Unconventional electrolysis

Alan Smith

I considered putting a 140mm computer fan, but in the end the problem occurred because the reaction did not operate as intended, producing a continuous short-circuit at the electrodes instead of intermittent plasma/noise-producing ones at a high rate. Under that mode of operation the coil does not seem to warm up much more above ambient temperature, but I haven't measured this precisely with temperature sensors.

In the past couple days I did a few more tests.

The most useful one for the purpose of reproducing the hissing noise is what I made today. I found that it helps to have a limited amount of water and a progressively decreasing gap (wider at the bottom and narrower at the top) as I think I confirmed that the origin of such noise appears to be the gap region exposed to air and traces of water and metal particles. I've made a few videos showing this. Here initially the gap was about >1 mm at the bottom and 0.3-0.4 mm at the top.

I recommend clicking the links to watch them in higher quality.

https://streamable.com/g65px

(Link)

I've also tried running the electrodes out of the acidic water bath. When they're in operating conditions they keep making the same hissing noise and high voltage bluish sparks for a while.

https://streamable.com/ch8yu

(Link)

Other videos in the same series:

001 Hissing loudly. The sudden noise was a steel cap getting magnetically attracted to the coil during current surges

002 Began hissing, then started drawing excessive current as the noise stopped

007 Running quieter

The biggest problem against reliability (the reaction evolving into an unrecoverable short circuit) is from conductive particles electrically connecting both electrodes in undesired areas, for example their sides. So in practice it would probably help covering them with heat/chemical-resistant paint like magicsound planned doing. I couldn't manage doing that efficiently just using thin mica spacers: the electrodes eventually fail working as intended.

I've also recorded directly (192 kHz audio) from the AM radio the hissing noise. I thought I had the radio tuned to 530 kHz, but upon inspection (as of writing) it seems I had it tuned to 1600 kHz.

Such files are too large to be uploaded here directly in their entirety, but here's a link to a representative excerpt (.WAV format) corresponding to the below screenshot, in full quality:

https://mega.nz/#!y9BkwIzB!Evx…L8iWzMOvTzn2JcZuFiIE_5sVc

At the beginning of the spectrum there is background radio noise, which then evolves into noise picked up from the jar.

EDIT: for those who want to analyze the entire audio file(s), here is the archive (2.56 GB, Audacity format):

https://mega.nz/#!OwIi0axL!h_4…vJ_DLuGabmWYQlK-bHBVxy82o

At the very least, noise reduction would have to be performed to remove the background radio and 50 Hz (?) signals.

I think in the end the above experiments do not show much more than an electrolytically-actuated spark plug-ignition coil system (if such thing actually exists), where transient short circuits caused by material deposited from the anode are the "switch".

There are videos showing spark plugs running outside their engine block where a noise somewhat reminiscent to what I'm observing is generated, like for example:

So what would be the differences in my case from ordinary spark plugs? I can guess at least as follows:

• Significantly higher currents involved
• Lower voltages (probably less than 1 kV but possibly still in the range of several hundred volts - depends on the coil and current)
• Electrode material is deliberately blown in relatively large amounts into the plasma region
• Significant presence of water
• Discharges occurring over a wide surface rather than a limited area
• No attempt nor desire in reducing the EMI produced in the process
• Probably more tolerance to overdriving faults - if anything due to to the larger amount of materials involved

EDIT: in the automotive field an oscilloscope is sometimes used to check out the ignition coil signal. From a section of the cleaned out radio signal where the spark rate was not very high I get a waveform similar to this:

I guess the individual events could be compared to a typical oscilloscope signal from an automotive ignition coil, here showing a single spark event. I think in my case the initial high voltage spike couldn't be properly recorded with the radio equipment in its entirety:

Source: https://www.interworldna.com/p…s._secondary_circuits.php

There's an explanation of the various points in the above graph:

Quote

The secondary waveform shows the voltage required to jump the plug's electrode (A), and (B) the length of time that the HT is flowing across the spark plug's electrode after its initial voltage to jump the plug gap. This time is referred to as either the ‘burn time’ or the ‘spark duration’.

In the illustration shown, it can be seen that the horizontal voltage line in the centre of the oscilloscope (C) is at fairly constant voltage of approximately 3 kV. This voltage is referred to as the Sparkline kV. This voltage is the voltage required to maintain the spark flow across the plug's electrode, and is determined primarily by the secondary resistance within the HT circuit. From the 0 ms point on the scope to point D is the spark duration, in this case around 1.0 milliseconds. The waveform is then seen to drop sharply into what is referred to as the ‘coil oscillation’ (E). The coil oscillation should display a minimum number of peaks (both upper and lower) and at least of 4 - 5 peaks should be seen. A loss of peaks on this oscillation shows that the coil needs substituting. An example of a faulty coil and the subsequent loss of oscillations can be seen in Fig 1.2. The oscillation seen at point (F) is called the ‘polarity peak’; this voltage will be of the opposite polarity to the plug firing voltage as this is created when the magnetic flux is initially built, or at the start of the dwell period.

The radio transmitting qualities are a side-effect, although with undamped discharges getting produced at a relatively high rate I kind of expected that significant radio noise would end up getting produced.

What I didn't initially expect was that roughly the same noise getting acoustically emitted from the jar/electrodes would also be present in the form of a radio signal at the frequency band at which the radio was tuned to (1600 kHz) and possibly above that - in an earlier test I could listen it even on some empty FM radio stations from the same transistor radio.

As the aim in these tests has been also producing discharges as powerful and abrupt as possible within the limitations of the power supply while also keeping everything as simple as possible, pretty much all measures against normally undesired radio noise as used in ordinary spark systems have been omitted.

If anything, having now established how it's actually operating (at least for the radio noise part) I'm interested checking out if anything changes/improves by making it work better as an ignition coil/spark system while still keeping it electrolytically operated. Perhaps the cheapest/easiest modification could be including a better metallic core to the coil rather than the bunch of ferromagnetic tools I have used for a while.

Paradigmnoia

I think these tests I'm doing are still relatively safe in that there are no extreme voltages (yet, at least) and arcing over large distances is unlikely to occur, although relatively low voltages at high currents can be dangerous too. It's not clear what kind of voltages can be achieved following contact separation (from the short-circuits blowing up electrodeposited material transiently forming thin conductive paths), but they seem at times large enough to apparently produce arcs jumping the entire gap width.

magicsound

That looks nicer than what I planned doing. I was thinking in simple terms of "bigger is better" and only planned to increase coil inductance as much as possible to make it more likely to produce discharges at a higher voltage.

I found challenging to say the least to make my contraption operate under stabilized conditions for prolonged periods of time; the pulse/discharge rate also appears to vary greatly (from less than 100 Hz to at least a few thousands Hz) depending on gap conditions, temperature, solution pH, etc. Likewise I expect that finding such resonant frequency (if it isn't a moving target) will be a difficult task as long as the electrodes are allowed to self-regulate as I've been doing so far.

It will be interesting to see with proper equipment up to what frequencies this resonating noise actually shows up.

DnG

I'm not sure what I'm looking at, but it appears there are several coils intended for high-voltage, high-frequency operation?

magicsound

In the latest tests performed on 2019-01-04, which produced the resonant noise almost right away compared to previous attempts, I used a wider gap at the bottom portion immersed in water (using a couple folded mica spacers obtaining a gap visually slightly larger than 1 mm), and narrower one at the top portion (using 4 mica spacers piled on top of each other, obtaining about a 0.4mm gap). The rationale for this was thinking that in the immersed portion conductive debris accumulation would be more likely to occur than in the portion exposed to the atmosphere and therefore that it might have needed a larger gap to mitigate unrecoverable, non-intermittent short-circuit faults/failure modes. Furthermore I thought that providing a variable-sized gap could have increased the chances of reproducing the previous results.

I think in the end what helped the most was having a significantly lower amount of water in the jar than usual, so that a larger portion of the gap would be exposed to just traces of liquid water, water vapor and oxygen-hydrogen gases instead of being completely immersed in water. Material deposition still occurs under these conditions and eventually fine conductive metal particles find their way to the top of the electrodes. EDIT: I think for a similar reason the electrodes seem to work better when they are completely immersed in water (producing a stronger plasma reaction) when the water is almost boiling or boiling. I think this is not due to the higher conductivity of water at high temperature (which seems to work against this process), but due to the voids formed by cavitation.

In retrospect (now that you mention it) by doing the opposite and providing wider gap at the top this could have been similar in some ways to a Jacob's ladder, but I don't think I have observed a similar effect. Sparks appeared to be produced mostly randomly over the inner gap surface.

From minute 1:03 in this video there might have been a few transient events reminiscent of that, but it looks like they were still composed of several brief randomly located discharges in quick succession.

Quote from can

I think in the end what helped the most was having a significantly lower amount of water in the jar than usual, so that a larger portion of the gap would be exposed to just traces of liquid water, water vapor and oxygen-hydrogen gases instead of being completely immersed in water. Material deposition still occurs under these conditions and eventually fine conductive metal particles find their way to the top of the electrodes [...]

It might or might not be entirely related, but possibly similar conditions have been mentioned in this relatively recent Steemit MFMP blog post covering certain water arc experiments using high voltage discharges:

https://steemit.com/steemstem/…nd-charge-metals-for-lenr

Quote

Water spark plug as a Strange Radiation generator

This circuit may be a good device to study strange radiation. The challenge is that it seems to be sensitive to the amount of water that is applied across the junction, too little and no effect, too much, not enough of an effect. [...]

Although there is a significant difference in that the experiments I've been doing recently operate because of electrode erosion and that I do not directly use high voltages (but anode material deposition eventually causes short circuits, which in turn may cause high voltage spikes), there could be some correlation with the above aspect.

Wyttenbach

Since electrolysis is indirectly what makes the observed reaction possible (through electrodeposition processes "bridging" both electrodes and making them transiently short-circuit), hydrogen should be present. Furthermore the discharges produced should be able to dissociate traces of water that is always inherently present in the electrode gap under the conditions of these tests.

Probably a larger amount of hydrogen would be present inside the gap if the electrodes were to operate into a tighter/constrained space rather than a relatively spacious jar.

Today I tried replicate the previously reported test to make sure that it was as simple as I assumed.

Good news is that it does seem that having a low amount of electrolyte (or perhaps a better definition could be "plating solution" in this case) seems to help greatly in producing the resonating noise. The bad news is that reliability is still an issue and that under my conditions it's difficult to keep the electrodes running for a prolonged period of time without them eventually unrecoverably short out.

I tried using a slightly improved core for the coil over the previous solution, but it didn't seem to bring noticeable changes. Perhaps it wasn't that much of an improvement after all.

This time too I used a narrower gap at the top of the electrodes (5 spacers for about 0.5mm gap) and a folded mica spacer at the bottom (obtaining roughly a 1.0mm gap). Note that in this and the previous test I have "flipped" the electrodes, so that the surface worn from earlier test is now out of the solution.

As usual, after the experiment starts (using distilled water in just enough amounts to wet the bottom portion of the electrodes), adding HCl drops substantially increases the deposition process and current through the electrodes (which quickly increases from tens of milliamps to several amps), and together with this the previously observed noise also arises. So far at least this seems to be a reproducible process.

Videos

• [001]
  • Noisy operation at the beginning of the experiment, just after adding suitable amounts of HCl. The electrodes short-circuited thereafter due to debris accumulation that couldn't get blown away by the available current. The beeping is from the current clamp, due to current exceeding the measuring range of 40A. When I set it to the 400A scale, it read 41.8A.
• [004]
  • Some time after the experiment started and I managed to restore normal operation. At this point for some reason the resonating noise wasn't as constant as earlier on. Note the darkened solution from the fine metal particles dissolved in it.
• [005]
  • Later on operation turned into an unusually acoustically "silent" mode. The AM Radio noise recorded at the time didn't seem to have had large changes.
• [006]
  • Further later on the resonating noise occurred at a rather low rate. The single discharge events could almost be counted by ear at times.

Other observations

• Accumulation of debris in unwanted parts of the electrode assembly, both on top and on the bottom or on the sides of the gap, causes unrecoverable short circuits.
  • The bottom clip can be omitted if a staggered gap is provided (narrower at the top), as long as the residual pressure from the top clip will keep the bottom spacer in place.
• Allowing the electrodes to dry causes serious reliability problems in that the particles deposited become very difficult to remove.
• Water got evaporated rather quickly today.

Wyttenbach

I'm not sure what you mean exactly with spark horn. Do you mean something like this (schematically)?

Or do you mean more like a Jacob's ladder?

(Source: the internet)

In general, as I answered earlier to magicsound, so far I haven't tried to intentionally do this. I'm not sure if it would work in my case, however. I'm not directly producing high voltage arcs with an external ignition coil that can easily jump through a narrow gap as in the photo above. High voltages in my case can get produced only after contact separation when the two electrodes are shorted (i.e. touch) and then quickly unshorted thereafter, which eventually spontaneously occurs in these experiments. The voltage directly applied to the electrodes here is only 12V DC.

Below is a schematic description of what happens with the electrodes (or at least what I think happens). The process occurs randomly 100~1000 times per second over the entire gap surface. It's worth pointing out that material is not only electrodeposited from the anode, but can also get transported and accumulated from the water present in the gap (which should be considered more like a water/metal/metal oxide slurry):

The voltage induced when the contacts forcefully separate can be calculated as [ -L * (dI/dt) ] as explained here: https://en.wikipedia.org/wiki/Inductor#Constitutive_equation

In my case L is probably in the range of 0.001-0.002 H, dI possibly up to 40A and dt can probably be measured in the order of hundreds of nanoseconds, but these values except for L will vary greatly over time depending on testing conditions.