[Discuss-gnuradio] VLF Application

[Joseph DiVerdi]

I've been working on a VLF Software Receiver for use in investigating certain 
ionospheric phenomena. These investigations involve measuring the *complex* 
amplitude of several terrestrial VLF transmitters which has been forward 
scattered off of the ionosphere. The complex amplitude is sensitive to the 
elevation of the reflecting/refracting layer which in turn is sensitive to 
solar wind, CMEs (Coronal Mass Ejections), and supposedly to  GRBs (Gamma Ray 
Bursts).

The primary goal of this work is to set up a multi-channel, VLF observatory 
operating in the range 10-100 kHz using software radio and digital signal 
processing techniques which can monitor several frequencies simultaneously, 
measure the complex amplitude of each received frequency, and report these 
results on-line and real-time using popular Web-based technologies.

To accomplish this an untuned, vertical, whip antenna (96") has been erected 
with a mast-mounted preamplifier. The preamplifier serves a number of purposes: 
(1) provides very high-impedance load for the antenna (‰1 MOhm), (2) provides a 
low-impedance, balanced output to drive the long wires into the 
workshop/laboratory (‰8 Ohm), (3) provides voltage gain (‰400), (4) provides 
frequency band pass filtering (-3dB beyond 1 and 100 kHz), and (5) protection 
for the workshop/laboratory equipment (spark-gap, crossed-diodes, sacrificial 
capacitor) .

n.b.: There is a considerable body of knowledge in the Ham radio community 
regarding installing and using elevated outdoor antennas (aka lightening rods) 
in a safe and responsible manner. If you're interested in this particular field 
please consider obtaining a copy of an ARRL Ham Radio Handbook, read it, and 
follow it. One of the points to be made is it is difficult to protect against a 
direct lightening hit so, as Don writes, disconnect and secure properly during 
storms.

Connection between the mast-mounted antenna/preamplifier and the 
workshop/laboratory equipment is made using simple unshielded, four-wire 
telephone wire. I have experimentally verified the transmission performance of 
‰50' (adequate for my current purposes) when driven and terminated at low 
impedance (use a pair of impedance transforming audio transformers, 1000:8 
Ohm). It is also important to terminate the 1000 Ohm side of the 
workshop/laboratory transformer with a 1000 Ohm resistor.

To date this "front-end" system has been briefly characterized. Due to my 
observatory's proximity to WWVB (60 kHz), the voltage observed across the 1000 
Ohm terminating resistor with an oscilloscope is dominated with WWVB's ASK 
(amplitude Shift Keyed) signal and in the current incarnation is ‰1Vpp. Some 
preliminary experiments intended to selectively observe other transmitters have 
been performed including (1) a so-called "lock-in amplifier" driven by a local 
VLF range signal generator which permits easy tuning and indication and permits 
baseband detection but suffers interference from "harmonic sensitivity" and 
limitations from the phase-noise of the signal generator, and (2) a high Q 
(50-200) band pass filter created using a classic, three op-amp, Bi-Quad 
circuit which performs no detection and results in an output at the carrier 
frequency.

Several external transmitters have been observed at 13.8, 25.2, 36.0, and 60.0 
kHz on a preliminary basis.

Current work has moved indoors to the workshop/laboratory. A Pentium II class 
computer has been built running (Red Hat) Linux v7.2. The GNU-radio software 
radio suite has been installed. An as yet unproven Sonic S3 sound card has also 
been installed for testing purposes.

I am currently searching for a used PCI-based ADC input board with at least one 
(single-ended or differential) input with at least 12 bit resolution and 200 
(preferable 250) kSample/second sampling rate.

The ADC's sampling clock signal will be fixed at 200 (or 250 kHz) and initially 
generated by a local stable crystal-controlled source. This data stream will be 
processed by complex multiplication with software local oscillator and low pass 
filtered and downsampled with a software filter to a roughly 1 kSample/second 
data stream. This stream will then be downsampled once more to around a 
one-to-ten sample/second rate which will be archived, processed into graphs, 
and fed to a local web server. This entire processing stream can be arbitrarily 
replicated (limited by CPU processing power) to monitor several different 
frequencies simultaneously.

Initially the data stream(s) will be scalar RF magnitude only because of the 
inevitable phase noise of the oscillator used for the ADC clock. At a later 
date the local oscillator will be derived from the 10 MHz WWV signal which is 
line-of-sight to my site. The expected reduction in phase noise for the ADC 
clock derived from this signal is expected to permit observation of each 
received RF signal's phase as well as its amplitude. Processing software will 
be updated at that time.

That's the summary of work to date. I hope to be "on-the-air" by the beginning 
of summer 2003 with a detailed web site describing the instrumentation, 
software, graphical data and text data. It is expected to mimic my current 
terrestrial magnetism observatory at http://xtrsystems.com/magnetometer.

Best regards,
Joseph
--
Joseph A. DiVerdi, Ph.D., M.B.A.
http://diverditech.com/           970.980.5868 (voice)
http://xtrsystems.com/            970.224.3723 (fax)
PGP Key ID: 0xD50A9E33