Atomic Time

[Joseph]

Hello all,
I have had an atomic (table) clock for about 6 months now. At first it needed to be turned toward Colorado each night to pick up the signal. Now, I can leave it facing south and it picks it up. Besides the possibility that the neighbor across the street has one and so now mine is picking up her RF amplified signal, has the NIST fired up any new transmitters in the US lately?

[Will]

Joseph,
I've had one of those about four years now - I live some 1200 miles or so from the Colorado broadcast site. During the first few weeks of using the device I found that it would work in some parts of the house and not others and at different times working and not working in the same place. My clock picks up the signal within a few minutes of being energised but then only switches on it's receiver daily at midnight - that's how I tracked it's performance - I watched it carefully at mifnight for a few nights.
I assumed part of the problem to be varying atmospheric attenuation and so I hung six or seven feet of wire from the ceiling as an antenna and ste the end of it under the clock - it has worked without interruption ever since (Which is more than I can say for myself !)

[dyarker]

It is VERY unlikely that another receiver across the street would have any effect.<p>www.boulder.nist.gov/timefreq/ doesn't mention any new transmitters.<p>The average height of the ionosphere changes from summer to winter. Right now you're probably getting a direct ionosphere bounce instead of a rebounce off a mountain or city building or something; or reverse of that. Expect reception to get worse again by the end of the month.<p>I like Will's solution.

[Dean Huster]

Dale, the ionosphere explanation would be perfect for the HF WWV transmissions, but I think that most of the clocks being sold now use the WWVB signal at 60KHz which has little ionospheric effects. <p>I can't tell what my RS atomic-controlled clock uses, but I suspect it's the 60KHz transmission because of what it's little external antenna looks like.<p>Will, I guess I never considered the fact that the clocks didn't constantly monitor the time signal. I suppose that they would just do a spot check to conserve power. Running that receiver 24/7 would eat up the batteries. Makes me rethink the possibility of tapping into the clock's oscillator for a phase-locked frequency standard.<p>The old Heathkit "Most Accurate Clock" used the HF transmissions and it was moody as all heck, sometimes taking overnight or two or three days to lock on and display time, even with an external antenna.<p>Dean

[dyarker]

A telescoping whip antenna would be HF, a ferrite bar with wire wrapped around it for an antenna would be 60KHz as you say.

[Joseph]

What had been happening is that since I can see on the LCD display when it is trying to receive the 60khz, I would face the ferrite rod toward Colorado. I figure that it is positioned horizontally inside the unit--the standard way. A couple of nights in a row recently, when I turned it, it did not succeed. But the next couple of nights I did not turn it and it did get the broadcast. Now, last night, without being re-oriented, it failed once again.<p>Dean, back in the '70's, I used to compete with my brothers in DXing commercial AM stations. We discovered that skip always happened to some degree every night, but varied. My brother reached our ultimate goal by picking out the call letters of a West Coast station (from the Ohio valley region).<p>
Dale, when I checked out that Web-site, I had wondered if it had not been updated. Your input resolved that question now. I believe that a few thousand meter wavelength was given for the 60khz signal. One thing I have always thought is that AM broadcast signals are absorbed by the ground. If multiple skips happen at night, then that idea was wrong. Maybe something as big as a mountain could reflect them too.<p>Will, I like your external antenna idea too. I am thinking of trying my old phone line which runs in the correct direction, N-S, already. <p>A couple of weeks ago, I used an idea I saw here and ran some scrap cat5 as a new line to the telephone pole wire feed junction, disconnecting the old network of modular lines in the process. My phone connection is much improved now.

[Dean Huster]

Joseph, I, too have spent a lot of long hours chasing BCB DX -- been doing it for over 40 years, in fact! The AM BCB does skip terribly, as you say, which is why we can hear clear channel WSB, WWVA and KOA in Poplar Bluff -- at night.<p>However, at 60 KHz, you're not getting the skip. You're getting the ground wave over the entire lower 48. There are some minor propagation effects caused by the ionosphere, not much, but enough that calibration labs aways checked quartz standards against WWVB over several days to average out the variations.<p>You do have to compensate for the propagation delay, however, if you want deadly-accurate time reference for calibration your own clock to the nearest few microseconds or better. It does take time for the "time tick" to finally make it from Ft. Collins to your receiver site. For our off-the-shelf radio-controlled clocks, that's no big deal. Most don't deal with that issue at all, worrying only about what time zone they're in. The Heathkit "Most Accurate Clock" did have a DIP switch on the bottom so you could set to the nearest available setting for that delay, giving you about 10 ms precision against WWV (5, 10 or 15 MHz).<p>If you do have a "system" clock, such as the old HF time standards using Sulzer or AN/URQ-10 quartz standards driving a digital clock and generating time signals to send throughout the system, you have to adjust your clock to "tick" so many hundreds of microseconds before the "tick" you heard over the HF WWV signal to compensate for the propagation delay. The local clock "tick" triggered an oscilloscope and you adjust the system so that the beginning of the 1000 Hz burst that comprises the WWV "tick" began the proper amount of time down the screen.<p>GPS-referenced standards are cesium-accurate, but there was always something really neat about a "manual" system clocked against WWV or WWVB. I designed such a system back around 1975, but never built the thing! Even had the circuitry, oven and crystal for a highly-stable crystal oscillator.<p>Dean

[dyarker]

Does 5000 meters ring a bell?
Speed of light in vacuum = 300000KM/Sec = 3.0e8M/Sec
Period of 60KHZ = 1 /60000Hz = 1.6666666667e-5Sec
Wavelength = 1.6666666667e-5Sec * 3.0e8M/Sec = 5e3M = 5000 meters<p>Cheers,

[Joseph]

Dean, does it mean that the 54-1600KHz band does not go far during the daytime because the sun energizes the ionosphere, causing it to change? Then the signal is simply interfered with? I remember that the ionosphere changes altitude between night and day, but I can't remember which way it goes.<p>Dale, 5000 meters makes sense. That is very long as far as radio waves are concerned.<p>[ September 12, 2004: Message edited by: Joseph ]</p>

[Dean Huster]

Actually, the ionosphere doesn't move up and down in altitude as the Earth turns. There's actually several layers to the ionosphere, each of which is activated or intensified as that area of the Earth turns toward the sun. Some of the lower layers are the most intense at noon in your area. A radio wave, like a flat rock on the water, has to be at a shallow-enough angle to skip off the layer and back to the Earth. These lower layers tend to cause too steep of an angle during the daytime and the radio wave just passes through rather than skipping, which is why we only receive the ground wave.<p>Lower frequencies are absorbed by the ionosphere while higher frequencies, up to the maximum useable frequency, skip. The best skip is at a frequency just below the maximum useable frequency. The VLF signals are pretty much absorbed, leaving the ground wave. As you may know from the U.S. Navy submarine communications, the lower frequencies tend to actually penetrate the Earth a bit. Of course (the skipping rock thing again), if you have a transmitter with a beam antenna, you can angle the signal so that it punches right through for no skip at all or angle it to alter the skip point just like a carom on a pool table. It's a pretty iffy science when it comes to antenna aiming.<p>Actually, most of the ARRL Amateur Radio Handbooks have good chapters on skip, the ionosphere, the 11-year sunspot cycles, layers, absorption, maximum useable frequency, etc.<p>Dean

[Edd]

ATOMIC_CLOCK_MUSINGS<p>Noticing the comments to Atomic Clock correction referencing receivers, I cued on the fact that my first generation Oregon Scientific unit had been showing its Lo Batt sign for months so I opted to explore its inner workings at battery replacement time. This unit utilizes a separately housed , orientable receive antenna rod element with the unit located at the end of a 5 ft length of RG214 size cabling. The antenna’s housing resembles a barbell in construction, but with only one “weight” centered on the bar but with that weight rotated 90 degrees to make up its profile. That odd / enlarged profile would not be necessary to the covering of a small center winding within the ≈ 5in X 1/2 inch overall housing dimensions. A cursory examination of the housing initially suggested an ultrasonically bonded longitudinal seam, however, a fingernail flexing revealed an opening of the seam. So an utilization of a 60 degree re-sharpened hi-den polyeth putty knife, along with toothpicks inserted to hold the opened seam in check, quickly opening, yet retaining a cosmetically pristine housing. Internally there was located a 3/8 in diaX 4 ¾ in L powdered iron/ferrite rod…..heavy on the ferrite at this low freq utilization. One quarter from its end was wound a three layer coil with # 30 wire, 40T+30T+20T for 90 total turns placed upon a wax impregnated sliding cardboard form with its final positioning / tuning being ≈ 1/4 from the rods end. The whole rod being mounted within the housing with double stick foam tape.
Moving over to the main receiver proper revealed a very nice 3 X 3 ¼ PCB with its green resist and gold flashed terminals and plated feed thru vias. Initially component inspection revealed very, very little on the board. A set of three micro/SOT/flat pack xstrs providing power amp interfacing to a mini ¾ in mylar diaphragm coned/voice coil design (50 &#8486 alarm speaker). The PCB’s fabricator's marking revealed a code date of 03 week of 2001….my initial 2x AAA cells I had dated at 8/2/01…so about a 3 yr lifespan was had. Down at the board bottom was a mini ½ in sq ferrite cored xformer..3 terms + 2 terms with 2 micro xstrs feeding one of its windings with its other winding going towards the front large LCD displays….2 side terminals. No massive electronics visible yet…moving around to the front of the PCB revealed 35 gold fingers interfacing to the large LCD unit….straightening 4 alum bezel fingers and pulling the LCD display away reveals the units custom electronics hidden behind it on the pcb……a micro flatpack Chip On Board design with about a ½ inch flat blob of black Hysol/Epoxy encapsulant. Possibly a 40 pin breakout…..obscurity precludes a positive count. Fully half of the ic’s top conns route upwards to interfacing with the display unit and the bottom half route off to: the snooze switch, the speakers audio amp cktry, the front adjustment p-buttons and the driver circuitry for the small transformer.
I found the ½ in xformers sec leads going to an electro luminescent panel on the back of the LCD display. Firing up the unit and activating the backlight pushbutton lights up the detached panel with a whitish blue hue… a la Timex. These electroluminescent panels were first seen in small AC plug in nite lite panels back in the ‘60’s. A scoping of the xformers secondary reveals a nice looking 80V p/p sine being developed as the el-lum panel drive. Quite good sine waveform enhancement was being provided by a resonating capacitor across that sec winding….…considering the primary driver xstrs receiving ≈4 Khz driving pulses from a section of the COB.
An inspection of the rear of the LCD display reveals it to have no silvering , just a matte ground glass surfacing of the back….permitting an indirect illumination of the el-lum panel…..although this LCD unit is of reflective design…not transmissive. Looking down to where the ext antenna comes into the main unit reveals the presence of 4 wires…Litz wires….using 4 color codes… to be specific…looks like 25/40 count/gauge wire.[4 wires here, into only two on the loop antenna winding, will not go.] Further inspection of the loops rod, and pulling it out, reveals a small 1 in sq PCB hidden behind it with a little ≈ 3/8 in COB IC centered on it. Immediately noticeable is one of the small 1/8 in dia by ½ in long metal cylindrical encased xtals… thinking of it…. there was also a like xtal unit back at the main receivers COB IC . A scoping and dumping the gated out of the scope into a freq counter reveals the main receivers xtal sitting right on 32,768Khz and mounted right to the side of the xtal is a little 27 pf xtal “squegging” capacitor. So what they are doing is running the clock just like the bulk of electric wall clocks/watches, using that divided 32,768 time base. Then just prior to/at 12:00 am the u/p decoded time/date stream info from the signal is used within the u/p to reset the clock and refresh with the correct data on the screen.
A check back at the antenna pod reveals 2 0f the the Litz wires to be swirled about a companion Litz wire providing shielding to them. One shield wire carries +3V to the “pre-amp” PCB while the other shield wire is power ground. The two shielded wires provide a floating differential transfer of the post amplified ( a whopping 5 Mv P/P) 60Khz WWVB to the main COB chips input for further conditioning and processing . Going back to the “pre-amp” board and inspecting the incoming signal revealed a weak signal requiring a cascading the output of scope Ch 1 back into Ch 2 input to acquire sufficient gain to view the microvolt signal level present. Now in reference to some query as to the critical rod adjustment…specifically having the rod antenna broadside to the reception path…for optimum reception. On further rotation I noticed minumal sig attenuation until I had the rod end pointed right at Boulder direction, then the signal had diminished by about 20% with the change occurring within a -10 /0/+10 degree rotation spread , whilst swinging the rod either side of the null. Placing metal in close proximity to the rod also dampened reception…such as you might create by placing the ant rod unit on a metal cabinet. (Or inside a metal trailer home/Airstrem camper, car, etc.)
Checking the coil inductance of the 90 turns revealed a reading of 740 μh. Computations reveal needing an .0095 μfd cap to parallel resonate @ 60Khz, an .01 poly unit was found inside, shunting the coil , so the winding must have been tuned to that caps tolerance /value by sliding its form on the rod to get that inductance value and then waxed into position. Checking back to the “pre amp” PCB and its xtal revealed a very weak signal across either side of the xtal, so the unit was not utilized as an oscillator function. A monitoring of the waveform across the xtal did reveal a weak 60Khz across it and a ever so slight mechanical nudging of the xtal case produced a coincidental enhancement of the signal nodes…sooo….looks like they are using it for a 60khz xtal filter…… midstream in the preamp circuitry….. to enhance and filter/cleanup the 60 Khz signal.
For a final test after reassembly, a small pocket size AM radio was tuned to a station free 520Khz setting of the AM band and used for signal sniffing of the harmonic rich radiations of the pulses used in the multiplexing of the LCD display, including a connection made to the earphone jack for scope analysis.
The ferri-loop antenna within that type of radio is made in the same manner as aforementioned. The difference being in its space saving design, with the ferrite rod cast as a flat bar, instead of a larger round rod. The norm being an ~2 in length by a 1/2 in width but with only an 1/8 inch bar thickness. Then its antenna coil is wound on its sliding form with its position ending up ~1/4 to 1/3 from the rods end…as determined in final tracking alignment.
If you were to bring that coil end of the AM radio up to the face area of the LCD display and move it around for max aural pickup, it will start picking up the harmonics from the multiplexing signals feeding the 7 seg displays. Listen closely and you can even detect the minor tonal nuances produced by the segment shifts on the changing segment displays. Monitoring on my Oregon unit there is a 5KHz pulse train intermixed with a 360 Hz one with the low frequency being the throaty dominant tone with the hi freq adding raspy overtones.
Monitoring the lower left front of my Atomic wrist watch reveals a 4 Khz pulse train along with a lower ~250 Hz one, with the lower being the dominant one also.
I also have one of the newer Atomic table top receivers as a snazzy, silver unit with its , modernistic slope styling, and branded as ….Sharper Image. It uses a small antenna pod at the center rear of the receiver housing,that can be snapped out and back about an inch further away from the rear case. The pod must house about a 2 ¼ long by 3/8 in dia smaller antenna rod. Its multiplex harmonics voicing sounds about the same as the watch.
One thing distinctively different to the old design Oregon unit is that it also is weakly superimposing two different ticks along with the prior mentioned mpx sounds. Making alternate strong and weak ticks…..just like an old mech grandfather clock, but giving 120 ticks per minute.<p>Joseph….To optimize Will’s long wire antenna supplication , you might do an Am radio probing of the the ferrite antenna’s pod….exposed, I assume… and localize which half the coil is wound upon and then use 1 strand of your Cat 5 wire to enter the battery door and make a neg ground terminal connector or you could use an external ground, separate from the receiver and then come over and make a 1-2 turn inductive coupling link over the pods external case centered in position with the internal coil and secure the wire turn(s) from unwinding from the pod exterior and then the wire routes on to connect with the “long wire antenna”.<p>Dale…indeed that is the 5000 meter band…….and I’m about to tune in for some of that Ft Collins 60Khz signal, as its only a paltry 227 wavelengths away from my Big D’s QTH (location).<p>73's de Edd
[email protected] ........(Interstellar~~~~Warp~~~Speed)
[email protected]........(Firewalled-Spam*Cookies*Crumbs)
<p>[ September 15, 2004: Message edited by: Edd Whatley ]</p>

[snelln2]

EDD, I read this bulletin board every morning and thoroughly enjoy your postings. Are you kin to Mr. Greenspan? Keep up the good postings Edd

[Dean Huster]

Just like DMMs and calculators, since we're buying these fancy little clocks for $20 or less these days, I wouldn't expect to find much electronics inside that wasn't dumped under a blob of epoxy or sandwiched inside one or two custom chips.<p>I'd like to know if there's an el cheapo GPS or OEM GPS front end out there that one could use to access the cesium-referenced frequency/time signal to phase-lock to it for a good frequency standard without paying the $5000 that Agilent wants for a similar unit -- even though that $5000 is a heck of a lot better than paying the bucks for a cesium standard.<p>Dean

[tonybackache]

I purchased from WalMart an Atomic Casio Wrist Watch. I can press one function and it shows the time when it received signal broadcast. Most days it is 5:03am. There has been a few days when the time was 4:03am & 3:03am. If watch is left in a dead spot inside house receiving indicator will display no reception. I have enjoyed this watch. I have DISH-TV and the time posted there & wrist watch time are exact all day every day.

[dyarker]

Dean,<p>In Sep 04 Servo, DARPA Grand Challenge, Robo-Magellan - Part 2, Michael Miller talks about using a Garmin GPS18 OEM unit. No mention of price, but less than $5000 I'd guess.
www.garmin.com/manuals/GPS18_TechnicalSpecification.pdf and www.garmin.com/products/gps18oem links given.<p>Cheers,