First experiments with cathode glow!

I have successfully done my first experiments with vacuum and gas discharge.. It was quite simple, the vacuum pump is much powerful I expected.. I work only with air, I have no noble gas available yet.. There is a short video I made today. First, I turned on a HV power supply, then the pump is turned on, as the pressure in a glass bottle gets lower, the glow changes..


Common glass bottle, one hole for air-evacuation tube, one for cathode wire, all sealed by hot glue. I used the lid as an anode.

Vacuum setup

I planned to use a charging hoses with integrated valves to control provide and control the vacuum in all parts of system.

But I found that the valves can’t hold the vacuum and let the air in ๐Ÿ™ So I cut the valves and currently I use only the hoses mounted on a glass tube.. I will probably build the setup on a glass joints and glass valves with vacuum grease. It is not the simplest way, but it is cheap and should reliable.


Vacuum pump running

I finally cleared the pump from mercury (at least what I found), placed new o-ring (what I destroyed myself) and refilled the original Ultragrade 19 oil. The pump is pretty silent (much more than I had expected), I am looking forward to test it..

Some photos..



Spot (tack) welder

We can’t use a soldering inside the vacuum tube, the tin would melt during the heating of the tube because of the tin low melting point. Instead of that it is necessary to use a welding, probably the best solution is small spot welder, I choose a capacitive version. Large capacitor is charged on 16-20V and then discharged via thyristor and electrodes into the welded stuff. You can find tons of info about the spot welder, I won’t mess the internet by another one ๐Ÿ˜‰ I recommend this one: Tab welder.

Just some photos of mine.. I had to add another capacitor, this one is weak, and some foot-switch..

I used 4AWG flexible copper wires, little thyristor ฤŒKD,audio capacitor 2 farad (0.28 Farad in real..)


The black cylinder is a capacitor, I didn’t expect it to be really 2F for that price ($30).

Brass rods as electrodes, soldered on the copper core and covered with heat-shrink tube.

Tips of the electrodes.

Borosilicate glass

As I mentioned before, I am going to use a borosilicate glass in combination with tungsten wire. This glass is used in labs for scientific glass blowing, so the first place where I started to look for that glass were e-shops with lab supplies. There ale lots of suppliers, but all want some minimum quantity to order – tens or hundreds tubes of each diameter. But I had a luck, I found a 60% sale on 1500mm tubes, diameters from 12 to 20mm and a some 42mm, there were no minimum quantity needed. The glass brand is Simax, it is czech manufacturer.

Now I am looking for 6mm tubes, I will probably have to buy huge amount of tubes ๐Ÿ™ ..

Vacuum pump

I bought a used Edwards E2M5 vacuum pump for $150. I had thought the vacuum pump would be the most expensive part of the system, what a luck to find a pump for that price.. The only problem here is that after unpacking the pump from the box, I found a few drops of mercury spilled in the inlet port. So I examined the pump and realized that someone in the previous lab probably made a mistake and let the mercury get in the pump ๐Ÿ™ I take it apart and caught all the drops I saw with the tin wire. When I assembled it back, I found one little rubber o-ring left ๐Ÿ˜‰ I had to disassembly it again and then I found I had destroyed a large 44mm o-ring ๐Ÿ™ Now I am waiting for new one from eBay. In the meantime, I got new oil, original Ultragrade 19.

I am still afraid of mercury vapours, I will probably place a hose on the outlet port and stick it out of the window..

The pump is able to achieve a 7.5×10-4 torr, reasonable minimum for the nixie tube before filling by gas. Alek is working with much higher vacuum, I don’t think that is necessary. The final pressure of the nixie should be (according to Alek) 20-30 torr after filling by gas.

Tungsten wire

I started searching for material by looking for tungsten wire.. I found several expensive offers on eBay, too much money for several inches of wire ๐Ÿ™ Then I wrote several wire manufacturers, there the price/length ratio was much better, but the minimal quantity I had to order was 10kg of wire ($1500) ๐Ÿ™ Finally, I tried to search for the tungsten wire as a “Wolfram wire” and found one seller in UK, who offered a wire for a reasonable price – 300cm/$5..

It is 0.6mm diameter wire, quite hard and thick, 0.2mm in diameter would be better choice.. The tungsten wire has quite high melting point (3422 ยฐC), so it is fire resistant, I don’t have to be afraid of exposing it into the fire during glass working. That is difference against the copper wires.

Vacuum gauge

For the rough measuring we can use a common manometer, for more exact measurement at lower pressures, it is necessary to use a gauge for high vacuum. As expected, such a thing is pretty expensive, so I look for some diy solution and I found ๐Ÿ˜‰ It is called Pirani gauge, it consists of one perforated bulb and some simple electronic. The electronics I chose was Sooty’s one. It is simple, cheap and easy to build.. I took a 24v/1.2W, cold resistance 46 ฮฉ. I wasted 3 of 4 I have at home ๐Ÿ™ I perforated the bulb using a tungsten wire (0.6 mm) by heating it up in flame and pricking it into the bulb. Don’t heat the bulb (at least not much) otherwise it deflate because of its lower pressure.

I tried to melt the bulb and tube together, but it cracked during cooling, probably because of different thermal expansion coefficient of glass tube and the bulb. So the next try (with another bulb) was to seal it into tube by hot glue. I hope it won’t evaporate any vapours when evacuated. It is just some kind of plastics, so who know..

Glass to metal seal

The essential matter when making vacuum tubes is good contact of electrodes and glass case, it is called glass to metal seal. It looks simple – just to melt glass, place the hot wire electrode in and wait until it cools down. But there is a problem of different thermal expansion coefficient (TEC), what causes the glass near the connection to crack. So You must use glass and metal with similar TEC.

Soda-lime glass + copper wire

One way is to take ordinary soda-lime glass (TEC 8.5) and choose metal with similar value of TEC, it is often copper (TEC 17) or Dumet wire (I think it is also based on copper). The advantage of using soda-lime glass is that it has lower melting point, so it is easier to form it in proper shape. The main disadvantage is high TEC, so the glass tends to crack easily. Alek zawada is going this way, he mentioned a lots of cracks when dealing with soda-lime glass and copper.

Borosilicate glass + tungsten wire

Borosilicate glass is glass with lower TEC (3.3), it is more resistant to cracking when cooling. A tungsten wire (TEC 4.5) could be used in combination with that glass – it has similar TEC, so it should theoretically produce no cracks ๐Ÿ˜‰ Disadvantage of the tungsten wire is that it can’t be welded to tube’s insides, I haven’t solved that yet.. Tungsten wire is also very hard and the work with it is no comfortable like with copper. But You can expose it directly to the fire when glass forming, it won’t melt unlike the copper. I chose this combination..




Thermal expansion coefficients


Material Linear coefficient, ฮฑ, at 20 ยฐC
Aluminium 23
Brass 19
Carbon steel 10.8
Copper 17
Diamond 1
Glass 8.5
Glass, borosilicate 3.3
Gold 14
Invar 1.2
Iron 11.8
Kapton 20
Lead 29
Magnesium 26
Mercury 61
Molybdenum 4.8
Nickel 13
Oak 54
Platinum 9
Quartz (fused) 0.59
Rubber 77
Sapphire 5.3
Silicon Carbide 2.77
Silicon 3
Silver 18
Stainless steel 17.3
Steel 11.0 ~ 13.0
Titanium 8.6
Tungsten 4.5

Source: Wikipedia

Feel free to contact me.