Barometric Compensation

At the clock change, just after midnight on 28 March 2010, a trial barometric compensator was installed. The idea is to correct for the slope of the "going vs pressure" graph. The slope is about -8ms/day per mbar and the main effect appears to be buoyancy caused by change in air density. The objective is to adjust the compensator to get a zero gradient. The theory of a barometric compensator is a bit subtle.
Here is the data from the first 4 weeks since compensation, compared with before compensation:

  After compensation (red lines +1.0ms/day per mbar)

  Before compensation (red lines -8.6ms/day per mbar)
Click on the graphs for the latest, interactive data. It'll take a few weeks to evaluate its performance fully.

Barometric compensator.
Each of the two brass masses weighs 750g.

Compensator as installed early in the morning of
28 March 2010.

Side view of the compensator attached to the pendulum. There's a bit of a wobble on the pendulum - needs straightening out.

Trinity College, Cambridge 2014