Traditionally, high quality reception of VLF natural radio signals has involved driving or hiking out to remote locations far from civilisation, in order to find an environment free of pollution by artificial radio signals.
I've lost count of the times I've hiked out into the countryside with a small portable VLF receiver in search of a quiet spot for a night's listening. All I ever seemed to get is sferics and tweeks, ... and hum, the ever-present signal radiating from the country's power distribution system. Even in the best of spots this hum never goes away completely, in my part of the world at least. Now and again a whistler manages to surface above it, but I never contrived to be in the right place at the right time to hear anything spectacular. On the whole, it was quite a disappointing experience.
The only course of action seemed to be to arrange the means of receiving VLF signals cleanly from home. Although I live in a rural location, the moorland around here is crisscrossed by big powerlines, as well as smaller ones leading out to the many farms. So when the receiver is switched on, most of what comes out of it is hum - a suffocating, mind numbing hum, which smothers everything but the loudest sferics. Just beneath the hum, there is another layer of noise - an unpleasant brew of signals produced by televisions and computers in the house.
So as my portable receiver began to collect dust on the shelf, I set about making it possible to receive from home.
I'm glad to report that I've had a fair amount of success with setting up a home receiver, and I can now listen to a completely hum-free rendering of the earth's natural VLF activity, more or less continuously. At times there is interference from machinery and so on, and although the antenna is now almost immune to wind and drizzle, it is still wiped out by driving rain or hail. But most of the time reception is first class, and I can listen to the signals anywhere in the house. They sound awesome played through the stereo in the living room - great clarity and depth.
This web page describes the various tricks I've used to achieve this.
It quickly became clear that the main source of hum was the overhead line at about 12kV which brings power to the property. This terminates about 100metres away at a pole transformer, and comes the rest of the way in at 240v, still overhead.
This is where a suburban location, in which powerlines tend to be buried, would likely be better than my rural site. I found that I needed to be 150 m from this overhead line before the hum field was no longer obviously influenced by it. Beyond this distance, the hum field settles to a fairly constant value - a level set, presumably, by the proximity of the next most distant source of hum: some huge overhead lines about 4 miles away across the moors. The hum level would also rise as I approached the house, but not too severely, and it appears that placing the antenna just 10 or 20 m from the house is sufficient to reduce direct pickup from the domestic wiring to a level at which the more distant hum field takes over.
I found many spots where the hum was very low, but unfortunately, so was the wanted signal too! This occured beneath trees, or near to walls, fences, outbuildings and so on. The important figure of merit is the signal/hum ratio, which is not so easy to measure. I couldn't think of a neat way to measure this, so I ended up fine tuning the antenna location simply by listening to the signals and estimating how often a loud sferic made it above the hum. In the end I settled on a spot on my neighbour's property, about 200m from the end of the 12kV line.
The answer seemed to be to beef up the grounding arrangements at the antenna site, so that interference coming from the house would be shunted away. Putting in several ground stakes reduced the ground impedance to around 100 ohms, but this did not seem to improve the situation much. The hum signals coming out from the house along the coax screen must be at quite a low impedance, and I soon came to the conclusion that a prohibitively substantial grounding system would be needed at the receiver site in order to dispose of these completely.
Bringing a coax out from the house seemed to bring with it all the domestic interference that I wanted to leave behind, and defeated the whole point of siting the antenna some distance away from the buildings. Therefore I decided to use a wideband radio link to retrieve the signals from a totally self-contained battery powered receiver. I could bring the coax up quite close to the receiver box without causing interference, providing I made sure that the coax screen was insulated from the ground near the antenna, and ensured that it didn't come anywhere near the antenna rod itself. This was close enough to allow the use of a small VHF FM oscillator, which could bridge the small gap from the receiver to a little insulated antenna fitted to the end of the coax. This oscillator gave a fully-quieting signal in a scanner back at the house, and preserved the original signal/hum level.
With that done, I made up a simple software comb filter set to 20mS delay. This took away most of the hum but gave the remaining signals an unpleasant drainpipe-like quality. Increasing the number of delay stages eliminated that problem but it became necessary for me to continually fiddle with the delay to keep the sharp notches lined up with the incoming spectrum of hum harmonics. Typically I would have to tweak the controls about once a minute to keep it on track, which was pretty tedious. Luckily, it turned out to be fairly easy to make the filter keep its own track of the hum period, and the result was a software program along the lines of the demonstration code in Humfilt.
So far so good. The hum was removed pretty much entirely, it tracked well, and there was no unpleasant coloration of the sound by the filter. Unfortunately this revealed a deeper level of noise coming from non-harmonic components of the power line voltage. These were not particularly strong, but they were at times numerous across the spectrum and consisted mainly of sidebands of the mains harmonics. Luckily most where well defined tones at more or less constant frequencies.
To deal with this non-harmonic power line hum, I added another stage of software filtering which transforms the signal to the frequency domain with an FFT, averages each bin, and recognises any persistent frequencies. Those that exceed a tunable threshold have their FT bins zeroed before the signal is reverse-FFT'd back to the time domain. This was immediately effective in removing virtually all the remaining constant power line noise, and the only thing that gets through now is broadband noise and the wandering tones coming from things like power tools and washing machines.
Unfortunately I could also hear a lot of hiss, along with a considerable number of
signals that had no business being in that part of the spectrum at all.
here is that GBR at 16kHz is a huge signal and the entire spectrum 16-25kHz
was being neatly translated to the range 0-9kHz by cross-modulation in the
receiver. This is not specifically an issue concerning domestic reception but
is a problem that any VLF receiver must overcome. To cut a long story short
I tried various arrangements of front-end and eventually concluded that in my case:
a) most of the intermod was occuring in the back end of the receiver, rather than the first stage;
b) my receiver noise floor was way too high, dominated by the RC filtering before the front-end.
I eventually settled on a configuration with just a bare minimum of RC low pass filtering before the first stage, and the first stage is followed by several poles of both low pass and high pass filtering. The noise floor is still dominated by the thermal noise from these filter resistors, and things could be even better if I used LC filtering instead. However, I was able to find a compromise of RC values which gave me a receiver noise which was substantially less than the genuine VLF background hiss coming in from the antenna.
Putting all this together, I built a whole new receiver which performed much better. The circuit diagram is available here. I added in a few features such as a moisture detector and a low battery warning - each of these inserts a warning tone into the receiver signal. Another very useful feature is a changeover relay that can open and close the receiver connection to the antenna. This relay is commanded by a 27Mhz radio control signal and when activated, the receiver just sees the thermal noise from the 10meg bias resistor. This can be calculated quite accurately and allows a reasonable end-to-end calibration of the entire signal chain to be made. See my notes on calibration here.
The spectrum of the receiver output (without any software filtering or equalisation) is shown below.
The green line is the thermal noise of a 10meg resistor shunted by about 37pF. The actual receiver noise is lower than this because when the antenna is connected (the red line) the bias resistor noise is further shunted by the antenna capacitance. It is nice to see that the overall response matches the background noise curves given in What and Where is the Natural Noise Floor? by John Meloy. The pronounced dip at about 4 kHz is present, as well as the broad peak around 10kHz and the rapid rise in amplitude between these two is reproduced. The mess of hum harmonics is very prominent, as is a whole bunch of very strong MSK signals between 16kHz and 24kHz. Considering the receiver has about 5 poles of low pass filter at a corner of around 10kHz, these signals are doing pretty well! They were the major source of intermodulation in the back end of the receiver, and in fact the GBR signal continues to set the limit to the overall dynamic range.
However, with the new lower noise floor and the greater receiver gain, the dynamic range was begining to frighten the VHF oscillator. The louder signals would overmodulate the FM, sometimes almost dropping it out from the scanner, which had a knock-on effect on the hum filter tracking. Turning down the gain simply meant that the noise floor of the scanner receiver used for the downlink began to compress the bottom end of the dynamic range. Not only this, but some remaining intermodulation was almost certainly occuring within the VHF downlink. Basically, the dynamic range and bandwidth of the VLF signal was just too much for the simple wide band FM downlink.
I did consider a more sophisticated downlink, using a pair of the cheap wideband FM microwave ISM transmitter/receivers which are common these days. But before doing this I decided to have one more bash at the direct connection.
As soon as I connected the coax screen to the receiver ground, a vast amount of hum and noise poured in. This was something like a hundred times the level of hum picked up from the antenna alone, and was totally overloading the whole system. As well as the vastly increased hum, a great deal of noise was introduced from the computers in the house, in particular one 21" monitor which seemed to chuck out loads of hash right across the spectrum - far more than a TV with the same size screen.
Clearly there was a lot of interference sitting on the domestic ground to which the bottom of the coax ultimately attached. There's perhaps a few tens of millivolts between the domestic ground and the receiver ground, compared with perhaps some few hundred microvolts of hum picked up directly by the antenna.
There are various technologies available to isolate the receiver ground from the domestic ground. Obviously the radio link is one option, but then the extra noise, limited bandwidth, reduced dynamic range, and opportunity for intermodulation, all have to be accepted. Another simple method is to place an audio transformer between the receiver and the coax. Employing a little driver transformer ripped from an old transistor radio to isolate the coax made a huge difference. Now when I connected the coax, the hum increased by only about a factor of two above that of the isolated receiver. The quality of the signal also went up quite a lot - which I wasn't really expecting. The VHF link was pretty good I think - you could play hi-fi music through it with no noticeable distortion, but even so it must have been compressing the VLF signal quite a bit, both in amplitude and bandwidth.
Following a tip, I put in a second isolating transformer at the bottom end of the coax, so that the entire coax is floating, and this did away with that last bit of extra hum introduced by the direct connection. The resistor R is anything from a few 10s to a few 100 ohms, and serves to dampen the resonance of the transformer windings with the coax capacitance. Now with this isolated coax connection, the sound of the natural VLF spectrum was just awesome. Much of the time the signal is completely free of local interference, and the clarity and depth is amazing at times.
I'm very pleased to say that this has almost completely disposed of wind noise. In a strong breeze the antenna just flops limply from side to side with no vibration at all. In very strong winds I can just hear a faint hiss from the receiver occuring with each sideways flop of the 2 metre long, 40mm diameter tube. By the way, I should say that the antenna is at ground level, not raised on a pole. The signal level could be increased by hoisting it up into the air, but honestly there's more than enough signal already and the challenge is to obtain a good signal/noise ratio. Since I have a fairly open space for the antenna, raising the tube would risk increasing the mechanical noises by more than the signal is increased.
Here's a quick summary of the steps I've taken along with a few tips:
And as a final warning, if you fall asleep listening to the VLF and wake up to hear a sound like the world ending, it's probably just a sheep scratching its backside on the antenna.