Mysterious green spot randomly appears in one of my nixie tubes

[greg...@hotmail.com]

Every once in awhile, a small green glow occurs in the rear area of one of my 5092 tubes. The glow stops when the digit is turned-off, and does not reappear again if the tube is re-energized even after multiple tries. Then, perhaps days or weeks later, it reappears when the tube is energized.

Unfortunately I can't easily check the current to see if it's significantly different when the green discharge occurs. The digit glows fully in it's normal color, so there's no obvious harm.

I see that neon has multiple spectral lines, including green, so perhaps this is a discharge between internal nodes of the tube, and for some strange reason it's favoring a different spectral line ???

Green is a shorter wavelength than the dominant orange, so I'm inclined to think there is a higher field strength present in that region, thereby causing excitation to a higher state hence the different color, but the fallacy in that idea is that the green glow is a fairly rare event. I would expect  that if the internals in that region were closer together, for whatever reason, that the green glow would always occur; it doesn't.

It's been doing this randomly for more than a year now.

It's entirely different than the blue dots I've seen on tubes that have been sitting on the shelf for years, because those go away after a few minutes or hours of usage.

I suspect that's from mercury that has collected locally, and is subsequently distributed after the tube is energized.

[Dan Hollis]

On Thu, 29 Aug 2013, greg...@hotmail.com wrote:
> It's entirely different than the blue dots I've seen on tubes that have
> been sitting on the shelf for years, because those go away after a few
> minutes or hours of usage.
> I suspect that's from mercury that has collected locally, and is
> subsequently distributed after the tube is energized.

blue dot is due to multiplexed driver design, not due to problem in the tube.

from dieter:
The blue spots are caused by the following circumstances:
If low voltage drivers are used, as example the 74141 drivers, the cold
cathode voltage (non glowing cathode voltage) drops down to the clamping
voltage of the output transistors (70-100 Volts).
In this state the mercury slightly breaks down, slightly ionizes, and
little current flows from the anode, through all 9 non glowing cathodes
through the Z-Diodes (Driver protection diodes) to ground. And that
inonized mercury can be seen as blue spots in the tube.
That's all.

-Dan

[greg...@hotmail.com]

My clock is direct-drive, using 400V NMOS (IRF740) separate drivers on each cathode. Yeah, that's 45 drivers for 6-digits heh-heh.....did someone say "overkill" ????

DC supply is 350V (from AC-line doubler) to all anodes. I like the higher voltage so that as the ionization voltage increases with age, they will still illuminate reliably. It also reduces cathode-current variations due to tube characteristics, line-voltage, etc at the expense of higher energy usage.

But I will check the leakage current of the dark cathodes. Datasheet says less than 1ua at 25C (<50uA at 125C but I'm nowhere near that temperature....).

When my cathodes are off, they essentially float (off-impedance of NMOS). I thought about tying them thru a big resistor to the HV supply, but I reasoned that the leakage current was negligibly low so there should not be any glow.

With regard to the green glow tube, it's definitely the tube, and not the driver, because when I rotate the tubes for uniform usage, it's the same tube with the green glow even though it's being driven from a different driver than previously.