The Centennial of the Crystal Radio Receiver

Laszlo Morocz
Copyright 2006 by Laszlo Morocz

Just over 100 years ago, Greenleaf Whittier Pickard (Fig. 1) received U.S. patent number 836,531 for one of the pivotal pieces of modern technology, the solid state silicon detector. This early silicon diode was the direct ancestor of all of today's silicon semiconductor devices, from the simple rectifier to the multimillion transistor CPU chip.

The centennial of Pickard's patent was 20 November 2006. In honor of the centennial, I decided to examine Pickard's patent to see what I could learn about early 20th century science and technology from the original filing.

Figure 1. Greenleaf Whittier Pickard (1877-1956) invented the silicon detector that made possible the crystal radio. This photograph is image number 1879 from the IEEE History Center. Used with permission and copyright (c) by the IEEE.

From the patent's title, "Means for Receiving Intelligence Communicated by Electric Waves," it can be seen that Pickard's emphasis was on radio communications. This is not surprising, for at the time he worked for the American Telegraph and Telephone Company.

The significance of Pickard's silicon detector is best demonstrated by a review of the history of early radio. At the beginning of the 20th century, the state of the art for radio receivers was an antenna system connected to some kind of radio wave detector, which was in turn connected to a relay, meter or telephone earpiece. Signals were telegraphy only. Operators heard clicks in their headphones, clicks from a telegraph sounder or saw a meter needle move. Some systems had the ability to print marks on a strip of paper. The distance between marks indicated whether a Morse code dot or dash had been received.

The most successful detector was the Marconi coherer. It was a glass tube with electrodes at each end (Fig. 2). The space between the electrodes was filled with metal filings.

Figure 2. Outline view of a Marconi coherer, a glass tube with electrodes at each end.


The electrodes were connected across a coil which was coupled to the receiver's antenna system. They were also connected to a battery and relay circuit so that the current needed to energize the relay also flowed through the coherer.

Normally the filings were loose and presented a high electrical resistance to current flowing through the device, so the relay stayed open. However, when a radio wave passed through the coherer, the filings cohered, i.e., they packed together. This lowered the coherer's resistance and allowed the battery current to flow through the relay. The relay activated a telegraph sounder and a loud click was heard. The relay also activated a tapper which tapped the side of the coherer's glass tube, shaking the filings loose. The battery current was cut off, the relay opened and the sounder made a second click. Just as in the railroad telegraph systems of the time, the time between the "down-click" and "up-click" indicated whether a dot or dash had been received.

An argument could be made that this device was the first glass tube-based amplifier. A small amount of power went in and a greater amount of electrical power, sufficient to operate the sounder, came out. It was all or nothing, so it could also be described as operating in saturation or digital mode.

The coherer detector worked, but it had many flaws. It was insensitive, finicky to adjust, used many moving parts and was slow. Because it was based on a mechanical action and required another mechanical action (the tap) to reset it, professional telegraphers of the day had to slow down their sending so the coherer could keep up with them. It could also interfere with itself because of the self inductance of the relay and tapper coils. With all these problems there was a lot of motivation to find a better detector.

Marconi himself had come up with another detector. It was the magnetic detector, or "maggie." This detector was based on a ribbon of fine soft iron wires arranged in an endless loop. The loop went around a pair of wheels so that it was in continuous motion. The wheels were driven by a clockwork mechanism which the radio operator had to occasionally wind or the radio stopped working.

As the ribbon went around the loop, it passed by a pair of magnets which induced a magnetic field into the ribbon. As soon as the ribbon moved past the magnets, it began to lose the magnetic field, but due to soft iron's magnetic hysteresis the field persisted for some time. While it was still magnetized, the ribbon passed through a coil of wire which was connected to a telephone earpiece. As the moving magnetic field passed through the coil, it induced a current which flowed through the earpiece, resulting in a soft hiss being audible to the operator.

A second coil was also wrapped around the moving ribbon, inside the earpiece coil. This coil was connected to the antenna system. When a radio wave passed through the coil, it caused the ribbon to lose its induced magnetic field. This resulted in the familiar two clicks in the earpiece, one when the field dropped out and the induced current stopped and another when the field came back and the induced current restarted.

The maggie was much more sensitive than the coherer. Since it operated on magnetic fields instead of discrete metal filings, its response time was very much faster than the coherer's. Being clockwork driven and giving off no sparks, it did not interfere with itself. However, it still had many moving parts and required very careful adjustment to balance its sensitivity against its internal noise (the hiss caused by the moving ribbon).

This was the environment in which Pickard began his search for a better detector. He concentrated on the so-called mineral detectors, pieces of natural or synthetic minerals which exhibited an electrical change in the presence of radio waves. He tested tens of thousands of materials before settling on the amorphous silicon which he claimed in his patent.

Pickard's patent not only claims silicon as a material suitable for detecting radio waves, it also describes a complete radio receiver based on the new detector material. In effect, he patented the crystal set.

The circuit is given in Fig, 1 of his patent, which is reproduced here as Fig.3.

Figure 3. Pickard's silicon detector patent includes this circuit diagram for a fully functional radio receiver.

The circuit in Fig. 3 is a basic crystal set still in use today. There's a loop antenna (called a "wave interceptor" in 1906), a series-tuned tuning circuit with a double-variable inductor (L) and a variable capacitor (C), the detector (TJ) and an earphone (T), which is bypassed with a capacitor (C'). The tuning circuit allows independent adjustment of the antenna tuning and the detector loading. This is a sophisticated setup which still works very well in today's AM radio environment.

This brings up an interesting feature of Pickard's circuit: it was audio ready. Just barely a month after Pickard's patent was granted, Reginald Fessenden made the world's first audio broadcast. On Christmas Eve of 1906, Fessenden broadcast readings from the Bible and performed "O Holy Night" on his violin. Anyone within range using a radio with a mineral detector was able to hear voices and music, instead of the normal clicks of telegraphic communications. The Marconi detectors described above were unable to detect the audio from Fessenden's broadcasts.

Pickard's patent shows the details of his silicon detector, and his drawing is reproduced here as Fig. 4. Anyone who has ever used a mineral detector crystal set will recognize the details.

Figure 4. Outline view of Pickard's silicon detector from Fig. 2 of his historic patent.

Figure 5 shows a contemporary crystal detector made by radio artist Tom Kipgen. The family resemblance is undeniable. Both share a ball joint that allows a spring-loaded contact to be positioned against the detector material. Many of these details survived almost unchanged into the 1960s when cheap transistors finally put an end to commercial production of crystal sets.

Figure 5. A contemporary crystal detector made by radio artist Tom Kipgen. Used with permission and copyright (c) by Tom Kipgen.


While the mechanics behind Pickard's detector survived the test of time, the science behind it did not. The detector undeniably worked. It was efficient enough to detect radio signals and render them audible without the aid of an external power source. However, Pickard's notions of how this was accomplished were totally wrong in 1906. They couldn't be correct since the theoretical basis for understanding the phenomenon had not been developed yet. Einstein had just published "On the Electrodynamics of Moving Bodies" the year before. It would be 20 years before quantum theory would be developed to the point where it could begin to explain semiconductors, and much of that work would actually be done almost 20 years after that during World War II as part of the effort to develop efficient and rugged radar receivers. Pickard would actually live to see this work accomplished.

But back in 1906 his patent stated that the detector worked through some kind of "thermo-electromotive force", being basically a thermocouple. He interpreted what he observed as a manifestation of a "thermo-junction." In his theory, the junction of the spring-loaded contact and the silicon converted the incoming high frequency radio signal into heat. The narrow contact point concentrated the heat. This heat was then converted by the same junction to a varying DC voltage which was then converted into sound by the telephone earpiece. He called it a "thermo-electric regenerative detector" because it was supposedly regenerating the heat from the radio wave into a DC voltage.

In his patent Pickard spends a great deal of space worrying about how to eliminate the counter electromotive forces (voltages) caused by the junction formed on the other side of the silicon. This was the side away from the point contact, the side which was connected to the earpiece instead of the antenna. If his thermal theories were correct, that junction should be generating equal and opposite voltages which would then cancel out the ones from the point contact. His way to get around these voltages was to make the contact area between the silicon and the supporting cup as large as possible to dissipate the heat and to prevent the formation of a counter electromotive force.

This was the correct thing to do, but for totally different reasons. Increasing the surface area of the contact on the back of the silicon provided less electrical resistance and let the weak currents pass more easily. In fact, Pickard himself wrote that, "In practice, however, it is impossible to obtain equal [thermal] resistance at each junction ...hence ... there will exist a sufficient preponderance of thermo-electromotive force of one sign to operate...." That is, the effect he was so worried about never actually appeared outside of his theory.

The narrow contact point was also the correct thing to do, though again for different reasons than Pickard thought. One clue to what was actually going on was the fact that the patent specifies the "massive amorphous or graphitic solid form" of silicon in preference to the crystalline form. The crystalline form was the purest of the three. The other forms probably had just enough manufacturing contamination to act as a dopant. In reality, Pickard was forming a point contact semiconductor junction.

So Pickard's patent was a mixed bag. Technologically it was a tour-de-force. He laid out the basic components of a relatively sophisticated receiver that is still usable today. He successfully harnessed silicon as a semiconductor junction diode which allowed him to demodulate radio signals, including AM audio signals. But his theoretical claims were completely wrong. Not due to any fault of his own, but because technology had outstripped theory and the Einsteins, Bohrs and Diracs had not had a chance to catch up yet.

His silicon detector was sturdy, cheap, had no complex moving parts or adjustments and needed no external power source (electrical or mechanical). For the first time in communications history, the solid state silicon semiconductor replaced a glass tube device. It would be overtaken by the glass tube in the very near future as vacuum tubes were developed, but it would come back into its own during World War II and win complete supremacy by the last third of the new century.

Pickard himself continued to make valuable contributions to communications technology and science. Besides marketing his patented detectors, he investigated the effects of solar eclipses and meteor showers on radio propagation. His long, productive life ended in January of 1956, by which time Bell Labs had developed the point contact transistor, a piece of doped silicon with three wires, not two.

There were many other important radio pioneers in the early 20th century, many doing similar work. H. C. Dunwoody received a patent on carborundum detectors the month after Pickard's patent was granted. But none of the other patents showed the construction of a workable, sophisticated solid-state silicon semiconductor junction-based radio system. Greenleaf Whittier Pickard's patent is unique and important in that respect.

Happy 100th birthday to the crystal set!


Bibliography

IEEE History Center Biographies: G. W. Pickard.
http://www.ieee.org/web/aboutus/history_center/biography/pickard.html

United States Patent Office Patent No. 836531, "Means for Receiving Intelligence Communicated by Electric Waves."
http://patft.uspto.gov/netacgi/nph-Parser?patentnumber=836531

W. H. Marchant, Wireless Telegraphy, 1914.


Useful Links

Tom Kipgen's Radios are featured at http://www.wynterarchtops.com/radios/index.htm This interesting site shows old technology radios built as modern pieces of art. This site is also a source of parts for those wanting to build their own radios.

The Xtal Set Society is at http://www.midnightscience.com? This site offers books, plans and parts for crystal set builders. They also have a good newsletter for members.