Making Nixie Tubes At Home

[Aleksander Zawada] makes vacuum tubes in his home. One of the most challenging builds he has taken on is to produce a working Nixie tube. He describes the process in a PDF (Internet Archive, updated 2024), covering his success and failure. It seems the hardest part is to get the tube filled with the proper gas, at the proper pressure, and firmly seal it. In the end he managed to make a tube with three digits (0, 1, and 2) that worked for about 700 hours before burning out.

[Aleksander] joins [Jeri Ellsworth] on the short list of hackers who can pull off extreme industrial manufacturing at home. Kudos.

[Thanks Duncan]

27 thoughts on “Making Nixie Tubes At Home

  1. thank’s Hack a Day this is great rescearch and good to see posted here , keep up the good work ,
    good article and inspiration to analogue tubes for valve amplifiers for example guitars , thumbs up and a great day to you guys ;O)

  2. from the pdf, this is the spirit of the hack:”It may seem to be crazy and impossible but I try. The writers in all available books have said: “Making homemade tubes is impossible.” I believed them for a long time, but I wanted to try.”

  3. I seem to remember seeing this PDF a while ago…
    Still, HaD is hacking paradise today. This is an exceptionally difficult project.

    I hope Aleksander is reading this, as I have some advice, because I was going to try this myself, and have done a ton of research. Forgive me if you know these already- it wasn’t clear from your PDF if you knew about these parts:

    The most difficult part is the glass to metal seal for sure, you got that right. 3 things that can make it much easier for you:

    1. Get a glass tempering oven. In my city, Pittsburgh, PA, we have the Glass Factory, a dedicated glass-blowing/casting facility and mini glassblowing college. I learned from doing some over work there glass will shatter far less if you have a tempering oven. The oven acts as a controlled cooling chamber that takes a lot of the internal cooling stress out of the glass, by slowly un-heating it over a time curve set for a glass type. If you can’t buy one, find a local shop that has one and rent time on it. Glass needs a day or more to cool properly to lower

    2. Use boro, or borosillicate glass. You didn’t mention what kind of glass you were using. Start with good, pure borosillicate or pyrex (trade name for a specific borosilicate glass)- and…

    3. This is where I think you’re having the most trouble- You’re using copper. The thing about glass to metal seals is, you need to match the thermal expansion curve to a metal with a similar curve, or one stays too big in cooling and cracks the other.

    Two metals with very close thermal curves to borosilicate glass are Kovar, and Tungsten. Both have been used for metal to glass seals, and believe it or not, I found both on ebay in wire spool form, perfect in size for making vaccum tubes, nixies and whatnot. Kovar is very expensive, apparently. Both should work much better than copper- sealing tighter to glass, and causing less breakage even without thermal tempering.

    I figure, you could always seat pins on top of the metal through the seal, and plug on your nixie internal asembly, as it were, if you wanted to. The glass connection is the tricky part. If you wanted it to last longer, you could get the internals platinum plated very, very thinly, or add a very very small drop of mercury metal to the inside of the tube to assist as a getter. Many nixies and triodes did that, I’ve read.

    I’ve found general Paschen’s curves for nixie gas mixture, but nothing with actual numbers on a scale to aid in plotting. Just a general picture like you found. Have you found anything more exact with values attached? This is what keeps me from attempting the nixie build, along with not having a vaccuum pump.

    Have you considered blow molding your nixie tubes for more consistent stresses and volume calculation?

    I applaud your success- you’ve gotten further than I, so forgive me if I said something you already knew and didn’t mention in your work.

  4. I forgot one thing- if you want to try using a steel-tungsten alloy for parts of this, and you have the capability of melting steel (who knows, maybe you have a homebuilt arc furnace- some here apparently do) try this:

    Tungsten melts at 6192 Degrees F, way too high to melt at home. BUT- one of my teachers has told me that you can actually dissolve tungsten into molten steel. It’s one of tungsten’s strange properties, it will completely dissolve in molten steel, and you can precipitate the tungsten out by dissolving the steel in acid. He used to work for a company that did just that apparently.

    It might be pointless to mention, but if you find a way to melt steel at home (build a simple arc furnace with some carbon arc welding rods), you could technically cast tungsten/steel mixes for digit annodes/cathodes, or make your own hybrid wire.

  5. Every one of today’s posts are excellent. Giant CO2 lasers, diy computers, vacuum/nixie tube building, and RF transmitters. Great stuff and I hope to see more like this from HaD in the future.

  6. Dude! That is just cool. And I’d like to echo the sentiment I’ve seen expressed numerous times lately – these latest hacks are great! True hacks! Nicely done – this is the kind of stuff I come here for.

  7. well i don’t have time to review the whole thing tonight but since my ‘brother from another mother’ posted i will at least take the time to chime in too.

    there is only a handful of people that could pull this off with home based resources and even fewer that would take the time to do so, that makes this a great hack!

    we see so many college group hacks and senior projects, not to downplay ANYONES work (i posses about half the knowledge to fully understand most the stuff here in the first place), but it gets a bit ehhhhh. you know how many RFID dorm locks can we see in between something really unique.

    and today has been a plethora. good job aleksander and the rest of the hackers today and thanks HaD for bringing it to all of us.

  8. Thanks for all!
    This pdf is rather old. Now I have better laboratory than 2-3 years later.

    Usually I use soda glass. The coefficient of thermal expansion is ca. 100E-6 K^-1.
    I have small electric oven with big thermal capacity. That’s why cooling is very slow.

    I use borosilicate glass, but only for aparature. Not for tubes. In Poland is kind of this glass called “termisil” and “silvit” (this is old glass). Unfortunatelly, the coefficient of thermal expansion isn’t good for vacuum-tight connection with any metal.
    Sometimes I use molybdenum (kovar) glass (In Poland it’s usually kind called “ZS5Na” or “ZS5”), but rather only for vacuum gauges.
    Pyrex glass is very expensive; it’s good glass for aparature.

    I prefer “soft” soda glass, although this glass “love” glass stress.

    I buyed a lot of special wires with copper for vacuum- tight connection with soda glass.
    Preparing glass wafers with wires isn’t any problem. I’d like to prepare glass disks, but it’s difficult.

    Sometimes I add mercury inside nixies, but I hate mercury, because it’s poisson:/.

    About gas mixture: I have some luck: In my job I found a few time ago, old steel bottles 10l with neon-argon mixture (Ne-0,1%Ar -if I remember well). It’s from ca. 1989, when in my job was exeperiments about PDP displays. In the old article form 1963 I found that the pressure in nixies should be 20-30 Torr. But I fill my tubes “on eye”- for the best lighting.
    Having this steel botlles with gas mixture, I haven’t any problem with volume of tube.

  9. If you’ve got the ability to alloy your metal for electrodes, an iron and thorium alloy might work too. The radioactivity from the thorium apparently makes the strike voltage a lot lower by pre-ionizing the gas a bit. Same reason tungsten welding electrodes often have thorium in them. I suppose if you were melting some tungsten into your iron, as suggested above, you might as well use the tungsten thorium alloy, for the benefit of both.

  10. interesting note re. tungsten.
    Plus its easy enough to get hold of in the form of welding rods for not that much $.

    has anyone run into the “microwave smelter” ? a while ago someone discovered that you can generate temperatures in excess of 2000 degrees using an unmodified domestic MW oven.


Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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