atomic clocks

I read this a few years ago, but it begs the obvious question. How do they know without an even more accurate clock to gauge it against? So their clock cannot be the most accurate - a paradox.

My wristwatch does once in 24h - 01:00 IIRC.

All the desk clocks I have sync every hour, and all sync on battery replacement.

A few years ago, the UK 60kHz MSF service moved from Rugby to anthorn, loads of people in the South have since been complaining about flaky MSF reception, including me - which is a little odd as I bought all my clocks in Lidl, they're all German made and almost certainly pick up the Frankfurt

"Midnight" is what //your// clock thinks is midnight. It doesn't have to be the least-bit accurate, because the clock will sync at the local time written in the firmware. Then the clock will be accurate.

The "more-accurate" clock is the NIST atomic clock. Once your clock syncs with the NIST clock, it is within 15 ms (or so) of the "absolute" time.

As usual, I disagree.

I beg to differ here. WWVB 60 KHz is vertically polarized with the H (magnetic) field running parallel to the ground. It's a magnetic loop, which works on the H field, not the E field. Pointing the loop up and down does not work. Turning it vertically is just as bad as aiming a bar end directly at Colorado: the signal drops right into the noise. So it is written, so it must be.

Also: "NEW IMPROVED SYSTEM FOR WWVB BROADCAST" (2010???) Starting on Pg 3, it describes the optimum antenna orientation. For optimum reception the ferrite rod should be oriented broadside towards Fort Collins. Also take care not to point the ends of the rod towards Colorado.

I've done measurements with various rod and loop antennas. There's no simple answer to optimum reception. Big antennas pickup more signal, but also more noise, resulting in the same signal to noise ratio as a relatively small antenna. What seems to work best is E field shielding to get rid of locally generated noise, and a high Q loop, to remove interference. However, I don't recall trying a vertically oriented ferrite rod antenna, but I'll see what it does (maybe tonite).

Also, it's quite useful to compare your reception with the official monitoring stations: (If the graphs don't show, fix your Java).

Local time, Ft Collins CO time, or UTC time? Optimum times and durations (dark path) vary depending on location and season: (left coast)

WWVB Radio Controlled Clocks: Recommended Practices for Manufacturers and Consumers (2009 edition) See Section 4.B. Synchronization by Radio at Assigned Times on Pg 12. Table 2 has the local times and duration for various receiver locations. Attempting synchronization on the hour at midnight, 1 a.m., and 2 a.m. guarantees a dark path at all United States

I again beg to differ.

WWVB Radio Controlled Clocks: Recommended Practices for Manufacturers and Consumers (2009 edition) See Section 2A on Pg 11 When a RCC is first turned on, it will begin looking for a signal and attempt to synchronize.

It takes a while for the PLL to find the signal, especially when the receiver is going on and off with the power saver. The algorithm is a bit complex, but it will sync within about 15 minutes if you have a signal. Check out the beginning of the C-Max CME6005 flow chart: Again note that it will try immediately to obtain a signal and sync. With a test generator and ideal no-noise conditions, it will usually sync in about 5 minutes after power on.

Or, just plug in the battery after midnight or when the NIST predicts there will be signal, and it will sync fairly quickly.

My Casio Waveceptor, that will actually sync to different time signals depending on where the watch is, it will check for a signal at midnight, and then if it doesn't sync, keep checking for a few hours on the hour until it does sync, or it stops for the day.

That is actually better than the clocks I have, which generally try to sync about 3 or 4am, and if that fails, doesn't try until the next day at that time.

The clocks and watches all need to be oriented right, I can think of only one time that I had the watch on at the right time where it sync'd up okay.

Michael

More: WWVB Radio Controlled Clocks: Recommended Practices for Manufacturers and Consumers (2009 edition) See Section 3.F. Pg 9. Antenna Orientation Most RCC antennas are directional and achieve maximum gain when they are positioned broadside to the transmit antenna in Fort Collins, Colorado.

I've never had that problem but since I'm probably close to the "limit" of the signal from Colorado, I do find I have to orient the clocks and watches the right way each night.

My watch starts trying to sync at midnight (and if it fails, every hour on the hour till about 4am), but the clocks only try to sync at about 3 or

One weird thing, and it seems to vary with the clock, I find the "sync" button may not work. Maybe it's just because I'm trying it at the wrong time, but late at night I can press the button and it's not sync'd up the next day. Taking out the batteries, and putting them back, does make them sync, three at least, and that gets the right time. It's an odd thing, because I've had some of the clocks go out of sync some times, and they don't resync until I take out the batteries. But they all have sync buttons.

Michael

I've twice had the antenna come loose, and on both occasions, restoring it to its original vertical orientation within the case brought back reception.

Let us know. I'm curious.

I have another atomic clock with an external antenna. I'll set it up so it's vertical, and see what happens over the next day or two.

Well, at the end of the WWVB chain, there is a cesium standard oscillator feeding the station.

Yep. Even the NIST doesn't like the term. WWVB Radio Controlled Clocks: Recommended Practices for Manufacturers and Consumers (2009 edition) See 9.B. Use of "Atomic Clock" Nomenclature Pg 34. ... we contend that use of the term "atomic clock" is technically incorrect and misleading to consumers, and its usage should be avoided. Unless there is actually an atomic oscillator inside the RCC (such as a cesium or rubidium oscillator), we recommend that the term "radio controlled clock" be used to correctly describe the product. Labeling products or documentation with the term "atomic timekeeping" is also considered acceptable.

Good luck stuffing this genie back into the bottle.

Incidentally, the easiest way to check your signal strength is to just cram an oscilloscope into the clock module and look at the decoded data. This should offer some clues as to what to look for: (Note that the loopstick is horizontally mounted.) According to the NIST, to obtain reliable updates, the decoded SNR should be 20 dB or better. That should be easily visible on a scope and what I'll try tonite.

One of my clocks has a "signal strength indicator". There's a symbol on the display, and though it looks more like a microwave dish, the more curved lines displayed the stronger the signal. If there's no indicator, it didn't sync up.

But other than orienting the clock so it does get the signal, rarely do I miss a sync, and that's almost ten years after I got the first atomic clock. It also depends on the clock, the $2.00 at a garage sale one seems finicky compared to the Radio Shack ones.

That said, remember they changed the transmission to some extent. They now transmit a phase modulated signal, an attempt to do away with noise problems "at the extreme points of the transmit area", and I'm not sure how that's worked out, or how common the clocks are that can take advantage of it. And I'm not sure how much, if any, the new method causes problems with old clocks.

Michael

Various station clocks are put up against portable clocks that are referenced to an average of multiple clocks. At least that's how they used to do it. You were not supposed to put two clocks next to each other for long. They would slew together. I used to sync my wind up wristwatch against a cesium standard. I also set my frequency counter against a 1mHz output right off the system.

I like my analog dial radio clock. The hands move real fast when it syncs from switches from dst. Stepper motor hands.

Greg

It's midnight. The clock synced at about 11:30 PDST (7:00 UTC) so I now have enough signal. Yep, that matches the monitoring station signal strength:

However, there's a problem. I don't have a storage scope or DSO handy. The sweep rate needs to be very slow (about 0.5 sec per division) in order to see the 1 baud data. I also have a nasty headache which means I'm not going to setup a better camera tonite. To be continued another evening.

I disassembled an Oregon Scientific, by Integrated Display Co of Hong Kong, "Time Machine". Connections were: +VE +5v but works at +3v. GND Ground DCF Data out PON Ground to Power On I moved the PON wire and soldered it to GND to continuously enable the clock electronics (in the antenna module). I soldered wires to the data out and gnd pads, and reassembled the receiver. The wires went to a scope input. Temporary photos:

When there was sufficient signal, the waveform on the scope was a clean and slow moving 1 baud signal. However, when the signal was weak, there were many more transistions, also known as noise. The difference was very obvious.

When the loopstick was horizontal and perpendicular to the approximate direction of Ft Collins CO, the signal was clean and slow moving 1 baud data. When either end of the loop was pointed at Ft Collins CO, it became quite noisy. It was a very pronounced change, but only over a fairly small (about +/- 10 degree) arc. Pointing the loopstick vertically was pure noise with no visible signal. Like I said, vertical is not going to work (trust me for now).

I'll try to produce some scope photos tomorrow evening. Two aspirin and some sleep first.

Argh. That should be: The clock synced at about 11:30 PDST (6:30 UTC)...

I just took one of these apart, it projects the time on the ceiling, which broke but the antenna is horizontal in the unit, not vertical.

Plus I don't think it has a regular "time" to sync the clock, it like the outdoor thermometer has a radio tower/signal indicator on the display when it's seeking the WWV, and it just seems to come on at random. Maybe once every couple hours.

Maybe it tries to sync all the time and the indicator comes on when it catches the signal or something.

-bruce snipped-for-privacy@ripco.com

I trust you. It's working for me (with the same unit) at the moment. I'll let it sit another couple of days.

Don't worry about it. There's no reason not to believe your measurements.

All my clocks start at midnight and consequently attempt to sync.

What I have noticed is, some you have to manually short the battery terminals to discharge the reservoir cap on the PCB so it doesn't remember the previous time. One clock in particular must have a reverse polarity protection diode, as shorting the battery terminals doesn't do any good. The only thing I can do is leave the battery out for a while to force a resync.

I haven't thought about that, but taht might explain why I have problems resync'ing one of the clocks (actually, part of a "weatherstation".

Michael

All my clocks try to sync on the hour every hour, my Casio Waveceptor tries to synch every 24h - if it fails, it will re try on the hour every hour until 4am.

If the dish tower symbol in the display isn't showing - the time is only as accurate as the onboard crystal and any cumulative drift since the last time it synched.