"N_Cook" wrote in message
news
More interesting than the other stuff i should have been doing.

From 24 inch circumference of the recording drum I make the key data
for that top image 0.0868 inches per pixel, vertical and horizontal,
as circular wheels. The clocks were 10 day , allowing 7 days of
continuous recording on one sheet of paper, much overlapping of
about 19 tide cycles, before having to change paper and re-synching.
Hence accurately inferred diameter dimension there.
I reckon, from 20 foot of tide range, 7 turns of tape ,then from the
following, 1 1/3 inches wide and 0.083 inches ,probably 5/64 inch,
thick tape.
The spigotted pulley , the measurement-active bearing face , between
11.08 in and 11.13 in diameter.
So for a jump between 2 and 2.5 inches, if perforations spacing 0.75
in between centres, then 46 spigots and jumped 3, 2.25 inches,
pulley diameter 10.98 in.
If 1 inch spacing, 35 spigots, jumped 2 , or 2 inches, pulley diam
11.14 in
If 1.25 in spacing ,28 spigots, jumped 2 or 2.5 inches, diam 11.14
in
If 4/3 inch spacing , 26 spigots, jump of 1 or 2 is outside the 2
to 2.5 inch range, but if so, then pulley 11.03 in diam .

For the tape width appearing to the left wide side on, from the
lower image , ratioed pixels to the diameter of the suspension band
counterweight and then used the upper image to get 1.33 inches.
The lower image shows the recess in the rim of the upper spigotted
pulley, assumed to have one thickness of tape on it.
From the twisty looking appearance the tape band is broken or cut,
and just draped through near and to the rear of the lower pulley,not
wound on it , same diameter as the upper one ,same recess , but no
spigots on that one.

The failure mechanism , prior to jump, as a band suspension.
12:1 gearing stops, so spigotted pulley stops rotating, the float
continues rising. Band disengages from pulley, counterweight takes
up the slack 1 ,2 or 3 times, until the float goes back down again.

Presumably the band was cut from sheet stock of a commercial
thickness. For such a low production count it would be hard to justify
not buying standard components wherever possible.

(36" / pi) - 0.08" gives 11.38" for a diameter that would move the
neutral plane of the band 36" per revolution. If the camera wasn't at
the level of the wheel its apparent height would be foreshortened but
the round drum's diameter wouldn't.
-jsw