12 August 2026 Eclipse—14 MHz recording in Castillo de Curiel Spain of Costas array spread-spectrum signal transmitted from Valencia Spain

Two files are in this set, a .wav file of approximately four hours and a .zip file containing 57 synchronized (by GPS) .wav files of four minutes each. The .wav and the multiple SYNC.wav files "zipped" together are simultaneous recordings from the same radio receiver but recorded from different points in the receiver's audio chain. Although they correlate well, the long file analog sampling is timed only by the receiver's internal crystal oscillator (it is timestamped accurately from the receiver's internal GPS receiver but the sample rate is not GPS disciplined). The individual SYNC files are both timestamped accurately from a GPS receiver and are sampled accurately as timed by the same receiver's pulse-per-second output. See documentation from the AudioMoth project for more details on the preparation of the synchronized files, AudioMoth GPS Sync | openacousticdevices. The signal source from Valencia was timed by its own GPS PPS system. The time difference measurement from Valencia to the receiving point thus contains information about signal time of flight along its total path. That path contains for each receiving station past the Valencia horizon at least one ionospheric bounce. The path length is thus longer than the ground-based great circle distance from transmitter to receiver and was presumed to vary as the varying solar illumination on the ionosphere changed the ionosphere refraction and virtual reflection height. The files were recorded simultaneously during the course of the 12 August 2026 total solar eclipse as the moon's shadow traveled from Iceland to Valencia, Spain. The spread-spectrum Costas array signals were sent within one communications channel audio bandwidth, upper sideband with a dial frequency of 14.110 MHz. Eclipse information may be founded in several web pages, including this: https://en.wikipedia.org/wiki/Solar_eclipse_of_August_12,_2026 The signal was recorded in grid square IN71wp, 41.6410356,-4.1039847 (high atop the castle, https://maps.app.goo.gl/iHFngk3Gxyuk45Lz7) The signal transmitted was of an n=21 Costas array, 40 millisecond time slots, frequencies linearly distributed from 300 Hz to 2700 Hz as an upper-sideband signal. Timing was determined by the PPS (pulse-per-second) signal from a GPS receiver. There is a 40 second silence at the beginning of each second. A 40 ms stabilization tone at 700 Hz follows; the 21 segments are send; another 700 Hz tone ends the sequence. The Costas hopping sequency is 3, 0, 8, 2, 18, 6, 15, 14, 9, 7, 20, 16, 19, 11, 17, 1, 13, 4, 5, 10, 12so the resulting frequency sequence as actually transmitted in the audio channel, in Hertz, is0, 700, 660, 300, 1260, 540, 2460, 1020, 2100, 1980, 1380, 1140, 2700, 2220, 2580, 1620, 2340, 420, 1860, 780, 900, 1500, 1740, 0 (0 signifies no signal. The first tone in the Costas array, 660 Hz, therefore begins at the transmitter 80 milliseconds after the beginning of the UTC second) (there is a small frequency offset from the direct digital synthesis system, never more than .003 Hz)The CW ID sent every 5 minutes is EA5KRP, the callsign of Radio Club Universidad Politecnica De Valencia.The transmitter's master oscillator (and therefore its transmitting frequency) was not GPS disciplined. It was a Kenwood TS-2000, kept at 10 watts output so that the heating would be kept to a minimum and the fan would not activate. The antenna was a log-periodic array beam aimed at 320 degrees (toward Iceland). The transmission was from EA5RKP , the amateur radio station of Radio Club Universidad Politecnica De Valencia. The receiver was an ICOM IC-705, also not GPS disciplined; it was frequency stable at the level of amateur radio equipment similar to the Kenwood transmitter described above. The receiver was in ambient temperature of approximately 40 C and variable, outdoors during the recording period. A wire antenna was about 30 meters above terrain, broadside to the transmitted signal. The .wav recording is with the internal soundcard recording system of the transceiver and was intended as a backup from the Audiomoth recording prepared with precise, GPS-determined timing as described below. The internal recording does, however, give better indication of an incidental beacon signal than does the Audiomoth recording, presumably because its analog sampling is before the receiver's audio amplifiers and thus frequency-shaped slightly differently. The file shows excellent reception of the Valencia spread-spectrum Costas array signal and its associated CW ID over several periods but likely not during the time the moon's shadow was between us. There may be some Costas signal there but I haven't identified it yet and I will be surprised if further analysis discloses any. A CW beacon was incidentally noted as present through much of the file at about 3100 Hz in the audio channel. Copying the Morse signal by ear, it is identified as amateur station HS0ZEA/B in Thailand, grid square OK05qo (15.604° N, 101.375° E, great-circle distance to the receiver about 10,093 km).The noise with level alternating every thirty seconds or so, variably, is likely from a Coca Cola cooler nearby on the castle roof.Work remaining is on better signal processing and on similar processing for the externally-recorded, precisely-timed signal.David