[time-nuts] What are the important features to look for in antime sync piece of hardware?

Lux, James P james.p.lux at jpl.nasa.gov
Mon Dec 1 21:11:22 UTC 2008


> -----Original Message-----
> From: Gretchen Baxter
>
> very true!
>
> but looking for a basic feature set.
>
>

Single most important thing.. Is there an accurate description of what the hardware does, what its performance is, and how that is verified?  That's what separates "real" gear from consumer electronics.

This is my standard comment when review articles in the consumer literature talk about things being "lab grade" (usually implying rugged, expensive, precise, or something else.. Maybe lots of knobs?).. My first question: "Does it have a uncertainty statement?"... You could have a foot ruler (scale, really) with 1/64th inch marks on it, but if it doesn't also say "accurate to X % at Y temperature", it's not lab grade.  On the other hand, you could have a foot long hunk of wood that says (1 foot +/- 1 inch) on it, and that would be "lab grade".. If I measured something with it, I could make a meaningful statement about the precision of my measurement (100,000 ppm, roughly). Granted, it's pretty imprecise, but at least it's known.

Take a "atomic clock" (referring here to the things that receive the 60kHz WWVB signal and set themselves).. The consumer gear (which meets many needs.. I have more than one at home) just says "atomic accuracy" and words like that, and might come with a few paragraphs explaining how to set the time zone.  On the other hand, a professional piece of equipment that does essentially the same thing (has an oscillator and clock inside that is disciplined or set by the WWVB signal) will come with a manual that tells you just how much you can expect it to drift when it doesn't get a signal, over what temperature ranges it will work, how much signal strength it needs to function, etc.

So, rather than comparing the consumer clock to the time displayed on your cell phone to tell if it's working, and going, "yep, looks about right".. You could look at the pro gear and say, it's "16:46 on 1 December 2008 plus or minus 10 seconds" and have reasonable confidence that you "trust" the time to the same accuracy over the next days, weeks, or months, and that if it failed, it might tell you that.



More information about the time-nuts mailing list