Building tubes

[robbiemcm]

Well, obviously tubes have to made of something don't they. It's not like they did them out of the ground the way they are, they have to be manufactured that way. I'm wondering whether it is possible to build your own tube effectively (although it may not be anywhere near as small, or similar in shape or anything). Is it even possible to build your own tubes, or would it be almost impossible to come by whatever makes it up. If you understood them well, and designed your own, would it be possible to make something that sounds better than stock tubes?

[octafish]

Sure, if you have money to burn and a glass blowing factory.




Errr, no. Especially in answer to your last question.

[SaBer]

Actually someone here once posted a link to a page with someones diy tube. It was huge and it didn't last very long, but it worked!

[Paul Perry (Frostwave)]

Tubes are horribly, horribly complicated. More so than transistors or chips, believe it or not. At least when it comes to getting them to do what you want in the real world. And tube materials are pretty hard to get... and a lot of the actual construction techniques are probably lost with the last of the old timers. It's not like recycling light bulbs!

[R.G.]

Tubes are a little like playing the guitar. The guitar is the easiest instrument in the world to play poorly, and one of the hardest to play really well.

Tubes are the simplest and easiest electronic amplifiying device to get to barely function. To get a *good* tube that amplifies well, that works reliably and lasts a long time is really difficult.

You don't even need to blow glass - there were metal jacketed tubes.

There was a guy who made his own tubes on the net a ways back. He took a glass tube, open at both ends. He inserted a cylindrical "plate" nearly the diameter of the whole glass tube, and put a tungsten lamp filament through the center as a heater/cathode. The grid was a spiral spring around the grid. Leads came out through stoppers on the ends. Evacuated, it worked long enough to measure gain. If he had used melted glass ends instead of neoprene stoppers, it would have worked for a while. You can definitely make tubes in your garage, just like you can ... almost... make germanium transistors.

But you can't make good ones. At least I didn't see any way for me to make good ones. Getting a truly permanent glass to metal seal is quite difficult. Apparently molten glass wets very few metals, so you need to put a plug of that metal right where the glass meets the leads. Then you have to seal that plug of metal to the actual leads, and it turns out that you need an interface metal to add some compliance to the actual mechanical leads to keep from breaking the seal by thermal expansion coefficient differences. So the actual wire that goes through the glass in a tube is actually a five-layer stack inside the glass - external metals on the outside, interfaces inside those, and wettable metal in the middle. That gets you a hermetic seal that lasts for perhaps a century, as witness we now have lots of 50 year old tubes. Thank goodness...

The interior materials all have to NOT absorb gasses and then release them slowly so the tube dies. They all have to be vacuum stable, and thermally stable because it gets hot in there. Precision stamped mica sheet is the material for holding things in place, and you definitely need things held in place because for high gain you need the grid quite close to the cathode, and with tiny openings.

The plates need to be special metal plated for suppressing secondary emission when electrons hit the plate. The cathode needs insulated from the heater to keep from emitting electrical hum, and still be heated up to 700C through the insulation reliably. The cathode in receiving tubes needs coated with barium and strontium carbonate which is burned out during evacuation to a coating of barium and strontium oxides and releasing carbon dioxide to be pumped out. Otherwise you need tungsten or thoriated tungsten cathodes heated to upwards of 1200C to emit enough electrons.

During evacuation, you have to heat the whole innards up to high temps to drive off gas that is adsorbed (hanging on to the surfaces) so it can be pumped out or otherwise it will poison operation. Then you need to reflow and seal the evacuation vent to keep gasses from getting in. Once that is done, you flash the getter, which is a high-activity metal, to a vapor. That's the silvery lining on the inside of the tops of tubes. That deposits a layer of the active metal that will chemically bind with any gasses as the tube ages.

None of this is anywhere near the complexity or materials difficulties of modern semiconductors. That's why tubes are mostly made in technically backwards places - they bought the old manufacturing machinery from US tube making giants, or in the case of the Russians and Chinese, made their own. A new-design modern tube could run rings around old tubes, and be manufactured in a much more automated and efficient way.

But it would not sell many units, so the cost would be high, so YOU wouldn't buy it, so the makers would go broke. Or more likely, the bright youn MBAs at the maker's place would take one look at the market forecasts and not make them in the first place, which is what actually happens.

whew... don't get me started on germanium transistors... 

[Gladmarr]

...wait.  You can recycle lightbulbs?!?  Now that's DIY.  I have to try that sometime!  

[sir_modulus]

recycle them.....I used to eat them!  (actually...it was one of my shortlived hobbies =P)

Cheers,

Nish

[chumpito]

Don't ever let anyone discourage you.  Take alook here http://home.earthlink.net/~lenyr/hm-triode.htm (he also did a diode)  Sounds better is subjective, try it for yourself.

[Eric H]

A new-design modern tube could run rings around old tubes, and be manufactured in a much more automated and efficient way.

But it would not sell many units, so the cost would be high, so YOU wouldn't buy it, so the makers would go broke.

Some of the tubes coming out of eastern Europe recently have shown signs of new design --at least small updates to old designs. Someone may be willing to take a chance. The market may be fairly small, but it's solid, and larger than 10 years ago --especially since the old stocks are about gone.

-Eric

[zachary vex]

why not build your own solar cells?  here's a kit... from bell labs, 1960's... still available!

http://www.bellsystemmemorial.com/belllabs_kits_se.html

and yes, those are asbestos sheets in the upper corner.

[Gladmarr]

...can I eat those?