rjr@mgweed.UUCP (Bob Roehrig) (09/19/85)
To Phil Biehl and other interested parties: Subject: receiving and using the 60 kHz WWVB transmissions I have a home built WWVB rcvr that I have been using for about 10 years with good success. My location is near Chicago. A couple years ago, NBS worked on their WWVB transmitting system and the signal strength improved here greatly. In order to use the signal as a frequency standard reference, you have to use it directly, so to speak. What this means is that you cannot use a superhet type design where the signal is mixed to another frequency by a local innacurate oscillator, so the receiver I use is a TRF type. I am not sure where the schematic of my unit is so a lot of what follows is from memory. The best antenna is a shielded tuned loop. Mine was made by winding about 22 turns of 22Ga plastic covered wire around a waste basket (about 3 feet diameter). The electrostatic shield was made by wrapping the loop in tin foil. The shield is connected to one side of the loop which is the ground side of the loop. Remember that the loop must be broken and not grounded at both ends or it will act as a shorted turn. The break in the shield can be either at the center or one end (it's not critical). The loop is tuned to 60 kHz with approx 2000pf mica cap. Tune it like you would a toroid inductor for audio: connect to a signal generator thru a 100k resistor and use a scope to monitor the signal across the loop. The loop feeds a 2 stage (FET/PNP) amplifier which is in a metal box. The output is from a tuned circuit that provides a low impedance output to the coax. The 12 volts to run the preamp is fed up the coax from the rcvr. The rcvr itself is several stages of J-FET tuned amplifiers using 1 mh inductors which finally feeds some untuned stages to build the signal up to TTL level. At the rcvr input I use a 60 kHz crystal as a filter which makes the bandwidth 1 or 2 Hz. While a narrow filter is not absolutely necessary (especially if you are in a strong signal area), it really helps. (H-P used a 15Hz wide mechanical filter in their VLF rcvr). I can provide circuit dwgs to anyone interested. Were I to do it again, I would certainly do some things differently. I have a much better preamp circuit (also simpler) and I would probably use a PLL chip as the main rcvr gain block. ---------------- There have been articles in ham publications regarding WWVB and using the signal but I disagreed with many of their techniques. It is best not to use the 60 kHz signal directly as a counter time base since the rcvd signal has jitter on it. Also there is the diurnal phase shift as well as the intentional phase advance/retard each hour which is used as their "ID". Commercial systems use the rcvd signal merely to compare a local "house" standard against. Comparison is done over a fairly long period (like days) for best results, or in some cases, continuously by observing the comparison on a chart recorder. My system consists of a 1 MHz oscillator in a proportionally controlled oven as the house standard. This osc is used as the time base for my counter, clock, and synthesizer. The 1 MHz osc is also divided by 10 to 100 kHz. The 100 kHz local signal is then multiplied by 3 using a tripler circuit ( made using a padded down 455 kHz IF transformer), squared up again, then divided by 5 (1/2 of a 7490) to provide my locally generated 60 kHz signal. This and the rcvd signal are both fed into a comparator, which is nothing more than a TTL flip-flop. One output of the flip-flop has a lowpass filter on it which gives and average DC voltage reading related to the duty cycle of the 60 kHz square wave. This DC voltage is observed on a voltmeter and also feeds a chart recorder running at 4 inches per hour. Fine corrections to the 1 MHz osc are made after observing the chart after a day or two to see the trend of drift. The present 1 MHZ osc needs some work since it is slightly affected by room temperature changes, however it does maintain an accuracy of at least 1 part in 10 to the 10th. The oscillator control is by a 10 turn pot with a counter dial which controls the bias on a varactor diode in the crystal circuit. The osc ckt is made from dual gate FET's and buffered twice within the oven. The loop antenna is mounted in the attic. Tests showed that it worked just as well in the ham shack, which is below ground level (but who wants a big loop in the way). One stage of the rcvr (prior to limiting) feeds an "S" meter to observe the signal strength. I also added an AGC circuit to detect the 10dB amplitude shifts which contain the BCD time code. I have not yet attempted to build a time decoder. At this time I divide the 1MHz down and feed a standard digital clock using a National clock chip (5314 maybe). So far the clock remains within one second between CST to CDT time changes. The clock and osc run off a 4 amp/hour 12 volt nicad if power fails. The rcvr/comparator run off the charger/supply. Would like to discuss the topic further with anyone interested and will be glad to furnish dwgs of my setup to those REALLY interested. Locally I am called "Bunky Bob's Batavia Bureau of Standards". Using WWVB as reference sure beats using WWV on HF as a frequency standard. At my location, the BEST results that can be obtained on HF are 1 part in 10 to the 7th. Bob K9EUI