Hacker News new | comments | ask | show | jobs | submit login
Hand Made Vacuum Tubes by Claude Paillard (2012) [video] (vimeo.com)
114 points by js2 9 months ago | hide | past | web | favorite | 52 comments

We had huge vacuum tube industry here at Czech Republic (Tesla tubes) most of it is now bankrupt, but I know couple of people who bought old equipment and still makes the high end audio tubes at small workshops.

I tried to find tour of my high school teacher company (the school itself was owned by Tesla company), but I am not unable to find those, but there is another neat workshop for Nixie tubes: https://www.youtube.com/watch?v=NRYcvzbaRGE

A lot of the Tesla equipment is still in use by JJ Electronic in Slovakia who are still making a lot of the popular tube types found in musical instrument amplifiers.

Oh yeah, when they disolved the Tesla tube company their equipment ended up in hands of many people that designed and manufactured audio equipment: back in comunist era it was almost impossible to get decent music/hifi gear so a lot of people went on designing and building their own gear and then went on creating small boutique companies after revolution.

JJ makes the most reliable tubes on the market now. They are fantastic.

I loved when a few years ago one of the motherboard manufacturers brought out a MB with a tube based audio stage.

Didn't JJ tubes buy the old Tesla tube factories? I have a bunch of JJ tubes for my guitar amps!

Tesla used to make many millions of tubes per year. There must be many companies that bought their machines + many more still collect dust somewhere. Even that nixie tubes guy mentioned buying tons of equipment and reselling those.

That's cool history to know. Seems a lot of this trade knowledge has been lost

similarly, there's this place https://www.youtube.com/watch?v=KwK8KSAJrGs

Side note, a question often asked is why do guitarists players insist on tube amps? The answer from a lot of unscientific explanations is "warm tone" and "great timber" which is pretty unsatisfying. After going down the rabbit hole of designing and building my own HVDC guitar amp, I can assure you it's much more complicated! They do "sound better", but it's the way vacuum tubes modifying the signal is imperfect. The extremely simplified explanation is this: Any modification of the source signal during amplification is called distortion. It turns out some types of distortion sounds "pleasing" to the human ear. This is because the distortion can produce harmonic artifacts musically related to the fundamental note. This is what violin and sax players called "overtones" and gives those instruments their signature sound. Guitarists can use a tube amps cranked into distortion (either from the pick attack or from having it set too loud) to create overtones, aka the sound of rock and roll.

Yes, you are correct, vacuum tube amps have higher levels of distortion (and noise in most cases) compared to solid state amps. For vacuum tube amps the characteristic curve of distortion is a quadratic function called the second harmonic (frequency double the original sound and an overtone factor produced by most musical instruments - eg. a stradivarius - that gives depth and richness to sound). Solid state amps on the other hand produce tertiary and quintic harmonics that are not as pleasing to hear (which is why they try to reduce distortion for solid state amps to zero).

It was my impression that vacuum tubes also tend to be more readily able to handle high voltages than transistors (whether absolutely or for a given price), which might also be a factor; the tube amp might just be able to pump out sound at a higher volume, which tends to be a big deal for live performances (especially outdoor performances with large crowds).

Not really :/ my tube guitar amp is 5w and has an EL84 output tube. The tube alone is $5-$20 and there's the transformer which can be $100 or more. A single MOSFET can do the same job for less than $1 and similar wattages without the transformer.

You can get MOSFETs that'll switch 500v for under $2 as well. Check out the irf820, I issued it to drive an fx loop in my amp design.

This changes at extreme power levels. There are few semiconductors that'll transmit in the thousands of watts. Tubes still are used in transmission technology for this reason.

The "extreme power levels" are what I was getting at, though I might've been mixing up voltage and wattage.

Ah, makes sense. Actually audio engineering is a few years ahead of software engineering as far as architectures go. The similarities are uncanny.

For a concert, its better to have 20-40 amplifiers rather than one big single point of failure (microservices vs mainframe). It's better to lose a few small parts than one big one. When it comes to distribution of source signals, the architectures live engineers are using with aes50/Midas or Dante are pretty danged cool. They just assume component failure is assured.

The background music is a loop of Teddy Wilson and His Trio playing "The Man I Love", on the LP "Mr. Wilson and Mr. Gershwin".

Wrote these notes on what I understood the video was showing. This man is hand-making vacuum tubes of an early-1920's design with some design simplifications.

Part 1

The video begins with the final product of his labors. This vacuum tube is a triode - it as three elements (or what he terms electrodes): a filament, a grid, and a plate. The elements are inside a glass envelope, which was sealed after evacuation on the top, leaving the little swirl of glass. The wires for the elements (or electrodes) are soldered into the pins of this blue plastic base he has. To me, it looks like the WD-11 vacuum tube of the early 20's, but I think he patterned it on something else.

0:12: Cutting the metal that will become the plate. Heh, odd choice of tool, but whatever works!

0:47: Bending little tabs on the end of the plate material. Later on he is going to wrap this metal on a mandrel into a cylinder, spot weld it closed, then spot weld a little wire to it (both supports it and conducts the electrical current from it). This forms the plate electrode. In this design the filament in mounted vertically in the center of the tube, and all of the elements are mounted concentrically to it, so its usually the outer surface of the plate you see when you look at tubes like this. This design uses the filament as the cathode.

1:58: Making the grid element. This is the element that you vary the voltage on to change the current flowing from filament and plate. He is using a basic helical design, thus the winding of the wire around a small mandrel. The grid is spaced widely to not impede so many of the electrons in the path from the filament to plate - but closely enough that he can still get enough control of that flow when he puts a voltage on the grid. The little loops on the ends of the grid are a nice touch - I don't think I've ever seen that on a real production vacuum tube. The grid wire is much thicker that would otherwise be used in a production tube - BUT, if you do make it with fat wire, it is self supporting and that saves a lot of work in adding in additional supports.

2:19: Making the filament element. It is very much like a incandescent light bulb filament but without the coiled-coil design you see in a typical bulb. He is straightening the filament wire before welding it to the supports. Early tubes used the filament as the cathode and the plate as the anode. Some of these used straight tungsten wire, which you would have to get quite hot to 'boil' off the electrons. Then they figured out that dipping the filaments in metal oxides reduced the temperature you had to heat the filament to. I don't think he did such coating here.

2:59: Now we are cooking with gas...he is beginning the glasswork to construct the envelope. Glass workers can correct me, but I believe he pre-warms the glass, the hits it to melt it enough to allow the tool to flare the end of the tube, then uses the carburizing flame to bring it back down to temperature slowly to prevent it cracking from shock?

4:23: Neat tool for 'cutting' glass tubing. He scores it first, then uses this loop heater to cause the glass to crack through, releasing the part. From here we see some special tooling for turning the flared tube into the base of the envelope, and squishing the end in preparation for the installation of the electrical leads that will pass through it.

6:19: He is assembling the leads for the elements. The wire he welds all of them to is just for support during the later steps. I wonder if the dark areas of the wire are part of a treatment he did to make a good metal-glass seal. Good sealing here is essential - if a microscopic gap opens up when the metal and glass heat up during usage, gas will get into the tube, change its characteristics and given enough of it - ruin the tube entirely.

6:49: Leads are sealed in glass envelope base. Placed into a little oven to anneal/cool slowly.

7:36: Making the body of the envelope. No special seals here, but some nice glasswork in making the rounded end and attaching the evacuation tube.

9:50 - A whole pile of glass tubes. Job well done.

Part 2

He is welding on the electrodes - first the filament, then the grid is slipped over that, then the plate cylinder, then the support/return for the filament.

0:37: It appears that he is doing a cleaning and surface preparation process here. Oils, organics, fluff, all cleaned away. Everything has to be clean as we move toward sealing and evacuating the vacuum tube.

0:44 - The base and body are joined.

1:20 - The evacuation. not only is he evacuating the tube, he is heating the tube to drive off moisture, gases stuck on the surfaces within the tube. You can really see this at 2:05, where he uses what looks like an RF induction coil to heat the internal elements. The digital display must be the output from some high vacuum gauge. Later on in the video you see some special pumps he has built, but it isn't clear what he used for this tube process. I think the sound in the background is the pumps at work. Perhaps a combo of a roughing pump and mercury pump?

2:15 - Lights up the filament for the first time, all in the effort to get everything besides the elements out of the envelope. Also a check that it keeps heating for an extended period of time and doesn't just fizzle.

2:26 - The sealing operation; a set of flames melts the little tube at the top of the envelope, locking in the vaccum and flavor.

2:58 - This part impressed me the most - he molded his own tube bases. He would have had to selected a material that can hold up to heat well, as the heat from the filaments will conduct right down to the base pins, melt and eventually degrade the plastic surrounding the filament pins.

3:20 - Testing the tubes, using a custom rig to plot curves much like you'd see in tube data sheets.

4:19 - All right, time to use the tubes! first, he has it in a little regenerative receiver, then a CW (for Morse Code) transmitter, then what seems to be a VHF transmitter using a bit of old 450 Ohm transmission line as the tank circuit.

5:01 - Finally the man himself, Claude Paillard, and some of his other work. Love the jacket at the mill.

I wrote to him about how awesome his tube project is and to find out what the heck that music playing in the background was, and he wrote back:

"About the triodes...they have numerous 'children'...in the 2004-2005 8 or 10 were on a very low power transmitter (around 100 milliwatts) with a quartz on 3580 khz...with a lot of 'local' QSO in France. One of these is in the A.W.A museum in Bloomfield N.Y... Around 2009 we go to one watt of RF always Xtal driven, on 3560 kHz...the one on the picture has made more than 300 QSO all over Europe...The last is with a pair of triodes running 5 watts output on 80-40-30-20 meters, with VFO of course...and worldwide received With a careful look, you can identify the 'box' as an old surplus 'BC375 Tuning Unit'

There also others use, but the main project was to return to 'good old time' of our fathers in Ham radio"

Check out his website at http://paillard.claude.free.fr/

These notes are exactly what I was looking for, thanks for writing them.

if you liked watching this, then maybe you'd also like these videos, which i watched some time ago:

making a miniature vacuum tube triode https://www.youtube.com/watch?v=dRI0ZLTP6_0

The Art of Making a Nixie Tube https://www.youtube.com/watch?v=wxL4ElboiuA

The nixie tube one is interesting because the guy was attempting to reboot commercial tube making. I wonder how he fared, is he still in business? Doesn't seem to have posted any new videos.

Yes, he is.[1] $145 for each Nixie tube.

[1] http://www.daliborfarny.com

You'll probably also like glasslinger's youtube channel[1]. They make a bunch of things, including Fleming Diode Vacuum Tubes to Nixie Tubes. Lots of extremely interesting things.

[1] https://www.youtube.com/user/glasslinger/videos

Very cool video. Vacuum tubes are still highly sought after by high end audio enthusiasts, as well as guitar players who still cherish tube amplifiers.

Scarcity is a problem. No wonder most Western countries rarely have even small boutique production labs these days - the materials and processes involved are (were) highly carcinogenic, and creates a lot of toxic waste pollutants. I keenly follow Dalibor Farny (mentioned elsewhere in this thread) - he seems to have devised a better way to make Nixie tubes (though not tubes like 12AX7s etc. that amplifiers use).

But one wonders, with the current excellence of amp modellers such as the Fractal Axe-FX and the Kemper Profiler etc. whether soon the digital emulation of that analog tube sound will be enough to finally put the last of the tube makers out to pasture?

Guitarists and audio enthusiasts are some of the most superstitious and bull-headed people I know. You can overlay spectrum graphs of digital vs analog audio output, do double-blind tests, etc. and most will still prefer the analog version of it; crediting intangible things like "bloom", "warmth", "air" and touch-responsiveness.

Not discrediting the market for digital modelers though; most fans of these systems that I know use them for less wear and tear for tours and gigs on their tube counterparts.

EHX buying NewSensor's factory ensures that tubes will be around for decades, even for this niche market. Plus the NOS market is still one any laymen can participate in.


> Guitarists and audio enthusiasts are some of the most superstitious and bull-headed people I know.

I do think audiophiles trump musicians when it comes to superstitions. Just recently found out there’s a market for cables costing hundreds of euro’s to interface hi-fi components, with lush descriptions of the soundstage afforded by such cables. Ignoring the fact any record one listens to will have its sound passed through 20 euro cables at some point in its production.

Agreed. And not just any cables, but polarized cables, with gold plated connectors, and the copper wire formed in oxygen-deprived environments, with individual strands following certain twist-pattern to optimize the electron flow, etc. To read the descriptions, one would think the ore were mined by elves on the moon! Nonsense, all.

I'm hoping these blokes never crack open the amp to look at the plain old PVC-insulated wiring within. And don't tell them the oil-in-paper caps are fading over time. Ditto with the carbon resistors.

Meanwhile, humidity of the air and barometric pressure has a measurable affect on the sound of paper-cone speakers. How long before an "audiophile" barometric chamber is marketed?

I think that another factor may be something I recently learned about called "the Ikea effect": the idea that a part of the attraction that people have for Ikea furniture, above and beyond the low cost, is that because they're required to assemble it themselves, the owners feel that they've put something of themselves into the furniture.

I can see something like this being a part of the audiophile attraction to these types of components: that every time the owner adds a new set of magic interconnect cables, magic power cables, or whatever, they imaging that they hear an improvement. In the end, they have a lot of themselves invested in their "rig", as they call them, and so find it easy to hear virtues that aren't really there.

This discussion always makes me link to [1] the fact that audiophiles can't tell the difference between a high-end cable and a coat hanger. It's amazing what our ears want to hear vs. what they actually do.

The weakest link always has an effect, as you mention. To some musicians, this is the effect. I.E. Hendrix preferring cheap coily-cables which rolled off the high-end on his Strat/Marshall combination nicely. Now coil cables are a premium.

[1] https://www.google.com/search?q=audiophile+coat+hanger+test&...

This is news? Hundreds of Euro's is the LOW END of boutique cable. While there is much nonsense in the audiophile world, I'm also not going to connect my $100,000 speakers with an ugly $5 cable. If you're already in that realm, $1000 cables don't mean shit in the balance sheet.

Some guitarists like to use old tube amps because they can be repaired, maintained, and modified much more easily than their modern counterparts. Old tube amps tend to have point to point wiring, which is much easier to fiddle with than reflow soldered PCBs.

Electric guitarists are often poor, clever, and self reliant. They can regularly find themselves out in the boonies with nothing but duct tape and baling twine to hold their rig together. Choosing a more robustly built older amp is not an irrational choice for such a person.

Experienced guitarists (well, I, anyway) can easily an instantly spot the difference between a modeller and a real tube amp. I also haven't found tubes too difficult to find; whether NOS or produced recently.

Audiophiles... I don't understand why you want tube amps. They are good for guitar players because of their imperfections (e.g. overdrive, "warmth", etc) which is the last thing I want when I'm listening to a quality recording.

>Experienced guitarists (well, I, anyway) can easily an instantly spot the difference between a modeller and a real tube amp.

It might be possible, but it surely isn't easy nor obvious. "Instantly," as you claim? Sorry, but I'd bet a sizeable stack of money that's not the case. Perhaps if you're talking about a 20-year-old 1st generation Line 6 Pod or Yamaha Fx500, then okay, I can see that. But modern modeling amps are exceedingly good at reproducing the nuances of a valve amp. The Kempler is remarkable -- I would go as far as to say astounding -- in its ability to reproduce the breakup, compression, warmth etc. of an amp at all volume levels. If you heard one in person you'd be a believer. At the other end of the spectrum, even a $150 Fender Mustang amps does a respectable job of sounding faithful [disclaimer - the factory settings are pure rubbish and almost embarrassingly poor, but tweak the settings in the computer editor and a whole wealth of sounds are available, some of them are brilliant).

And why would someone want a valve amp for playback of recordings? The same reason you claim to hear the "warmth" of a valve guitar amp. The same reason guitarists use an EP3 (echoplex tape delay) with the tape out -- that preamp circuit compresses, clips, and enhances the waveform in a way that is very pleasing to the ear. Some notes increase in clarity and bloom in a pleasant way.

A valve amp doesn't imply distortion. Leo Fender was a radio tech before forming his amplifier/instrument company. He knew the value of having lots of headroom.

>If you heard one in person you'd be a believer.

The key is blind testing. Everyone thinks that they can tell the difference between a Kemper and a valve amp.


The gap has closed rapidly. I've gone to modeling totally, sold all my tube stuff. The digital stuff is about 90% of the way there, and the thing is I can have 90% of about 50 different amps. So nice to be able to go, from, say, a Vox AC30 to a Marshall stack in the flick of a switch.

I can't speak for you, but I am sure that most guitarist actually cannot tell the difference.

I've posted many threads over the years at TheGearPage, with real amp recordings vs hardware/software modellers, and in nearly all cases, the guesses are just that - guesses, which fall into a 50/50 distribution. That is of course, when the guys who are purported 'experts' actually participate in the polls. Most of them tend to disappear whenever they are challenged to voice in on them.

I have some lovely tube amps here, and I also own a couple of Axe-FX's, and for me, when travelling, the Axe-FX is my 'go to' purely for convenience. At gig volumes, and in a band mix, it is pretty hard for me to tell the difference, and I enjoy the tone the same.

I know a lot of guitarists say that they have to have the 'feel', but for me, the 'feel' is there anyway (I usually feed my Axe-FX through the power amp stage of a real amp). I normally play fairly clean, blues/rock based stuff, and the last gig I flew across the country for, the rented backline actually had a Mesa Rectifier III for me. I've never played a Rectifier style amp before, but I had my Axe-FX, with a Fender Bassman amp model preset, through the power amp stage of the Rectifier, and it sounded amazing. The other bands were asking me how I was getting such crystal clean tones out of a 'metal' amp!

> Audiophiles... I don't understand why you want tube amps.

The ode to the way things used to be I think drives this market. Combine that with vinyl and you have a medium and listening experience that's superseded entirely.

Nope. Things have been this way even back when it was "the time when things used to be that way." When vinyl was popular, and before CD's were a glimmer in Philips's eye, the audiophiles held out RTR tape machines as the holy grail.

Sounds like an ode to the way things used to be. Yearning for the past, no matter the present. The folks who grew up on tape machines are fewer and fewer. I'm sure there were gramophone enthusiasts saying the same sort of things.

I think the essence is in the word ‘emulation’. If a thing exists, and then another thing exists whose sole purpose is to sound like the other thing, this will definitely add to the prestige of the initial thing. I think it’s the trend for digital software and hardware manufacturers to focus on emulations of analog hardware that’s keeping analog hardware desirable for the current generation.

Make 20000 of these and you can start building your own version of ENIAC!

diy mass production plant :)

Here's the original page where the video was posted, with some context information, in French : http://paillard.claude.free.fr/

If you're interested in this sort of thing, I found other similar videos by another unique individual, here : https://www.youtube.com/user/glasslinger/videos

This is a beautiful video. I have a general concept of what's going on - maybe someone with more knowledge can correct me: the video shows the guy building a tetrode, which is a way to either control voltage or current depending on how its used. In the testing portion of the video he shows the current response as he varies the voltage through the tube... (?)

Fascinating video

CW Was CQ DE _ . _ . _ _ . _ _ .. .

(Seek You from) normally a call sign would follow K0MGD

The amount of skill involved in this is mind boggling.

I watched at first thinking DIY? and then saw all of the custom equipment involved - Not for me.

It does make me reconsider the value of several boxes of older vacuum tubes I have upstairs in storage, thanks to my father-in-law - I would probably best be served by going through first them before discarding...

just recently saw this, where you see two custom 275 lbs monoblock tube amp-- I thought it was extreme but cool: http://www.stereomojo.com/Small%20Speaker%20Shootout%202007/...

Vacuum Tubes are amazing - they switch around subatomic particles called electrons. But they are simultaneously low tech.

That's a beautiful vacuum(?) pump @15:10. Anyone know more about it?

I'm unsure, I think it's custom or something vintage (both).

No one uses mercury any more as a working fluid in the pump. you use vacuum rated oil instead. And he's definitely using mercury. Also looks like a two stage piston pump. And mostly people use liquid ring vacuum pumps as roughing pumps.


Difficulty the roughing pump only gets you to a few microns at best. Good enough for neon work. But tubes require a hard vacuum. Need a diffusion pump (hot oil vapor) or turbo molecular pump (ultra high speed 'fan') after the roughing pump.

I don't think he's using a diffusion pump or turbo molecular pump.

There is a page on the pumps on the author's website (in french): http://paillard.claude.free.fr/triodes/pompes/pompes.html

It seems he made all the specialized equipment he needed to make the tubes.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact