Bart Lee - Marconi Article 2005

A Meditation on Marconi's Mercury Detector and his Transatlantic Triumph of 1901: Nothing Ventured, Nothing Gained

By Bart Lee, KV6LEE for the Poldhu Amateur Radio Club, 21 Nov 2005

Bart is CHRS General Counsel Emeritus

Branly's Coherer before 1901. The French scientist Edouard Branly (1844-1940) invented the coherer as a device for electrical experiments in the 1880s. The device holds filings of one or more metals such as silver and nickel between the faces of two cylindrical rods, in an evacuated chamber, usually glass. It is thus the earliest vacuum device after the light bulb and the Edison-effect device of a light bulb and a plate inside of it. It pre-dates the vacuum tube diode "valve" of Dr. Ambrose Flemming by about 20 years and the vacuum tube triode "Audion" of Dr. Lee deForest by a couple more years. (Pictured, above left: A drawing of a metal filings coherer)

The principle of operation of the coherer is not electrons in a vacuum and it is not semi-conductance as we know it in silicon transistors and chips. When radio-frequency alternating voltage appears across the two cylinder faces between which rest the filings, a direct current can begin to flow between the two cylinders. The radio-frequency voltage seems to drop the surface resistance among the filings, in the same way a detergent or surfactant such as soap decreases the surface tension of water molecules. The particles seem to cohere tegether, and thus conduct electricity better. When the resistance among the filings goes down, the direct current can flow through them. The resistance among the filings remains low after this change, whether or not the radio-frequency current continues, and irrespective of how long the direct current flows. The coherer is thus a one-shot, on-only detector of radio waves, and the first detector of radio waves after Dr. Heinrich Hertz's loop with a spark gap. It is like an electro-chemical and mechanical relay. (Pictured, left: One of Marconi's coherers)

The coherer was adapted to signaling service by application of feedback. The direct current that flowed through the coherer was routed through not only a recording device (such as a telegraph inker with continuously running paper tape) but also through an electrical bell mechanism. That current caused the tapper arm of the bell to move, as it would to ring the bell. That tapper arm was positioned to tap the coherer, gently, which tap de-cohered the filings. These filings then resumed their natural relatively high resistance to electrical current. It is the need to tap the coherer after each pulse that implies that the device continued to conduct direct current whether or not the radio frequency voltage ended.

If and when another pulse of radio-frequency energy came along, the filings once again cohered together, the current once again flowed to the recording device and the tapper, and the coherer thus prepared itself for yet another pulse of radio energy. Inasmuch as the tap happened as soon as the pulse went through the filings, the beginning of each signal restored the coherer. Using the Morse code, the signals pulsed out in radio energy provided different patterns in time, with greater or lesser delay between a pulse and the next pulse, depending on whether the pulse was a long one, a "dash," or a short one, a "dot." The dash and dot appearance of telegraph signals came from the ink recorders in use on land-lines and marine cables, in which the length of the pulse determined the length of the inked line. (Pictured, left, a close up of the coherer tube holding the filings, under it the tapping mechanism)

A Morse code dot or "dit" is here represented by an exclamation point: "!". The absence of a dot/dit, in a given time period a dit in length, is represented here by a small "o". In Morse telegraphy, a dash or "dah" is in time-length three dots or dits long. This is usually shown in writing as an "em" dash, a longer dash, here " ---" . In the text that follows, beware the ambiguity between the letter "O" as in " S-O-S" or the words "On"or No," and the small "o" here used only as a place holder. This information, namely that the signals pulsed out in radio energy provided different patterns in time, suggests that the coherer would thus present an "S" ( usually printed as three dots, thus "···" but here as "!!!") as three closely spaced direct current pulses followed by an absence of a next pulse. This would be something like "!!!o," where the glyph " o " is merely a place-holder for a dot-length of time with no signal in it. That use of a glyph is analogous to ! the glyph zero ("0") as a place holder in calculation. Thus the regular English alphabet letter "O" (usually printed as three dashes, thus "--- --- ---" ) would be heard through a coherer as three widely spaced, i.e. , longer-in-time direct current pulses. This would look and sound something like: !ooo!ooo!ooo. The distress signal "S-O-S" repeated would thus be:!!!ooo!ooo!ooo!ooo!!!oooo!!!ooo!ooo!ooo!ooo!!!oooo!!!ooo!ooo!ooo!ooo!!!oooo etc.

Other letters had various unique patterns of close and wide spaced pulses. The rhythm of telegraphy was steady enough, as if sender and receiver were operating at the same "clock-speed," so that different letters could be distinguished from this data. In this aspect, looking only at the beginning of each pulse, however long, Morse code is an entirely binary code of on and off at a given rate of time. In each interval, there is either a "dit" or there is nothing.

Soon after the coherer came into use, wireless operators found that they could copy the messages without an inker from the sequence and timing of the sounds of the tapper. The tapper's sounds resembled that of a land-line telegraph "sounder," although that latter device made one sound at the beginning of the pulse, and then a different sound at the end of each pulse. The coherer's tapper made its tap at the beginning of each pulse whatever its length. (pictured, left: another Marconi coherer device)

Soon telephone receiver earpieces were employed to hear the combination of the direct current flowing through the coherer and the radio-frequency's audio "hash" modulation. Originally, Branly had used a galvanometer and battery circuit to detect the change in the coherer's resistance, and hence the radio-frequency pulse. Some experimenters used telephone transmitter carbon granule microphones as radio-frequency detectors. Others used carbon battery rods on knife edges. By 1907, "crystal" detectors of Carborundum and Silicon, providing audio to a telephone earpiece, had replaced the coherer and the devices that followed it, such as the Flemming valve and Marconi's magnetic detector. With the invention of radio-frequency regeneration by Edwin Howard Armstrong in about 1914, the deForest Audion became the commercial detector of choice. In 1909, Braun shared the Nobel prize with Marconi, but maybe Branly deserved one as well. As late as the 1950s, a Japanese radio control toy used a coherer circuit.

Branly's coherer was available to Marconi and he used it. In Newfoundland in 1901, however, he employed a Mercury detector. Chanda Bose invented it, and Solari of the Italian Navy developed it after reading about Mercury as a detector of radio frequency energy in the British scientific press.

Marconi's Bose/Solari Mercury Detector of 1901.

All of the literature that raises issues about how Marconi could possibly have heard Poldhu makes two assumptions. The first is that ionospheric reflection was not then possible on the frequency Marconi used, about 833 kilohertz. As previously reported, the facts that sunspots were exactly zero in December 2001, that Marconi carried out his experiment near the Winter Solstice, and about local sunset, greatly enlarges the possibility of ionospheric skip reception. But the second assumption is just as important: that Marconi used a passive receiver and thus had no gain available to him from his circuit. Modern propagation models suggest that a passive receiver simply would not work even if there had been ionospheric reflection.

The Mercury detector was then known as an auto-coherer or self-restoring coherer. Contemplating the operation of Branly's coherer and the Mercury detector suggests that Marconi did enjoy gain, considerable gain, in his Newfoundland receiver. What both coherers need to operate is a direct current (DC) voltage across them. Radio frequency (RF) energy from the antenna, on its way to ground through the coherer, changes its DC resistance from high to low. Once the resistance goes low, a current of any amount may flow through the coherer (governed of course by Ohm's Law). It was common to use a battery of between one and two volts for this DC. Such a DC voltage across a device is often known as a bias voltage.

Marconi's Mercury detector patent application shows a bias voltage across the Mercury interface with Carbon and Iron electrodes. Pictured at left is Marconi's original Mercury detector on display at the British Science Museum, London (author's photo).

The possibility of significant gain arises from the fact that the power into the device as radio frequency (RF) energy is minuscule. Yet if it is even just barely enough to change the electro-chemical properties of the coherer's internal interface, it permits a great deal more power (as DC current) to pass though. It is this DC current that operates a transducer. Initially Marconi (and others) employed landline telegraph inkers as a visual display of a message. Marconi, in Newfoundland, used a telephone receiver earpiece to hear the pulse of DC from the bias voltage applied to the Mercury detector.

The Poldhu transmitter had been engineered by Flemming as a double-spark system to produce a sharp and very short pulse of RF energy into its antenna, at 13 kilowatts or more. All that the receiver in Newfoundland had to do was make known a series of very short pulses. The hearers in Newfoundland, Marconi and Kemp, heard the sequence of pulses, the "clicks," as the letter "S" repeated ( e.g., !!!ooo!!!ooo!!!ooo!!!ooo etc.). They reported hearing this pattern of three dots 38 times in two days. What made the diaphram of their telephone receiver move or "click" was the current of a pulse of the DC bias voltage (either from zero to a higher value or from a lower value to a higher value). What enabled that DC pulse to flow was the very much smaller power of the RF energy coming down the antenna and changing the DC resistance of the detector. This is a classic example of "amplification" as defined in cybernetics. It is not linear amplification as is known in audio and radio electronics. It is the amplification of a pulse, which was already well known in landline telegraphy and performed by electro-mechanical relays. It is as if the Branly coherer (with tapper) and the Mercury detector were simply momentary-on relays, controlled by the small RF input but producing a large DC output.

Thus, I suggest that Marconi did not employ a passive receiver, but rather a sensitive gain receiver. It remains to be seen from experiment if the Mercury detector operates as such a pulse amplifier. It is possible that it operates merely as a diode, without gain but requiring a bias voltage. It is possible that it operates as a negative resistance device, like a tunnel diode. Zinc oxide (or a related compound) is reported to work as an amplifying detector by reason of negative resistance, and the oxide coating of the surface of Mercury may do the same thing. Circa 1907, the carborundum detector required a bias voltage, as did the Perikon detector using a Zinc compound as one of its two minerals. The Galena detector, on the other hand, did not require a bias voltage. It is not now known if any of these biased detectors provided gain (Galena did not). It is known that early wireless operators found them to be more or less sensitive, and it is possible that a more sensitive detector was, one way or another, providing some amplification. All that would have been perceived at the time was that that particular detector was more sensitive, which was exactly Marconi's perception of his Mercury detector, the very reason he employed it with an audio transducer for his undoubtedly successful 1901 transatlantic experiment.