Talk:Arc converter

Why was the anode water-cooled?

I believe I understand the reason for the magnetic field and the hydrogen atmosphere, but does anyone know what role cooling the anode played in raising the frequency of the arc converter into the radio range? Why the anode? I could understand cooling the cathode; that might have some role in shutting off thermionic emission of electrons quicker to stop the arc. But the anode doesn't emit anything, it just absorbs electrons. Or maybe cooling the anode had nothing to do with increasing the frequency, it was required to dissipate the heat due to the striking electrons to prevent the anode from melting. --Chetvorno^TALK 01:40, 12 November 2014 (UTC)

It is a spark transmitter?

The lead sentence says:

The arc converter... was a variety of spark transmitter used in early wireless telegraphy.

I feel this sentence is inaccurate in two respects:

• The arc converter should not be classified as a "spark transmitter" (spark-gap transmitter). Although it worked by an electric discharge between conductors, this was not an intermittent spark caused by the discharge of a capacitor, but a continuous arc. This distinction is important historically. The early history of radio is divided into the "Spark era" in which transmitters produced damped waves which could not transmit audio but communicated by wireless telegraphy using Morse code, and the "Continuous wave era" in which electronic oscillators produced sinusoidal continuous waves, which could be modulated with audio in amplitude modulation (AM) as well as transmit Morse code by continuous waves (CW). These two technologies are distinguished both in sources from that time and contemporary wireless histories. The arc converter was an electronic oscillator, one of the first transmitters which produced continuous waves. It should not be considered part of the spark era.
• Similarly, as mentioned above, the arc converter was not solely used in "early wireless telegraphy". Although widely used for radiotelegraphy, it was also used to transmit sound (radiotelephony) by amplitude modulation, and was the transmitter used in the first AM broadcasting stations.

These are understandable inaccuracies but can confuse readers so they need to be corrected.--Chetvorno^TALK 02:01, 14 December 2015 (UTC)

• Quick comment. I'm not disagreeing, but I'm not certain. The typical spark gap transmitter was a spark impulse that rang for a few cycles before the next one came along. That said, the Poulsen arc is not a continuous arc; it is intermittent. Is that a spark? There's the (late-to-the-party) 1911 Chaffee gap continuous wave generator (Poulsen with the hydrogen and cooling but minus the magnet) that Morehouse lists as a spark transmitter. Glrx (talk) 02:05, 16 December 2015 (UTC)

The Poulsen arc was intermittent? Good to know; I've always been confused on that point. I seem to remember some authors describing it as a "continuous arc" (but maybe they were describing the DC current to the arc) and others saying it was intermittent. It makes sense that the magnetic blowout could extinguish the arc momentarily during the cycle. Due to the high Q of the tank circuit the oscillator would still produce sinusoidal waves.

Regardless, my reading of the sources is that the engineers of the time distinguished between the two classes of wireless transmitter on the basis of the waveform they produced, not the generating device; "Damped wave" vs "undamped wave" (i.e. CW) transmitters: [1], [2], [3], [4], [5], [6]. "Spark" transmitters were "damped wave" transmitters, while the Poulsen arc was clearly classified as "undamped wave". It's going to be confusing to general readers to say in the lead sentence that the Poulsen arc is a "spark transmitter", and then he follows the link to the spark-gap transmitter article and reads that spark transmitters generated damped waves. --Chetvorno^TALK 13:51, 21 December 2015 (UTC)

• Another quick comment. Yes, the Poulsen arc is intermittent. Furthermore, the Poulsen arc isn't the high-Q continous tank circuit that you suggest. For most of the cycle, the arc is conducting and there's a parallel RC tank doing its usual thing. When the arc dies, the LC tank connection is broken, and and plenty of ugliness happens. It's sort of like L gets replaced with a much larger L'.... Although some features of the circuit remain continuous (the current through the inductor and the voltage across the capacitor), not everything is continuous. See, for example, Pederson figure 8 page 266, with its transient steps at turn-off and turn-on. The frequency determination is not a simple LC.

I agree that engineers of the time distinguished between damped and undamped waves, but I don't think the sources unequivocably say all spark transmitters are damped wave. That's why I gave the Morehouse description of the Chaffee gap continuous wave generator. Morehouse does have a continuous wave chapter (including Poulsen arc and alternators) that is separate from the spark gap chapter. The sources are not unambiguously grouping spark generator technology with damped wave. Yes, the typical spark transmitter would generate damped waves, but that doesn't mean some spark generators could not produce undamped waves. The undamped wave was the better technology, and traditional spark transmitters had difficulty operating at desired fundamental frequencies.

Glrx (talk) 20:29, 21 December 2015 (UTC)

By the way, have you ever come across a detailed modern explanation of how it worked? I mean why the hydrogen atmosphere, magnetic field, and water-cooled anode increased the upper frequency limit? The Bureau of Standards pamphlet [7] seems pretty authoritative, but it was written in 1922. The article could use an explanation of how the arc works. I was thinking of drawing up a simple diagram of the arc chamber in Inkscape ( after the holidays! ). --Chetvorno^TALK 06:58, 23 December 2015 (UTC)

Uh, no, I never come across a good explanation (ancient or modern) of how the Poulsen arc works. It is clear that lots of magic is involved. Pedersen IRE 1917 has the most detail that I've seen. He points out that electrode gap is small. IIRC, he describes the carbon electrode as cup-shaped to create arcs at the edge of the carbon electrode. He qualitatively describes the influence of the magnetic field on the spark and even quantitatively describes the arc moving a mm or so before dying. If the field is too weak, then multiple arcs form. If the field is too strong, the arc is blown out too early. He didn't put it this way, but there's miniature Jacobs' ladder perpendicular to the magnetic field; the magnetic field determines how fast the arc moves. The arc gets longer and longer and finally dies. The field is chosen to get the arc lifetime right. I don't recall if Pedersen addresses the fast positive ion issue; I think he skipped it because it wasn't the thrust of the paper. Somebody said hydrogen gas was used, but it was troublesome to have the gas bottles around; consequently hydrocarbons were used. I also think Pedersen is silent about the purpose of the cooling. I don't recall anybody going into that issue for the Poulsen arc, but somebody must have. NBS may have talked about the carbon electrode being consumed/burned up (and also rotated to even its wear), but NBS only hints at a rectifier action (arc doesn't restart in other direction like spark-gap transmitter); Morehouse identifies a rectifier action for similar spark-gap devices but not when discussing the Poulsen arc directly. If you think about it, a carbon electrode can have a thermionic hot spot to supply electrons for the next arc; the other electrode (anode) is cold so it isn't a good thermionic emitter. That gives you a rectifier for the fast carriers. I haven't seen any discussion of chamber temperatures, electric fields, mean free path lengths, and ionization. Well, there was a safety tip about not opening up the chamber while it is still hot because it will explode. Glrx (talk) 07:33, 24 December 2015 (UTC)