BLEAKFEED · all stories

I Wind a Church Clock and It Gains Four Minutes a Week on a Pendulum That Is the Right Length

Read in episode 039, Deranged Countryside Horror Stories, from 26:26

A pendulum clock keeps time by the length of its pendulum and nothing else, and ours is measured and has not been touched. It gains four minutes a week, checked against the radio time signal, and it has gained the same four minutes since the 1890s, on three different pendulums.

I wind the church clock in our village and I have done it for twenty two years, after my uncle, who did it for thirty.

How a turret clock keeps time.

The clock is driven by a weight on a cable. The pendulum swings and lets the escapement release one tooth of a wheel each swing, and the length of the pendulum sets how long each swing takes. A pendulum of a certain length swings at a certain rate, and the arithmetic for it is three hundred years old and has never been wrong.

To regulate it you screw a nut under the bob up or down a fraction, which shortens or lengthens the pendulum by a hair, and the clock gains or loses a few seconds a day. That is all there is.

I check it against the radio time signal every Sunday when I wind it.


It gains four minutes a week.

Not roughly. Between three minutes fifty and four minutes ten, every week, with the pendulum where it is.


Four minutes a week is a pendulum about a centimetre too short, and a centimetre is enormous. You cannot miss a centimetre on a regulating nut and my uncle could not have missed it either.

So I measured the pendulum. From the suspension to the centre of the bob, and I did it with a tape and then with a surveyor's staff, and it is the length the clockmaker's plate on the frame says it should be, to within two millimetres, and two millimetres is a few seconds a day and not four minutes a week.


What has been done, over a lot of years.

Regulating it. You can, of course. Screw the nut down until it keeps time, which puts the pendulum a centimetre longer than its design length, and it keeps time for about a fortnight, and then it is gaining four minutes a week again with the nut where you left it.

I have a book of that going back to 1961, in three hands. Every one of us has regulated it and every one of us has watched it go back.

The pendulum itself. Replaced in 1978 when the old rod was found to be cracked, and again in 2009 when the suspension spring went, and both times the clockmaker who fitted it set it up, timed it in his workshop against his own regulator, and brought it here keeping time.

Both times it was gaining four minutes a week within a month.

The escapement, the wheels, the weight, the cable. A worn escapement will make a clock gain, and a clockmaker has looked at ours three times and found it in fair condition for its age and not the cause.

Temperature. A pendulum rod expands in warmth and a clock loses in summer and gains in winter, and this is real and it is why good rods are compensated. Ours is not, and it gains its four minutes in July and its four minutes in January.


The thing in the book that I did not notice until I read it all the way through.

The gain is written down every week, back to 1961, and before that on loose sheets my uncle kept from his predecessor.

It is not exactly constant. It creeps. In 1961 it was three minutes and forty seconds. It is four minutes and five now.

Twenty five seconds in sixty years. A clock is not supposed to change its rate at all, and one that does changes it because something wore, and a thing that wears goes one way and then gets fixed. This does not get fixed. Three pendulums, two suspensions, one escapement overhaul, and the number has walked steadily upwards through all of them as if none of them had happened.


What we do about it.

I put it back four minutes every Sunday, which takes a moment, and the village has a clock that is right on Sunday morning and four minutes fast by the following Saturday night, and it has had exactly that clock for as long as anybody alive can remember.

People set their watches by it. I have told them not to. They do anyway.


What I do.

I wind it, I check it, I write the gain in the book.

The one thing I have changed is that I do not put it back at the frame.

You can set the hands from the clock room, at the frame, by releasing the motion work and moving them, and that is what you do. Or you can stop the pendulum for four minutes and start it again, which is the lazy way and is exactly as good.

I stop the pendulum. I hold the bob, I look at my watch, and at four minutes I let it go.

I started doing that in 2014, after a Sunday when I had the motion work released and my hand on the minute hand, moving it back, and it moved forward against me. Steadily, at the rate the clock runs, with the pendulum stopped and the train locked and my hand on it, for what I counted as four minutes, and then it let me.

Also in this episode