From: Geoffrey Kolbe
Date: 2017 Dec 28, 05:11 -0800
It was a beautiful clear, cold, day yesterday here in the Scottish Borders and with the sun sliding sliding towards the horizon, the waxing moon was clearly visible to the South West. I took the opportunity to 'take a luner' and wandered out into the field with my little Troughton & Simms pocket sextant and a stop watch. Being small and light and having no telescope to restrict the field of view, taking lunars with these little sextants is quick and easy.
I took a limb-to-limb measurement and started the stopwatch, then went back to the kitchen where there is a radio 'atomic' clock on the wall. At 14:10:00 precisely I stopped the stopwatch and it read 3 minutes and 16 seconds, so the time I had taken the lunar was 14:06:44. I carefully read the tiny scale on the sextant and the observed lunar distance was 104 degrees 16 minutes (zero index error). Then I sat down and calculated what the observed lunar distance should have been at my position (55N15, 02W43). The calculated lunar distance was 104 degrees 15 minutes. Looking at the almanac, I saw that the moon's longitude was changing at a rate of almost exactly 12 degrees a day, so a one minute difference in lunar distance corresponds to 2 minutes in time. I concluded that the radio 'atomic' clock was 2 minutes slow.
So, my question to those who know about these radio clocks - how is it that a clock that is supposed to be accurate to within a millisecond is actually two minutes slow?