Time is a frequency

How I Became A Time-Nut

John Ackermann   N8UR  

Everyone knows that time is money.  But did you know that it’s also frequency?  Time and frequency are different ways to look at one of the fundamental dimensions of the universe (to be more precise, the ferquency of a radio wave is defined by the number of its cycles per second, and the second is defined by the frequency of a specific atomic interaction – mathematically, time and frequency are inverse measurements of the same thing).

 

Ham radio is all about frequencies – we need to know where to tune our radios in order to find each other’s signals, and to stay within the bands where we are allowed to transmit.

 

From my earliest days as a ham, I was always interested in frequency measurement.  I remember that one of the hams in our club had an “Eldorado” frequency counter with orange nixie tubes for its display.  I was fascinated by that box.  A little later, I built a much less fancy counter from a kit, and then started bringing home piles of test gear from hamfests.  I still have the first “frequency standard” (very accurate crystal oscillator, with ovens to keep the workings at a constant temperature) that I found back in about 1985.  I used it with a military surplus counter, that like the Eldorado had nixie tubes, but had even more knobs and buttons (and its fan made a lot more noise).

 

Once you have a frequency standard, you need a way to calibrate it.  This is where radio comes in.  Of all the fundamental quantities (like length, mass, etc.), frequency is the only one that can be measured at a distance, via radio.  In the US, the National Institute of Standards and Technology (NIST) broadcasts standard frequency and time signals with a very high degree of accuracy.  By receiving these signals you can calibrate your own equipment.

 

So, my flea market searches expanded.  I already had my ham radios that would receive the WWV time and frequency signals, but NIST has another station that transmits on a much lower frequency; for various reasons, that can be measured several times more accurately than the WWV signals.  So I dragged home a 40 pound box that contained a WWVB receiver.  It plotted the difference between my frequency standard and the NIST on a mechanical chart recorder that made a resounding “clank” every 10 seconds.  My wife loved it…

 

My second flea-market frequency standard had a beautiful electric clock attached.  Now, I could directly see how time was related to frequency (frequency is defined as the number of radio cycles per second), and I started figuring out how to set my clock as accurately as possible, as well as keeping it running at the right speed.

 

It would be boring to relate all the buys, sells, and swaps from that point on.  I’ll just say that eBay significantly contributed to the filling of my basement!

 

While my interest in time and frequency started with the simple desire to be able to measure my ham transmitter’s frequency, it led me into nooks and crannies of precise measurement as a hobby in and of itself.  I’ve become a “time-nut,” one of those people for whom measuring to trillionths of a second is a fascination.  There are over five hundred of us on the “time-nuts” mailing list (https://www.febo.com/mailman/listinfo/time-nuts).

 

One of my fellow time-nuts, Tom Van Baak, did perhaps the best “science fair” experiment ever.  Tom has some of the best clocks in the world in his collection, and he decided to see if he could prove that Einstein was right about that relativity stuff.  According to the general theory of relativity, the “speed” of time is affected by gravity.  If a clock on the earth gains altitude (moves further away from the earth’s gravity), it will run fast and gain time.

 

In his lab near sea level, Tom measured the time of three extremely accurate atomic clocks.  He then put those clocks, along with his three kids, into the family van and drove to nearby Mt. Rainier, 5400 feet above sea level.  They spent two days there and then returned home.  Einstein’s theory predicts that, given the time spent at altitude, the clocks would gain 22 billionths of a second during the trip.  When Tom compared the traveling clocks to others that had stayed at home, he found that they had gained, on average, 23 billionths of a second.  Tom and his kids proved that Einstein was right!  You can read about Tom’s experiment at http://www.LeapSecond.com/great2005/index.htm.

 

I haven’t done anything quite as exotic as Tom’s experiment, but I was able to capture data that confirmed the impact of a powerful solar flare.  On December 5, 2006, a large spot on the sun exploded, creating one of the strongest solar flares in the last 25 years.  I was recording data from a GPS receiver at the time (GPS not only tells you where you are, it can also serve as one of the most accurate time signals available), and recorded a “blip” in the strength of the GPS signals that coincided directly with that flare.  A story about the solar flare is at http://science.nasa.gov/headlines/y2007/22feb_nosafeplace.htm and my data is at http://www.febo.com/pages/gps_solar_flare/index.html.

 

Since I am a computer as well as radio geek, my interest in time led me to explore the Network Time Protocol (http://www.ntp.org) which is used around the world to keep computer clocks synchronized.  I started running my own NTP time servers, and soon found myself unhappy with the performance that a standard PC can provide.  Another time-nut, Poul-Henning Kamp, found that an inexpensive single-board computer with no hard disk had some special features that allowed it to serve time more accurately than any standard PC.  However, even that box would benefit from some hardware changes to give it a more stable clock.  That led me to design two products that are now being sold by TAPR (http://www.tapr.org): the Clock-Block frequency synthesizer (http://www.tapr.org/kits_clock-block.html) and the FatPPS signal conditioner (http://www.tapr.org/kits_fatpps.html).  The end result is perhaps the most accurate NTP server hardware in the world.  My web page at http://www.febo.com/pages/soekris/index.html describes that hardware, and http://www.febo.com/time-freq/ntp/stats.html shows how well it works.