Connect with us

Building DIY Vacuum tubes

Discussion in 'General Electronics Discussion' started by George Schmermund, Jun 19, 2016.

Scroll to continue with content
  1. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    I've been playing with DIY vacuum tubes for a while now and the latest effort is in the area of simplifying the electrode geometry. Cylindrical electrodes are classical in design and work well as triodes, but the assembly is more difficult to align reproducibly. I've switched back to the parallel plate style of the WE 101D and its ladder grid, though the ladder is more demanding to build compared to the Audion's serpentine grid. I did make the first ladder grid, as shown in the photos, by just hand holding the pieces to weld them. This method is easy, but tedious. I'll use an aligning fixture for the next build.

    The next part that needs attention is the filament. The original WE 101D's that I've looked at on eBay show an 'M' shaped filament. This is a simple arrangement if you're using a glass arbor to support and position the electrodes. I'm going to avoid using an arbor and just utilize the extra pins in the button pinch and run three filaments in parallel. I'm looking forward to getting one of these tubes processed and tested. I expect the electrical characteristics to be ostensibly the same as the original 101D. We'll see.
    1-P1040534.JPG 2-P1040531.JPG
     
  2. hevans1944

    hevans1944 Hop - AC8NS

    4,560
    2,132
    Jun 21, 2012
    Your biggest problem will be obtaining (and holding) a really good vacuum. Are you planning on using a filament-heated getter after pump-down and seal off? Those pesky electron critters don't like colliding with residual gas molecules, although it does make for a nice glow with appropriate external circuitry.
     
  3. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    hevans - Thank you for the comment. Producing a useful high vacuum today is much easier than it used to be. Harbor Freight sells 2 stage mechanical pumps at very reasonable prices. There are many sources for mechanical pumps used for refrigeration service. Some good pumps can also be found on eBay for reasonable prices. A 2 stage pump in good condition and with clean oil is capable of routinely reaching vacuum levels of less than 10 microns (mTorr). If a simple DIY diffusion pump is used in series with the mechanical pump the vacuum level can get down to less than 10^-6 Torr.

    As you mentioned, getters can be used to cleanup stray gas molecules and still continue to pump on the sealed tube for many years. I've had very good results using low vapor pressure epoxy to seal the stem to the bulb before tube is baked and then sealed off from the pump. My preference these days is to use a torch to seal the bulb directly to the stem before pumping on the tube.
     
  4. hevans1944

    hevans1944 Hop - AC8NS

    4,560
    2,132
    Jun 21, 2012
    I once worked with a scientific glass blower who made his own quartz-enclosed mercury diffusion pump for high-vacuum pumping, backed by a belt-driven Welch vacuum pump. I tried to stay away from his lab for fear that I would accidentally breath some of the mercury vapor, but the risk was probably minimal. Still, I prefer the Varian "coke bottle shaped" oil diffusion pumps with a liquid-nitrogen (or at least water-cooled) cold plate at the top of the pump. You might be able to find miniature versions of these for sale on eBay, air-cooled instead of water-cooled and with Quik-Flange fittings... one of the scientists I worked with bought one, but AFAIK he never used it. Good for checking Bayard-Alpert ion vacuum gauges. We also used a large cryo-pump to "polish" the vacuum at the end of our particle accelerator beam line, but I never saw any utility there. The system was sealed with "O"-rings which limited (because of outgasing) the ultimate vacuum we could achieve. But my all-time favorite for ultra-high vacuum work (10^-9 or better) is a turbomolecular diffusion pump humming along at +40,000 rpm. I don't think, as a hobbyist, that I can afford to buy any of these high-vacuum pumps, but I do have a two-stage direct-drive rotary pump gathering dust down in my basement. Maybe I'll try to fire it up someday... but I think first I have to disassemble and clean it, replacing seals as necessary. That's a lot of work just to suck some air!

    What do folks use in DIY oil diffusion pumps? I sure can't afford the synthetic stuff we used in the Varian pumps. Does clear mineral oil have low enough vapor pressure and a big enough molecule to work?

    What do you use for a filament? Tungsten or maybe molybdenum wire comes to mind, but I think commercial vacuum tube filaments were coated with low work-function oxides for better electron emission.

    Wow! Homemade vacuum tubes! I am soooo impressed! Nice pictures too. Have you measured any characteristic curves for your triodes? Any idea what kind of plate and grid power dissipation they can handle? I bet you could build a kick-ass 400 kHz induction heater with that triode!

    Hop
     
  5. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    4-P1040031.JPG Hop - It's good to see that you're obviously no stranger to the world of high vacuum. Having your own good 2 stage pump is nice for starting out doing electronic vacuum experiments. Unless the pump that you have has been abused and sat for a long time with no oil in it, you may get by with just a flush and an oil change. The most important additional instrument that is needed for doing experiments is a good vacuum gauge that has been calibrated. Without a gauge there is no way to successfully know where the pressure is. It's like trying to make voltage measurements with a meter that has no display.

    DIY vacuum tubes are not that hard to make and even the most primitive attempts can be used as detectors in simple radios. De Forest demonstrated this when he had his first Audions made. The vacuum was typically so poor that the tube would start to have a glow discharge at 40 volts. But they worked!

    I'll attach a photo of one of the 'Audion' style tubes that I put together back in the mid 90's. It is one of many that have sat in the back of a drawer until last year when I got to playing with vacuum tubes again. As can be seen, the tube is sealed with epoxy and is still vacuum tight after 20+ years and can still hold off 200V with no trouble.
     
    CDRIVE likes this.
  6. 73's de Edd

    73's de Edd

    3,050
    1,288
    Aug 21, 2015



    73's de Edd
     
    hevans1944 likes this.
  7. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    Claude Paillard's video has certainly become one of the cult classics in DIY vacuum tubes over the years. I still find it fascinating to watch.
     
  8. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    Hop - The usual choice for pumping fluid in DP's used in most labs is DC 704 silicone oil. There are other more expensive oils that are readily available on the internet, but they offer little improvement to the average science hacker doing general vacuum experiments.

    The filament wire of choice for me at the moment is 3.1 mil thoriated tungsten wire. There is a good selection of tungsten wire available on eBay. The thoriated tungsten variety has 1000's of times greater emission than plain tungsten, but the tube processing needed to get this high emission is more critical. Either way very good DIY tubes can be made with tungsten. Coated filaments and indirectly heated cathodes are also doable at the DIY level.

    A diffusion pump is a good addition to any home lab and it enables the experimenter to easily get down to the 10^-6 - 10^-7 Torr region fairly easily. Of course a good mechanical pump is required to rough down the vessel and also back the DP.
     
  9. hevans1944

    hevans1944 Hop - AC8NS

    4,560
    2,132
    Jun 21, 2012
    Before I "retired" I usually purchased DC-704 silicone diffusion pump liquid from Kurt J. Lesker, but a 500 ml bottle (enough for one large Varian) was somewhere around US$100. Other workers in my lab opted for the more expensive grades, thinking this would improve the vacuum in their systems. They knew diddly about vacuum systems, where cleanliness is essential and metal-to-metal (Conflat type) copper seals are absolutely necessary for 10^-7 Torr or better vacuum. We had an ancient Varian helium mass spectrometer leak detector, but I was the only one who ever learned how to use it. We had residual gas analyzers, but again I was the only one who ever used it.

    A lot of the things we did involved applying thin-film coatings in a reactive gas atmosphere. So the drill was to pump the chamber down to as good a vacuum as possible (in a reasonable period of time, say, thirty minutes; then turn on the gas reactant flow through a mass flow controller and fiddle with throttling the gate valve to the diff pump and the mass flow rate until the process pressure was obtained, usually a few tens of milli-Torr; then run the process whether that be magnetron sputtering, filtered or un-filtered cathodic arc, or ion-assisted deposition... whatever was on the menu for the day.

    Most of the time I operated a tandem linear particle accelerator that we used to implant oxygen ions to graded depths in gallium arsenide integrated circuits. We were using the ions in lieu of chemical etching to isolate heterojunction bipolar transistor (HBT) structures epitaxially grown on the original GaAs wafer. You probably have an HBT in your cell phone, but these circuits were used for military electronics.

    A tandem accelerator accelerates negative ions from ground potential to a positive terminal potential, about +1.7 MV in our machine. Inside the terminal, the negative ions pass through a stainless-steel nitrogen stripper-gas tube at low pressure. The extra electron is stripped off, along with up to four outer-shell electrons, to create a positive ion. In larger machines, all the electrons are removed, leaving just a positively charged nucleus. Our machine didn't have enough energy to do that. Anyway, the positive ions leave the stripper tube and are now repelled by the terminal potential through a second (tandem) accelerator column. The total energy the ions acquire from the creation of a negative ion at the source to the final energy leaving the accelerator depends on three factors: initial negative-ion injection energy (usually 20 keV), terminal potential (up to 1.7 MV), and positive-ion charge state. So, with +4 ions and 1.7 MV terminal, we get about 1.7 Mev ions going in plus 4 x 1.7 MV = 6.8 Mev going out for a total of 8.5 Mev ions (neglecting the 20 kev injection energy).

    We never used that much energy on the GaAs wafers because the epitaxial layers were only a few micrometers thick. Didn't hurt to go deeper, but it would have been a waste of time so we didn't. Later, we implanted oxygen ions in a pattern of narrowly-spaced lines on un-doped semi-insulating GaAs wafer sections. This forms a Photo Conductive Semiconductor Switch (PCSS), for which my boss and I applied for a patent. Don't know what happened there, as I signed whatever rights I might have had to my former employer. So, yeah, I know my way around vacuum systems.

    My negative ion source used cesium sputtering. A coiled molybdenum heater wire heated the back side of a tantalum cup exposed to cesium vapor produced in an external oven and conveyed to the ionizer through stainless steel tubes. The tubes had about forty or fifty amps of AC current passed through them to keep them hot so the cesium wouldn't condense inside. The liquid-cooled sputter target was whatever material we wanted that would produce negative ions when sputtered by positive cesium ions. It was a complex process. The cesium ions, accelerated through a 5 kV field, sputtered the target surface and some remained embedded in the surface. Freshly sputtered atoms or molecules would sometimes grab an electron from the deposited cesium and emerge as negative ions with a -1 charge. The efficiency was all over the map. Some things, like gallium oxide powder mixed 50:50 by volume with silver powder, produced hundreds of microamperes of negative ions. Other materials would struggle to make it up to 5 microamperes. It was fun while it lasted (1996 to 2014) but now I am free to pursue other things in "retirement".

    Hop
     
    (*steve*) likes this.
  10. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    Hop - With a career background like that it should be a trivial task for you to build a working triode. This could get very interesting if you were to support the idea that DIY vacuum tubes are really doable in the home workshop. If you need a small DP I can post a couple of photos of a very simple pump that I've designed and built to pump on small experiments like tubes. It can be constructed in an afternoon with simple tools if the materials are at hand. The design defies the normal concept of how a DP works and dramatically reduces the complexity and number of parts required.
     
  11. hevans1944

    hevans1944 Hop - AC8NS

    4,560
    2,132
    Jun 21, 2012
    Working, yes, but high performance... probably not. I grew up puttering around with vacuum tubes before I was a teen-ager, and learned to appreciate the complexities of them before getting involved with semiconductors in the 1960s. In the Air Force I worked on defensive fire-control systems that had state-of-the-art equipment that used both vacuum tubes and semiconductors, but mostly vacuum tubes... tiny ones about the same diameter as a #2 pencil with teflon-insulated wire leads soldered to forced air cooled terminal boards and shock mounted in silicone. Those puppies were designed to survive the EMP from a nearby nuclear weapon going off in the vicinity of a B-52H heavy bomber, itself heavily loaded to capacity with nukes and hell-bent on mass destruction if the USA had to go to war. Those were scary times, what with a President and his brother and a civil rights leader getting assassinated and a crazy Russian pounding his shoe on a table in the UN. But we were prepared to respond in massive kind if the Big One went up. My EWO counterparts used bigger vacuum tubes: slide-out jamming racks filled with dozens of 4CX250B ceramic beam power tetrodes arranged as distributed RF power amplifiers. They would sometimes fire those puppies up and eliminate communications over a huge radius by broadcasting "white" noise. Amazing how much electrical power you can suck out of eight turbo-fan jet engines and still keep flying.

    When it comes to power electronics at high frequencies, no one has come up with a replacement for magnetrons, klystrons, TWTs and BWOs for microwave power handling. Sure, you can push GaAs technology and create microwaves with Gunn diodes and whatnot, but for real "reach out and touch someone hard" kind of power you need vacuum tubes. Semiconductors are catching up (finally) but AFAIK it is still cheaper to use vacuum tubes for serious power from DC to microwaves. However for really high power, switch-mode semiconductors have totally replaced thyratron gas-tube switches.

    In our shop we used a water-cooled triode as a series voltage regulator for a 12 kV DC power supply delivering several amperes to an electron-beam hearth. This tube failed twice during the sixteen years I worked there, so I sent it to a guy in California who rebuilds them. Both failures occurred when cooling water failed and the machine did not shut down, not the fault of the tube. My fault after the first failure for not realizing there was no flow-sensor interlock on the return line of the water supply. Fixed that after the second failure.

    I had a pair of 4PR1000A beam power tetrodes configured as screen-grid controlled push-pull variable-amplitude power oscillators, running at 40 kHz or so, driving a Cockcroft-Walton voltage multiplier to provide 1.7 MV terminal potential. The guy I replaced was changing these out every few years when he couldn't get the terminal voltage above one megavolt, but I discovered that no one had bothered to check under the chassis and clean out fifteen years accumulation of gunk, like the stuff that would accumulate on the back of the CRT in old televisions. That's when I discovered that both ceramic wire-wound grid bias resistors had finally burned out. It would sort-of work with just one grid bias resistor because of the transformer cross-coupled feedback, but only with fresh new tubes. Replaced resistors, cleaned gunk off everything and it has worked fine ever since. Vacuum tubes are tough if you treat 'em right. Well, at least they now have a half-dozen replacements on hand, unless the old tubes were discarded after I left. Someone told me people swarmed the lab and stole all the "goodies" I had stock-piled over the years I was there.

    It would have to be a well-equipped workshop to do it right, as Claude Paillard's video demonstrated. Nothing complicated or out of the reach of a determined amateur though. I thought it was a nice touch to use an induction heater to heat the plate to incandescence while pumping on the envelope to encourage outgasing. Didn't look like Claude used any sort of getter with his tubes. It is absolutely necessary to use a capacitor discharge welder to assemble things, but those are easy to make too. Not sure if he was using quartz or boro-silicate glass, but a tube annealing furnace is a good idea to stress-relieve the parts. I actually built one of these furnaces a few years ago from off-the-shelf components available from Omega Engineering. I used mine to melt aluminum to test the durability of our thin-film coatings applied to steel core-pins used in die-casting molds for aluminum engine blocks. So, yes, a determined amateur could make triode vacuum tubes as good as, and probably better, than Lee DeForest ever thought possible. That wouldn't be me though. I am too caught up in PIC microprocessors and learning how to do SMD assembly for now. A gave away most of my vacuum tubes to a fellow who restored "antique" radios, but kept a few power tubes "just in case" I feel the need to play around with tubes again.

    I would be interested in seeing your simple pump design. This doesn't involve using aluminum beverage cans does it? Please post some photos!

    Hop
     
  12. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    651
    May 8, 2012
    When I read the title to this topic I involuntary rolled my eyes:rolleyes: in anticipation of reading a nube's over ambitious attempt to do something far beyond their understanding and capabilities.

    How wrong I was! I can't speak for everyone here but these are some of the emotions this thread evokes in ME.

    Totally Awesome!
    Gloriously Ambitious!
    Talk about the ANTITHESIS of typical EP topic!
    WOW!
    I love this guy!

    Admittedly, when Hop starts reminiscing about cool things like heavy bombers, awesome ordinance and Niki's footware, it can't help but dress up a thread. :D

    If we ever have a monthly or yearly Project contest this will get my vote!! ;)

    Welcome to EP George!

    Chris
     
    Last edited: Jun 23, 2016
  13. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    Chris - Thank you for your kind words and warm welcome to this forum. I'll keep posting any progress in the triode builds as long as there are folks here that are interested in following along.
     
  14. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    Since I mentioned DP designs I'll toss out this steampunk project for consideration. Some time back I accepted a friendly challenge to make a scratch built diffusion pump (DP) with stuff that can be easily found at places like Target, Home Depot, and most any hobby shop. While the skeptics were snickering I was out collecting the pits and pieces.

    I have a very well outfitted machine shop in the garage, but not everyone does. Therefor, to level the playing field, I decided to build the entire project out on the back patio using only the most common of hand tools. A propane torch was used to braze some of the components together, but I considered a torch to be a hand tool, too. The total construction time was about 20 hours. This could qualify as a weekend project.

    Here are some pics of how parts of it went together. There are more pics detailing various steps, but these here should give you the flavor of the project. For scale, the cocktail shaker is about 9" tall. This design was successful enough to produce a vacuum of 2 x 10^-5 Torr on its first pump down . With some tweaks it would probably get down to the low 10^-6 Torr, but it wasn't worth the effort compared to ease with which the newer and better mini DP design can be built. I offer up this project only as an amusement for those of you following this thread. file.jpg 2-DSCF1624.JPG 3-DSCF1636.JPG 4-DSCF1650.JPG
     
    CDRIVE likes this.
  15. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    651
    May 8, 2012
    George, I like it when we know a member's name instead of only a handle. That said I hereby dub you 'MacGyver'.

    We have a member (Miguel Lopez) that lives in Cuba. What most of us find common and easy to obtain are unreachable for him. If you lived on the island you would be his best friend!

    I will definitely be following your vacuum tube progression and looking forward to seeing the electrical tests results.

    Chris
     
  16. hevans1944

    hevans1944 Hop - AC8NS

    4,560
    2,132
    Jun 21, 2012
    Me too! A very welcome indeed to EP, George! I am sorry I cluttered up your thread with my reminiscing.

    It seems there are a lot of hobbyists out there reproducing triode vacuum tubes, so your efforts are certainly appreciated here. I would love to encourage others to do similar things, although most of the drive-by newbies closest experience with a vacuum tube is the magnetron in their microwave oven, which they normally forsake to play around instead with the high-voltage transformer. There might be some Darwinian selection at work there.:p

    Edit: I forgot to include this Hackaday link in my original post. Lots of interesting comments and links at the end of the article.

    Anyhow, please bestow on our forum more photos and comments on what you are doing or plan to do.

    I finally looked up the circa 1920 Western Electric WE 101D tube and compared its plate characteristic curves with the one you posted. Looks like your version is a wee bit short on emission, but still a nice job! Do you think your tube could be built into a regenerative AM receiver?

    upload_2016-6-21_15-40-50.png
    Plate characteristic curves for WE 101D triode extracted from Western Electric Vacuum Tube Data, Part 1, 1941

    Hop
     
    Last edited: Jun 22, 2016
    CDRIVE likes this.
  17. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    Hop - The tube that was posted earlier along with its curves was about as simple as it gets. I wanted to test the idea of making a tube with the least amount of equipment that would still make a working tube. The geometry follows the earliest de Forest 'single wing' Audion. The single wing design is notoriously inefficient and is good as a detector, but little else. The photos attached in this post are the same tube with the filament powered up. The point of these photos is to demonstrate that a working triode can be constructed without the need for a spot welder of any sort. This tube was built using only a wire wrap technique which can be mastered in about 30 seconds. The filament uses a pinch method for attachment. There seems to be a wall of intimidation surrounding the hobby of tube building and my hope is to bring the bar down to a level where basically anyone wishing to give the project a try won't be scared off. 4-P0001383.JPG 3-6-P0001383.JPG
     
  18. hevans1944

    hevans1944 Hop - AC8NS

    4,560
    2,132
    Jun 21, 2012
    I think you have succeeded. The low vapor-pressure epoxy sealing technique truly does bring this into reach of the rankest amateur who can get their hands on a decent vacuum pump.

    Hop
     
  19. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    The 'double wing' Audion was a much better tube and could be used in an Armstrong regenerative radio circuit. The photo shows an example of the ones I made way back when. These tubes were built using a spot welder because I have always had them them in the shop and a welder does make the assembly go smoother. The wire wrap technique would work for assembling this tube, too. The double wing style is a much more efficient tube and I've built several regenerative sets using them. In the end they're still primitive tubes, though.

    The WE 101D style that I have in the works now should have about the same characteristics as the original commercial tube. The envelope on mine won't be a tennis ball type, just the straight walled ones that I make. An interesting point to me, as far as following the basic design, is that the phone company was still using the 101 type tubes as late as 1977 as a repeater tube according to the literature. Not a bad legacy tube to build.
    3-P1030946.JPG
     
    Last edited: Jun 23, 2016
    CDRIVE and hevans1944 like this.
  20. George Schmermund

    George Schmermund

    14
    7
    Jun 19, 2016
    I had a chance to do some machining over the weekend. What I have now is the first attempt to make a fixture for aligning and spot welding the grids for the the new triode. The material for the ladder is a 20 mil wire called Blendalloy which is an alloy of nickel and iron. The side posts are 40 mil nickel plated steel wire. Fortunately both wires are magnet.

    I was able to machine a piece of 1/4" thick mild steel plate into a suitable shape and add a series of slots into the plate to accommodate the grid parts. A magnet on the underside of the fixture was used to keep the wires in their assigned slots. A piece of aluminum (nonmagnetic) could then be clamped down on the wires to keep them in place for welding. The aluminum was clamped down with nonmagnetic SS bolts. This arrangement, when assembled, dramatically increases the ease and speed with which the grids can be welded and then shaped to fit between the plates.
    1-P1040558.JPG 2-P1040559.JPG 3-P1040554.JPG 4-P1040556.JPG
     
    CDRIVE likes this.
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-