CLOCK MAKING TIPS AND TRICKS
Or, Things I Wish Someone had Told me Before I Started
Disclaimer: The following does not purport to describe the right
way to do anything. It is just a description of what I do, and 40 years
of marriage have taught me that I am often wrong.
Don't assume that the thinnest scroll saw blades do the best work. Very
fine blades, such as "puzzle blades" are best only when you
need a really thin kerf or extremely tight radius, but they are slow,
break easily, and are not as easy to "steer" on straight lines
or gradual curves. For pieces too large for my bandsaw I use a raspy
hook-tooth blade and cut wide of the line. Otherwise I use medium blades
consistent with the smallest radius I need to cut. The Flying Dutchman
Ultra Reverse #9 is narrow enough for tight turns, thick enough to hold
up well, but not wide enough for smooth steering. It can be tiring to
use reverse blades, because you have to hold the work down firmly, and
for most clock parts, a little tearout on the back that the reverse
teeth are designed to prevent is not visible.
Except for inside or intricate cuts requiring a scroll saw, or long
straight cuts suited to a table or radial arm saw, a band saw is easier
(and safer) than any other kind of power tool. You can buy a 9"
tabletop reconditioned Ryobi band saw for about $100. Quarter inch blades
are better for straight lines and heavier work, such as frames, if the
radii are not too tight. I use a 1/8" blade for rough cutting the
faces of gear teeth -- faster than any blade on a scroll saw -- then
sand to the line. Not all suppliers offer these short blades, but I’ve
found the the Morse blades available on Amazon cut way better than the
generic blades I started with.
For laying out patterns, I don't use messy spray. Instead I go to an
art supply source (like Blick’s online) and get
A medium-size can (with brush) of Best-Test rubber cement
A pint can of the same for refilling
A pint can of Bestine rubber cement thinner (don't try to use mineral
A rubber cement pickup block (looks like a block eraser)
A small flip-top squeeze bottle or spray-top bottle, to fill withBestine.
Before cutting, use the pickup block to lift any cement from the lay-out
so that wood dust will not adhere when cutting. To remove the paper
after cutting the wood, if it wants to tear when peeling, spray or squeeze
the Bestine onto the paper, wait 30 seconds for it to soak through and
soften the cement, and just lift off. Bestine dries rapidly, so work
fast. Two minutes later the wood will be dry again, and the remaining
cement can be easily rubbed off with the pickup block.
For a slower and cheaper method, use mineral spirits,
but that must dry for 10 minutes after pattern removal or it will be
gooey on the pickup block. I hear a heat gun works well by softening
up the cement, but I don’t have one so can’t comment. Remember:
never use mineral spirits as a rubber cement thinner! You’ll ruin
I like the minute or two of working time wet rubber cement
provides to adjust the pattern before it sets up, but to get a very
firm, fast bond, use a thin coating on both surfaces and let dry before
sticking them together.
Some of Clayton Boyer’s larger and more curvaceous
frames are may require a glue-up of two pieces together. And, the shapes
of the separate pieces may be devilish hard to accommodate clamps. So
plan ahead, and add to each side of the pattern “ears” or
flats sticking out with parallel surfaces to facilitate the clamping.
After the glue sets, then cut off the ears to the final profile.
Of course, while rubber cement is wonderful for paper, it is no good
for joining wood pieces together. There is no one adhesive best for
Polyurethane glue (like Titebond II) expands
as it sets and causes a foamy squeeze-out that must be scraped off
when firm but not cured – 30 or 40 minutes. Excess can be cleaned
with alcohol when wet, but that won’t prevent squeeze-out. It
requires moisture to act, so in dry conditions, the wood can be moistened
lightly before application. This expansion absolutely requires clamping
to prevent the pieces from being pushed apart, but it is a distinct
advantage when joining end grain or pieces with uneven surfaces or
voids. Stronger than wood.
Elmer’s glue is also stronger than wood
and is easier and less messy, because it is water-soluble when wet.
It benefits from clamping, but since it doesn’t expand when
setting, it’s the stuff to use if clamping is not practical.
The white kind dries clear, and never reaches 100% hardness, so I’m
told it can creep over the long term under heavy load. The yellow
kind does not dry clear, but it reaches full hardness. It sets up
faster than white, a property which I like.
Super glues are fast and easy. I am using
them more and more, especially for non-porous materials. Keep an alcohol
swab handy to clean up spills before they harden, which happens fast.
Always buy the more expensive grades of sandpaper, and never bother
with the old-fashioned cheap stuff.
I find the most useful all-around sandpaper to be the
yellow 220 grit, but when significant smoothing is required, don’t
try to save time by starting with too fine a grade. You’ll spend
more time and get less done than if you work down with two or three
grades of increasing fineness. Don’t bear down hard with sandpaper.
I do not know why, but a light pressure not only makes the paper last
longer without filling but also seems to remove more material than heavy
I also love the 3/16”-thick sponge-backed “All
Purpose Sanding Pads” made by 3M in three grades. They follow
curved features nicely and on flat surfaces with a first coat of shellac
to smooth out, they are less likely than sandpaper to sand through to
bare wood at the edges.
I found the cheap 1” belt sander sold by Harbor
Freight and others lousy for any purpose – too loose and unpredictable.
But if you have a band saw, you can buy band saw sanding belts incredibly
cheaply from customsandingbelts.com. These work so well I have stopped
trying to cut teeth to the line and now sand to the line instead. I’m
using the 180 grit, ¼” width belts but perhaps a wider
belt would be more easy to control.
For manual correction of dental and other detail I usually use a flat
sanding stick. Take a regular paint stir stick and sand the profile
down so that one edge is sharpened to an acute angle and the other remains
flat or is rounded to whatever radius you like. A sanding stick sharpened
at the end is useful for working on clocks that are already assembled.
Use rubber cement to hold 150- or 220-grit sandpaper so it bends snug
around the edge of the stick. When the paper fills at the edge, peel
it up and shift the edge to fresh paper, which will often adhere again
to the old adhesive.
For flats, edges, and the ends of rods and tubes, a belt-disk
sander is essential. A cheap 4” model such as the smallest Craftsman
is just fine for clock work. Don't be concerned about some user complaints
that the Craftsman is underpowered – those users are just trying
to push it too hard.
For sanding the insides of curves, Ridgid makes an absolutely
marvelous spindle/belt sander, a recent luxury purchase I would not
now want to be without.
WHEELS AND PINIONS
When cutting wheels and pinions, I start with a 1/16" or 5/64”
hole in the center regardless of the final arbor size. It’s easier
to locate a small bit accurately than a larger one, which will then
center itself nicely in the pilot hole. Cut the outside circle of the
wheel or pinion a little larger than the final diameter. Either before
or after cutting the tooth side profiles, sand the circumference to
the line on the belt sander. If concerned about concentricity, you can
mount the gear on a 1/16" rod set in a piece of scrap wood clamped
to the table of the disk sander and turn it down to exact radius relative
to the center. To assure smooth running prior to assembly, you can optionally
set wheel and pinion pairs on a fixture of scrap plywood with the 1/16"
pins at the correct spacing per plan. You can then check the mating
of the two and make any adjustments to either teeth or final spacing
before committing the frame. Usually, I do this step on short brass
stubs on the open back frame itself, after drilling holes to the final
size but before doing any assembly.
To drill out the 1/16" holes to final diameter while
keeping true the center location, you can use brad point bits, but be
careful that the point is right on center -- they can be bent, or even
come from the seller off-center. In general, I don’t much like
them, but they have the virtue of reducing tear-out if they’re
sharp. Forstner bits are better than brad points, but I’m not
sure whether you can get them in the odd 64th sizes often needed. An
interesting tool for enlarging holes is the "center bit" available
from buydrillbits.com and others. They are made for metal lathes, I
think, but are great in a drill press if you’re mindful that they
overheat easily. They’re very robust and will not bend or drift,
but are available only in a few sizes. The small or “pilot”
tip diameter is of minor importance, as this bit tends to center nicely
in just about any pilot hole. I’ve just started using this nifty
bit, and wish I had known about it earlier. Do not confuse this tool
with the "self-centering drill bit" sold by Rockler and others.
The one I mean is short and double-ended, without much by way of grooves.
When enlarging pilot holes, I used to clamp the work
down, tap it carefully into place, and use a brad point bit, but all
that didn’t always keep the bit in the center. Now I just use
a regular or center bit, get it spinning, hold the work to the table
by hand, and let hole find the bit as the bit comes down. It sounds
mighty sloppy, but for me it works better, and it’s easier and
Sometimes a clock will run reliably from the beginning and need little
attention except for winding and occasionally regulating. But a wood
gear clock is next thing to a living organism, and it can take weeks
or months to work out its little issues. Even with simply regulating
(adjusting the pendulum screw), it is helpful to know when the last
adjustment took place and what that adjustment was. I keep an index
card near each clock and use a notation system to record its history
as I go to wind it each day. My notation system is:
. [period] New day
, [comma] Set hands. If you find five dots between commas, you probably
don’t need to regulate the pendulum – it’s pretty
accurate. If you see a comma occurring after every one or two dots,
then you probably want to give the pendulum screw ¼ turn and
see if it starts keeping better time.
+ [plus sign] Running fast
- [minus sign] Running slow
~ [tilde] Running true. Combinations can be used: a tilde above a
plus would mean “Running slightly fast.”
Vertical zigzag or sawtooth Sanded a problem tooth or other part
“A” plus a fraction Adjusted pendulum. A ¼ means
I turned adjusted screw ¼ turn
S plus a time of day Clock stopped at a particular time indicated
by its own hands
// Double slash Serviced clock in a way that its wheels no longer
show the same positions at a given indicated time.
Many would not bother to keep this little log, but I would
never not bother. It helps me see trends and decide what to do next
to keep the clock running true.
TROUBLESHOOTING STUCK CLOCKS
The first order of business, before even starting up for the first time,
is to mark one location on each wheel and pinion pair where they are
going to meet every time you assemble the works. There’s no point
in fixing a problem spot and then lose the fix by changing the relative
positions later on.
I keep a log to record the times of day when a clock
tends to stall, and by referring to the periodicity of the various wheels,
the precise location can often be identified.
Running problems usually occur in the "going train"
where big gears ("wheels") are driving little gears ("pinions")
. As noted in Ward Goodrich’s classic, The Modern Clock (modern
as of 1905 anyway) the business of such "spur gears" is accomplished
in general by the addendum (top half of the tooth) of the wheel acting
on the dedendum (bottom half of the tooth) of the pinion. Problems in
the meshing of gears, on the other hand, frequently occur when the less
important addendum of the pinion collides with the less important dedendum
on the wheel on the way in, instead of grazing into position. This is
called butting. What this means, in practical terms, is that if you
find stoppage problems where butting occurs, the first order of business,
after marking the location, is to adjust the upper part of the pinion
tooth (its addendum) and/or open up the part of the wheel tooth that
lies toward the valley (its dedendum). As explained above, this should
not affect the integrity of the working surfaces, assuming you have
shaped those correctly. Also blunt any sharp top corners of the teeth
on both wheel and pinion, usually a good practice across the board.
I have found that some wood clock plans seem to incorporate gear profiles
that are theoretically correct but not forgiving enough for the vagaries
of wood and woodworkers. So I’ve overcome my fear of removing
too much material when shaping gears and sand fully down to the line.
Gears normally won’t stop if they’re a little loose.
The above paragraph applies only to cases where the wheel
(the larger one) drives the pinion. Where the opposite is the case,
as in the motion works (gearing down to move the hour hand), all these
rules are void -- but you rarely find a problem in this part of a clock.
In no case should the top of any tooth meet the “bottom
land” of any opposing tooth: If this happens, your cause may be
bad centering of the wheel and/or the spacing on the frame. Even if
the frame is already drilled, dowling and redrilling if needed may be
a better idea than going around messing with tooth profiles everywhere.
Stoppages are not always due to gear problems. Failing
to balance the escape wheel (by drilling and inserting brass plugs)
can definitely stop a clock, I’ve found. If the crutch pin is
too tight or too loose in the pendulum slot, you’ll have trouble.
You should have pin clearance of about the thickness of a dollar bill.
Rarely, riding turns can occur on the wind spool, decreasing torque.
Checklist for troubleshooting a stuck clock, based on
things that have gone wrong for me:
Butting teeth – usually pinion addendum entering
against wheel dedendum
Tooth jamming between opposite teeth -- tooth too fat or valley too
thin (relieve the non-functional surfaces of tooth or valley)
Wheel out of round, no clearance with pinion
Out of balance arbor, especially escape
Side of wheel scraping something
Not enough drive weight
Too much drive weight (excess friction)
Pins on contrate wheels not of even length
Hands scraping or hitting other parts
Tube binding on rod, sometimes from excessive setscrew pressure
Arbor too tight in frame, or not co-linear with both arbor holes in
the frame. (The latter issue can happen even if the arbor seems to
spin nicely at each end when frame is disassembled). To make a nice
curette for gently shaving out a tight or misaligned arbor hole, take
a suitable size brass tube and grind an oblique lengthwise flat through
one end, making a blade form perfect for enlarging or cleaning the
arbor hole of shellac or varnish.
Arbor compressed lengthwise (no endplay)
Crutch pin too tight or too loose
Damaged or misplaced pendulum pivot
Riding turn on wind spool
Pallets out of adjustment
Fouled weight cord
Fouled weight cord pulley
I do not think I have ever had a stoppage that did not
fall into one of the above categories.
RODS AND TUBES
One of the rookie gaffes of my first try at clockmaking was using the
hacksaw to cut tough stainless rod (took forever) and delicate brass
tube (I don’t even like to think about that). I quickly realized
that the Dremel abrasive cutoff wheel was better than a hacksaw. But
I didn’t like the way it walked and jumped and soon found an even
better answer. Harbor Freight has a 6” cutoff saw, including motor,
for $29.95. Well worth it. The chopper makes a rough end that can be
finished nicely on the disk sander. Arbors should be placed in the drill
press and polished with a sequence of medium to ultra fine sandpapers,
or with wood or leather clamps.
Bearings are seldom specified in clock plans, but I now use them on
all parts with heavy loads – normally the first one or two arbors.
I used to use KMS plastic bearings but after a couple of failures now
mostly use VXB metal ones. KMS makes nice plastic spool bearings for
rigging weights, though.
I won’t say much about this, because I am not very good at it.
A good grade of birch plywood is the usual choice for
making wheels and pinions, and for many other parts requiring strength,
but it doesn't finish very well. American birch ply grades sold locally
are often lousy: weak and full of voids. I have settled on www.dndplywoodonline.com,
which offers many other hardwood plywoods as well, as my regular supplier.
Or try to get “Finnish birch,” which has more plies and
is better quality than generic Baltic. With any plywood and with some
other woods as well, stains and other finishes may not apply evenly
and tend to blotch. To minimize this, you can rub in (not brush on)
some 1-2 pound dewaxed shellac before applying any other stain or finish.
If it doesn't say dewaxed, don't buy it. Some top applications like
polyurethane will check (crack) over ordinary canned store grades.
I use dewaxed shellac flakes from shellac.net: super blonde or “platina.”
Darker shellacs like garnet tend to turn out too yellow or orange. Plan
ahead if using flake: depending on strength of mix, temperature, and
frequency of agitation, it can take up to several days to fully dissolve
in a half-pint canning jar. In that size container, use one ounce of
flake per “pound” strength desired. I store flake in the
freezer sealed in several layers of plastic bag, and keep mixed shellac
in the fridge. Moisture ruins shellac, so use only the freshest alcohol,
and always let mixed shellac come to room temperature before opening
to avoid water contamination through condensation. Very inflammable
– keep it away from a gas range or other flame. Always strain
through a paper towel before using.
My method is to use platina shellac (from flake) for
all steps except the final coat. Alcohol-based, it dries fast and can
be mixed in all different ways. The first coat of shellac will seal
(and optionally stain) the wood and raise the grain. When thoroughly
dry, the raised grain is brittle and can be easily sanded or steel-wooled
(with synthetic wool, not metal) down to a smooth base for the next
coat(s). Never allow any finish to get on gear teeth, except perhaps
a soft paste wax, though even that might raise the grain.
I’v recently started using dyes called Transtint.
Forrest Burnett, kinetic sculptor and clockmaker par excellence, has
described a method of mixing three cups sanding sealer with 2/3 cup
of lacquer thinner, then adding tint drop by drop until reaching the
desired shade. He uses a costly airbrush-type sprayer and gets splendid
results. I don’t have such a sprayer. In my limited exploration
of the subject, I have found that the brown shade of Transtint in a
thin 1-2-pound shellac-alcohol is quite good for darkening birch to
a handsome faux walnut grain, and the colors like blue or red are marvelous
for allowing the grain to show through a beautiful, bright hue. None
of the other staining gels or liquids I’ve tried can come close
to the Transtint results. I always conclude with one coat of the most
expensive satin polyurethane I can find, using one high-grade brush
for application and a smaller dry artist or chip brush to pick up the
sags and globs before they set.
If you don’t have a router but have a Dremel tool, get the cheap
little Dremel router table and the small quarter-round router bit. It
will apply a lovely little radius to such elements as your frame and
your wheel cutouts. It’s a cheap, sloppy little tool, but it’s
little inaccuracies can easily be sanded out by hand.
You’ve got to be able to take down the clock for service. If you
have to use any kind of permanent anchor to mount it to the wall, consider
what I did with my Marble Strike (which you can see on YouTube). I made
a mounting frame slightly larger than the original back frame of the
movement, countersank T-nuts into the back facing forward, and then
attached it permanently to the plaster wall with wing anchors. The clock
now mounts and dismounts easily from the mounting frame with bolts joining
the original back frame to those T-nuts.
HEALTH AND SAFETY
Wood dust can disable and even kill. Really, look it up if you don’t
believe that. If you can smell it, you are breathing it. Larger-scale
woodworkers like cabinetmakers spend thousands on dust control. Before
you start making dust, make quite sure that this problem is addressed.
If you work in a shed or other location suitable to exhausting air directly
to the outside, flexible ducts leading from the tool via a strong blower
(like the Delta) to the outside make the cheapest solution. Whether
you choose that or an expensive vortex filter system, the suction should
be concentrated on the tool you’re using, preferably with a separate
overhead filter box nearby to capture ambient dust. I don’t have
one of those. Instead, I use two hoses: one connected to the tool’s
dust port and one near the work to draw away dust thrown up from the
table. Blast gates can route the suction to different areas of the shop.
A truly correct procedure would include a dust mask, but the ones I’ve
used interfere with my eyeglasses or vice versa.
Never stand edge-on to a spinning cutoff wheel or sanding
wheel. I have tried several different pairs of goggles for eye protection
– the kind with an elastic strap around back of the head. I almost
never used them because they were so fussy and uncomfortable. Now I
have slip-on safety glasses of the kind used by many dentists, and I
use them all the time. They fit over eyeglasses and are easy to wear.
A fine inexpensive product is the Encon 1400 available from Practicon.com.
Never place a foot, hand, or cat under a drive weight.
Weights drop – that’s their job, and sometimes they do it
at the wrong time.