Skip to content

Do you really want to know what time it is?

O-oscillator-1

There is an old saying that has become known as Segal’s law: A man with a clock always knows what time it is. A man with two clocks always wonders what time it is.

This is a nerd post. Sorry, non-nerds.

Normal people, I’m going to guess, are perfectly OK if their clocks are within a minute or two or three of the correct time. To nerds — and to much of the technology that you use every day — even a single second is an eternity.

Let’s look at a little history.

When I was a child, the two most accurate clocks that I can remember were the clock on our kitchen stove and the electric clock that sat on the shelf above my grandmother’s rocking chair. Because I was a nerd child, I found out how those clocks kept time. They did it by counting the 60 Hertz cycles on the AC (alternating current) power line. Even in the 1950s, power companies had pretty good methods of keeping the alternating current that powers our homes alternating at a stable rate. So, to be accurate, those electric clocks only needed to be able to count those 60 oscillations per second on the electric power line with a synchronous motor.

Time and frequency are very closely related. You can confidently say that something oscillates so many times a second only if you know, very precisely, how long a second is. By 1920 or so, our technologies had reached the level at which the accurate measurement of time and frequency — and the coordination of time and frequency from one place to another — became very important to economic and technical development. One of the first ways of solving this problem was with special radio stations operated by the National Bureau of Standards. Shortwave radio listeners were very familiar with these broadcasts: “At the tone the time will be, X hours, Y minutes, Coordinated Universal Time.” These broadcasts continue today, though GPS has now made the broadcasts largely obsolete. We’ll talk about GPS in a second.

Nerds knew (though the general public didn’t care) that those radio broadcasts were disseminating not only a time standard but also a frequency standard. If you have a shortwave receiver than can be tuned to 10 Mhz (Mhz = megahertz, or a million cycles per second), you’ll hear a click once per second and a voice announcing the time once a minute. But the frequency of the carrier signal also is very precisely controlled. The frequency of the broadcast is as close to precisely 10 Mhz as modern technology can get it, and that’s pretty close. So if you had a need to accurately measure frequencies, you could tune some special (and very expensive) radio equipment to these radio broadcasts and use the frequency of the carrier signal as a 10 Mhz frequency standard. Other expensive equipment, called frequency counters, if given access to a highly accurate 10 Mhz reference signal, could accurately measure all other frequencies.

Do you remember in the 1980s when “quartz” watches became a big deal? Quartz oscillators have been in use at least since the 1920s, but in the 1980s it became possible to make quartz oscillators small enough and cheap enough to fit inside a watch. Oscillator is a nerd word, but it refers to a simple circuit that produces a sine wave (alternating current) at a certain frequency. When you turned the dial on an old radio, you were changing the frequency at which the receiver oscillates. You were tuned to a station when the receiver oscillated at the same rate as the transmitter of the station that you wanted to listen to. The receiver used a variable frequency oscillator, because turning a dial changes the frequency.

A quartz crystal oscillates at a known and fairly stable frequency. Let’s say that there’s a quartz crystal in your watch oscillating at a frequency of 32,768 cycles per second. An easy circuit to build into a watch is a “divide by 2” circuit. If you send the signal through a “divide by 2” circuit, you get half the frequency, or 16,384. Divide it again and you get 8,192. If you do this division 15 times, you get a signal that is pulsing at once per second. Use that signal to move the second hand, and you’ve got a watch.

Though quartz watches are about ten times more accurate than mechanical watches, they’re still far too inaccurate for demanding requirements. Over time they drift, probably by seconds per week. Oscillators based on the properties of rubidium or caesium are more accurate than quartz, and both are used in expensive time and frequency equipment.

But, these days, how does our technology handle the need for highly accurate time and frequency at a reasonable cost? These days it’s done with GPS.

How GPS works is fascinating in itself, but let’s save that for another day. The important thing is that for GPS to work, the GPS satellites must send an extremely accurate time signal to the GPS receiver — your smart phone, for example. If your GPS device displays the time, you can count on it to be highly accurate.

These days, anyone who needs an accurate time and frequency standard uses special GPS receivers. Cell phone towers need both accurate time and frequency. Electric power generating stations now use GPS timing to coordinate the “phase” of the current on the power grid. The computer systems used by banks, or by stock trading systems, require accurate timing and are using GPS references. Most 911 call centers now use GPS references to ensure accurate time records, which are required by regulations.

Nerds like me often have apparatus at home for accurate time and frequency measurement. The cheapest way to go is to buy on eBay equipment that is considered obsolete by commercial users but which still works perfectly well.

In the top photo, a GPS receiver is tracking satellites and displaying the time. The output labeled “10 Mhz” is carrying a 10 Mhz sine wave “disciplined” by the GPS time signal so that the 10 Mhz reference signal can be trusted to be highly accurate. In the lower photo, that same 10 Mhz is being sent to my Hewlett Packard 5335A frequency counter, which is saying, “Yup, that looks like 10 Mhz, when compared with my built-in reference oscillator.” However, the GPS 10 Mhz reference signal is more reliable than the frequency counter’s internal oscillator, though that Hewlett Packard internal 10 Mhz oscillator is very good. The frequency counter has a plug in the back for a 10 Mhz external reference signal. So if the cable from the 10 Mhz output on the GPS device is plugged into the back of the HP frequency counter, then the frequency counter will use GPS as its frequency reference rather than the less accurate internal oscillator.

So, if you were at my house and wanted to know what time it is, I might suggest that you ignore the clock on the kitchen stove and go look at the GPS clock. It will be right within a millionth of a second, and that’s good enough for a nerd, most of the time.

O-oscillator-2

9 Comments

  1. M. K. wrote:

    I am by no stretch of the imagination a nerd but found this post fascinating. Now I know where to get the most accurate time in our neck of the woods.

    Wednesday, July 6, 2016 at 7:20 pm | Permalink
  2. Phil wrote:

    Not sure if I’m drawing a correct conclusion, but does your post imply that the time on our cell phones is very accurate?

    Wednesday, July 6, 2016 at 9:01 pm | Permalink
  3. daltoni wrote:

    Hi Phil … The time on your cell phone probably is coming from an NTP server over the Internet. NTP = Network Time Protocol. Still, NTP is pretty good and should be off no more than a largish fraction of a second.

    Apps on your cell phone that get their time from GPS satellites rather than NTP would be more accurate.

    A device such as a Garmin GPS device would have no source of time other than GPS, so it should be very accurate.

    The differences in accuracy here are likely to be too small to be perceptible by a human, but to machines that can accurately split time to milliseconds or nanoseconds the differences could be large.

    Wednesday, July 6, 2016 at 9:32 pm | Permalink
  4. Phil wrote:

    Very interesting… Thanks!

    Wednesday, July 6, 2016 at 10:30 pm | Permalink
  5. Henry wrote:

    Fascinating Trivia…what about atomic clocks? Isn’t that a standard or protocol too?

    Friday, July 8, 2016 at 7:25 pm | Permalink
  6. daltoni wrote:

    Hi Henry… Caesium or rubidium oscillators are atomic clocks. Both types have been used in GPS satellites for accurate time keeping, plus the clocks on board the satellites are monitored and regularly adjusted from the ground. Atomic clocks are always evolving, and new GPS satellites are regularly launched. I think that newer GPS satellites will use “hydrogen maser” atomic clocks. Scientists are constantly working on more accurate clocks, and GPS satellites are among the most critical places for their deployment.

    See: http://theconversation.com/sharper-gps-needs-even-more-accurate-atomic-clocks-38109

    The cool thing about having GPS-based time and frequency equipment is that we nerds don’t have to have such fiercely expensive clocks. GPS, which is relatively inexpensive, takes advantage of the best time-keeping standards available.

    Friday, July 8, 2016 at 7:52 pm | Permalink
  7. Henry wrote:

    Thank you for the explanation. I’ll be sure to go to the “conversation”

    Saturday, July 9, 2016 at 12:03 pm | Permalink
  8. Henry wrote:

    I meant to ask regarding GPS – why do some folks get lost, especially in the desert?

    Saturday, July 9, 2016 at 12:06 pm | Permalink
  9. daltoni wrote:

    Hi Henry… People lost in the desert are more likely to be a map error than a GPS error. Maps of low-population areas or rarely traveled roads are often wrong.

    GPS errors are rare but not unknown. Back in February, GPS clocks some were wrong by 13 microseconds, which caused all heck to break loose in some places:

    http://www.bbc.com/news/technology-35491962

    Saturday, July 9, 2016 at 3:24 pm | Permalink

Post a Comment

Your email is never published nor shared. Required fields are marked *
*
*