Wellington VHF Group, Inc: Online.

Expecting to battle atomic powered megacycle measuring maniacs and other assorted time nuts the 58th Doug Gorman Frequency Measuring Contest went to air at 9 pm on Monday evening 6th September 2010. That’s 2100 NSZT, 0900 UTC, 967798815 GPS time which is fifteen seconds ahead of UTC, or UTC plus 34 seconds for TAI* to you and 1283763600 Unix time**. Hey, you need to know these things when measuring frequency to a third of a part per billion. But enough of that already!

The contest is named after the late Doug Gorman MBE ZL2IY who filled various roles for NZART including that of General Secretary and Officer Commanding AREC. This year the Wellington VHF Group Branch 74 offered to take over the contest from long time organizers, Branch 63.

After taking a moment to remember those amateurs and their families in the Canterbury Area affected by the weekend’s earthquake, the contest rules were read out on 3895 kHz in voice before switching to Morse for the remaining formal parts of the contest proper. The rules are published elsewhere on the site and results will appear in the near future.

While contestants had the fun of first finding five frequencies at the top end of the 80 metre band and then measuring them it was the organizer’s job to radiate these frequencies with nothing less than atomic precision! And it turns out that even with today’s modern technology of high performance GPS disciplined frequency standards that's not easy. (More after the jump).

A commercial transceiver, the Icom F7000, was used as the transmitter because it allows an external reference input to the PLL system and because it was compatible with the antenna tuner used for the high performance NVIS antenna at ZL4JY. This antenna is a 28 metre per leg cantered doublet fed with 600 ohm open wire feeder erected over a specially constructed ground plane of seven No. 8 galvanized wires (yes Don, the ground plane also works as fence). The F7000, when equipped with two 120 mm fans, can run at 50 watts key down indefinitely. This last point is important in a contest where the transmitter need to run for more than 40 minutes continuously.

Let's get technical! The reference for the F7000 needs to be about 32 MHz while the GPS disciplined standard produces 10 MHz, so where was the other 22 MHz going to come from? A signal generator was the answer but there were a couple of special requirements. The F7000 is only able to be set in 100 Hz steps which would allow contestants to guess the exact frequency with ease. By using a frequency reference for the F7000 that could be altered in 0.01 Hz steps, the transceiver's 100 Hz steps could be interpolated in steps of about 0.001 Hz given the ratio of reference frequency to 80 metres. This also allowed correction of a small flaw in the F7000. In this radio the first IF of 12 kHz is generated by a digital signal processor (DSP) whose clock is not locked to the PLL reference but instead uses a 30 MHz un-compensated crystal oscillator. In normal use small errors in the 30 MHz DSP reference are not important as a 1 part per million error, for example, results in an IF frequency error of only 0.012 Hz but for the contest this was just not good enough. This error could be allowed for however by offsetting the PLL reference by a similar amount. Close inspect of the signal generator in the picture will show that a reference of 31.99999997 MHz was required for the F7000 to be precisely on frequency.

The HP E1938A ovened hockey puck crystal, the heart of the GPS disciplined oscillator (GPSDO) frequency standard.

To confirm the exact transmitted frequency an HP 5334A frequency counter was connected to the transmitter output via a resistive divider. The 5334A has a special math feature in that an offset can be subtracted from the displayed frequency to enable more resolution. Thus instead of seeing the counter display 3895000.00 Hz for example we can subtract 3890 kHz and have the counter display 5.000000 kHz, improving the resolution from 0.01 Hz to 0.001 Hz with a 10 second gate time. With the counter and sig gen both locked to the GPS standard the transmitted frequency could be maintained to better than 0.001 Hz or 1 mHz (that’s a millihertz) over the 40 minute contest period. The measurement uncertainty is the absolute value of +/-1 count +/- reference error. As the reference error of the GPS standard is about 0.1 parts per billion or better, the total uncertainty will be less than +/- 1.4 mHz.

PC is showing the GPSDO status screen and the Morse sender program CWtype. The sig gen (LCD screen) is the Icom F7000 PLL reference and below that is the HP 5334A counter.

The Agilent HP3815A GPS disciplined standard works by receiving time signals from typically 8 GPS satellites and then carefully adjusting a rather special hockey puck 10 MHz crystal to almost the right frequency. Keeping all this gear stable and protected from mains failure required a small 24 V DC plant consisting of a rectifier set and a pair of AGM style lead acid batteries.

OK so that’s the batteries what’s all this atomic stuff? Well each GPS satellite carries four caesium atomic clocks which work by counting the hyperfine transitions of caesium atoms which oscillate roughly 10 billion times per second … the definition of a second, more or less. More or less? Well Einstein’s theory of General Relativity predicts that the GPS clocks at in orbit above the earth will tick faster by about 46 microseconds per day due to the weaker gravitational field than the same clock on the surface of the Earth. In addition Special Relativity predicts that clocks moving at GPS orbital speeds will tick slower by about 7 microseconds per day. To allow for this the GPS clock are set to run slow before launch!

We've come a long way from the BC-221.

Megacycle measuring maniacs and time nuts.

John ZL4JY

*Temps Atomique International
**Just for Mark ZL1VMF