I am assuming you are repeating the claim on the wikipedia page. Or maybe (hopefully) you have another source? I saw the claim on the wiki article and my first reaction was bullshit. My second reaction after stumbling across dead links in two of the three the citations was bullshit. After re-reading the article and the one functioning citation without seeing a accuracy/precision/stability figure my third reaction was also bullshit.
The strontium clock at Boulder loses one second in five billion years. [1][2][3]
Does anyone have any information supporting the pulsar clock claim? I have a sneaking suspicion that the author of this claim --just like the catalog in the chair pocket on airplanes--thinks my radio clock that syncs to WWVB/DCF77 is "an atomic clock."
Some atomic clocks are considerably less accurate. I can't find a citation, but IIRC, older cesium atomic clocks are accurate to about 1 second every 300 years.
