Wednesday, April 22, 2015

Marconi Rides Again!

This was a very recent and interesting re-discovery. On April 26, 1901 patent 777 was issued to Marconi's Wireless Telegraph Company. This memorably numbered British patent was ambiguously titled "Improvements in Apparatus for Wireless Telegraphy."  It's more vague that manly criminally trolled software patents today. In 1904 the U.S. grated Marconi a matching patent numbered less memorably as 763,772. See above.

The patent is a difficult read even on a scale of 19th century electrical patents. It's a mix of wrong assumptions, archaic terminology and technically pedantic. But this one uniquely contains new and only inadvertently patented technology. Yes. Marconi drew, but did not describe a unique feature within the patent. Fixated on building coil configurations upon his prior patent 586,193 he missed something big.

At Cambridge, Gehan Amaratunga and Dhiraj Sinha led a group of engineers looking into that very detail. The drawing depicts a transmitter linked to an antenna which is connected to a coil. But the coil  had one end grounded (marked "E" for Earth) and the RF signal was clearly fed to the middle of the coil. That detail inspired Amaratunga's team to develop a way to reduce the size of a GHz antenna using dielectric material to emit radio waves.  More here. Early testing with piezoelectric films at 1575.42 MHz have already proved that the technique can be used to radiate 1 watt at 60% efficiency from a tiny chip.

Bear in mind how amazing this omission was. This patent was at the center of a legal dogfight back in 1900.  The patent addressed the endemic problem of RF interference. It resolved the issue by tuning all the primary components: coils, aerials etc. That's why the patent is so specific with the number of turns of wire in each layer on each coil.  There was no other way to tune a circuit at the time. However the solution was stolen directly from an 1897 patent held by Professor Oliver Lodge. Furthermore, Tesla patent 645,576, granted in 1900 described a four-circuit system for a transmitter and receiver with all four circuits tuned to the same frequency. In 1892 William Crookes first predicted tunable transmitters would be invented. He didn't predict that  Marconi would then steal it. More here. But my main point there is that three highly competent radio engineers looked that that configuration and didn't see what Amaratunga saw.

Previously it was not understood why or how a piece of dielectric material could radiate signals when a RF signals were applied at one end. It's completely counter-intuitive. All of our understanding began with 100 year old equations developed by Maxwell. The IEEE refers to the behavior as "asymmetric coupling between the spark generator and the antenna." It allows the transformation of the RF electric signal into electromagnetic radiation.  The concept is borrowed from quantum theory and it's the first explanation of how Marconi’s transmitter emitted electromagnetic waves over a century ago. More here.