A Copper Curiosity

So what follows is likely wrong.  Not that it is intended to be or that the logic and premise are flawed, just that the available information is limited and some assumptions had to be made…..and you know what happens when you make assumptions.  But.  Even if this can be proven incorrect, as likely it can, the exercise is still of value from a learning standpoint….and, even if wrong, you will learn something from the background information and process of reaching the conclusion.  And who knows?… maybe the resulting answer will prove correct afterall!

Georges Claude, the French inventor, developed what we know as neon signs in 1910.  He filed a French patent on March 7th, 1910 and then gave a public demonstration of the technology at the Pairs Motor Show in December of that year.  Shortly after, he obtained a US patent as well.  This was US patent number 1125476 applied for on November 9th of 1911 and granted on January 19th of 1915.

Aside of the basic process and gas used, one item of particular interest that stands out in this patent is in regards to the electrodes.  In practice, a luminous tube in operation will eventually see a decrease in pressure, with subsequent rise of operation temperature due to the transformer maintaining the constant current value, sputtering of the electrodes, and eventual failure.  This is due in part to the gas reacting with the electrode metal (in the case of gases previously used, like Carbon Dioxide, Nitrogen, etc which are reactive)  or in the case of inert gases such as Neon or Argon, the process is less of a reaction and more of an apparent absorption type process.  The gas in the tube, eventually “cleans up” and does not leave enough free molecules in place to carry the current and illuminate.  As the process progresses, it accelerates.  Now, in a properly processed tube with correct fill pressure and running on the correct amount of current, this process will take decades.  Yes, decades.  Neon is perhaps the longest lasting artificial light source yet devised.  There are examples that still function 70 plus years later.  One reason for this relates back to Claude’s original patent.

In Claude’s patent, he specified that the electrodes needed to be a minimum of 1.5 square decimeters per ampere of current used in the tube.  This equates to 15,000 mm^2 per amp of current. (10,000 mm^2 / 1 dm^2)  No  doubt he arrived at this value by a combination of calculations and experimentation.  Since he did not elaborate in the patent on any upper limitation, it can be assumed that the benefits continue to increase but it is likely that this is with diminishing returns and so he felt that beyond the 1.5 value there is unlikely any great practical advantage.

The above specification is historically interesting because of the effect it had upon the industry in its early years.  This patent protection, and Claude’s willingness to pursue litigation against any unlicensed use of it, effectively killed much of the competition during the 1915 through 1932 period.  Sure, others who did not purchase a license from Claude could make neon tubes, but in order to not infringe upon his patent, they had to use electrodes with a smaller surface area and their product would therefore not last as long.  There are ways around this, however…some with sound basis and some with “snake oil.”  Additionally, in the 1935 book, “Neon Signs” by Samuel C. Miller & Donald G. Fink, part of the electrode chapter mentions the expiration of the patent and the mass availability of electrodes meeting the Claude specification.  It goes on to state that while Claude specified 1.5 dm^2 per ampere, commercial practice usually dictated a need for between 9 and 12 dm^2 per ampere.  This is a pretty large discrepancy and one must guess that Claude’s figure was under ideal laboratory conditions and perhaps before adequate use in the field had been experienced, afterall, 2 decades had elapsed from the issuance of the patent and the publication of that book.

If sparkplugs were the quack medicine of the nascent automobile industry, for the neon business, it was electrodes.  If you can imagine it, someone probably tried it.

During this period, others began selling so-called “non-infringing” electrodes.  These were made below Claude’s surface area threshold for a given current rating but were made in different shapes, or with other materials and coatings to improve upon their efficiency and lifespan.  Some of these techniques, such as electron emission enhancement coatings, proved valuable in their own right and were incorporated into subsequent designs even after the expiration of the original patent.

Today, and certainly since the 30’s and 40’s, many common electrodes are made of a nickel plated steel or iron type alloy which then incorporates a coating inside to improve emission efficiency and in some cases a ceramic collar to prolong the life and help prevent ionization taking place on the outer surface. (“outside ignition”)  These are some of the tricks of the “non-infringing” models that were retained even after the surface area patent was no longer in force.

Here, however, is an old electrode.  It is made from Copper.   This means it was intended for a tube filled with Neon only.  It could not have been used with a tube containing any mercury as Hg will form an amalgam with the copper, not only softening the copper but also trapping the mercury and preventing it from doing its job in the tube.  This electrode is a bit of a curiosity.  When was it made?  Likely during the late 1920’s to early 1930’s.  Was it one of the “non-infringing” models?  Maybe.  Maybe not.    There are no markings on it to indicate a manufacturer or a definitive date range.  But, there are some ways we can try to pin it down a little better.

I obtained these examples from a fellow bender who pulled them from their original sign.  This sign dated to the 30’s and had other neon units equipped with the same electrode types that were still functional.  The unit these examples were on had been broken.   He repaired the broken section and wanted to reprocess and refill the tube while keeping these original electrodes in place in order to preserve as much of the sign’s originality as possible.  I completely approve of the desire and attempt.  …But…It was unlikely to be successful, as reusing old electrodes is not a good idea due to the emission coating being previously activated in bombardment and then whatever contamination may have affected it after the breakage….but…the effort was worthy and commendable to those of us who value the history and try to keep our restorations as authentic as possible.  But, I am not surprised it did not work.  In the end, he had to remove these electrodes, fit new ones, and then process the repaired tube as usual.

Other than the copper shell material, this electrode looks much like those we see today.  It has no odd shapes, no extra internal glass jacket like some of the early Claude electrodes…all in all, pretty standard looking.  Except for that bare, heavy, copper construction.  What else can we learn from a closer look?

I took the “worst” looking one of the two and did a partial dissection in order to get a better look and take some measurements.  I found the shell length, not including the tapered portion, to be about 43mm and its diameter to be tad over 9mm and the inside diameter to be 7.874mm.  There was a strip of mica used on the inside of the glass and a pair of Dumet metal leads crimped into the back of the shell and sealed through the glass with the usual pinch.  All and all, nothing unusual, save for the copper.  Taking a look at them with a polarizer, I found little residual stress, clearly showing that they were well made and properly annealed to prevent cracking and leakage.  So I moved on to doing some rough calculations based around Claude’s numbers from the patent.

Although I did a calculation of the outside surface area out of curiosity (hey, curiosity is the basis of this entire website!)  it is of little significance to my immediate purpose because in practice, the inside surface area is what matters.  Electrodes are made to avoid that “outside ignition” and even the emission coatings are primarily applied to the inside only.  The intent being that the inside is the portion that shall be the point of initial ionization of the gas in the tube.  And when you look at one up close, that is exactly what you see happening during operation.  In addition, for simplicity, I disregarded the tapered end portion from my measurements as there was no practical way for me to measure the area of the inside of that taper and a calculation based upon the outside minus the wall thickness would have been flawed anyway due to the uncertainty of the material thickness in the crimped and drawn part due to the manufacturing process.  So, right off the bat, there is a significant margin for error here.

So, taking the 7.874mm inside diameter and 43mm overall length, we get an internal surface area of 1063.14 square millimeters.  (A=2 X pi X R X L, or 2×3.14×3.937×43)   1063.14 mm^2 is equal to 0.106314 dm^2  Getting back to Claude’s value of 1.5 dm^2 per ampere, if converted to mA, which is the common range of typical neon tubes, we find that his specs would require 0.0015 dm^2 per mA of current.  Typical neon transformers have been made in sizes of 18, 20, 30, 45, 60, and 120 mA over the years.  And, typically we see electrodes sold for ranges of 20-30, 45-60, and 120-180.  So, as it happens, a 60mA rated electrode would, by Claude’s patent, need at least 0.09 square decimeters of area.  Thus the 0.1063+ value is certainly in the ballpark for a 60-70mA electrode.  And, physically, the overall size of it compares similarly to that of modern electrodes that are sold for the 30-60mA ranges.  Now, if one were to look at this in comparison to the statement from Miller & Fink that closer to 9-12 dm^2/A of uncoated material was required then at best, this becomes roughly a 10-15mA electrode.  This seems very unlikely, as it is a sizable item and rather heavy.  It is worth noting that while Claude specified surface area, he made no mention of overall mass.  This electrode shell has a wall thickness and mass that is much greater than what is seen today.  Additionally his numbers are based upon non-coated electrodes, and as we know, a good emission coating will greatly improve the efficiency.  So perhaps this added mass was a design feature to overcome limitations of the time.  And, since I was unable to account for the internal area of the taper, there, too, is a little more wiggle room for a better current rating.

Those assumptions, while not 100% perfect, are certainly reasonable and could be valid.  What is known supports the vintage as being 1930’s, possibly even early 30’s.  Since the Claude patent protection ran out in 1932, anyone could make electrodes based upon his calculations after this year.

If one wanted to make a case for this being a “non-infringing” electrode made prior to the expiration of the patent, it would be on thin ice….but I’ll give it a try…

Nowhere is it written that you cannot make a tube with 120mA electrodes and run it on only 60mA.  In fact, many of us do very similarly in order to build a unit with a longer life or to just stock fewer sizes and types of electrodes.   The manufacturer’s suggested current ratings are based upon their designs, calculations, and tests.  But, if a manufacturer was trying to get around a patent based on the current for a given surface area, it would be easy enough to build an electrode that passes muster at 60mA, according to the patent, but instead call it a 90mA or 100mA electrode for sales purposes.  Now, if you actually ran it at 90 or 100, it would not last as long (months or years vs decades, really) …..but….you could sell it as a 90 or 100 model.  Just say that your coating or shape or whatever special magic your sales people come up with makes it good at that current.  Did I mention that no one here was selling 90 or 100mA transformers at that time?   Well, typically they weren’t.  So now you could sell your patent infringing 60mA electrode but call it a 90mA model and it would be “non-infringing” based upon the surface area criteria.  ….and in the real world market, lacking a 90mA transformer, the sign people would have attached it to a 60mA or maybe even a 30mA one……and it would live a long and happy life.  Did anyone actually do this?  I don’t know.  But they could have.  Or I could be wrong.  But you still learned something today.