Transference of Electric Power – Single Wire vs Telluric

In this new experiment on transference of electric power a comparison is made between power transfer through a single wire and through a telluric transmission medium, using a cylindrical Tesla magnifying transformer (TMT) apparatus. The TMT apparatus and linear generator is the same used in the High-Efficiency Transference of Electric Power series  both over 1.5m and 11m, and these new experiments are a continuation on those previously reported. This experiment is also the first in a new series on telluric transmission of electric power, and whilst I have experimented with telluric transmission over the years, none of this fascinating area of Tesla research has yet been reported here on the website. One of the pictures in the main slider at the head of this website shown here, shows telluric reception experiments made in 2017 at the upper parallel mode of a nominally 2Mc, 160m amateur band, flat coil. In the experiment reported in this post, the TMT transmitter and receiver are housed in different buildings of the lab, and can be connected by a 30m single wire, or a telluric channel ~18m point-to-point between the two ground systems, and 26m in total length including the cables. There is no special consideration of the ground/earth/soil between the two buildings, although the transmitter ground system used is specifically designed and constructed to provide a low impedance connection to ground.

Wireless transmission of power at a global level appears to have been one of Tesla’s greatest vision’s and endeavour’s, and one that he appears to have invested so much of his time, effort, and money. From early experiments in his New York laboratory, to larger scale experiments at Colorado Springs, to the grand-scale transmitter at Wardenclyffe, which unfortunately does not seem to have been operated in earnest before being dismantled. Tesla communicated this work mainly through his patents[1,2], demonstrations and presentations[3,4], and personal research notes[5]. In more recent years his life and work have been discussed and considered in a lot of detail, and there are many different perspectives online regarding all aspects of his endeavours, from whether there was/is any basis for this TMT system to work at all, all the way through to detailed analysis of how such a system was constructed, how it was intended to be operated, and the kind of results that could be accomplished in power distribution through this method. What is much more rare is solid experimental evidence, measurement, and subsequent consideration and analysis of what can be experimentally accomplished in transferring power between a transmitter and receiver in a TMT arrangement through the earth. This specifically includes what power levels can be transferred over what distance, at what frequencies, with what level of losses, and through what transmission principles and modes, and in addition, what the impact of this would be for the surrounding environment and life in general.

What follows are my own considerations and perspectives on Tesla’s Wireless Power, and what I feel are some of the most important considerations for these types of experiments. I will over a series of posts be demonstrating aspects of these principles, and looking at the type of results that I have been able to accomplish so far in this field. For me, Tesla’s “wireless” power as a description of the field is somewhat misleading, as in my perspective it never really was “wire-less”, in other words it never involved no “wires” between the transmitter (TX) and receiver (RX). By this I mean that the TMT apparatus, to transfer even the tiniest amounts of power between TX and RX, requires a single transmission medium of lower impedance than the pervasive surrounding medium, and connected from the TX secondary coil lower-end, to the RX secondary coil lower-end. So if we assume that the pervasive surrounding medium is air, then the single transmission medium of lower impedance might include, for example: a single metallic wire, a telluric channel through the ground, or even a gas discharge tube that has been ignited by the potential gradient across the Tesla coil secondary. If this single transmission medium is not present then only minute levels of power can be transferred between the TX and RX, consistent with transverse electromagnetic propagation from a radio transmitter to a radio receiver.

I could conjecture that Tesla may have seen his TMT approach to power distribution as distinctly “wire-less” when compared to other electrical systems of the time, like Edison’s DC power distribution, that required two conductors to make an electric circuit between the generator and load, and hence the normal losses that occur in a closed loop electrical circuit. In simple comparison, Tesla’s system appears as an open loop electrical circuit relying on the potential gradient of the “cavity” established across the secondary coil of the TX, the transmission medium, and the secondary coil of the RX. In this cavity it has been suggested that a different mode of transmission can be established, the longitudinal magneto-dielectric (LMD) mode, which is again distinctly different from the transverse electromagnetic (TEM) mode, and in principle can lead to very high-efficiency of power transfer, over very large distances, and with very low losses. These modes have been proposed and explored in detail by electrical researchers such as Eric Dollard[6-10] , and other aspects of wire-less power by researchers including Tucker et al.[11] and Leyh et al.[12], and also in my own experiments in the Transference of Electric Power series on this website.

Another important aspect that has been widely discussed is the requirement for the ground systems used at both the TX and RX in a TMT system, to present the very lowest impedance possible, or resistance at resonance, to connection of the TX and RX coil to the telluric channel. This would appear to be common sense, at least for the TEM transmission mode, where the lowest losses in the system will occur when the impedance of the ground connection at the TX and RX are at their lowest, combined with the lowest impedance of the single transmission medium between the two. However, this is not necessarily the case for the LMD mode, where in my own experiments and particularly the first in the sequence on High-Efficiency Transference of Electric Power, it is demonstrated that the efficiency of the power transfer increases as the impedance of the single-wire medium increases. In particular it was demonstrated that more than 500W of power can be transferred through a single wire no thicker than a human hair, a 40AWG (0.08mm or 80 microns) nickel plated copper wire, where the power transfer efficiency could be measured up to 100% according to the limits of experimental accuracy of the measurement equipment.

It was suggested in this previous experiment that … “Power transfer of this order through such a thin wire is possible as the dielectric and magnetic fields of induction are contained or guided around the single wire. Removal of the single wire from the receiver end prevents any power transfer to the receiver, which shows that when driven by a linear sinusoidal generator, a lower impedance transmission medium, (in this case the single wire), is needed to guide the induction fields between the transmitter and receiver coils.” So in principle it could be conjectured from this unusual result with a very fine single wire, that provided the correct LMD mode is arranged in the cavity of the TMT system, the lower impedance of the ground system may not be as important as previously suggested. Certainly for the TEM mode in the transmission medium large losses will occur from higher impedance connections and the single wire medium itself, as well as radiative losses along the length of the single wire, and reflections from impedance mismatches and transitions across the cavity. From my results I conjecture that the combination of the TEM mode in the TX primary, LMD mode in the cavity formed by the TX secondary, transmission medium, and RX secondary, and the TEM mode in the RX primary, leads to the highest efficiency in the transference of electric power, and is discussed in detail in High-Efficiency Transference of Electric Power – 11m Single Wire.

The LMD mode also removes the requirement for every part of the TMT system to be in principle at the same resonant frequency. It seems to be widely thought that the highest efficiency of transmission of power takes place in a TMT system when all the sections are arranged to resonate at the same frequency, hence forming one continuous minimal impedance coupled resonator system. Whilst again this would most likely be the case for the TEM mode, from my own measurements it is not so for the conjectured LMD mode. I have measured that at highest efficiency of power transfer the LMD mode in the single-wire is not the same frequency as that measured in the primary of the TX or the RX. Furthermore, there is spatial coherence of the LMD mode but not temporal in the cavity. In the TEM mode there is temporal coherence across the cavity but not spatial, measured, presented and considered in detail in Transference of Electric Power – Part 1. These experiments and their results, suggest that there are significant differences between the TEM and LMD modes, and how a TMT system performs when it is arranged to operate in one mode or the other, or in a combination of both modes, which I conjecture and have in-part confirmed through measurement, is actually the optimum arrangement for the highest efficiency of power transfer.

This experimental post consists of two video experiments one based on a single-wire 30m TMT system, and the other with the same TMT system connected by a telluric channel. Telluric is often used as a description of the transmission medium in Tesla research when the ground/earth/planet is used to form the “single wire” and hence the cavity between the TX and RX. In this case the impedance of the Telluric channel is much higher than that of the single metallic wire, and hence we make a comparison as to the likely power that can be transferred through the channel, what modes of transmission are involved in the system, and what the magnitude and mechanisms of the losses are involved in the channel. The generator used in both experiments is the same linear amplifier generator featured in the High-Efficiency Transference of Electric Power series, and is explained in detail in those posts, and is used to drive the TMT system at the fundamental series resonant mode. In addition to measuring power transfer in both mediums, the small signal impedance characteristics of the TMT system are measured, and then tuning and matching the generator to the apparatus to ensure maximum power transfer to the experiment with the minimum losses.

The first video experiment of a cylindrical TMT system with a 30m single wire demonstrates and includes aspects of the following:

1. A Cylindrical TMT experimental apparatus using a 30m single wire transmission medium between the transmitter (TX) and receiver (RX) coils.

2. Setup, matching and tuning, and operation of a 1kW linear amplifier generator, adjusted to drive the TMT experiment at the available series resonant modes, and further adjustment during operation to maximise the power transfer efficiency, and minimise reflected power.

3. Small signal ac input impedance characteristics Z11 from the perspective of the generator, and showing tuning of both the series and parallel resonant modes to establish optimum experimental starting conditions.

4. At 1.920Mc using a balun feed to the transmitter the maximum power transfer efficiency was measured at ~ 34%.

5. At 1.890Mc without using a balun feed to the transmitter the maximum power transfer efficiency was measured at ~ 40%. This frequency and drive method produced the highest efficiency observed during the experiment.

6. The optimum power transfer was accomplished with the maximum number of four primary coil turns, and balanced parallel modes, at both the TX and RX coil.

7. Extension of the single wire from 30m to 40m, close to the quarter wavelength of the generator drive frequency, did not change the maximum measured power transfer efficiency of ~ 40%.

8. It is discussed and conjectured that the TEM transmission mode is dominant in the experimental setup, and as a result large losses occur through radiation from the single wire.

9. It is conjectured that the LMD transmission mode was not adequately established in the single wire over 30m or 40m, and hence the much lower power transfer efficiency than expected from previous experiments with 1.5m and 11m single wires. In previous experiments with an 11m single wire transfer efficiencies up to 96% were measured, and it was conjectured that the LMD mode was adequately established as the dominant transmission mode.

Video Viewing Note: The video control bar has a “Settings” cog icon where you can select video quality, which by default is set to “Auto”. For clear viewing and reading of the VNWA software characteristics and text on the computer screen, “1080p” video quality is recommended, and may need to be selected manually from the settings icon once playback has started.

The second video experiment of a cylindrical TMT system with a telluric channel demonstrates and includes aspects of the following:

1. A Cylindrical TMT experimental apparatus using an 18m Telluric transmission medium between the transmitter (TX) and receiver (RX) coils.

2. Setup, matching and tuning, and operation of a 1kW linear amplifier generator, adjusted to drive the TMT experiment at the available series resonant modes, and further adjustment during operation to maximise the power transfer efficiency, and minimise reflected power.

3. A custom ground system, using copper water pipes driven into the ground, and consisting of a main RF ground and a reference test ground.

4. Small signal ac input impedance characteristics Z11 from the perspective of both the TX and RX, and showing tuning of both the series and parallel resonant modes to establish optimum experimental starting conditions.

5. Large signal tuning using a small breakout flair at the top of the telescopic tuning aerial attached to the top-end of the TX secondary coil.

6. Signal reception tuning, using a Sony ICF-2001D radio scanner, to calibrate the proportion of signal transmitted through the radio-wave and the telluric-wave from the transmitter to the receiver.

7. At 1.860Mc 10W of input power at the TX resulted in ~0.55mW via the radio-wave, and ~0.7mW via the telluric-wave, and a total of ~1.25mW at the RX coil, into a HP435B power meter with an 8481H 3W thermocouple power sensor.

8. At 1.860Mc 500W of input power at the TX resulted in a total of ~80mW at the receiver through the radio-wave and telluric-wave combined.

9. It is discussed and conjectured that almost all of the transmitter power is absorbed into the earth around the ground system, and radiated from the secondary coil in the TEM transmission mode. This diffuse absorption and radiation around the transmitter system results in very little power incident on the RX system, and hence at 1.860Mc in the 160m amateur band, radio communication appears possible through the telluric system, but significant transference of electric power does not appear possible at this frequency.

Video Viewing Note: Again “1080p” video quality is recommended, and may need to be selected manually from the settings icon once playback has started.

Figure 2 below shows the schematic for the experimental apparatus used in the video experiments. The high-resolution version can be viewed by clicking here.

Experimental Apparatus and Operation

The schematic and principle of operation for the experimental apparatus used in the video experiments is a variation to that used in the High-Efficiency Transference of Electric Power series. Much of the equipment used, and a detailed explanation of the linear amplifier generator are covered in that series of posts. The power measurement meters have been changed from the Bird 4410A analogue thruline power meters, and replaced with 4391A digital readout thruline power analysers. The digital readout of the Bird meters makes them easier to read both during the experiment, and on the video. The 4391A power meters were both calibrated using the same inline method previously presented, at 500W input power for direct comparison on a single range, and with a limit of experimental error of <0.5%. User uncertainty and errors in reading the analogue dial during the experiment is further reduced through using the digital readout. The other significant measurement additions are for the telluric transmission experiments where the received powers are much smaller and hence different instruments have been used. Radio signal strength is measured using an Sony ICF-2001D radio scanner, which has been adapted to allow for a direct BNC input for external antenna connection, as well as the integral telescopic aerial mode. Direct received power levels are measured using a Hewlett Packard HP 435B power meter with a HP 8481H 3W thermocouple power sensor.

As demonstrated in the video the 30m single wire transmission experiment is initially setup using the VNWA and these results are discussed below. This provided a tuned starting point for the complete TMT system, where the TX and RX coils are arranged to resonate at the same frequency. The fundamental series resonant mode was initially set at 1.92Mc, but then subsequently empirically adjusted to 1.89Mc for slightly increased transfer efficiency across the single wire. The parallel modes of both the TX and RX coils were balanced, using their primary coil tuning capacitors, to equal magnitude of impedance, before connecting the 500W load to the RX coil output. Care was taken to keep the operating frequency of the TMT system within the 160m amateur radio band, and also where the lab is in a remote setting to minimise any operation interference on adjacent radio bands. With the initial conditions set the power was increased gradually from 10W up to over 500W, whilst minimising any reflected power back to the linear amplifier by adjustment of the Palstar antenna tuner. In this way power could be passed to the transmitter across the 30m single wire and into the receiver to power the load.

It is important to note from the generator tuning and operation in part 1 of the video experiment the differences that arise in power measurement at the linear amplifier output, (as measured by the MFJ-998), and that measured at the input to the TX coil by the Bird 4391A. The 4391A measures forward and reflected power right at the input to the primary coil which depends on  the match between the antenna tuner output impedance and the TX input impedance. At resonance the input impedance of the TX coil is predominantly resistive and the SWR as measured by the 4391A varied in the range 1.5-2.7 dependent on the fine tuning of the antenna tuner. The power measured at the output of the Kenwood linear amplifier by the MFJ-998 is now on the input side of the Palstar antenna tuner, where the tuner is transforming the impedance of the TX coil to be as close to 50Ω as possible, minimising reflected power back to the linear amplifier, and allowing maximum dynamic range and output power utilisation from the linear amplifier. So when the MFJ-998 SWR is minimised in the experiment as close to 1.0 as possible, the MFJ-998 is used to measure the power supplied from the linear amplifier into 50Ω impedance, and the 4391A is user to measure forward power into the impedance presented by the TMT system at its primary coil input at the TX. These two meters will then read a different power when at the minimised SWR of the linear amplifier, and will read more closely the same as the SWR is detuned at the antenna tuner to correspond to the TMT input impedance.

For consistency in experimental measurement of the input power and output power of the TMT system the forward power measured both by the 4391A at the TX, and the 4391A at the RX was used to assess the power transference efficiency across the TMT system. This was then compared at two tune conditions of the antenna tuner, firstly when minimising the SWR presented to the linear amplifier which leads to different power readings on the MFJ-998 and TX 4391A, but minimal reflected power at the linear amplifier output. The second with detuned SWR between ~ 1.5-1.9 presented to the linear amplifier which leads to close match between the power readings on the MFJ-998 and the TX 4391A, but with slightly reduced efficiency and increased reflection at the output of the linear amplifier. As demonstrated in part 1 of the video, the experiment was initially operated using a 1:1 current balun at the output of the 4391A in order to properly convert the output from the unbalanced output of the generator, to the balanced input condition of the primary coil. However this appeared to reduce power transfer efficiency by up to 5% and was subsequently removed from the experiment when it was empirically retuned to 1.89Mc.

For the telluric measurements where the RX coil is not visible to the TX VNWA measurement the initial tuned conditions were set to 1.87Mc using the VNWA small-signal fundamental series mode of the TX coil connected to the telluric ground system. This was slightly empirically adjusted to 1.86Mc when tuning using the large-signal generator drive, and corresponded with the maximum neon brightness at the top-end of the TX secondary coil. A small breakout flare was generated at high TX input power > 700W which also was maximised around 1.86Mc. Care needs to be taken only to use this as a large-signal tuning check, as any breakout at the top-end of the secondary coil will reduce the top-end impedance of the coil to the surrounding-environment effectively increasing the quarter wave length of the coil, and hence reducing the series resonant frequency of the TX coil. By both empirical tuning methods 1.86Mc was determined to be the optimal large-signal generator driving frequency to the telluric connected TX coil with 39cm defined telescopic aerial extension. This tuned telluric experimental frequency keeps all the experiments in the 160m amateur band with a very high-Q TX and RX coil, and hence very tightly contained transmission bandwidth using only CW and morse-code for radio call-sign identification.

Another key measurement in the telluric experiment which needs some consideration is the process of measuring the radio-wave of a radio transmission. For all radio transmission, and as transmitters are almost always grounded down to earth, there major component of the transmission, that is the propagating TEM wave from the radio transmitter antenna to the receiver antenna. In relation to the telluric part of this experiment, we cannot assume that all the power transferred from the TX to the RX coil is via the telluric channel through the ground, as there will also be a radio-wave component at the receiver. We also cannot simply remove the bottom-end ground connection of the RX coil to measure the this radio-wave component, as this will change the wire-length of the secondary cavity, and hence change its fundamental series resonant frequency, and any connected receiver which is tuned to the transmitter frequency will erroneously show no received signal, simply because the RX coil is not correctly tuned to the transmit frequency.

To accomplish the radio-wave part of the experiment, and as demonstrated in part 2 of the video experiment, the telluric ground connection is removed from the RX coil, and is replaced with a single wire 10m in length which is NOT connected into the ground or to any other grounded end-point. The telescopic aerial at the top-end of the RX coil is now fine adjusted so that the series mode resonant frequency of the RX coil matches the transmit frequency. This is accomplished by maximising the received signal at the receiver at the correct TX frequency, and then cross checked by VNWA measurement to confirm correct tuning of the RX coil. In this way the RX coil is now tuned to the correct frequency for receiving the transmitted signal, but is also not connected in any way to the ground. The signal strength now received on the radio receiver, or power meter, is a result of the radio-wave contribution only, and is less than the combined radio-wave and telluric-wave, as can be seen in the video experiment. The proportion of radio to telluric wave can also give a good indication as to the dominant transmission mode involved in the transference of electric power between TX and RX coil. Equal radio and telluric components tend towards a dominant TEM mode of propagation between the two, or with a combination of TEM and LMD, with the TEM mode dominant. A much larger telluric wave can indicate a dominant LMD mode, and this will be demonstrated in the Telluric Transference of Electric Power series. In this experiment the radio-wave and telluric-wave contributed about equal proportions of the received power in the telluric part of the experiment.

Figures 3 below show a range of pictures of the experimental apparatus, measurements, and some of the key setup conditions for both the single wire and telluric experiments. It is interesting to note that in fig. 3.3 the phone on the top of the MFJ-998 shows the live image of the remote camera setup in lab2 to monitor the RX coil and apparatus. The remote camera is connected through local WiFi in lab2, and then to the router in lab1 by wired LAN connection between the two labs. This live remote video monitoring allows operation of the TX system whilst monitoring directly the RX system, and to produce the live inset video in both parts 1 and 2 of the video experiments.

Telluric Ground System Design and Construction

The ground system associated with a TMT system has always been considered as a critical part of the engineering required to make a successful telluric transmission system, with minimal losses and maximum transferred power between the TX and RX coils and the telluric transmission system. Tesla himself noted that it is necessary to get a firm grip on the ground if it is to be resonated by his wireless power system, and a lot of effort was poured into minimising the impedance, or resistance at resonance, of the connection between the ground electrode and bottom-end of the Tesla coil[5]. Subsequently in conversations with Eric Dollard he has pointed out that it is imperative to get as much copper into the ground as possible for any telluric experimentation and get as close to 0Ω as possible, in other words, to minimise the contact resistance of the Tesla coil secondary bottom-end to the transmission medium in the ground. In addition for a true Tesla transformer, as the bottom-end of the secondary is connected to ground by the minimum impedance possible, the top-end of the secondary needs to present the highest possible impedance of the coil at an elevated position above the ground, and preferably with a top-end load such as a metal sphere, ball, or toroid.

Arranged in this way, and according to conjectures and postulation on the LMD mode,  the Tesla transformer ot TMT system forms a complete longitudinal cavity from the high-impedance top-end of the TX coil, through the telluric transmission medium, and up to the high-impedance top-end of the RX coil. It is in this condition that the coherent LMD mode facilitates the very high efficiency transference of electric power between the generator and the load. The Tesla transformer also fulfils the key step of transforming the TEM mode in the primary to the LMD mode in the secondary cavity, meaning that the generator TEM mode is transformed to LMD mode in the single wire or telluric cavity, and then back again to the TEM mode in the primary of the receiver. It is conjectured that the LMD, or longitudinal mode as it is often referred to, forms a standing wave across the cavity with one or more defined null points  in the cavity, and hence as such, is not subject to the same losses as a propagating transverse electromagnetic wave. Some of my own experiments in the Transference of Electric Power series, and the High-Efficiency Transference of Electric Power series, appear to support the existence of the LMD mode, and that indeed very high power transfer efficiency can be established between the TX and RX coils of a TMT system in the close mid-field region.

It remains to be experimented and tested to see if this LMD conjecture can be extended across far-field distances and does indeed result in lower power transmission losses, and hence higher efficiency of power transfer. As a point to note, I do also myself debate the necessity for the lowest resistance connection to ground when the LMD mode is properly established. The coherence across the entire cavity of the LMD mode should not necessarily require a low impedance connection to ground, or even a low impedance telluric transmission medium. This would certainly be necessary if the transmission mode is by TEM propagation, where any higher impedance, mismatch of impedance, absorption and reflection of power, and radiation losses will make for huge loss of power across the distance of the transmission medium. All these factors are certain for TEM power transmission, but not all may apply for a coherent LMD mode properly established over the transmission cavity. This conjecture remains to be confirmed or refuted through experimentation.

Accordingly in my own experiments, and as a starting point for my telluric experiments, I designed and constructed a ground system within the available space, materials, and budget that are accessible to me at this time. From the perspective of getting as much copper into the ground or the lowest resistance to ground, this appeared as the best place to start, that is, to enable maximum possibility of receiving a signal through the ground at distance whether it be by TEM or LMD transmission modes, or a combination of both. The design uses 22mm copper water pipe which gets a good quantity of copper into the ground, and with reasonable surface area, by simply drilling holes and driving in tubes, as opposed to having to dig or excavate large pits in the ground. The essence of the water pipe is that at any time water can be piped through the ground system and down to where the contact between the copper and ground is actually occurring. In addition small holes where drilled along the length of the underground copper pipes to allow water to escape along the length and hence irrigate the soil around the pipes extend into the ground as well as at the end of the pipe. This gives the possibility to prepare the ground system before experiments to irrigate the surrounding underground earth and reduce the ground system resistance to the earth to as low as possible. This water irrigation works well as intended, and after about 1 hour of irrigation the ground system impedance falls significantly.

In order to make measurements of the ground system performance I included a single copper pipe reference where I can measure the impedance between the main ground system and the reference ground system. As this experiment is an introduction to my telluric experiments so far, I will include these measurements in the start of the reported series on Telluric Transference of Electric Power. The preparation of the ground system involves connecting the reference ground to the main ground via plastic water hose, and then first connecting the top-end of the reference to the main water supply. This quickly allows water to flow into an fill up all the pipes in the system, and before the water has a chance to soak away from the ends and through the holes into the surrounding earth. When this is done the water direction is then reversed to fill from the bottom-end connection of the main ground system and left for up to an hour to back-fill all the pipes, soaking away into the ground and reducing the contact resistance between the earth and the copper. The total exposed (above ground) length of the main system is 3m, and the underground copper is ~ 15m. The total physical size of the ground system is less than one-tenth of the wavelength of the generator at the 160m amateur band.

The main and reference system have a copper electrical feed point which is soldered into intimate contact with the copper pipework and does not disturb the water flow within the pipes. The main system is connected by a 4m 0AWG micro-stranded silicone coated cable to the lower end of the TX secondary coil, and for the reference to measurement equipment via a 4m 12AWG micro-stranded silicone coated cable. With irrigation for ~ 1hr, and in the winter months with good rainfall, so the water-table is at its maximum in the area, the impedance of the main ground system to the earth can be as low as ~ 15Ω @ 1.86Mc. The total impedance of the main ground system, as measured with the reference system, and including the 4m 0AWG ground cable between the bottom-end of the secondary and the ground system terminal, is ~ 40-60Ω @ 1.86Mc, dependent on season and irrigation. This is approximately one-third to one-half of the resistance of the secondary coil at resonance, and as such presents a reasonably low and solid connection for the TX coil to ground at the frequency of operation, and was practical to construct and build in the space available. I would have preferred a centre fed star arrangement for physical construction, consistent with many preferred ground system arrangements used by radio amateurs in the MF and HF bands, however my available space did not allow for this, and I adopted a straight design with the same amount of copper underground. All in all the ground system so far has proved to be effective, and I have been able to measure telluric transmission of power and signals over significant distance from the transmitter, which will be presented in the Telluric Transference of Electric Power series.

It should also be noted, and as indicated in the schematic in fig. 2, that the linear amplifier generator is NOT connected itself to the RF main or reference ground system used in the telluric experiments. This is arranged in order not to introduce uncertainty into the source of any measured telluric transmission through the earth. For generator safety during operation the equipment and components of the generator are connected together by their earth chassis connections and then in-turn connected to an isolated line supply earth. This continues to protect the generator equipment and components in the event of an electrical fault, whilst isolating the earth connection from the telluric RF ground system, and hence not influencing or confusing the measured results.

Figures 4 below show pictures of the final main and reference ground system outside lab1, and some from its construction in 2019. For the telluric comparison in this experimental post lab2 uses a dedicated RF ground through a single copper coated steel ground rod, and is typical for use in amateur radio work when connecting a linear amplifier transmitter and receiver. Clearly in this experiment the TX and RX ground systems are quite different in size, copper under the ground, and hence contact resistance between the telluric transmission medium and the RX coil.

Small Signal AC Input Impedance Measurements

Figures 5 below show the small signal ac input impedance Z11 measured directly on the experimental system, and using an SDR-Kits VNWA vector network analyser, as used on many experimental pages on this site. The measurement setup, equipment, and connection to the experimental apparatus is shown in fig. 3.2.

To view the large images in a new window whilst reading the explanations click on the figure numbers below.

Fig 5.1. Shows the input impedance Z11 over the range 100kc to 5Mc for the TX coil primary connected to the VNWA, and with balanced parallel modes with the primary tuning vacuum capacitor set to 396pF. The bottom-end of the secondary coil is connected directly to the main ground system as would be the case in the telluric experiments, and the top-end telescopic aerial is set at its default length of 39mm that sets a wire-length that corresponds to ƒS = 1.87Mc @ marker M2 for the fundamental series resonant mode. This same point was empirically adjusted to drive at 1.86Mc for optimum large-signal tuning. The TX coil resistance at the series mode M2 presents a resistance of 24.1Ω which is conveniently very close to one-half of the optimum generator system output impedance of 50Ω. This could ideally be connected to the generator directly using a high-power 1:2 current balun with minimal if any antenna tuner transformation to the linear amplifier. For flexibility in tuning for this experiment the Palstar antenna tuner was used directly to transform the 50Ω output of the linear amplifier to the 24.1Ω at the TX coil primary input. The lower and upper parallel modes from the TX primary and secondary coil are impedance magnitude balanced, with lower mode ƒL = 1.62Mc @ M1, and the upper mode ƒU = 2.24Mc @ M3. It should be noted that the RX coil for this measurement is also connected and correctly tuned to its own ground system at lab2 in the reciprocal arrangement, but cannot be “seen” at all in this VNWA measurement.

Phase change is consistent with a typical high-Q, loosely coupled and loosely wound, Tesla coil, and the series and parallel modes all occur at a phase angle of ~ 0° consistent with a resonant circuit mode. This characteristic presented in fig. 5.1 forms the base small-signal impedance characteristic for the telluric experiment presented in this post, and also for experiments presented in the Telluric Transference of Electric Power series in the 160m amateur band in the MF band. Lower frequency telluric experiments in the LF band have very different characteristics and will be presented in future posts. For best match to the linear amplifier generator the fundamental series mode ƒS is used as the optimum driving point where most power can be coupled directly into the secondary cavity. At the receiver in telluric experiments both the series and lower parallel mode can be tuned to the 1.86Mc and both are useful for different aspects of the measurement. For signal strength experiments using the radio scanner the parallel mode is best as it presents a high-impedance to the output of the RX coil, which is well suited to maximum incident voltage at the input to the radio tuner. For absolute power measurements using a 50Ω power sensor, in this case the HP 435B with HP 8481H sensor, tuning the RX coil to the series mode is necessary for making power measurements, where the transfer of power between the RX coil and sensor input impedance is best optimised.

Fig 5.2. Shows the characteristics for the complete TMT system connected by the 30m single wire, and without the 500W load connected at the primary of the RX coil. The low impedance of the single-wire transmission medium allows the VNWA to “see” the characteristics of the RX coil reflected into the input impedance measurements. This is particularly useful to accurately setup the TX and RX coils, at least for the TEM modes, where their fundamental series resonant modes can be matched, and the parallel modes can also be matched. Here in this characteristic the parallel modes are shown as balanced, and the series mode is that of the TX coil dominant at ƒS = 1.88Mc @ M4. The system is unloaded at the RX coil and hence this is the highest-Q measurement of the TMT system, where the parallel modes at both the TX and RX are very sharp and also split to give two peaks at the lower mode, and two peaks at the upper mode. The primary tuning capacitors CPTX = 354pF and CPRX = 498pF have been adjusted to bring about the best empirical balance between the parallel modes, and hence equal influence of the parallel modes in all four coils, two primary coils, and two secondary coils. I have discussed and conjectured in Cylindrical Coil Input Impedance – TC and TMT Z11 that balance of these four parallel modes in a TMT system is the optimal starting point to maximise the generation of the LMD mode across the TMT cavity, and that the LMD mode can be further fine tuned by adjusting the parallel modes at both the TX and RX coil.

It should be noted that the frequency split in the upper and lower parallel modes is quite narrow, (as compared to say the TMT system measured in the close mid-field region in High-Efficiency Transference of Electric Power over 1.5m, and show in fig 3.2), which shows the reduced coupling between the TX and RX coil over the longer distance of the 30m single-wire. Over the 1.5m single-wire the lower parallel modes where split by ~ 70kc, whereas here they are split only by 30kc. There are also low impedance series points at M2 = 1.66Mc, and M6 = 2.32Mc which could be alternative driving points for the linear amplifier generator. Both points have significantly higher impedance presented to the generator, and hence M4 remains the best point to drive the TX coil for maximum transference efficiency across the TMT.

Fig 5.3. Here the 500W load has been connected at the output of the RX coil primary, and the series fundamental modes have been finely tuned and balanced using small changes in wire-length affected through the telescopic aerial at both the TX and RX secondary coils. The final tuned lengths of the aerials are TX = 39cm, and RX = 37cm, and these were used as the base tune when needing to reset to a known starting condition. Adding the 500W load has collapsed the parallel modes at the RX coil, although of course they remain part of the actual electrical system at the receiver. The close tuning of the series modes leads to frequency splitting through beat frequencies between the two resonators which results in the double phase relationship seen at markers M2, M3, and M4. Two fundamental series resonant modes, and upper and lower, are now present at ƒSL = 1.85Mc @ M2, and ƒSU = 1.92Mc. The upper series mode formed the starting frequency for the 30m single wire experiment where the input impedance is resistive, RSU = 51.5Ω @ M4 and very close to the untuned system output impedance of the linear amplifier generator at 50Ω. This driven point was subsequently moved to M3 at ƒS = 1.89Mc which yielded a slight increase in transfer efficiency. The parallel modes of the TX coil remain largely unaffected by the split series modes and are balanced with slight adjustment to CPTX = 371pF.

Fig 5.4. Here the matched series fundamental modes have been detuned by increasing the wire-length using the telescopic aerial at the TX coil from 39cm to 50cm. This reduces slightly the lower series frequency, ƒSL = 1.82Mc @ M2, which then becomes the dominant mode with respect to the generator drive. This dominant series mode reduces the input resistance of the TMT system, RSL = 35.2Ω @ M2, and slightly imbalances the parallel mode tuning at M1 and M5.

Fig 5.5. Here the matched series fundamental modes have been detuned by reducing the wire-length using the telescopic aerial at the TX coil from 39cm to 17cm. This increases slightly the upper series frequency, ƒSU = 1.96Mc @ M4, which then becomes the dominant mode with respect to the generator drive. This dominant series mode reduces the input resistance of the TMT system, RSU = 26.9Ω @ M4, and slightly imbalances the parallel mode tuning at M1 and M5, the other way from fig. 5.4. It should be noted that the centre drive point of the upper and lower series modes at M3 remains less impacted by the frequency detune of the series modes, and hence represents the optimal stable drive point over the dynamic range of the experiment with ƒS = 1.89Mc @ M3. The higher impedance of point M3 requires further tuning using the antenna tuner, or is also suitable for 1:4 current balun at the input to the TX primary coil.

Fig 5.6. Shows the effect of increasing the single wire length from 30m to 40m, which also makes the single wire almost exactly a quarter wavelength of the generator drive frequency. The increased wire-length in the cavity has increased the overall wire-length of the TX and RX coils at their bottom-ends, and hence the five resonant points of interest indicated by markers M1-5, have all shifted down slightly in frequency. The centre drive point at M3 now being at 1.86Mc rather than at 1.89Mc. Otherwise the TMT system impedance characteristics remain largely unchanged, and the quarter wavelength length of the single-wire does not have such a big impact as might be at first expected given the complete impedance transformation from a short circuit to open circuit across a quarter wavelength wire. And this is an important point to note, that the length of the cavity is now defined by the quarter wave TX and RX coil plus some of the wire-length at the bottom-end and the top-end that is within the magnetic coupling distance of the coil. For example, if we take just the TX coil and add a single wire at its bottom-end of say 1-2m, this will have a very distinct change on lowering the fundamental series mode frequency ƒS. If we now add a further 5m to the single-wire this further reduces ƒS, but not to the same amount. Adding a further 10m has even less impact on reducing ƒS.

So the impact of adding single-wire length to either end of the coil has diminishing impact to ƒS with increasing length, and this is the product of the wire-length which is within the magnetic field coupling of the coil. And this is what is happening with the increase in single-wire from 30m to 40m. Only the wire length up to about 5m from the bottom-ends of the TX and RX coil have a significant impact on reducing the ƒSL and ƒSU, whilst the 20 or 30m in the middle makes much less difference to the TEM frequency characteristics of the TMT system. So increasing from 30m to 40m single-wire is not really about the quarter wavelength impedance transformation, but rather simply an increase to the middle section of the transmission medium, with only slight impact on the frequency of the five resonant points of interest. This would continue for increasing length of single-wire with diminishing impact on the frequency characteristics until the TEM losses along the wire length collapse the coupling of TX and RX coils.

Single Wire Comparison at Lengths 1.5, 11, and 30m

In this experiment with a 30m single wire in the TMT cavity the best result obtained at 1.89Mc was 200W supplied by the RX coil to the load, for 500W supplied to the TX coil by the generator, yielding a power transfer efficiency of 40%. This is very much lower than that obtained for the 1.5m @ 99%, and 11m @ 96% in the High-Efficiency Transference of Electric Power series. The biggest loss mechanism in this experiment is expected to be radiative losses from the single wire, as the single-wire did not heat up, and the components at the TX and RX did also not heat up. Some power would have been lost in resistive losses along the single-wire length, but the majority of the power would have been radiated from the single wire acting as a long-wire antenna between the two coils. This means that the TEM mode was the dominant transmission mode in the cavity, whereas it has been conjectured in the 1.5m and 11m single-wire experiments that the LMD was dominant and resulted in very low losses along its length, and particularly in the case of the 11m single-wire.

If the LMD mode conjecture is developed further for this experiment, then it is clear that despite careful tuning and adjustment of all the series and parallel modes in the TMT system, and the careful adjustment and exploration of frequency around these modes, it was not possible to engage the LMD transmission mode as the dominant transmission mode, and for as yet unknown reasons. Without the LMD mode the TEM mode leads to considerable radiative losses at the frequency used from the wire length, which is of course why power transmission at high frequencies over large distances using single-wires is impractical when only the TEM mode is involved. It is unclear why the LMD mode could not be engaged in this setup as per the 11m single wire, as nothing else has significantly changed in the experimental apparatus, operation, or measurement method. I do not currently see the increase from wire length from 11m to 30m to have such a substantial change on the LMD mode conditions that would be required to be established, but nonetheless there are clearly other unknown factors in the setup and balance of these modes over single-wires of increasing distance.

Single Wire vs Telluric Transmission Medium

One of the central aims of this experiment was to make a side-by-side comparison of a TMT system, where the transmission medium between the TX and RX is a direct connected single-wire, or a telluric channel through the earth, and where both channels were of comparable distance between the TX and RX. The total length of the telluric channel in this experiment was, 4m TX earth wire + 18m telluric point-to-point + 4m RX earth wire,  or minimum length of 26m. This actual length of the channel may be longer than this, if we consider that the telluric channel may not be a direct point-to-point path between the two ground systems. Nonetheless a 26m telluric transmission channel was considered comparable in length to the 30m single-wire. What we see from the measured results in this experiment is orders of magnitude difference in the transmitted power between the TX and RX with the two different transmission mediums. As already discussed the best result so far for the 30m single wire ~ 200W from 500W efficiency 40%, whereas for the 26m telluric channel the best result was ~ 80mW from 500W or an efficiency of  0.016%. For the 80mW received at the power meter with an optimum impedance match of almost 50Ω between the RX coil output impedance and the power meter sensor, 35mW ~ 44% is from the radio-wave with no connection to the telluric ground system, and 45mW ~ 56% is from the telluric-wave via the ground system.

This enormous difference in power transfer through the telluric ground system implies that almost all of the power at 1.86Mc has been absorbed into the ground, in other words to heat up the ground around the main TX ground system, with very little of it being transferred to the RX ground system. It is expected that the impedance of the telluric connection between the two ground systems for the TEM mode is likely to be much higher than the single-wire. Whilst the telluric system does not have the same radiative losses as the single-wire the power is easily absorbed by the transmission medium, and especially at the higher frequencies being used for this experiment. The reasonably close balance between the radio-wave at 44% and the telluric-wave at 56% suggest to me that the TEM transmission mode is again dominant in this experiment. This is an important point to note, that at the frequency used, we would expect the telluric medium losses to be very high, which they are, but we are also interested in the dominant mode of transmission in the medium. It can also be conjectured that the balance of the radio-wave and telluric-wave can also be used as an indication of the dominant mode. With approximately balanced radio and telluric waves I conjecture that this indicates a dominant TEM mode, whereas with a much stronger telluric wave without loss of received power could indicate a dominant LMD mode. I raise this conjecture here as I have measured much larger imbalances in the radio and telluric-wave in other telluric trials over longer distances which will be presented in subsequent posts, e.g. at 2 miles the radio to telluric-wave proportion was measured to be ~ 1:5 for only 10W of generator input power.

In the 1.5m single-wire experiment in High-Efficiency Transference of Electric Power it was demonstrated that an increase in the impedance of the transmission channel using a single wire no thicker than a human hair, a 40AWG (0.08mm or 80 microns) nickel plated copper wire, actually increased the efficiency of power transfer at 500W. So the concept of increased impedance in the telluric channel is not necesarily a limitation to high efficiency power transfer, provided the LMD mode of transmission is the dominant mechanism. Even if this were the case in the current experiment and the LMD mode was dominant, I would still expect high power loss from absorption into the earth at the frequency being used in the HF band at 1.86Mc. There has been considerable discussion in the field regarding the best frequency for telluric power transfer and/or communication, what frequency the earth is electrically resonant at, and what is the earth’s impedance and admittance to different modes of transmission both over the surface, and deeper into the body of the earth. I will look at these areas in more detail in my next post on Telluric Transference of Electric Power.

Summary Conclusions and Next Steps

In this post transference of electric power has been explored and demonstrated, using a TMT system with two distinctly different transmission mediums between the TX and RX coils. Tuning of the different series and parallel modes of the TMT system have been well explored, and demonstrate many aspects of the TEM characteristics of single-wire transmission line systems. Telluric transference of electric power has been introduced along with the apparatus, method, and forms of measurement required to characterise this fascinating area of Tesla research. From the experimental results and measurements presented the following observations, considerations and conjectures are made:

1. The maximum 30m single-wire efficiency that could be established in this experiment was 40%, where the losses along the single-wire are predominantly radiative from the long wire acting as an antenna, and some resistive losses along the wire length.

2. From the results obtained the predominant transmission mode along the 30m single wire is expected to be transverse electromagnetic propagation, the standard TEM mode of transmission.

3. It is conjectured that the LMD mode, for as yet unknown reasons, could not be tuned as the dominant transmission mode in this experiment, which also led to high losses, and huge reduction in power transfer efficiency. This is directly in contrast with the results obtained in 1.5m and particularly 11m single wire experiments, where the LMD mode was established between the TX and RX coil, with a null node in the centre, and maximum electrical intensity at the top-end of each of the TX and RX secondary coils.

4. The tuning and matching of the series and parallel modes of the TMT system are conjectured as important to establishing the LMD mode, and so far in this experiment, the correct balance of these modes has not yet been established. It is considered that it is possible to establish the correct setup for the LMD mode to be dominant, and that this may result in a much higher transfer efficiency between the TX and RX coils.

5. The 26m telluric transmission channel resulted in very high losses, with a transfer efficiency of no more than 0.016%. These losses are expected to be predominantly through absorption of the transmitted power into the earth at the frequency used of 1.86Mc.

6. The proportion of radio-wave to telluric-wave in the telluric experiment was 44% : 56%, and so it is conjectured that the TEM transmission mode was again dominant between the TX and RX ground system.

7. It is conjectured that the high impedance of the telluric transmission medium , and the connection of the TX and RX coils to the ground, is not necessarily a limitation to the high efficiency of power transfer when the LMD mode is dominant in the transmission medium.

The next step to the single wire part of this experiment would involve working with the TMT tuning and setup conditions, in order to attempt to resolve conclusion 4, and establish the LMD mode as the dominant transmission mode, and in a similar way as was accomplished for the 11m single wire. If this cannot be established then the conditions for the LMD mode, and its limitations, need to be studied in more detail.  For the telluric transmission medium I will be presenting more experiments and results for progressively further distance from the generator and out into the far-field.


1. Tesla, N., System of Transmission of Electrical Energy, US Patent US645576A, March 20, 1900.

2. Tesla, N., Apparatus for Transmitting Electrical Energy, US Patent US1119732A, January 18, 1902.

3. Tesla, N., Experiments with alternate currents of very high frequency and their application to methods of artificial illumination, American Institute of Electrical Engineers, Columbia College, N.Y., May 20, 1891.

4. Tesla, N., Nikola Tesla on his work with alternating currents and their application to wireless telegraphy, telephony and transmission of power: an extended interview, 1916 Interview – ISBN 1-893817-016, Twenty First Century Books, 1992.

5. Tesla, N., Colorado Springs Notes 1899-1900, Nikola Tesla Museum Beograd, 1978.

6. Dollard, E., Condensed Intro to Tesla Transformers, Borderland Sciences Publication, 1986.

7. Dollard, E., Theory of Wireless Power, Borderland Sciences Publication, 1986.

8. Dollard, E. & Brown, T., Transverse & Longitudinal Electric Waves, Borderland Sciences Video, 1987.

9. Dollard, E. & Lindemann, P. & Brown, T., Tesla’s Longitudinal Electricity, Borderland Sciences Video, 1987.

10. Dollard, E., A common language for electrical engineering – lone pine writings, A&P Electronic Media, 2013.

11. Tucker, C. & Warwick, K. & Holderbaum, W., A Contribution to the Wireless Transmission of Power, Electrical Power and Energy Systems 47 p235-242, 2013.

12. Leyh, G. & Kennan, M., Efficient Wireless Transmission of Power Using Resonators with Coupled Electric Fields, Nevada Lightning Laboratory, 40th North American Power Symposium, 2008.

13. A & P Electronic Media, AMInnovations by Adrian Marsh, 2019,  EMediaPress

14. Dollard, E. and Energetic Forum Members, Energetic Forum, 2008 onwards.


 

Telluric Transference of Electric Power – MF Band 2-8 Miles

This experimental post is a follow-on from the Telluric experiment presented in Transference of Electric Power – Single Wire vs Telluric. In that previous experiment a Tesla Magnifying Transformer (TMT) apparatus, consisting of TX and RX cylindrical Tesla coils, were connected together via a 18m point-to-point telluric transmission medium, and with ground connection cables 26m in total between TX and RX secondary coils. In the medium-frequency band (MF) at 1.86Mc, in the mid-field region, 500W input power to the TX coil generated ~ 80mW of output power at the RX coil, from a combination of the telluric-wave and radio-wave. In this new experiment the same TMT apparatus and generator is used, and the telluric transmission medium is extended into the close far-field region at 2 and 8 mile field locations from the TX coil. In both locations natural water features were used as the telluric ground connection for the RX coil, and the transmitted signal could be clearly received, and was shown to result from the combination of a telluric-wave component through the ground, and a radio-wave component above ground. It is conjectured that at the 2 mile location the longitudinal magneto-dielectric (LMD) transmission mode was dominant in the telluric cavity between TX and RX, and the transverse electromagnetic (TEM) mode was dominant at the 8 mile location.

The video experiment demonstrates and includes aspects of the following:

1. Portable Tesla receiver (RX) setup and tuning, using a cylindrical coil tuned in the 160m amateur radio band, for radio-wave and telluric-wave field experiments in the close far-field region.

2. Telluric ground connection using a submerged aluminium metal plate, firstly in a natural lake connected to a river 2 miles from the lab transmitter (TX), and secondly in a man-made reservoir 8 miles from the TX.

3. Small signal ac impedance measurements using a vector network analyser to tune the RX Tesla coil to the series and parallel resonant modes.

4. Fine tuning to different modes, and optimal received signal strength at 1.86Mc, using a telescopic aerial at the top-end of the RX secondary coil.

5. Comparison of radio-wave and telluric-wave measurement by re-tuning the RX coil from the Telluric ground plate connection, to an ungrounded single wire bottom-end extension.

6. At both 2 and 8 miles the CW audio tone could be received and heard at only 10W TX input power.

7. At 2 miles, 6 bars of signal strength were measured at 10W TX power at 1.86Mc for the telluric-wave and radio-wave combined, and 1 bar for the radio-wave only.

8. At 8 miles, 4 bars of signal strength were measured at 400W TX power at 1.86Mc for the telluric-wave and radio-wave combined, and 2 bars for the radio-wave only.

9. The lower parallel resonant mode of the RX Tesla coil was found to receive the maximum signal strength at both 2 and 8 miles.

10. The lower parallel resonant mode was found to be much more sensitive to body and object proximity than the series resonant mode.

11. It is conjectured that at the 2 mile location the longitudinal magneto-dielectric (LMD) transmission mode was dominant in the telluric cavity between TX and RX, and the transverse electromagnetic (TEM) mode was dominant at the 8 mile location.

Video Viewing Note: In the video the telluric-wave (in the ground) is referred to as the ground-wave, and the radio-wave (over the ground) is referred to as the sky-wave, and not to be confused with the amateur radio definitions of ground and sky wave.

The experimental apparatus, generator and operation, and the TX ground system, is exactly the same as that used in Transference of Electric Power – Single Wire vs Telluric, and is discussed and presented in detail, along with the full experiment schematic, in that post. Operation of the generator in this field experiments is via a research colleague at the lab, and setup, tuning, and operation of the generator can be viewed in detail in the single-wire experiment video presented in the aforementioned post.

A key measurement in the telluric experiments which needs some consideration is the process of measuring the radio-wave of a radio transmission. For all radio transmission, and as transmitters are almost always grounded down to earth, the major component of the transmission is the propagating TEM wave from the radio transmitter antenna to the receiver antenna. In relation to a telluric experiment, we cannot assume that all the power transferred from the TX to the RX coil is via the telluric channel through the ground, as there will also be a radio-wave component at the receiver. We also cannot simply remove the bottom-end ground connection of the RX coil to measure this radio-wave component, as this will change the wire-length of the secondary cavity, and hence change its fundamental series resonant frequency, and any connected receiver which is tuned to the transmitter frequency will erroneously show no received signal, simply because the RX coil is not correctly tuned to the transmit frequency.

To accomplish the radio-wave part of the experiment, and as demonstrated in the video experiment, the telluric ground connection is removed from the RX coil, and is replaced with a single wire 10m in length which is NOT connected into the ground or to any other grounded end-point. The telescopic aerial at the top-end of the RX coil is now fine adjusted so that the series mode resonant frequency of the RX coil matches the transmit frequency. This is accomplished by maximising the received signal at the receiver at the correct TX frequency, and then cross checked by VNWA measurement to confirm correct tuning of the RX coil. In this way the RX coil is now tuned to the correct frequency for receiving the transmitted signal, but is also not connected into the ground.

The signal strength now received on the radio receiver, or power meter, is a result of the radio-wave contribution only, and is less than the combined radio-wave and telluric-wave, as can be seen in the video experiment. The proportion of radio to telluric wave can also give a good indication as to the dominant transmission mode involved in the transference of electric power between TX and RX coil. Equal radio and telluric components tend towards a dominant TEM mode of propagation between the two, or with a combination of TEM and LMD, with the TEM mode dominant. A much larger telluric wave can indicate a dominant LMD mode, and this is demonstrated at the 2 mile field location.

Small Signal AC Input Impedance Measurements

Figures 2 below show the small signal ac input impedance Z11 measured directly on the RX coil of the TMT system, and using an SDR-Kits VNWA vector network analyser, as used on many experimental pages on this site.

To view the large images in a new window whilst reading the explanations click on the figure numbers below.

Fig 2.1. Shows the small signal ac input impedance Z11 of the RX cylindrical Tesla coil, connected via the aluminium grounding plate submerged in a natural river-fed lake at the 2 mile location. The grounding plate is connected to the bottom-end of the RX secondary coil via an 8m 6AWG micro-stranded, silicone coated cable. The RX coil was tuned by adjusting the length of the secondary top-end telescopic aerial, as shown in the video, and in this measurement shows tuning to the lower parallel mode, (in this case the parallel mode of the secondary coil), at ƒL = 1.86Mc @ M1. The RX coil is setup without using balanced parallel modes, as with very small signal reception experiments the additional capacitive loading appears to reduce the amplitude of the measured signal via the Sony ICF-2001D radio receiver. At ƒL the input impedance, (output impedance presented to the radio receiver), is RL ~ 1719Ω. The higher impedance of the lower parallel mode is more suited than the low impedance of the series mode, to directly feeding the Sony radio AM external antenna input, and hence the input impedance of the super-heterodyne first stage receiver in the Sony. Maximum signal reception results were consistently accomplished in the field using the lower parallel mode tuned to the transmit frequency of 1.86Mc.

The fundamental series resonant mode here occurs at ƒO = 1.99Mc @ M2, and again can also be tuned to 1.86Mc by longer extension of the telescopic-aerial. A comparison of the receiver measurements were made in the video against the lower parallel and series modes, and it was determined that the lower parallel mode produced the best results for measurement with the Sony radio receiver, and the series mode would be better for direct power measurements using the HP435B with HP8481H thermocouple power sensor which has a 50Ω input impedance. For the most accurate direct power measurements the output of the RX coil should ideally be matched to the 50Ω input impedance of the sensor, ensuring maximum power transfer from the RX receiver coil to the HP power measurement system. If and when higher powers can be measured using direct power measurement, then a 2:1 current balun would be suitable to affect quite a good match between the RX coil primary output RS ~ 29.6Ω, and the HP power sensor at 50Ω. The upper parallel mode ƒU = 3.96Mc @ M3 originating from the primary coil, cannot be used in this particular experiment as it cannot be tuned down sufficiently low to 1.86Mc using either additional wire length (lowering the series mode), or loading the RX primary coil directly with parallel capacitance.

Fig 2.2. Here the RX coil at the 2 mile location has been tuned to the lower parallel mode ƒL = 1.86Mc @ M1 with the 10m ungrounded single wire at the bottom-end of the secondary coil, and adjustment of the wire-length of the secondary via the telescopic aerial length from 39cm to 45cm. The Q of the RX coil is noticeably higher from being ungrounded and the lower parallel resonant mode impedance is higher at RL ~ 2666Ω. The series resonant mode ƒS = 1.99Mc @ M2 is slightly stronger, and has a lower impedance RL ~ 17.5Ω. Otherwise the characteristics are very similar to when the aluminium telluric ground is being used. This tuned characteristic using the 10m ungrounded single wire was used to measure the radio-wave component of the received signal, which at the 2 mile location, was much lower than the telluric-wave component.

Fig 2.3. Shows Z11 of the RX coil connected via the aluminium grounding plate submerged in a reservoir at the 8 mile location. The parallel mode is here tuned to 1.85Mc rather than 1.86Mc, and there is a consistent 1Hz tuned error throughout this experiment at the 8 mile location. When checked the 1Hz difference did not make a discernible difference to the received signal strength or reception at the field location when using either the lower parallel or series resonant modes. It is interesting to note that the Q of the RX coil system is higher at the 8 mile location, and is more similar to the 10m single wire result in fig. 2.2, than the telluric-plate result in fig. 2.1. It could be considered that this may indicate that the telluric connection to the earth was not as good at the 8 mile location, something which was certainly reflected in the much reduced received signal strength measurements.

Fig 2.4. Here the series mode is now tuned at 1.85Mc, and it is interesting to note that the series mode impedance is again not much higher than that for the 10m single wire results in fig. 2.2, again suggesting that the telluric connection is not as good at the 8 mile location. So both the lower parallel mode and the series mode are closer here to the 10m single wire results achieved at the 2 mile location, and that may suggest that the 8 mile location was more suited to reception of the radio-wave, and less to the telluric-wave. This was indeed what was measured, that the telluric-wave and radio-wave contributed almost equally to the received signal strength at this location, and a lot of transmitter power was needed to get a well-defined signal strength measurement.

Fig 2.5. Shows the balanced mode of the RX coil, and with the series resonant mode tuned to the transmitter frequency. Note that for clarity the magnitude of the impedance scale, |Z| (blue) has been increased from the previous 500Ω/div to 2000Ω/div. The parallel modes from the primary and secondary coil were balanced using a primary loading capacitance of CPRX = 282pF, and this balanced condition in a TMT has been shown to be beneficial to achieving a very high transfer efficiency in single wire mid-field region experiments in the High-Efficiency Transference of Electric Power series. In this telluric experiment, in the far-field region, this balanced condition was found to introduce too much loading in the RX coil given the very small signals being received, which led to reduced signal strength measurements.

The capacitive loading in the primary coil was removed, and appears sub-optimal for these types of very low power level telluric reception measurements. If and when higher power can be transferred via the telluric transmission medium, the balanced mode may be necessary to maximise the LMD transmission mode, and hence the received telluric-wave. It should be noted that the TX coil is tuned and driven by the generator at its series fundamental resonant mode at 1.86Mc, and with the lower and upper parallel modes balanced using primary capacitive loading CPTX = 403pF, which was found consistently to be the most efficient setup for the TX coil and linear amplifier generator, used in both in this telluric experiment and the experiments presented in Transference of Electric Power – Single Wire vs Telluric.

Fig 2.6. Here it was tested to see the maximum balance that could be accomplished between the upper and lower parallel modes, and whilst keeping the lower parallel mode tuned to the transmitter frequency. This characteristic was tuned using a primary loading capacitance of CPRX = 60pF, a significant reduction in loading capacitance from the full balanced mode in fig. 2.5. This produced better signal strength results than the full balanced mode, but still not as good as the unloaded results with no additional primary tuning capacitor. At these very low reception powers it was concluded that the balanced mode simply attenuates the signal too much, and especially in the case were the telluric-wave is not very strong, and the LMD mode is not dominant.

Telluric Transmission in the High MF Band Far-Field

In the first field location 2 miles from the transmitter it was possible to clearly receive with 6 bars of signal strength at only 10W TX power at 1.86Mc for the telluric-wave and radio-wave combined, and 1 bar for the radio-wave only. The attenuation of the signal at 1.86Mc under the ground appears enormous, and it was considered in the previous experiment Transference of Electric Power – Single Wire vs Telluric that this loss is dominated by absorption of the transmitter power by the earth directly surrounding the main telluric ground system in the high medium-frequency band. In the previous experiment only 18m from this telluric ground system the measured power had already dropped from 10W TX power to 1.25mW at the RX coil.

So transmitted power in the earth surrounding the telluric ground system has already reduced by almost 4 orders of magnitude even before it is only 10s of meters away from the ground system connection. When we consider the result achieved 2 miles away the power would have dropped into the micro-watt level to produce the kind of signal strength received by the Sony radio receiver, and so we can conjecture that the transmission over the 2 miles was actually more efficient, than the transmission from the TX secondary coil through the ground system and over the distance of a few 10s of metres. This may also imply that there is very considerable power losses in the interface between the copper of the ground system and the earth, and even with significant water irrigation of the ground system, and relatively low measured impedance at the transmitter frequency.

It is very interesting in the 2 mile location that there was also a large difference in the received combined telluric and radio-wave at 6 bars, and the radio-wave at 1 bar, where in both cases the RX coil was tuned at the lower parallel mode to the transmitter frequency through adjustment of the coil wire length. Again in the previous 18m telluric experiment the proportion of telluric-wave to radio-wave at 10W was 0.7 mW : 0.55 mW, where both components are much closer and contributing approximately equally to the transmission of power from TX to RX, with only slight emphasis on the telluric-wave. In the 2 mile field location the ratio of signal strength telluric to radio is 5 : 1 which we can also conjecture may result from a more dominant LMD mode across the telluric cavity formed by the TMT system.

We do also need to consider the possibility that the radio-wave encountered significant obstacles in the 2 mile TEM propagation, reducing significantly the radio-wave component at the RX coil, but I would suggest that the combination of the two results regarding the better power transmission efficiency over the 2 miles distance than the 18m distance, the relatively close far-field distance, and the large signal strength ratio 5 : 1, could point towards a dominant LMD mode, and a preferential telluric transmission channel, over and above the TEM mode radio propagation channel.

In contrast at the 8 mile man-made reservoir location, although the signal tone could just be detected at 10W TX power, it was necessary to use up to 400W of TX power to get reasonable signal strength up to 4 bars. It was also noted that the ratio of telluric to radio-wave components was again around 1 : 1, and the far-field transmission distance had not significantly increased by going up to 8 miles at the transmitter frequency at the top-end of the MF band. It is considered here that the telluric channel/connection at the RX coil end was not as good as for 2 mile case, and especially in taking into account that the water-body used for the telluric ground was both man-made and may not be so well connected to the earth’s aquatic system. It is conjectured that the LMD mode was not established as dominant in the TMT transmission cavity, and that power reception at 8 miles was dominated by the TEM mode of far-field radio-wave propagation.

It must also be considered that the two field locations presented so far were not selected for any special water-table, river inter-connection, underground aquatic properties or channels, or for specific earth and rock type and composition. Both locations are in limestone regions and both are connected to water bodies, the 2 mile location being a natural river-fed lake, relatively close to the underground source of the river (a further 2 miles, so approximately 4 miles to the river source from the transmitter). The 8 mile location, being a man-made reservoir with a river tributary feed and outlet, is a further extension in the same direction from the transmitter. So the 8 mile location is essentially 6 miles further on from the 2 mile location, and 4 miles further on from the natural river-source of the 2 mile location.

Summary Conclusions and Next Steps

In this post, telluric transference of electric power has been explored and demonstrated in two different field locations in the near far-field region from the transmitter at 1.86Mc in the high MF-Band.  In both field locations signal strength could be measured at the transmit frequency in both the telluric-wave and the radio-wave at only 10W generator power. There was a vast difference in power required in each location to achieve approximately the same measured signal strength readings, 10W TX power with 6 bars at 2 miles, and 400W with 4 bars at 8 miles, with all other aspects of the TMT apparatus kept constant other than the field location telluric ground connection, and the over-ground terrain profile between the TX and RX. From the experimental results and measurements presented the following observations, considerations and conjectures are made:

1. The LMD mode is conjectured to be dominant in the 2 mile location based on the the large ratio between the measured telluric-wave and the radio-wave, and on considerations on telluric channel/cavity losses both for this experiment, and the previously considered 18m telluric channel.

2. The TEM mode is conjectured to be dominant in the 8 mile location based on the equal ratio of the measured telluric-wave and the radio-wave, and the large input power of 400W needed to get adequate measured signal strength, and on comparison with the very similar telluric experiment results in the 18m telluric channel.

3. The telluric connection quality to the earth through the type of water-body, is conjectured to be the most likely difference between the very different results of the two field locations. The difference in distance of 6 miles is not considered to be the major factor in the large difference in the location results.

4. The underground water inter-connection between the TX and RX is considered to have a significant impact on the quality of the telluric transmission medium between the two ground systems.

5. The impact of the earth soil and rock type and composition is as yet unknown on the telluric channel quality.

6. High losses will occur in the ground system to earth interface, and the telluric transmission channel/cavity with higher transmitter frequencies in the MF band. 1.86Mc appears far too high for any significant power transfer by the LMD mode in a telluric cavity.

7. Telluric transmission via the LMD mode is conjectured to be more efficient than by the TEM mode, and that with a sufficiently low frequency and a properly arranged LMD cavity in the TMT apparatus, it may be possible to transfer larger quantities of power in the far-field with better efficiency than could be accomplished using an overground wireless mode or radio-wave.

Next steps are to further explore Telluric Transference of Electric Power at different field locations both in the close far-field, and then at further distances from the transmitter, both at the same presented high MF-band frequency of 1.86Mc, and then at lower frequencies, and ultimately if possible down into the LF-band where Tesla was working with his own experiments. Lower frequency experiments present considerable challenges, including TMT size and scale, generator type and compatibility, radio regulation and licensing, availability of field locations, and resourcing and funding. If these challenges can be overcome then it may be possible to finally confirm or refute the possibility of high-efficiency telluric transference of power, and understand in much greater detail and accuracy the legacy that Tesla has left us to explore.


1. A & P Electronic Media, AMInnovations by Adrian Marsh, 2019,  EMediaPress

2. Dollard, E. and Energetic Forum Members, Energetic Forum, 2008 onwards.