How To Make A Pilotron, The Forgotten Tube

The vacuum tube is largely ignored in modern electronic design, save for a few audio applications such as guitar and headphone amps. The transistor is smaller, cheaper, and inordinately easier to manufacture. By comparison, showing us just how much goes into the manufacture of a tube, [glasslinger] decided to make a wire-element pilotron – from scratch!

To say this is an involved build is an understatement. Simply creating the glass tube itself takes significant time and skill. [glasslinger] shows off the skills of a master, however – steadily working through the initial construction, before showing off advanced techniques necessary to seal in electrodes, produce the delicate wire grid, and finally pull vacuum and seal the tube completely.

The project video is an hour long, and no detail is skipped. From 2% thoriated tungsten wire to annealing torches and grades of glass, it’s all there. It’s enough that an amateur could reproduce the results, given enough attempts and a complete shop of glassworking equipment.

The pilotron may be a forgotten design, but in 2018 it once again gets its day in the sun. Overall, it’s a testament to [glasslinger]’s skill and ability to be able to produce such a device that not only looks the part, but is fully functional on an electronic level, as well.

There’s a few people out there still building valves the old fashioned way, and we’d love to see more – tip ’em if you got ’em. Video after the break.

[Thanks to Morris for the tip!]

20 thoughts on “How To Make A Pilotron, The Forgotten Tube

      1. There seems to be some confusion of Pliotron and Pilotron. The Wikipedia page the article links to is about the PLIotron, while there is a reference to the GE PILotron here http://collection.sciencemuseum.org.uk/objects/co8362031/general-electric-ca101-pilotron-triode-mcvitie-weston-valve-collection-1918-thermionic-valve

        Since the are both described as Triode Tubes could it be that they are in fact the same device and an early typo resulted in two names for the same device? Either way its a fascinating article and the work in the video is extraordinary.

  1. I heard “a Tungsten filament with 2% Thorium” in the video and all isotopes of thorium are radioactive. Ionized particles does lower the voltages required, which is the reason it is used. But it got me wondering if you used a sensitive enough current meter could you actually use a powered off tube/valve as source of randomness.

    1. Of course there is much randomness in the precise current outputted.

      Both thermal in relation to the precise temperature of the filament and thermionic emmision and due to the number of passed electrons.

      But as far as I am aware these effects are brought out more easily in other configurations

      1. The randomness at very low currents has a name – shot noise. https://en.wikipedia.org/wiki/Shot_noise

        Thorium is used as it lowers the work function of the tungsten so you can get more current at a lower temperature (and power into the filament). The resulting work function is lower than either one alone. Radioactivity of the thorium has NOTHING to do with that whatsoever. There are no deliberate ions whatever in a tube – it’s why they call them vacuum tubes. All the alpha particles (helium) thrown off by the thorium decay (which is about the longest half life of all naturally occurring radioactive elements – 14+ billion years) aren’t enough to mess up the tube operation, and in fact, due to partial pressures, actually leak OUT through the glass faster than they are created (I work with high vacuum…and tubes, and noise issues). At the higher currents tubes usually run at, shot noise due to the discrete charge on an electron is more or less gone, averaged out in the 6.242 × 10^15 electrons/second/milliamp. Scale is important to understanding things!

        1. I do agree with you on the 14 billion years but (there is always a but after an agree) there is 0.02% that has a shorter life (75400 y) and other trace amount isotopes with a half-life in the order of days ( https://en.wikipedia.org/wiki/Isotopes_of_thorium ) And some of their decay products have a short half-life as well.

          So 0.02% of 2% or 0.0004 % which is a tiny number but considering the number of atoms in say a 1mm cube of tungsten with 2% thorium (~343 Quintillion (10^18)), which may be more than the metal in the filament but this is just a back of the envelope starting point, would be in the order of 1,372,000,000,000,000 atoms that are not tungsten and are not the isotope of thorium with a half life of 14 billion years. Again I’m using back of the envelope maths, so my calculation will be off by a few orders but the general tune should be the same. So if I run with that number and assume that all those have a half life of 75400 years I get that 1372000000000000/2/75400/365.25/24/60/60 every second ~288 decays events per second which would be more than normal background radiation, but still relative negligible. It would not generate a easily measured current. I was originally thinking of it as a form of Alphavoltaic/Betavoltaic power source with a low enough current to see the decay events, but high enough to be easily detectable. It would be neither.

          TL;DR – After a bit more thinking, any effect in a powered off device is not an easily measurable quantity.

  2. “The vacuum tube is largely ignored in modern electronic design, save for a few audio applications such as guitar and headphone amps.”

    Is it?

    I worked in broadcast radio in the late 1990s into the early 2000s. I guess that’s a long time ago now but it’s still well past the time that most people would think of vacuum tubes as being a common thing. Anyway, back then broadcast transmitters tended to use a vacuum tube or two as their final stage. Solid state transmitters did exist but building a transistor that could operate at such high powers and not immediately convert itself into a puddle of slag was difficult. I say transistor because ICs that can handle 10s or 100s of thousands of watts are not really a thing. I don’t think they ever could be!

    Besides being more expensive transistors were also much more susceptible to static discharge. That is kind of important when you connect them to an antenna that is several 100 feet in the air. Just the wind blowing across the antenna produces a lot of charge and lightning strikes are a regular event! i was told that transistors usually had higher fail rates than vacuum tubes in that application.

    Of course these “tubes” would be totally out of place in the back of Great Grandpa’s console radio. They were ceramic pucks with big metal heat sinks permanently attached that looked similar to today’s CPU coolers.

    So.. is this still the case nearly 20 years later? Or have transistors improved so much that they finally overtook even that market?

    1. I don’t know if they’ve taken it over yet, but they’ve probably started to make some significant inroads these days. We’ve got really nice high powered IGBTs and SCRs that can probably handle the kinds of abuse that used to make them unattractive for those applications. A lot of the RF side of it has come from satellites and microwave links for communications/internet where getting tubes to operate at 10s of GHz can be just as difficult as making good silicon.

  3. Check out Vinyl Savor on the internet for some nice Tube of the Month articles. I am waiting for a giant 110 dB/W horn speaker article on here. Many audio nuts like low power amps and high efficiency speakers. I honestly think you haven’t lived until you hear a RCA ’45 triode amp. Amazingly, linear devices. Most ‘too much hum in tube amps’ people have too little B+ filtering not heater hum. Mine are dead quiet.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.