Francois Dion - Rocket Science - Water Radio

The case of the Water Radio

I've been wanting to put this information online for the past 5 months, but just procrastinated until today. Early this year, I was working on some sound generating circuit. I hadn't done the amplifier stage yet, so I was feeding it to the microphone input on my Creative Live Drive, with the gain to the max. I had connected my circuit using an aluminium connector and a copper pair back to the microphone input jack. After testing, I disconnected the circuit. Eventually, the connector fell to the ground.

A few days later, I was making a saline solution (water with a lot of salt) and I spilled some on the floor. The solution ran down and touched the connector that was on the floor, still connected to the computer. The loudest 60Hz buzz came out of the speakers. That was expected. What followed wasn't. Just as I was about to pick up the connector, it went silent, like if the pair shorted. I put the connector back on the table. About a minute later, I started hearing what at first sounded like random noise, then I noticed some kind of pattern in it. Before I could start recording it went dead.

I had experimented with VLF tuners back when I was at the university and that sounded just very familiar. But how could a copper wire and salty water be a receiver of some sort? I started searching the web and altough I was finding plenty of VLF hobbyists, crystal radio web sites and others, I didn't find anything that explained or matched my "accident". In the meantime, I tried my experiment again, but all I used was a piece of aluminium foil, a little puddle of salty water and dropped the copper wire from the microphone in, right in the solution. At first, nothing happened, but eventually, something happened and I got it in recording.

The first sample is typical of what I recorded for almost an hour, familiar noise in MP3 format. Sounds like something, it is not random noise. Could be AM radio (you do need a rectifier for that - i made note to investigate this later if it prooved to be the case), but at that point, that was just a gut feeling. I tried appliying filters, looking at spectrograms, but in all, didn't see anything obvious. Another thing that would happen from time to time is complete silence (in MP3 format), right in the middle of this "noise". If you look at the waveform it is pretty interesting too.

I was mentionning the weird signals I was picking up to some people I knew to see if they had any ideas, but nothing came out of that. Then one day in the middle of march, I found the web site of Nyle Steiner, specifically, his Peculiar sounds from aluminium page. Could it be that somebody has figured out what's going on? He uses a 1uF capacitor and a resistor in his case, but it is a similar concept, it seems. Altough, no explanation, just more questions. Were we dealing with a signal caused purely by a chemical reaction, generating a varying voltage, or was the "circuit" acting like a tuner of some sorts? More questions and no answers. I figured that once I had put some of my stuff on the website, I'd contact him and see what he tought of my own recordings.

Every day, I'd record a bit, varying the concentration of the solution. Then, one day I decided to use an aluminium heatsink from an old power inverter instead of just aluminium foil. I'm sure my wife tought I had gone crazy... Finally, on March 26 2005, I had recorded almost 3 hours of output, not listening to it. The next day, when I looked at the spectrogram tough, I noticed this (JPG image). That cant be noise!

You can bet I got real excited when I saw the spectrogram. But I was even more surprised when I listened to this (in MP3 format). This is 1:40 seconds of pretty clear radio reception, starting at 9:07PM eastern time. Altough I am in North Carolina, this radio broadcast appears to be coming from Tennessee. No idea on the frequency or what, but really exciting nonetheless!

I got some theories, like maybe the alu-cu junction in a saline solution acts like a crude rectifier, a semiconductor and this contraption works like a "crystal" radio. I also have more graphs. I have to mention that I was not able to get any other broadcasts since then, but I haven't done any recordings in the past 3 months[that was in July, this is no longer true. I've done many recordings of radio broadcasts since then]. As I upload this page also, I'll be sending a link to Nyle, and see what his toughts are on this. At any rate, if you, the reader, have the patience, do try to replicate this experiment and share your results. It would also be nice to identify the voices and/or radio station on the recording. Stay tuned (!) for the next episode of the Water Radio. Oh, and why Water Radio? Well this radio is copper, aluminium, salt and water, I just picked one of the elements.

Water Radio Episode 2

James Fogg contacted me earlier this week with this information:

"It sounds like you have an accidental diode detector. Add an LC network
or Pi network and you might be able to tune to specific frequencies.

With an AM signal generator and an oscilloscope you might be able to
determine if you have a diode detector. In my youth (30 years ago - my
god!) I played with crystal and cat's whisker detectors and learned some

surprising things will detec,t. A ham friend of mine who used to operate
at legal maximum (1500 watts) using AM told me his neighbor complained
that their toaster used to detect his signal and they had a "talking
toaster". There have also been documented cases of dental fillings
detecting and the person hearing the audio signal (there was a TV show
called the Partridge Family that had an episode about this).

The presence of AL and CU in a brine could do several things, but a
chemist would know. Some things I'd look for are galvanic action,
oxidation (an easy semiconductor source for a diode) and possibly a

electro-chem reaction that makes a battery (two dissimilar metals in
acid will do this, maybe the brine in contact with one of the metals
changes the PH)."

That's what I was thinking, but what I wanted to figure out is what makes the diode. That's the real interest for me. I'm wondering if maybe the salt creates a "liquid crystal" in between the alu and cu. Or maybe as James suggests, it starts working when there is enough oxydation and that would be the diode detector. The galvanic action is also something that needs to be researched.

Many other people suggested similar hypothesis and stories. Jacques Hamel (VE2DJQ) in fact had mentionned I should measure the RF signal itself with a sensitive digital frequency meter. Unfortunately my frequency meter would require an RF preamp to detect. Jacques also suggested I look into late 19th/early 20th century radio experiments. So I started to look at what I had in my library, in particular in "La science illustree" a french scientific publication from the late 19th century. I didn't find anything there. I had Popular Mechanics from the late 30s and 40s, but nothing from before that era, and absolutely no Radio publications from 1900-1910. I had a gut feeling this is where I might be able to find something. What do you search for...

A few hours later, replying to my post on the Geeks list, Bob (WA2EGP) mentionned something that caught my attention in his email. First he said that I had:

"... created a non-linear junction which can mix the
"carrier" and sidebands of an AM radio station and get the audio out
of it.  There's the infamous "carborundum in the filling" story where
some one supposedly could receive stations on their teeth (although
I really think it's an "urban" myth).  I know of several cases of TVI
not caused by a ham radio operator but a bad contact between pieces
of metal that produced harmonics of the ham's transmission.  Considering
that AM stations can run really high power, it's no wonder that a
meager "crystal radio" can receive them.  During WWII, GIs used to
make foxhole radios out of razor blades and scrounged parts."

Did you notice? Radios made from razor blades? How does that work? A bit of searching for foxhole radio and WWII and I came up with

"According to an article in Popular Mechanics of October, 1944, the
blued steel surface of the blade gives the rectifying action needed
for detection without crystals."

from DJ Adamson's Antique Radio Page.

Alright, fair enough. Blueing, like black oxyde finishing, adds a fine oxyde coating to the metal. Or, alternatively, I'm sure a non blued but rusted blade (quite likely what the GIs used as new blued blades, quite likely rationned, would be used for grooming oneself, IMHO) would work in the same way. When I got home, I dug into my pile of PM, and found the October 1944 Popular Mechanics (They republished a slightly enhanced design in October 1962). Still, we are very far from the end of the 19th century or beginning of the 20th. Jacques suggested much earlier material.

After quite a bit of searching, I came upon a technical magazine that started in 1909: Modern Electrics. Apparently, 1909 was a great year for radio hobbyists as on page 17 of the April issue of Modern Electrics, it is mentionned that the "night air full of wireless". Going back a few months, in the January issue, an article mentions the use of a razor blade with carbon pieces to act as a detector. I need to find a library that has this magazine and scan the article in question. What is really interesting here is that the publisher is a young immigrant named Hugo Gernsback, the first publisher of SF. Reality IS stranger than fiction! It is also very interesting that Gernsback himself wrote in Radio-Craft (part 2) about the foxhole radios, in September 1944, pointing out he had "been there, done that" publishing Clark Pettingill's razor radio, 35 years earlier...

Gernsback was one of them obviously and fairly visionary (he had foreseen the tubes going away, replaced by semi-conductors, in 1924). But who is that Dr. Pickard he mentions in this article? Wasn't he the one who invented the "crystal" radio? I had to find out more about his semiconductor experiments.

Greenleaf Whittier Pickard was born in 1877 and patented the crystal receiver in 1906 (fascinating read, BTW). So, I have to search in the 1904-1906 time frame, I would think. Time to hit the libraries and antique book stores...

Of course, up to now I'm assuming a semiconductor detector. But what if it was really another type of detector, possibly a form of electrolytic detector (and this)?

To be continued...

The multimode detector

How about a boa that ate an elephant, or maybe a multimode detector? In this third "episode" of the Water Radio, I have more recordings to share. But first, a bit more detail on the "circuit".

The device is a piece of aluminium. I dont think the shape has much bearing at this point. I've used aluminium foil, an old heatsink with a flat area, an aluminium candle holder etc. This is the black horizontal bar on the picture. On it, a puddle of water with salt diluted in it (it's the gray part on the picture), roughly the size of a quarter.

A copper wire (red on the circuit) goes from the aluminium (I use an aligator clamp to hold it) to the ground (ring) of the microphone input on my soundcard.

The other copper wire (orange) goes from the signal (tip) of the microphone input on the soundcard, with the other end touching the aluminium, in the water. The angle at which the wire is touching the aluminium is important: you want to minimise the surface that is in contact (think cat whisker contact with galena, if you've built a galena radio [radio a Galene si vous etes francais] set before). Beside being of copper, the wire I've used is not solid core (stranded) and of guage 18 or so, with insulation on, if that makes a difference.

In this configuration, the Water Radio works in a mixed mode. What I mean by that is that the circuit acts like a "crystal" or "foxhole" radio, picking up RF signals (up to SW frequencies at least), and at the same time, because of the conductive solution bypassing in part the detector (see the next mode of operation also for the implication of this). This is what makes this circuit ultimately simple, yet oh so complex. It is going to require a lot more experimentation before I get a better understanding of the exact principle at work here (and that means free time, somewhat of a rare commodity around here).

In the meantime, I've gone thru and listened to all that I had recorded, and here is more from the radio program I got from Nashville, TN on march 26 2005. This is right about 9pm eastern time. I had to boost by 15dB to hear the voice. But it is pretty clear (hear the mp3). The station is on 5.935MHz (emitter #2, 100KW, according to their website) so it is shortwave, and not the commercial AM band that I tought I had been receiving.

This is the same circuit, but a little clamp is holding the (orange in the picture) wire suspended in the solution, not touching the aluminium. In this mode, there is no rectification going on and all it does is convert RF to an electric signal. What it means is that I'm only able to receive the frequencies that the soundcard can sample, and at 44KHz, that means 22KHz and below (Nyquist et al).

Are there RF emissions below 22KHz? Yes. First of all, there are signals that are transmitted at greater frequencies, but end up affecting the 22KHz and below range. Like LORAN. Another source is because of the weather. It shouldn't be too surprising when we know that a spark gap will broadcast all frequencies from DC to MHz and beyond. Similarly, lighting strike does something similar. And when these are bounced off the atmosphere, some form of waveguide effect filters and creates interesting artifacts in that frequency band. It is called VLF, btw.

Listening to VLF is not something new, as I mentioned in episode 1. It is reported as early as the late 19th century. Also, NASA has the INSPIRE project and they make available a tuner kit. Many schools around the country have signed up. If your school hasn't been able to get a kit due to money issues, or yourself, then today is your lucky day. With this (at a cost of a few cents), and an inexpensive tape recorder with a microphone in (or the microphone in of the soundcard on your computer) you can start experimenting. [NOTE: it is not confirmed yet that this mode of operation truely receives VLF signals. Shortwave reception is confirmed, see next section.]

I'm sure you all want to hear something. Well, here is a small MP3 sample of session I did on August 9th, at around 10pm: Reception 8/9/05 2153h to 2156h. What is really interesting is how immune to 60Hz hum from powerlines this circuit really is. I'm in the middle of the city. With a typical VLF receiver, usually you have to be very far from any power lines.

In that sample (as I reported to the VLF_Group on Yahoo) you can hear the RF from a lightning strike at 35 seconds into that recording, again at 44 seconds and 1:35. Bird chirps type of sound from 2:30 to 2:42.

RF or bust

As I mentionned in the previous section, it is not confirmed yet if I'm receiving VLF yet. I will be building a spark gap emitter and see if I get something, to confirm if I'm picking up RF.

Also, I think I need to pick up an actual VLF receiver and do an A/B comparison. An update for those that want to buy a VLF kit from NASA:

From: "Gallagher, Dennis" 
Date: Thu Aug 25, 2005  6:25 pm
Subject: INSPIRE VLF3 kits
I finally got a message back reporting that the VLF3 kits can still be
purchased. Bill Pine continues to run the INSPIRE project and I am sure
we will find another NASA supporter at Goddard Space Flight Center who
will attempt to pick up after Bill Taylor's loss.

Dennis

=================================
Dennis L. Gallagher
NASA Marshall Space Flight Center
NSSTC, XD12
320 Sparkman Drive
Huntsville, AL 35805
(256)961-7687 (office)
(256)961-7249 (fax)
=================================

The opinion on the VLF group on Yahoo is divided as to whether I am picking up VLF. On one hand I have gotten opinions like this:

From: "INSPIRE Team I-1" 
Date: Fri Aug 12, 2005  4:17 am
Subject: Re: [VLF_Group] Digest Number 992

Hello Francois,

Welcome aboard! Your first audio example (foil_radio_1.mp3) is a great example
of electrostatic discharge (also referred to as coronal discharge or brush
discharge). This type of signal is received directly beneath heavy thunderstorm
clouds as the electric field strength increases in intensity. The sound can
range from clicking and popping to scratching sounds and even whines, screeches
and squeals. As a side note, these are also the type of sounds people report
audibly hearing with their ears during northern light displays.

Your second example (reception_8_9_05_2153_2156.mp3) sounds like a second
example of electrostatic discharge. There are a couple short samples of
electrostatic discharge on the vlf group (static02.mp3 &
static_charge_9Sept2003_1528CT.wav). Often the whining and squealing will stop
if lightning strikes nearby and then begin once again as the electric field
strength increases once again.

ps.- I've burned out a couple transistors by listening to a VLF receiver under
conditions like this (before adding protection to my VLF receiver).
Most whistlers are generated by lightning out of state so I rarely run the risk
of listening when storms are nearby.

All the best,

Shawn

But then I've gotten replies like these:

From: pan@...
Date: Fri Aug 12, 2005  4:54 am
Subject: Re: [VLF_Group] New to list

Francois Dion wrote:
> I just joined ...

Hi Francois, welcome to the list.

I had a listen to reception_8_9_05_2153_2156.mp3.

This doesn't sound like any radio signal at all.

Those squealing noises sound just like electrical breakdown
across a dielectric gap. For example, corona discharge,
leakage across a capacitor, or a dry soldered joint.

From your website:
: What is really interesting is how immune to 60Hz hum
: from powerlines this circuit really is.

Which tends to confirm that you're not receiving VLF
radio signals, otherwise mains related signals would be
the most prominent.

Nyle Steiner did email me back. First, he confirmed that in his experimentation, he did not have the wire touching the aluminum, it was just touching the water. Regarding VLF, he's expressed doubt that it is picking up VLF but mentionned:

"It is a good observation on how it is similar to lightning signals but a better way to hear them would be to use a coil of about 40 turns and 3 feet in diameter. "

So here ya go. Get some magnet wire, 40 turns on a 3 ft diameter circle and connect that to your microphone in and that'll get you VLF. So the only advantage of the Water Radio, if it were picking up VLF is 60 Hz hum rejection. On top of that, if it is not VLF, what is it?

So maybe we dont have confirmation on the VLF, but I have plenty of proof of shortwave reception. One of the longest stable reception I've recorded was done on August 11th at night. On that day, I tuned in a broadcast from europe, talking about vacations for the Swedes, and an interview (with some other station fading in and out). I do not have the frequency on that one.

As you remember, my first ever RF SW signal that I was able to confirm as such (due to getting a radio program and a station identifier) was a signal from Tennessee, on 5.935MHz. That's roughly 50 meter in wavelength. Obviously, I'm nowhere near even a 1/4 wave antenna with about 2ft of copper wire... In that regard, it's pretty impressive that I get such a good reception.

I've put some more audio recordings from the Water Radio, in the download section. I'd be curious to hear from other people what station they've received. I'd also would like to identify that european program I received. I've received WCCR many times and I have a theory as to why, but I'll get into that another day. I will also be experimenting with longer and shorter wires.

Standalone Water Radio

I was asked by a reader if I could come up with a way that would not require a computer and soundcard. Something easily available and cheap. Earlier, I had suggested a tape recorder with a microphone input. And if you visit Nyle Steiner's Peculiar sounds from aluminium page you'll see a solution using an amplified speaker. Nyle didn't include where to get this and what part # so I tought I'd put this down to make an "official" kit.

The Radio Shack speaker pictured here and its thousand of other uses are detailed on my "Electronic duct tape" page. We will need a cable with a 1/8" mono jack on one end and a pair of alligator clips on the other end. To hold the wire in the electrolyte we will also need something like a "helping hand":

• RS # 277-1008 mini amplified speaker ($12.99)
• RS # 64-2991 Kronus helping hands ($12.99)
• RS # 42-2421 6' shielded cable ($1.97)

"CELL" DETAIL

For the cable, you just need to add an extra piece of copper wire (as pictured above) which you will clamp with the red alligator clip, or you can cut the red alligator clip, and expose some of the copper wire.

Finally, a piece of aluminium (like a small heatsink or foil - regular not the non stick kind) will also need to be procured. The thicker the piece, the more immune this is to 60Hz hum. The size in the other two axis dont seem to matter as much.

For the liquid, tap water, salty water or peroxyde can all be used (other electrolytes too) - all items found in your kitchen or bathroom. Note that the wire is not touching the aluminium. In the above pictures, peroxide was used and a lot of bubbles can be observed along with copper deposit. Using a heatsink made from anodised aluminium allows contact between the heatsink and wire and prevents copper deposit.