Edited Manuscript
Recognised text
Text below was produced by optical character recognition and may contain errors. Check it against the original.
, ,
A Radioscientist s Reaction to Marconi s first
transatlantic Wireless Experiment — Revisited
John S. Belrose
Radio Science Branch
Communications Research Centre Canada
Ottawa ON K2H 8S2
Introduction
On 12 December 1901 signals from a high power
spark transmitter located at Poldhu, Cornwall,
were reported to have been heard by Marconi and
his assistant George Kemp at a receiving site o n
Signal Hill, near St. John’s, Newfoundland. For
this reception experiment Marconi used a kite
supported wire aerial, an untuned receiver, a
detector of uncertain performance and a
telephone receiver. The signals if heard would
have traveled a distance of 3500 kilometres. Even
at the time of the experiment there were those
who said, indeed there are some who still say,
that he misled himself and the world into
believing that atmospheric noise crackling was
in fact the Morse code letter ’S’.
A little later, in February 1902, when Marconi Fig. 1 Photograph of the original conical antenna
was returning to the North America on the system installed at Poldhu, Cornwall (after BAE
SS Philadelphia, using a tuned ship-borne Systems Marconi Research Centre, Chelmsford,
antenna, he received signals using his filings Essex).
coherer from the same sender up to distances of
1120 km by day and 2500 km by night. Even
these distances are rather remarkable considering
the receiving apparatus he used.
This paper revisits that first transatlantic
experiment.
The Poldhu Station
Marconi’s ambition at the turn-of-the-century t o
demonstrate long-distance wireless comm-
unication, and develop a profitable long-
distance wireless telegraph service, led to his
pragmatic proposal in 1900 to send a wireless
signal across the Atlantic. He conceived a plan t o
erect two super-stations, one on each side of the Fig. 2 Photograph of Marconi s fan monopole,
Atlantic, for two-way wireless communications, December 1901 (after BAE Systems Marconi
to bridge the two continents together in direct Research Centre, Chelmsford, Essex).
opposition to the cable company (Anglo-
American Telegraph Company). For the eastern shown the plan, did not think the design to be
terminal, he leased land overlooking Poldhu sound. Each mast was stayed to the next one, and
Cove in southwestern Cornwall, England. For the only to ground in a radial direction, to and away
western terminal the sand dunes on the northern from the center of the mast system. He was
end of Cape Cod, MA at South Wellfleet, was overruled, construction went ahead, and both
chosen. aerial systems were completed in early 1901.
However, before testing began catastrophe struck.
The aerial systems comprised 20 masts, each 6 1
The Poldhu aerial collapsed in a storm o n
metres high, arranged in a circle 61 m in diameter,
17 September; and the South Wellfleet aerial
see Fig. 1. The ring of masts supported a conical
suffered the same fate on 26 November, 1901.
aerial system of 400 wires, each insulated at the
top and connected at the bottom, thus forming an At Poldhu Marconi quickly erected two masts
inverted cone. Vyvyan [1], the Marconi engineer and put up an aerial of 54 wires, spaced 1 meter
who worked on the 1901 experiment, when apart, and suspended from a triadic stay stretched
between these masts at a height of 45.7 m. The 18.8 microhenries (µH) for a turns ratio of 1:7.
aerial wires were arranged fan shaped, insulated at Scaling directly to a 50.8 cm square frame,
the top, as was his conical wire aerial, and multiplies all reactances by the ratio 50.8/20.3,
connected together at the lower end, see Fig. 2. which yields inductance values of 2.6 and
This photograph has been published and
47 µH.
republished, and clearly one can see only
12-wires with blobs at the end of the wires (which
This author experimentally modelled (scale
might be insulators?) -- but the view generally
factor 75) a 54 wire fan monopole, wires
held is that the aerial system as described above
connected to the triatic, and measured the
by Vyvyan is right, that is there were 54 wires.
resistance and reactance for this model on a large
The photograph has been touched up to show
elevated ground plane. The experimental model
wires, and for emphasis insulators are shown at
exhibits resonances (frequencies scaled to full
the end of each wire. It is believed that all wires
scale), see Fig. 4, at 935 kHz, and 3.8 MHz; anti-
were connected to the triatic, certainly 5 4
resonance at 2.4 MHz and 4.8 MHz; and
insulators would be very expensive.
approaching an anti-resonance between 7 and
In fact a recent study of archive documents 8 MHz; but no resonances above 3.8 MHz.
reveals that no true photograph exists of the
Poldhu fan aerial. The photograph in Fig.2 i s
clearly a modified reprint Fig. 1, showing two
masts exactly as originally sited and rigged, with
the others painted out [2].
The antenna was driven by a curious two stage
spark transmitter (Fig. 3). There were many
problems in getting it to work at the high power
levels desired [3]. Our principal concern here i s
the frequency generated by the Poldhu station.
The oscillation frequency is determined by the
natural resonant response of the antenna system,
tuned to a lower resonant frequency by the
inductance of the aerial jigger (Marconi’s words). Fig. 4 Impedance vs. frequency for Marconi s
temporary fan monopole December 1901
(experiment and simulation)
The author has also numerically modelled, using
NEC-4D, a 23-wire fan, included a realistic sag for
the triatic wire. Both numerical and experimental
modelling give essentially the same resonant
frequency (base load inductance 47 µH included)
for the antenna system (see Fig. 4), but the high
frequency self impedance response for the
Fig. 3 Schematic showing component values for experimental model is not reproduced by the
discharge circuit of the Poldhu transmitter i n numerical simulation, see Belrose [5].
December 1901. Note the additional components
CL (extra coils), and LT (a long adjustable tuning While the high frequency resonances and anti-
inductance of 40 turns using 40 feet of wire). resonances of this multi-wire fan aerial are of
From Fleming s Notebook (after Thackeray [3]). academic interest, the antenna system impedance
for frequencies above the lowest resonant
frequency is dominated by the inductive
The inductance values for the oscillation/aerial
reactance of the aerial jigger. The aerial jigger i s
jigger transformer, have long been debated, since
in effect a base-loading-base-tuning coil, which
the original transformer is lost, there are n o
drawings, and reports about it differ. Fleming’s (for an inductance value of 47 µH) brings the
notes record that the primary was 2-turns resonant frequency of the antenna to about
paralleled, and the secondary had 9-turns; but 500 kHz.
Entwistle [4] said there were 7-turns. Possible
limits for the size of the windings range from Note: the antenna circuit itself was not separately
45 to 60 cm on a square former. Thackeray [3] tuned, but resonance was established. Fleming
has postulated on possible values for this tuned the oscillator circuit by varying the value
transformer, based on measurements by George of the discharge condenser, a parallel series
Grisdale in 1985. Grisdale on a facimile of the connection of 24 condensers, to maximize the
Science Museum’s 20.3 cm open jigger, measured RF aerial current. Vyvyan has given the value of
primary and secondary inductances of 1.05 and the discharge condenser for resonance as 0.037
mfd. Using the postulated value for the
inductance of the primary of the jigger (2.6 µH), consisted of a small blob of mercury between two
the resonant frequency of the oscillatory circuit end plugs. Wetted contacts behave ohmically,
is 511 kHz. like a resistance. Non-wetted contacts seem t o
behave like a metal/oxide rectifier, some of the
Since the resonant frequencies of the oscillatory time.
circuit and the antenna circuit (according to the
author s model studies) are closely the same, the Vyvyan gives rather quite a different account i n
author postulates that the radiation would be his book [1]:
spread about a single frequency of about
500 kHz. "It was impossible to use any form of syntonic
(tuned) apparatus and Marconi was obliged t o
Marconi himself has been evasive concerning the use the next best means at his disposal. He
frequency of his Poldhu transmitter. Fleming i n therefore used a highly sensitive self-restoring
a lecture that he gave in 1903 said that the coherer of Italian Navy design, simply connected
wavelength was 304.8 m or more (984 kHz). with a telephone and the aerial, and with this
Marconi remained silent on this wavelength, but simple apparatus on Thursday 12 December,
in 1908 in a lecture to the Royal Institution he 1901, he and one of his two assistants
quotes the wavelength as 365.8 m, 820 kHz. But (reportedly) heard the faint ’S’ signals".
in a recorded lecture in the early thirties he says
the wavelength was approximately 1800 metres Vyvyan is quite definite in his statement that i t
(166 kHz) and the power about 15 kW, see was the Italian Navy coherer which was used.
Bondyopadhyay [6]. Vyvyan’s account casts quite a different light,
too, on technical matters, as it is clear that if the
Reception on Signal Hill letter ’S’ was heard, it was due not to coherer
action, but to the unwitting and unrecognized
For his transatlantic experiment, Marconi use of a simple diode rectifier. Thus we conclude
decided to set up receiving equipment i n that if Marconi did hear a signal, his mercury
Newfoundland. In December 1901 he set sail for coherer by chance [7] happened to be in the
St. John’s, with a small stock of kites and rectifying mode just at the time when signals
balloons to keep a single wire aloft in stormy were heard. The unreliability of the device i s
weather. made clear by present day experiments [2,7], and
by Marconi himself. In his June 1902 lecture
A site was chosen on Signal Hill, and apparatus delivered to the Royal Institution he stated:
was set up in an abandoned military hospital. A ’These no-tapped coherers (Marconi referred t o
cable was sent to Poldhu, requesting that the the device as a type of coherer) have not been
Morse letter ’S’ be transmitted continuously from found to be sufficiently reliable for regular
3:00 to 7:00 PM local time. commercial work. They have a way of cohering
permanently when subject to the action of strong
On 12th December, 1901, under strong wind electric waves or atmospheric electrical
conditions, a kite was launched with a 155 m disturbances and have an unpleasant tendency
long wire. The wind carried it away. A second toward suspending action in the middle of a
kite was launched with a 152.4 m wire attached. message’.
The kite bobbed and weaved in the sky, making i t
difficult for Marconi to adjust his new syntonic Despite the crude equipment employed Marconi
(tuned) receiver, which employed the Italian and his assistant George Kemp convinced
Navy coherer. ’Difficult’ one can certainly agree themselves that they could hear on occasion
with, but how he would determine the frequency three clicks more or less buried in the static. And
of tuning for his receiver is a mystery. Whatever, clicks they would be, not unlike atmospherics,
because of this difficulty, Marconi decided to use because of the very low spark rate of his two-
his older untuned receiver. History has assumed stage spark transmitter (estimated to be only a
that he substituted his metal filings coherer few sparks/sec). Marconi wrote in his laboratory
previously used with this receiver for the newly notebook: Sigs at 12:30, 1:10 and 2:20 (local
acquired Italian Navy coherer, but Marconi never time). This notebook is in the Marconi Company
really said he did, see Phillips [7]. Marconi archives and is the only proof today that the
referred only to his use of three types of coherers, signal was received.
as if there were something to hide: ’---one
containing loose carbon filings, another It is ironic that the low-PRF transmitter was
designed by myself containing a mixture of compromised by Marconi himself, when i n
carbon dust and cobalt filings, and thirdly the Newfoundland he put a telephone receiver to his
Italian Navy coherer containing a globule of ear to listen for three dots from Poldhu. Fleming
,
mercury between two plugs’. must certainly have known about Tesla s widely
published 4-circuit arrangement (2-circuits in the
Clearly there are uncertainties concerning the transmitter, a single stage spark transmitter, and
type and reliability of the detector used with the 2- in the receiver [8,9]). A higher spark rate
land-based receiver. The Italian Navy Coherer
which could have been easily achieved with contemporary scientists and engineers about the
, experiment that continues today, and kick started
Tesla s transmitter, would have given the
received signal, assuming that the mercury the race, by Marconi himself and by Fessenden
detector was working like a rectifier, a buzz-buzz (see for example reference [9]) to achieve reliable
sound, or a more musical sound, depending o n transatlantic wireless communications.
the spark rate, rather than just a click,
References
indistinguishable from an atmospheric.
1. Vyvyan, R.N., "Marconi and his Wireless, EP
The Enigma Publishing, 1974. First published in 1933 as
"Wireless over 30 Years", Routledge and Keegan.
Today we know that signals (depending o n
frequency used) can indeed travel across the 2. Pegram, T.W., R.B. Molyneux-Berry, and A.G.P.
Atlantic, and far beyond. But in 1901, anyone Boswell, The Coherer Era — The Original
who believed that they could, and did, believed Marconi System of Wireless Telegraphy , GEC
so as an act of faith based on the integrity of one Review, 12, No. 2, 1997, pp. 83-93.
man -- Marconi. 3. Thackeray, D., "The First High-Power
Transmitter at Poldhu", The AWA Review, 7,
If 500 kHz was the frequency of the sender, the 1992, pp. 29-45.
tests took place at the worst time of day, because
the entire path would have been daylight, and the 4. Entwistle, W.S., "Year Book of Wireless
daytime skywave would be heavily attenuated, Telegraphy & Telephony", Wireless Press, 1922,
even though it was a winter day, in a sunspot pp. 1245-1258.
minimum period, and there were no magnetic 5. Belrose, J.S., Revisiting wire aerials at the
storms at the time, or before the experiment. From turn of the century , Proc. Of the 4th Symposium
a knowledge only of propagation conditions, on Antennas and EM Theory, Xi an, China, 19-22
reception on Signal Hill is consistent with the August, 1997, pp. 73-78.
observed limiting ranges of reception on the ship
only if the untuned land-based receiver was 6. Bondyopadhyay, P.B., "Investigations o n
47 dB more sensitive than the tuned receiver o n Correct Wavelengtrh of Transmissions of
the ship, see Table 1. The author considers this Marconi’s December 1901 Transatlantic Wireless
unlikely. Table 1 establishes rather well the Signal", IEEE Antennas and Propagation Society,
sensitivity of the tuned receiver on the ship. International Symposium Digest, 12, 1993,
pp. 72-75.
Historians have speculated that the transmitter 7. Phillips, V.J., "The ’Italian Navy Coherer affair
might also have radiated a high-frequency signal turn-of-the-century scandal", IEE Proceedings A,
as well, since an HF signal would have been more 140, 1993, pp. 175-185.8.
suitable for transatlantic communications, c.f.
Ratcliffe [10]. But based on the author s better 8. Anderson, L.I., Nicola Tesla on his Work with
understanding and more detailed modelling Alternating Currents and their Application t o
studies of the antenna system used, one must Wireless Telegraphy, Telephony, and the
conclude that the Poldhu spark-transmitter Transmission of Power , Sun Publishing, Denver,
system radiated efficiently on only the CO, 1992.
fundamental oscillation frequency of the 9. Belrose, J.S., On the Birth of Wireless
resonanted antenna system, about 500 kHz. Telephony , IEEE AP-S Magazine, (in press),
2001.
In summary
10. Ratcliffe, J.A., "Scientists’ reaction t o
It is difficult therefore to believe that signals Marconi’s transatlantic experiment", Proc. IEE,
could have been heard on Signal Hill, since: 121, 1974, pp. 1033-1006.
TABLE 1
(1) the receiving equipment consisted of a long-
wire antenna, coupled to an untuned receiver Distance Field Strength dB µ V/m*
which had no means of amplification whatsoever; kilometers (for 1 kW radiated)
(2) the type of detector used was less sensitive Ground Skywave Total
and its performance unpredictable (compared Wave Field
with Fessenden’s barretter detector, devised i n Limiting Ship Ranges
c1902, or the galena crystal detector which 1120 km Day 23 19 24.5
evolved a few years later); and, (3) the untuned 2500 km Night -25 27.3 27.3
land-based receiver would have had to be 47 dB Signal Hill, Newfoundland
more sensitive then was his tuned ship receiver.
-- -21.5 -21.5
3500 km
* Computations are based on CCIR (now ITU
Whether Marconi heard the faint dots or not i s
Radiocommunications) Report 265-7, Reports of CCIR
really unimportant at this time in history. His
1990, Annex to Volume VI, Propagation in Ionized
claim "sparked" a controversy among
Media, Geneva, 1990, pp. 212-229.