[time-nuts] Chronometer contest sponsored by IEEE Spectrum

John Ackermann N8UR jra at febo.com
Thu Nov 29 15:30:27 EST 2007

David Forbes wrote:
> At 10:42 AM -0500 11/29/07, p.ross at ieee.org wrote:
>> IEEE Spectrum, the magazine of the Institute of Electrical and Electronic
>> Engineers, is issuing a D-I-Y challenge that may be of interest to the
>> members of this group. The goal is to build the most accurate possible
>> digital clock for under $100, accuracy to be measured before disciplining
>> with GPS, WWVB or other external aides. It is required that all parts be
>> generally available; for instance, you can use an OCXO (by all means!) but
>> only if it?s currently available on the open market. Display should be
>> seven-segment LEDs?we?re looking for a usable appliance, not just a
>> science project. Winner?s photo and project will be published on the
>> magazine, its Web site, or both.
> Phillip,
> That sounds like a really fun DIY project. However, this group could 
> spend 3 weeks debating what is and isn't OK to do. I hope there are 
> more rules than that!

Building off of David's comment (and I know there have been some others 
since his...):

Cool idea that it is, I think Spectrum needs to be really careful in how 
they define this contest.  As David noted, the folks on this list are 
capable of arguing endlessly over extraordinarily fine details, and 
without carefully defining what "most accurate possible" means, more 
heat than light will be generated.

Also, the requirement for LED displays is really inconsistent with the 
design goal; something that uses the eye as a measurement trigger is 
more limited by the interface than the resonator -- unless you're 
planning to measure the accuracy over a period of years.

To get meaningful results, at a minimum you'll need to specify (a) the 
temperature range over which the clock has to operate, and (b) the time 
period over which the measurement will be taken.

Allowing commercial OCXOs will take away most of the challenge, as that 
one element contributes almost all of the clock's frequency stability 
and accuracy.  I think that requiring that the oscillator start with a 
discrete crystal (or other resonator), and that all temperature, 
voltage, and other environmental control mechanisms be designed and 
built by the submitter, would make for a really interesting competition 
that encompasses both electrical and mechanical design (and craftsmanship).


More information about the time-nuts mailing list