Radio observation of the 22nd August 1998 solar eclipse

I. Introduction

The 22nd August 1998 annular eclipse is viewable from Sumatra (Indonesia), Malaysia, Sulawesi (Indonesia) and Papua New Guinea. Details of this event can be found in this Wikipedia page [1]. The Williams College's site lists at least 7 reports of this eclipse by visual observers in several countries [2]. The value of this work lies in it being the only radio observation of this event. Additionally, this work has the distinction of being the first radio astronomy experiment in Malaysia and furthermore, it precedes the first professional observation [3] by 7 years.

Fig. 1 this eclipse path lies close to the equator (graphic from [4])

At my location in Penang, Malaysia, which is ~4 degree north of the path of greatest eclipse, an 83% obscuration is expected (fig. 2). Expecting the moon to attenuate solar radiation, I set out to observe the eclipse at 408 MHz.

Fig. 2 Time-table published by the Astronomy Club and the Astronomy section of the Science Museum, Universiti Sains Malaysia (USM) , Penang, Malaysia. Local time = GMT + 8H.

II. Direct observation at 408 MHz

A. Measuring arrangement

The measuring setup is very rudimentary - consisting of a "generic" UHF TV Yagi aerial, mast-head pre-amplifiers (LNA), spectrum analyzer and power meter (fig 3). The aerial has 18 directors, 1 driven element and a wire-mesh reflector (fig. 4), but pertinent RF characteristics such as gain, beamwidth, and VSWR, etc. are not known. Hence,  it is not possible to determine the absolute value of the solax flux - only its relative changes can be measured. Because the aerial is intended for receiving terrestial TV, its mounting is fixed at 0 deg. elevation. Fortunately the sun is close to the horizon during the eclipse; i.e. 7-9am local time. The aerial is then manually pointed at the sun. The homebrew pre-amplifier is positioned just below the aerial in order to overcome cable loss and the spectrum analyzer's poor noise figure. The spectrum analyzer's settings are: 408MHz centre frequency, zero span mode and maximum (1MHz) IF bandwidth. At this combination of settings, the displayed noise floor is -90dBm - corresponding to a 24dB overall noise figure. The gain between the RF input and the IF output port is 51.7dB. The power is sampled at the spectrum analyzer IF output port using a power meter. Time-interval logging of the power meter readings was done using HP (now Agilent) Vee test automation software running on a PC.

Fig. 3  408MHz automated measuring setup

Fig. 4 A no-brand UHF Yagi aerial similar to the one used in the 408MHz setup (this photo is taken in the store where it was sold)

B. 408 MHz results

The received power graph has a flat-bottom trough that coincides with the expected eclipse period (fig. 5). Moreover, the trough is distinct and has steep sides that demarcate the beginning and end of the radio eclipse as opposed to the visual one. Unfortunately, setting up delay costs us the data between the 7.10am visual onset and the 7.27am measurement commencement. The radio eclipse's duration is ~56% of the visual one. We did not anticipate the duration to be so different for radio and visual. Possible confounding factors are solar radiation fluctuation and the sun moving out of the aerial's beamwidth.

The trough is 3-4 dB lower than the post eclipse value after 8.50 am. We have expected a trough with gradual slopes and so, cannot explain the flat bottom and steep sides.  It is possible that the flat bottom is a measurement artefact caused by the noise floor of the measuring setup.

Fig. 5 The distinct trough in the received power vs. local time graph clearly correlates with the eclipse's time table

III. Indirect observation by measuring ionospheric change in the Medium Wave band

A. Measuring arrangement

We are also interested in finding out if the eclipse could be indirectly observed via its effect on earth's ionosphere. To detect the ionospheric changes, we monitored the signal strength of a 783 kHz medium wave (MW) broadcast during the eclipse. The 1996 edition of the World Radio TV handbook  lists 14 East Asian transmitters on this frequency, but we suspect the station is 'The Voice of Vietnam' (VOV) because 'Hanoi' was mentioned several times. VOV's 783 kHz transmitter in Nghe An (19N 105E) is ~1600 km away from our site.

The measuring setup comprises a consumer-grade car radio receiver (Blaupunkt) and a non-resonant long-wire aerial (fig. 6). The receiver's high impedance FET input improves signal transfer from an aerial that is very short compared to the wavelength (~383m). The signal strength is measured at the AM tuner IC's "Received Signal Strength Indicator" (RSSI) pin which outputs a DC voltage that is proportional to the RF signal (RSSI is also known as 'Field Strength').

Fig. 6 Measuring arrangement for observing in the medium wave band

B. Data format

Due to PC being occupied with the aforementioned 408 MHz observation, the 783 kHz reception strength data was recorded using an ancient Philips PM8143 strip chart recorder. Unfortunately, this recorder's slowest paper speed is 10s per cm, which means that the entire length of an A4 graph paper is only sufficient for ~4 minutes of recording. Since this duration is too short to cover the eclipse, multiple traces are required to extend the observation duration. To allow the amplitudes of these disconnected traces to be compared, a calibration mark is inserted at the beginning of each trace.  The calibration mark is generated by stepping the amplitude of a signal generator from -80 to -90 dBm over 15s to 20s. The two power levels cause the trace to change by 16 to 17 mm on the graph paper (the graph graticules are 1mm in size). A manually operated switch allows the radio's input to be switched between the calibration source and the aerial.

There are also several minute gaps between traces due to the need to manually re-position the recorder's pen to the beginning of the next trace and to repeat the aforementioned calibration process.

C. 783 kHz results

However, no clear pattern indicative of the eclipse modulating the ionosphere could be discerned in the experimental results. The average signal strength is much above -80dBm over 7.23 - 7.54 am (fig. 7) but drops below -90 dBm over 8.08 - 8.12am (fig. 8, top trace). This large variation in signal strength within the eclipse duration prevents any meaningful pattern to be inferred. Moreover, the results are opposite to our expectation; i.e. we have expected the signal strength to peak at the eclipse maximum, but the data in the 8.18-8.22am period (fig. 8's 2nd trace) which corresponds to the eclipse maximum is at least 10dB lower than the 7.23-7.27am period (fig 7, top trace) which corresponds to the eclipse onset. The signal strength results at the tail-end and post-eclipse (fig. 9) also do not appreciably differ from one during the maximum (fig. 8's 2nd trace beginning at 8.18am) as they both average around -90dBm.

The failure to discern a pattern in the results that can be attributed to the eclipse may be due to the ionosphere being in a state of flux at sunrise; i.e. the ionosphere changes from radio-reflective to transparent in the transition between night and day. It could also be that the transmitter site in Vietnam is too far away from the eclipse path.

Conclusion

Using amateur-grade equipment, the 22nd Aug. 1998 solar eclipse was successfuly detected at 408MHz but not at 783kHz.

Acknowledgement

The author thanks C. W. Kong for programming the automated data collection and K. T. Phoa for loaning the spectrum analyzer.

References

[1] Wikipedia, "Solar eclipse of August 22, 1998," [Online] Available: http://en.wikipedia.org/wiki/Solar_eclipse_of_August_22,_1998

[2] "The Annular Eclipse of August 21, 1998," [Online] Available: http://www.williams.edu/astronomy/eclipse/eclipse1998/1998annular/index.html

[3] Z. Z. Abidin and Z. A. Ibrahim, "Radio astronomy research in Malaysia: Past, present and future," Proc. 2009 Intl. Conf. on Space Science and Communication.

(c) This work is protected by copyright and may not be reproduced in any form without the expressed consent of the owner.