Tube plate voltages

[Alan0354]

In high voltage world, insulation resistance is really not as important. Most insulators like Teflon or vacuum are plenty good. They are all over 400V/mil. It's the surface creepage that is the main problem. I don't remember the exact number, but the surface creepage distance is at least 10 times less than insulation.....like 40V/mil or less. Surface creepage is the current travel along the surface of material. Any dirt, moisture or contamination on the surface will promote a conductive path.

It's not the vacuum that is not as good an insulator, it's because you have a surface connect between one electrode to another electrode that create a creepage path. When I was working in mass spectrometers, I saw people had to really clean the surface before putting into the vacuum. That's where potting comes in. When you pot the high voltage circuit in the potting material, you get rid of the surface. Then you only deal with insulation resistance. Then things get a lot easier. Even a piece of paper can stand off a lot of voltage if you can eliminate the creepage path. But you cannot pot the vacuum tube as it needs the vacuum for electrons to flow. Also you have to be very careful when potting. Surface has to be clean, pour in the potting material and then put in an aspirator to create a vacuum to suck out any air trapped inside and on the surface to get rid of any surface inside so the potting material and the surface become one and eliminates the creepage surface.

This is where nano vacuum tube might run into problem. I don't know the latest technology anymore. But it does not sound easy to produce a nano tube that you have to seal a vacuum in, at the same time you need to have enough creepage distance to standoff the voltage. For something that run in over 500GHz, distance is the biggest enemy. The wavelength of 1GHz is 30cm in vacuum. So the wave length of 500GHz is 3/5 or 0.6mm. in order to eliminate the effect of impedance change due to wave propagation, the connection has to be less than 1/20th of the wave length. This means all connection has to be shorter than 0.6mm/20= 30 micron!!!! That is awfully short!!! You might be able to play some tricks to make the connection length longer. Say if you make the length exactly 1 wavelength, then the impedance is exactly the same as if there is no lead. But you better be very exact on the length as any deviation will change the impedance drastically. The length has to be control to within 30 microns. Then the whole thing about the heat of the filament that might rule out most of the common silicon based substrates.

I am sure if there is a will and market, there is a way to do it. Just how much people willing to pay.

[Tom Phillips]

I worked on a system that generated high voltages and was mounted inside an evacuated housing. We had trouble with arcing between points that worked fine on the test bench. Our solution was to reduce the vacuum inside the housing. That's when I realized why piston engine airplanes that can reach very high altitudes used "pressurized" magnetos. IIRC dry air is a better insulator than a high vacuum.

[Silvertone Jockey]

To get equal power, running at lower voltage requires higher current. I suppose it is much easier to build a high voltage tube than a high current one.

Exactly. Volts times amps = watts. Much easier to make a tube with a high voltage rating for power than a low volt one. Conductor sizing would be drastically different for one, another reason is that low voltage plates don't attract the negative ions given off from the cathode like high potentials do. In high volt tubes you have to be careful not to apply plate voltage before the tube is warm because the attractive force will literally strip the emission material from the cathode if you apply voltage too soon

[Alan0354]

I think it depends on what vacuum are you talking about If you just use a normal aspirator or even roughing pump, you still have a lot of air which has gas molecules and even water molecules which is polar. That's worst than dry air. When we talk vacuum, we are talking down to 10EE-9 atmospheric pressure. You really need high vacuum for it to work. We don't even try to turn on any voltage until the pressure goes down to the operating point.

Don't quote me on the numbers, in the mass spectrometer, we use roughing pump to pump the pressure down to 10EE-5 or 10EE-6. Then we turn on the turbo pump to pump it down to 10EE-9.

Someone correct me if I am wrong. People use dry inert gas in the tube before it gets pumped down. So even if the tube does not acquired high vacuum, the gas molecules are inert gas that does not ionized to create a conductive path. It's not easy to get high vacuum.

[soundmasterg]

In the examples I linked, (read them, they are quite interesting and a novel concept that is different from vacuum nanotubes) they use a cold cathode, and shrink the distance between elements to smaller than an air molecule. And instead of a vacuum they have it sealed in helium.....helium is wider spaced than oxygen/nitrogen so the chances of an electron hitting a helium molecule are low, and it behaves as if it is in a vacuum. They're running at lower voltages too...if I remember right 15V or something like that. They do have obvious advantages for someone like NASA as they wouldn't have to spend two years radiation proofing a computer that they want to use in space...they could just use these vacuum transistors. The thing I wonder about them that I couldn't find though is traditional vacuum tubes aren't bi-polar....they are limited to ground and up.....I wonder if these vacuum transistors would have that same limitation?

Greg