VERIFICATION OF THE TIME ACCURACY OF A MAGNOMETER BY USING A GPS PULSE GENERATOR

The time accuracy of geomagnetic data is an important specification for one-second data distributions. We tested a procedure to verify the time accuracy of a fluxgate magnetometer by using a GPS pulse generator. The magnetometer was equipped with a high time resolution (100 Hz) output, so the data delay could be checked directly. The delay detected from one-second data by a statistical method was larger than those from 0.1-sand 0.01-s-resolution data. The test of the time accuracy revealed the larger delay and was useful for verifying the quality of the data.


INTRODUCTION
Some magnetic observatories have been exchanging one-minute geomagnetic field data in near real time as part of a program of the International Real-time Magnetic Observatory Network (INTERMAGNET).Recently, however, one-second data distributions and their requirements have become a hot issue.For example, requirements for resolution, absolute measurement accuracy, noise levels, and so on have been discussed among the relevant participants in INTERMAGNET.A standard of time accuracy is one of the outstanding problems, but no common method to determine the time accuracy of the output of a magnetometer has been established.Rasson (2009) examined the time accuracy of a magnetometer with magnetic signals generated by a GPS pulse generator.To detect the delay of a signal that is much smaller than the time resolution, a one-second, statistical method was used to calculate the time accuracy.
The Japan Meteorological Agency (JMA) has been recording one-second data and 0.1-s-resolution field data (0.1 second data) at magnetic observatories at Kakioka, Memambetsu, and Kanoya (Minamoto, 2009).In addition, a new fluxgate magnetometer with a 100 Hz sampling channel was introduced in 2010 at Kakioka, which allows us to directly investigate the delay of its recorded signals with a GPS pulse generator.
In this paper, we describe the procedure and present the results of a verification of the time accuracy of the magnetometer by using a GPS pulse generator.

EXPERIMENTAL
The GPS pulse generator (Figure 1) was made by the Geomagnetic Laboratory of Natural Resources Canada (www.nrcan.gc.ca),Ottawa, Canada.The pulse generator system consists of a GPS receiver, microcontrollers, and a pulse generator.This system outputs rectangular pulses synchronized with GPS signals.The cycle time of the rectangular pulses can range between 2 and 3,276,810 s.

Figure 1. The GPS pulse generator
Once a coil is connected to the pulse generator, the magnetometer gives out precisely timed magnetic signals.
According to the technical documentation of the GPS generator, the potential in the coil reaches 80% of its maximum at approximately 5×10 -3 s and drops to 20% of its maximum at approximately 2.4×10 -5 s compared to the top of the second with a coil of 626 mH and 154.4 Ohms.To produce the magnetic signals, we used a simple 44-turn coil, which was handmade (Figure 2) with a diameter of 116 mm, a length of 126 mm, and an inductance of 5.5×10 -5 H was measured by an LCR meter (Inductance (L), Capacitance (C), and Resistance (R)).
A resistance of 460 Ohms was inserted to adjust the current.The inductance of the coil was so small that the time lag between the direct current rise and the magnetic signal was negligible.It dropped to 20% of its maximum current in approximately 1.9×10 -7 s.   4).The magnetometer had made a trial run and no noise was included in the published data during a trial run of the magnetometer.The periods adopted for the GPS synchronizing pulses were 4, 8, and 16 seconds.

RESULTS AND DISCUSSION
The calculated delays of one-second values were determined by the magnetometer from 1000 waves by using a method based on the technique of linear least squares parameter estimation (Rasson, 2009).The results show that the delay of the recording of the fluxgate magnetometer was 0.385 second.It was the same for each period (4, 8, and 16 seconds).
As mentioned above, the new magnetometer can obtain magnetic data with a higher time resolution, so the delay of the recorded signals could be investigated directly.Figures 5 and 6 show the changes of the 0.1second values, and Figure 7 shows the changes of the 0.01second values.

Figure 2 .
Figure 2. Coil used for the test

Figure 4 .
Figure 4. Setup of the timing accuracy test of the new fluxgate magnetometer.The coil was mounted north of the sensor.

Figure 5 .
Figure 5. 0.1 second values of the X component output of the magnetometer with magnetic signals obtained by the GPS pulse generator.The period of pulsation was 8 seconds.The output (vertical axis) has not been converted into magnetic flux density (nT).

Figure 6 .
Figure 6.Enlarged plot of Figure 5 near the digital switching