Cogging a mill.
I have been interested
in industrial archaeology and heritage for many years, but only now
have the time to really spend some energy on it.
I visited Path Head Watermill with my wife and grandchildren and was
very impressed with the work that had been done to restore the pond,
buildings and mechanism. On closer inspection, it appeared that the
cogs in the drive train were the worse for wear, with several missing.
Having a fully-equipped workshop and become the possessor of a couple
of apple boughs, I stupidly offered to make replacements for the missing
cogs...
Trevor Underwood (whose baby the mill is) informed me that because all
the existing cogs were so worn, it was no use replacing just one; they
all had to be done as a set! My wife immediately said that this was
'no problem, my husband can do that easily'! Doh!
Only later did the scale of the task become really apparent...
I obtained a couple
of the existing cogs to use as patterns and produced two sample cogs
- one for each of the large drive gears.
It was glaringly obvious that they were woefully inadequate; the teeth
were too wide, not deep enough and the wrong profile; the tenon/root
was flapping in the mortice; and the locking taper angle was far too
large.
In addition, I was using valuable timber for experimentation. So ...
I decided to take
proper measurements, draw up proper plans and make trial cogs out of
pine.
I purposely made the tenon oversize so that the cogs could be made to
fit snugly, made the tooth profile a proper involute curve and reduced
the undercut on the locking taper section. Success!
Until an over-eager blow with the mallet detached a section of the overhanging
tooth.
This highlighted a drawback in the construction; whilst the grain runs
in the right direction for the tenon and taper sections of the cog,
the tooth really needs to have the grain running along the width of
the face to prevent the tooth corners being easily broken. Enter Grand
Designs from More4...
Trevor mentioned
the use of new materials during the construction of a modern Viking
longhouse, which set me thinking about making composite cogs using modern
methods and materials.
Why not use the expensive and rare apple wood for the tooth, kiln-dried
pine for the root section, and attach them using steel screws and high-performance
adhesive?
This would have several advantages:-
• The grain can run in the optimum direction in both face and
root of the tooth, which means less warping.
• Stronger & longer face means better load distribution and
less wear.
• More economical use of valuable timber - 8 to 12 teeth from
the same amount of wood that made only 2 previously.
• Cheap or recycled timber can be used for the root.
• Softer root timber absorbs shock better than the dense fruit
wood which performs better as a facing material.
• Ease of construction - lends itself to mass-production methods,
which increases accuracy and hence longevity of teeth.
• If a tooth breaks or wears prematurely, it can be replaced quickly,
by unscrewing it from the root without having to extract the whole tooth
and neighbouring wedges.
• More screws may be added if deemed necessary in highly-loaded
applications.
Purists may gag
at this 'heresy', but, as most millwrights understand, any improvement
to performance or maintenance will be eagerly embraced - as it has been
for centuries.
The 64 cogs for
the upper bevel gear were made and installed. The pinion was dropped
into position and shimmed to mesh correctly then the water turned on
- nothing! The sluice was opened more - still nothing, until with a
sudden lurch, the whole mill exploded into a frenzy of noise and vibration.
The gears went into overdrive, just a blur of speed while the sluice
gate was frantically closed.
Only then did it become apparent what the problem was:-
In meshing the new cogs, a pulley on the pinion shaft was brought into
close (very!) contact with the cross-beam, effectively acting as a disc
brake! This meant that excessive power was required to start things
moving, but once they did, it was at top speed. A bit of work with a
chisel created enough clearance and things resumed as normal.
Once the cogs had bedded in, it became clear that the inner tips of
the teeth were bottoming, while there was quite a gap at the outer tips.
Oops! I had made a classic error when drawing the teeth; I drew the
originals based on an actual sample, then modified them when I decided
to go composite, so extended the face by 40mm, but forgot to modify
the profile, so the inside depth was about 4mm too big, and too wide.
An hour's work with a chisel and spokeshave had everything approaching
the optimum. More adjustment of the pinion mesh and things were much
smoother and quieter (still needs a bit of work there I feel).
Now to the great
spur gear cogs. I managed to sever the top of my left thumb during the
cutting of the teeth, which put me out of commission and things on hold
for several months, but once the sensation came back I finished all
80.
However, during construction, I was not happy with one of the cogs,
so I tried to separate the two parts with a hefty blow from a club hammer,
and again ... and again! No way would it part, proving the claim that
the glue is 'stronger than the wood' and giving a certain amount of
confidence in the composite construction theory.
They were installed in about an hour and a half. When the water was
turned on, everything went smoothly into action, with very little fuss,
noise or vibration. An inspection showed that some of the teeth were
already starting to get chewed up! That was obviously not right, so
an investigation revealed the culprit: - in its previous life, the stone-nut
(pinion) had been driven in the opposite direction, so the face that
was now being driven had never been used before and was as rough as
when it was first cast, whereas the non-driven face was polished to
a shine!
Out came the angle-grinder to smooth up the rough faces and create a
proper gear tooth profile. On the next operation, the ragged wooden
tooth surfaces were smooth and actually becoming polished, so things
were definitely improved. This also highlighted a problem in the gear
meshing, the tops of the wooden teeth were in mesh, but the bottoms
were rarely, if ever. Adjusting the pinion bottom bearing sorted that
out and now most of the teeth were in mesh over the full face.
Next I realised that the pattern I had been given to draw the teeth
was too tall to mesh correctly.
So ... remove 160 screws, plane off 6mm from the tops of all the teeth,
re-countersink the screw-holes, and replace the screws. Then adjust
the pinion for correct mesh, turn on the water and ... joy unconfined!
The teeth were mating across the whole face, meshing deep into the root
of the teeth and just kissing the back face on exit, perfect!
Also, the gouging (that had started to appear at first) was now smooth,
with no evidence of any very bad meshing and very quiet; in fact the
sound was like a room of pensioners eating dinner - the clacking of
false teeth. Better than I had dared to hope!
Once again, I had
made the classic mistake of accepting a pattern as being correct, without
going back to first principles and measuring the correct data - but
we live and learn, and hopefully don't make the same mistake twice!
The upshot of the
experience is that the theory of composite cogs seems to work in practice,
and they are much faster, cheaper and easier to produce than standard
cogs.
A
labour of love by Tony Watson
