the Annulus and the Carrier,

Virgo
A year earlier Thiout described the drunken fusee in Figures 10 and 11, and he stated that it was
designed by Virgo.4 Massotéau claimed to have invented this, but it is significantly different to his
design in Figures 8 and 9. In that design the switch sw acted on the sun s and the carrier c, but
Virgo’s design has the switch acting on the carrier c and the annulus a.
Again, to achieve the desired effect the carrier c must be loose (linked to the rest of the mechanism
by the click 1' and the switch sw), and the annular gear a is rigidly attached to the fusee cone; the
annulus is probably made as a separate piece and pinned to the fusee cone. Then:
(a) Drunken, clockwise winding: When the fusee arbor and sun gear s are rotated clockwise,
they causes the planet gear p to rotate anti-clockwise and simultaneously attempt to turn
the planet carrier c clockwise. But the click 1' locks the carrier and so Tc = 0 and NaTa = –
NsTs. That is, the annular gear and the fusee cone rotate anti-clockwise winding the chain
onto the fusee.
Using Thiout’s tooth counts, Na = 18 and Ns = 8, and so Ts = - 9/4 Ta. Again it takes more
than 2 turns of the key to turn the fusee once. (Note that in the illustration Na = 19 which
is incorrect; it should be 18.)
(b) Sober, anti-clockwise winding: When the fusee arbor s is rotated anti-clockwise, it causes
the planet gear p to rotate clockwise and simultaneously attempt to turn the planet carrier
c anti-clockwise. But the switch sw locks the carrier to the annular gear through its ratchet
and so Ts = Tc = Ta and the fusee cone rotates anti-clockwise to wind the watch and
the planet carrier c rotates anti-clockwise. As the planetary gears are locked, Ts = Ta and
winding is the same as in an ordinary fusee. Again the click 1' has no effect and the first
wheel is free.
(c) Running: When the watch runs the fusee cone, with the annulus a attached, is the
driver and rotates clockwise. The ratchet on a meshes with the switch sw and forces the
loose planet carrier c to rotate clockwise with a. So Ts = Tc = Ta and the carrier ratchet
meshes with the click 1' mounted on the first wheel 1 and so forces the first wheel to turn
clockwise. As the planetary gears are locked and rotate together the fusee arbor s rotates
with them and the planet carrier c rotates clockwise.
Like Massotéau’s design, maintaining power is only provided to the first wheel 1 during drunken
winding, when the carrier c attempts to turn clockwise and the click 1' presses against the carrier.
The similarity of Virgo’s mechanism, Figure 11, to Massotéau’s design, Figures 8 and 9, is not
obvious because of the different drawings. To make it clearer, Figure 12 is a different version of
4 Thiout, Antoine, 1741, Traité de l’Horlogerie Mécanique et Pratique, Paris: Charles Moette, Pere Prault, Hyppolite-
Louis Guerin, Pierre Clement, Pierre-Andre Debats, Louis Dupuis and Charles-Antoine Jombert, volume 2, page
383 and Plate 38. (See also Origins of Self-Winding Watches, http://www.watkinsr.id.au/Origins.html)
8
 The Art of Being Drunk

Figure 10

the drunken fusee used by Virgo in the same format as the diagrams of Massotéau’s fusee. Instead
of a separate ratchet on the annulus, the switch sw acts on the teeth of the annulus, simplifying
the design. Now it is clear that Virgo’s fusee is identical to Massotéau’s fusee except the switch has
moved; where it previously acted between the carrier and the sun, it now acts between the carrier
and the annulus.

Figure 11 Figure 12

9
 The Art of Being Drunk

On Switching the Sun and the Annulus

Not long after publishing this article, I found a description of another design, apparently created by
Thomas Moore.5 Britten’s description is:
Thos., Ipswich; ... “ Whereas Thos. Moore, clock and watchmaker in Ipswich, have for many
years observ’d the misfortunes which very frequently happen to pocket-watches of all sorts—viz.,
by sometimes coming into unskilful hands, &c., and often into the hands of servants (in absence
of the owners thereof ) they, endeavouring to wind up the watch, have turned the wrong way,
and by so doing, they have broke the work (and the like often happens when Juice of Grape
Predominate). Therefore this to give notice. That the above Thos. Moore have now made up
several curious Silver and Gold Watches (and will continue to do so) so curiously contriv’d, that
let the watch be wound up which way they please, either to right or left, they cannot fail of
winding up the watch, with more safety than if wound but one and the common way, and are
to be sold at very reasonable price. Any person or persons may have the freedom to seeing any
of the said watches at his house in Ipswich
aforesaid ” (Ipswich Newspaper, 1729). At
the Horological Institute is a fusee of one
of Moore’s watches, which was sent from
Syracuse, New York, by Mr. Geo. R. Wilkins;
the great wheel and larger end of fusee are
shown in the accompanying cut, and the
action will be understood by those conversant
with watch work; the device was described
by Thiout in 1741, some years after Moore’s
advertisement appeared. Figure 13
Figure 13 is the original cut. Figure 14 is the same as Figure 13, except that the view of the base of
the fusee has been changed to a view from above and labels have been added. Note that the annulus
a has ratchet teeth on the outside, as in the fusee designed by Virgo. In addition, there are four
strange features:
(a) The switch sw acts between the fusee base, and consequently the annulus a, and the sun
s. Part of this click is underneath the annulus and presumably the fusee cone has been
hollowed out to create space for it; a must be a separate piece pinned to the fusee cone.
As a consequence, it
appears that the sun is
below the level of the
annulus and the planet
p, which is impossible.
So the sun must be
about twice the height
of that in other designs
so that sw acts on the
bottom half and p acts
on the top half.
(b) The click 1' and the
area around it (green) is Figure 14
5 Britten, F.J., 1922, Old Clocks and Watches and Their Makers, London: E. & F.N. Spon, pages 744-745. (The
diagram does not appear in the 1894, 1956 and 1990 editions of this book, and perhaps it only appears in the 5th
edition.)
10
 The Art of Being Drunk
mysterious. If the green area is part of the base of the first wheel 1, then no delineation
is necessary. The green area cannot be part of the carrier c, because then 1' could not act
usefully; 1' and c would simply rotate together. Likewise, it cannot be a separate, loose
piece for the same reason. So it must be part of the first wheel.
(c) From its position, the extra click x must act between the first wheel 1 and the annulus a.
(d) The feature y is presumably a bridge, mounted on the first wheel, to support the carrier
c. But it would prevent the sun from meshing with the planet! And, as the carrier rotates,
the planet p would butt against it! Anyway, as the pieces are sandwiched together it is
unnecessary. In the following I will assume that y does not exist.
Figure 15 summarises the design, with the sun gear shown about twice its normal height. For
convenience, the switch sw is shown acting from the annulus.
(a) Anti-clockwise (sober) winding: The sun s
rotates anti-clockwise and click sw forces the
fusee cone and the annulus a to rotate with
it, winding the watch. The planetary gears are
locked and rotate anti-clockwise, with both
clicks 1’ and x sliding over the teeth of the
carrier and the annulus respectively. The first
wheel does not turn and has no maintaining
power.
(b) Clockwise (drunken) winding: The sun s
rotates clockwise and click sw has no effect.
The planet p rotates anti-clockwise and
attempts to turn the carrier c clockwise, but the
click 1' prevents this happening (and provides
maintaining power to the first wheel). So p
rotates the annulus a anti-clockwise to wind
the watch. Again click x slides over the ratchet
teeth of the annulus a.
(c) Running: The first wheel rotates clockwise Figure 15
and both clicks 1’ and x slide over the teeth of
the carrier and the annulus. So the planetary
gears are disconnected and do not rotate.
It is now clear that the click x has no function and should be omitted.
Before discovering the above design, I had thought that forcing the sun and the annulus to rotate
together would be very difficult, because those two gears are separated by the space occupied by
the carrier and the planet gear. So I created a design based on Tappy’s
fusee (see page 15).
So far I have concentrated on what happens at the base of the fusee,
where the planetary gears and the first wheel are located. But what
about the top of the fusee? As we can see from Figure 3, at the top of
the fusee we find the fusee arbor protruding through the fusee cone,
and it is possible to put a switch at this point. As a switch consists
of a ratchet and a click, a ratchet has to be put on the fusee arbor
(and so the sun) and a click on the cone (and so the annulus). This is
illustrated in Figure 16, taken from Andriessen’s drawings (see page
15). Figure 16
11
 The Art of Being Drunk
Figures 17 and 18 are identical to Figures 9 and 12 except that the switch sw has moved and
consequently the ratchet x1 is needed on the top of the fusee. Then:
(a) Drunken, clockwise winding: When the fusee arbor and sun gear s rotate clockwise the
switch sw has no effect and the planetary gears behave normally. The sun gear causes the
planet gear p to rotate anti-clockwise and, because the teeth of p mesh with the annulus
a, p attempts to pirouette around the sun and turn the planet carrier c clockwise. But the
click 1' locks the carrier and Tc = 0. So NaTa = – NsTs. That is, the annular gear and the
fusee cone rotate anti-clockwise winding the chain onto the fusee.
(b) Sober, anti-clockwise winding: The fusee arbor and the sun s rotate anti-clockwise and
the switch sw forces the annulus to rotate with them. So Ts = Tc = Ta, and a rotates anti-
clockwise (winding the chain onto the fusee). Now the click 1' has no effect and the first
wheel is free.
(c) Running: During winding the sun s is the driver and the annulus a is driven. Now a
becomes the driver, rotating clockwise as the chain is pulled off the fusee. It attempts to
rotate p clockwise and p in turn attempts to rotate s anti-clockwise. But the switch sw
locks the annulus a to the sun s so that Ta = Tc = Ts and the clockwise rotation of the
fusee cone is transmitted to the first wheel through the click 1'. As the planetary gears are
locked and rotate together the fusee arbor rotates with them.
At this point we know that drunken fusees can be constructed with planetary gears and a switch
between any of the three components, s and c or a and c or s and a.
What has not been stated is that locking the planetary gears turns the fusee into an ordinary fusee.
Because the fusee arbor turns with the cone (annulus) and the carrier everything can be removed,
the carrier attached to the cone and the arbor attached to the carrier. That is, when the switch
operates the fusee is identical in action to the fusee in Figures 3 and 4.

Figure 17 Figure 18

12
 The Art of Being Drunk

On Switching the Planet and the Carrier

This design is again identical to most of the previous designs, except that the switch has moved and
it now operates between the carrier c and the planet p.
As shown in Figures 19 and 20 the three actions are now:
(a) Drunken, clockwise winding: When the fusee arbor and sun gear s rotate clockwise the
planet p rotates anti-clockwise, so the switch sw has no effect and the planetary gears
behave normally. The sun gear causes the planet gear p to rotate anti-clockwise and,
because the teeth of p mesh with the annulus a, p attempts to pirouette around the sun
and turn the planet carrier c clockwise. But the click 1' locks the carrier and Tc = 0. So
NaTa = – NsTs and the annular gear and the fusee cone rotate anti-clockwise winding the
chain onto the fusee.
(b) Sober, anti-clockwise winding: The fusee arbor and sun s rotate anti-clockwise and the
switch sw stops the planet p from rotating clockwise; it always faces the sun around which
it revolves. So the carrier is forced to rotate with the sun, Ts = Tc = Ta, and a rotates anti-
clockwise (winding the chain onto the fusee). Now the click 1' has no effect and the first
wheel is free.
(c) Running: The annulus a becomes the driver, rotating clockwise as the chain is pulled
off the fusee. It attempts to rotate p clockwise unsuccessfully, because of sw, and so the
annulus and carrier rotate together, Ta = Tc = Ts. The click 1' links the carrier to the first
wheel and the watch runs.

Figure 19 Figure 20
I have now described the four possible designs that have the carrier separate from the first wheel.
I have ignored the two designs where the switch acts between the planet and either the sun or the
annulus because they are impossible; the planet pirouettes around the other gears and cannot be
linked to them by a click.
Finally, I expect some minor variations on the above designs may have been created and used in the
past, but they must be based on the same principles.
13
 The Art of Being Drunk

On Always Being Drunk, Huygens

In the designs we have considered so far, the carrier and the first wheel are separate pieces joined by
a click. Now we will look at the situation when the carrier is integral to (the base of ) the first wheel.
In 1683 Huygens described a fusee for use in clocks and
designed to provide maintaining power during winding,
Figure 21.6
The carrier c, on which the planet gear p is mounted, is
the base of the first wheel 1. The annulus, attached to the
fusee cone, is a disk with ratchet teeth on the outside and
the inward facing annulus teeth a on the inside. The carrier
and annulus are linked by the switch sw. The sun gear s is
attached to the fusee arbor.
When the watch is wound by turning the sun clockwise,
the carrier is locked because the watch train prevents the
first wheel 1 from turning, and so Tc = 0 and consequently
NaTa = – NsTs. That is, the annulus and the fusee cone Figure 21
rotate anti-clockwise winding the watch. (In detail, s
rotates p anti-clockwise and the teeth of p, meshing with a, make the carrier 1/c attempt to rotate
clockwise. But this is impossible because 1/c cannot rotate due to the watch train, and so the
rotation of p forces a and the fusee cone to rotate anti-clockwise, winding the mainspring.) As the
cone rotates, the switch sw, attached to the first wheel, slides over the ratchet teeth of the annulus
a, which is attached to the fusee cone, and has no effect.
However, for the planet gear p to turn anti-clockwise and rotate the annulus anti-clockwise against
the force of the mainspring, its teeth necessarily creates a clockwise force on the base of the first
wheel, the carrier, providing maintaining power to the first wheel 1/c.
When the watch runs, the fusee cone and the annular gear a rotate clockwise. Now, as the annulus
a rotates with the fusee cone, the switch sw forces the first wheel 1, and thus the carrier, to rotate
with it, and both turn as a single unit. So Tc = Ta and all three parts of the planetary gears rotate
together as in an ordinary fusee.
Although this design of Huygens was not meant to be used by drunks, we should also see what
happens when a sober person turns the key the wrong way, that is anti-clockwise, because the wrong
way is what, in ordinary fusees, would be the right way!
When the sun is rotated anti-clockwise the planet rotates clockwise and it attempts to rotate the
carrier clockwise, which is impossible because the carrier is the first wheel; consequently Tc = 0 and
NaTa = – NsTs and the annulus attempts to rotate clockwise. But this is impossible because the
switch sw and the ratchet teeth on the annulus a prevent that motion. So the key cannot be turned
anti-clockwise without forcing it and breaking something.
The fusee designed by Huygens includes only one feature, the planetary gears, and there is no
switch to lock the gears so that sober winding is not possible. To use a fusee with the carrier c and
the first wheel 1 joined requires some changes to the design.
Note that this design is identical to Virgo’s fusee, Figure 11, with c and 1 joined and the unnecessary
click 1' removed.
6 Huygens, Christiaan, 1934, Oeuvres Complètes de Christiaan Huygens, Vol. 18, pages 621-622, The Hague:
Martinus Nijhoff, 1888-1950. (Extract from Manuscrit F, 1683, page 175.) Available from http://ia902704.
us.archive.org/17/items/oeuvrescomplte18huyguoft/oeuvrescomplte18huyguoft.pdf
9
14
 The Art of Being Drunk

On Switching the Fusee Arbor Between the

Sun and the Annulus, Tappy
The fusee of Huygens is included because it is the first example of a mechanism where the first
wheel and carrier are joined, the carrier being the base of the first wheel. It illustrates the problem
that occurs with sober (normal) anti-clockwise winding.
The following description is based on drawings made by P. Andriessen in 2013.7 He produced
three drawings of a watch made by Abraham Tappy (born in Switzerland in 1735, died 1773); the
drawings show sober (normal) winding, drunken winding and running.
Figure 22 is Andriessen’s drawing 2 with new labels: FIG. 1 is the top of the fusee with the arbor/
annulus switch sw; FIG. 2 is the bottom of the fusee with the sun gear; FIG. 3 is the bottom of the
fusee with the planet and annulus gears; and FIG. 4 is a side view of the fusee f. There are two other
features, K and P. In this fusee the annulus is a separate piece and the pin K on the annulus enters
a slot P on the base of the fusee f that aligns the annulus and makes it rotate with the fusee cone.

Figure 22

7 P. Andriessen and J.H.Midavaine, “Abraham Tappy, un horloger vaudois expatrié aux Pays-Bas, à Veere et Goes”,
Chronométrophilia N° 76 Hiver/Winter 2014, pages 79-92 and Chronométrophilia N° 77 ÉTÉ/Summer 2015,
pages 17-39.
French translation of P. Andriessen and J.H.Midavaine, “Horlogemaker te Veere en Goes Abraham Tappy,”
Tijdschrift 3/2013, pages 24-32 and Tijdschrift 4/2013, pages 4-19.
15
 The Art of Being Drunk
Figures 23 and 24 are based on Andriessen’s drawings, with Figure 23 showing the ratchet and click
at the top of the fusee and Figure 24 showing the ratchet and click at the bottom. Although similar
to the previous mechanisms, there are two important differences:
(a) The carrier 1/c is the base of the first wheel 1. This is counter-intuitive because the previous
mechanisms require the carrier to rotate both clockwise and anti-clockwise requiring the
carrier c to be separate from the first wheel 1.
(b) In the previous mechanisms the sun gear s is an integral part of (or squared onto) the fusee
arbor. In this design the sun gear is loose and the fusee arbor x has two ratchets x1 and
x2 that are integral parts of (or squared onto) the arbor. These ratchets face in opposite
directions so that: When the arbor x rotates clockwise x2 and s' (at the bottom of the
fusee) force the sun gear s to rotate clockwise. When the arbor rotates anti-clockwise x1
and sw (at the top of the fusee) force the fusee cone and hence the annulus a to rotate
anti-clockwise. That is, there are two switches sw and s' !
The action is:
(a) Drunken, clockwise winding: The arbor x rotates clockwise. The ratchet x1 slides under
the switch sw so that the arbor x and the fusee cone a (and its annulus a) are not directly
linked. At the same time ratchet x2 rotates clockwise and, because the switch s' locks into
the ratchet x2, the sun gear s rotates clockwise with the arbor x. Now the behaviour is of
normal planetary gears.
The sun gear causes the planet gear p to rotate anti-clockwise and, because the teeth of
p mesh with the annulus a, p attempts to pirouette around the sun and turn the planet
carrier c clockwise. But because the first wheel and the carrier 1/c (attached to the train)
cannot rotate Tc = 0 and NaTa = – NsTs. That is, the annular gear and the fusee cone
rotate anti-clockwise winding the chain onto the fusee. As a is rotating anti-clockwise the
click ca has no effect.
Tappy’s teeth counts are Np = 6, Ns = 15 and Na = 30. As Ts = - 30/15 Ta, it takes exactly
two turns of the key to turn the fusee once.

Figure 23 Figure 24
16
 The Art of Being Drunk
However, from the theory of planetary gears (page 3), the correct value for the annulus
is Na = 27 and excessive friction will occur.8
(b) Sober, anti-clockwise winding: The arbor x rotates anti-clockwise. Now ratchet x1 rotates
anti-clockwise and the switch sw forces the fusee cone and annulus a to rotate anti-
clockwise with the arbor x. The chain is wound onto the fusee. At the same time ratchet
x2 slides under the switch s' and the arbor x and sun gear s are not directly linked. Again
as a is rotating anti-clockwise the click ca has no effect.
The planet gear p is mounted on the 1st wheel 1/c and meshes with the annulus a mounted
on the fuse cone. As the annulus is rotating anti-clockwise the click ca on the carrier (1st
wheel) 1/c rides over the teeth of the annulus and has no direct effect on the carrier 1/c. At
the same time the annulus causes the planet gear p to rotate anti-clockwise and it in turn
forces the sun gear s to rotate clockwise so that its click s' rides over the teeth of x2 and has
no effect on the arbor x. In this case the sun gear s is driven by the planet gear p and the
carrier is driven by the annulus.
(c) Running: When the watch runs the fusee cone, with the annulus a attached, is the driver
and rotates clockwise. The click ca forces the first wheel and carrier 1/c to rotate clockwise,
running the watch.
Because the annulus a and carrier 1/c are locked together, Ta = Tc = Ts and the sun gear s
rotates clockwise with them. The switch sw acting on the ratchet x1 causes the arbor x to
rotate clockwise, and the arbor causes the sun to rotate clockwise through the ratchet x2
and the switch s'. So everything rotates clockwise.
Maintaining power occurs during drunken winding because the planet gear p tries to turn the
carrier 1/c clockwise. During normal winding the click ca slides over the teeth of 1/c and there is
no maintaining power.
In the previous four fusees (designed by Massotéau, Virgo, Moore and me) the carrier c rotates
anti-clockwise during sober (normal) winding because the switch locks the gears together and
consequently Ts = Tc = Ta. And c rotates clockwise during running; it also tries to rotate clockwise
during drunken winding, but is prevented from doing so. Most importantly, during winding the
sun gear is the driver and during running the annulus is the driver.
In Tappy’s fusee the joined first wheel and carrier 1/c rotates clockwise during running, when the
fusee cone (and the annulus) is the driver, and tries to rotate clockwise during drunken winding
when the sun is the driver. If the behaviour described above is compared with my fusee in Figures
17 and 18 it will be seen that drunken winding and running are identical to that fusee and, most
importantly, the sun gear behaves as through it is rigidly attached to the arbor; the ratchet x2 acts
in both situations.
But 1/c cannot rotate anti-clockwise during sober winding, as happens in Figures 17 and 18, because
it is part of the first wheel.
So the purpose of the ratchet x2 is to prevent the carrier from rotating during sober winding by
disconnecting the sun so that it no longer drives the planetary gears. Instead, the planet p is driven
by the annulus a clockwise and it is free to rotate without affecting the carrier. That is, Tappy’s
fusee uses planetary gears when the arbor rotates clockwise and it de-activates, but does not lock, the
planetary gears and uses a separate mechanism when the arbor is rotated anti-clockwise. And if we
ignore the planetary gears, this separate mechanism turns the fusee into an ordinary fusee.
8 As occurred in John Arnold’s watches; see The Origins of Self-Winding Watches, pages 79-82. (Available from
http://www.watkinsr.id.au/Origins.html)
17
 The Art of Being Drunk

On Other Ways of Getting Drunk

Before looking at the ways to get drunk, Tappy’s design leads to an improvement to the fusee of
Huygens, Figure 25.
This fusee is identical, in principle, to Figure 21, except
it now has a ratchet x2 and switch s' attaching the sun
s to the arbor x.
When the arbor is turned clockwise, the ratchet and
the switch cause the sun to rotate clockwise and,
consequently, the planetary gears function and annulus
and fusee cone a rotates anti-clockwise, winding the
watch; the click ca slides over the teeth of a.
And when the arbor is rotated anti-clockwise (normal
winding) the ratchet and switch disconnect the arbor
x from the rest of the mechanism and its rotation does
nothing. As a result, the fusee, or other parts of the
watch, cannot be damaged by forcing the arbor in that
direction.
During running, the click ca locks the carrier to the
annulus and everything rotates clockwise. As the sun
s and the switch s' rotate clockwise, the arbor x is free Figure 25
and does not rotate.
This mechanism is identical to that in FIG 2 and FIG 3 of Figure 22, but without the ratchet and
switch in FIG. 1. Although interesting, the additional complexity of this design is probably not
justified.
Other designs for drunken fusees can be created by making unnecessarily complicated designs.
Figures 26 and 27 are the same as Tappy’s design, Figures 23 and 24, but with a separate carrier, and
so an additional click 1' is needed to join the first wheel to the carrier.
When the arbor x is turned clockwise for drunken winding, x1 and the switch sw are not active, but
s' and x2 rotate the sun s clockwise. Then p rotates anti-clockwise and tries to rotate c clockwise,
but it cannot because of the click 1', so a rotates anti-clockwise and the click ca is inactive. The
planetary gears behave normally.
When the arbor x is turned anti-clockwise for sober winding, sw rotates a anti-clockwise and ca is
not active, so the carrier is not turned. Also and s’, x2 are not active. The annulus a drives p anti-
clockwise so it turns s clockwise which is possible because click s' is not active. Actually, the carrier
can rotate anti-clockwise and which of the sun and/or carrier rotates is arbitrary.
When the watch runs, a is the driver rotating clockwise. Now ca is active and locks the carrier c to
a and the click 1' locks the first wheel 1 to the carrier. Consequently, as Tc = Ta then Ts = Ta and
the sun s rotates clockwise and the arbor x is free.
This is an absurd design developed by mindlessly (drunkenly) modifying Tappy’s fusee to have a
separate carrier and hence an additional click. It is actually the same as the fusee in Figures 17 and
18, with the sun loose and so requiring the addition of three parts, x2, s' and ca.
Similar, over-complicated fusees can be created from Tappy’s design where sw acts between the
carrier and the sun (see Figures 8 and 9) and between the carrier and the planet (Figures 19 and 20).
So none of these designs are actually new and none are interesting.
18
 The Art of Being Drunk

Figure 26 Figure 27

Can drunken fusees be constructed in any other way? In particular, the table on page 6 lists
three designs that have not been considered; they are when the carrier and first wheel are joined
(the carrier is the base of the first wheel) and the switch links the carrier to one of the annulus, the
sun or the planet.
Before considering these cases, the following table gives the number of concentric arbors and the
number of clicks in each design:
Design Arbors Clicks Linked Not linked
Massotéau 4 2 s/x, c, a 1
Virgo 4 2 s/x, c, a 1
Moore 4 2 s/x, c, a 1
Me 4 2 s/x, c, a 1
Tappy 4 3 s, c/1, a x
All require 4 concentric arbors:
(a) The first four designs have the fusee arbor x and sun s linked (they are a single piece) and
the sun s, carrier c and annulus a are linked by the planetary gears. The first wheel 1 is
separate and connected to the rest of the mechanism by a click. If the first wheel had to
rotate in both directions then there would need to be two clicks.
(b) Tappy’s design has the sun s, carrier c/1 (a single piece) and annulus a linked by the
planetary gears. The fusee arbor is separate, but because it must rotate in both directions
there must be two clicks linking it to the rest of the mechanism. This requirement is
illustrated by Figure 25, which is Tappy’s fusee, Figures 23 and 24, without sw and x1.
The silly design in Figures 26 and 27 has 5 concentric arbors (x, s, c, a and 1) and 4 clicks.
The first four designs work because they lock the planetary gears during sober winding. In contrast,
Tappy’s fusee works because the planetary gears are never locked. So during sober winding, when
the annulus rotates anti-clockwise, the sun is disconnected allowing the planet to rotate it arbitrarily
while Tc = 0.
Can the other three cases (linking the carrier to one of the annulus, the sun or the planet) be
constructed with the planetary gears always unlocked?
19
 The Art of Being Drunk
First, we have already seen that Virgo’s design, Figure 11, reduces to the fusee of Huygens which
is not drunken. Similarly, Massotéau’s fusee, Figure 9, Moore’s fusee, Figure 15, and my design,
Figure 19 all work for drunken winding, but sober winding is impossible. This is because, in all four
designs, sober winding requires the planetary gears to be locked resulting in Ts = Tc. But the carrier
cannot rotate because then the first wheel must rotate with it, which is impossible.
This explanation is not necessary, because in every design the purpose of the switch sw is to lock the
planetary gears so that the anti-clockwise rotation of the sun s is converted into the anti-clockwise
rotation of the annulus a and, consequently, making Ts = Ta = Tc. But as Tc = 0, Ts = Ta = 0 and
nothing can rotate.
Consequently, when the carrier and first wheel are a single piece, the fusee requires a design like that
used by Tappy, where the planetary gears are not locked. That is, the switch acts between the fusee
arbor and the carrier or the planet:

(a) x and 1/c : Figure 29 is based on Figure 23

and shows a possible arrangement where the
switch sw is moved to act between the carrier
and the sun. And x1 has been moved to the
bottom of the fusee and is attached to the
arbor just above x2.
Drunken winding, when the switch sw does
not act, is normal. And running, when the
annulus a is the driver, also works.
However, once again sober winding does not
work. When the arbor x rotates anti-clockwise,
x1 (through the switch sw) attempts to
rotate the carrier 1/c anti-clockwise, which is
impossible because Tc must be 0. So Ts = 0
and consequently Ta = 0.
Figure 29

(b) x and p : The only other possibility is to link the planet and the sun, but that is impossible
because the gears are not concentric, the planet pirouettes around the sun, and a click
cannot work.

That is, the five designs described in this article, and minor variants of them, are the only possible
designs.

20