An investigation into better ways to do cathode poisoning prevention - 1000 fps slow motion included

[Spirit's lab]

While designing my own clock, I decided to investigate the cathode poisoning prevention methods utilized by most clocks, and I discovered that there's room to improve and experiment.

Here's a short video: https://www.youtube.com/watch?v=skBwGGQ58MI

If you can't or don't want to watch: Switching between cathodes with a delay of 1-2 milliseconds is going to provide the same cleaning effect as the "slot machine", except without the extreme flicker which may be annoying to some.

As for my design - it's two HV5522s in the PLCC package connected to an ESP8266(for driving the HV5522s the 3v3 signals are shifted to 5 - that works up to a supply of 12.9V) and Yan's NCH6100HV boost board.

It would be interesting to see what everyone here thinks.

[Jeff Walton]

Question:  is the cycle long enough to effectively heat the surface to drive off the contamination? Are the pulses being driven at a high current for the milliseconds that they are on?  By high current, I'm speaking of a rate that exceeds nominal by say- 2x to 3x?  My understanding of the physics of cleaning the surface requires heating on either an instantaneous or continual basis to be effective in driving the contamination off of the cathode to expose a clean surface.   

I like the approach and would love to see a couple contaminated nixies (of the same type) subjected to some period of time using your method and conventional cycling and compare the outcome after months of operation. 

Nice video! 

Jeff

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/d11e2b91-22ab-47a4-b81b-da27cb3e1bed%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Allen Dutra]

The high speed cycling makes a lot of sense when the "slot machine" effect isn't desired.

I guess the harder question is how much a rapid cycling is enough. 

From the standpoint of creating a look and feel for the clock, I would implement 20ms of cycling (10 x 2ms) done every second when second digit transitions. This way the clock has a small flicker on every transition (second, minute, hour transitions will all have 20ms of cycling). Keeping flickering consistent and tied to the transitioning seconds digit makes the flickering easier to explain to those that notice. The most reductive answer to those that notice is "Nixie tubes are less stable than modern displays this is why there is a little flicking". This leaves out cathode poisoning as a concept and the fact that the flicking/highspeed cycling used to reverse cathode poisoning. Yet the answer still gets across the truth that because Nixie tubes are flawed there will be some flickering.

@Jeff

Generally if a tube has heavy cathode poisoning, over driving the tube with 2 to 3 times the rated current is often needed to try and recover / reverse cathode poisoning in a nixie tube. Yet assuming a nixie tube doesn't have any cathode poisoning the going theory is; if every cathode is used regularly, cathode poisoning never should never occur in the first place. Simply creating neon plasma around every cathode on a regular basis should prevent cathode poisoning all together. What qualifies as "regularly" in duration and frequency is a tricky question but doesn't change the theory.

[Spirit's lab]

@Jeff The heating should only be enough to prevent cathode poisoning starting with a healthy(ideally new) tube. I seriously doubt it would do anything to fix already existing cathode poisoning. It's driven using the same, standard current(4-6mA for an IN-18) for normal operation.

To fix existing cathode poisoning you do indeed have to seriously overdrive the tube and keep every cathode needed on, sequentially.

@Allen I made this with the intent of running it more often - a few seconds every 5, 10, maybe 15 minutes. If cycling on every digit change, it could be quite distracting, because the perceived brightness of the tube increases dramatically, and it would just keep flashing. Doing it every 30 to 60 seconds should be about the minimum time period acceptable, but it really depends on your tastes.

[Jeff Walton]

Still wondering if the CPP process can also be used to clean a contaminated tube over time as opposed to just trying to prevent poisoning.  I’m wondering about applying a higher power pulse for a short duty cycle so that the CPP goes a little beyond normal operation.   A cathode can poison even when the tube is in storage due to impurities in the gas and all cathodes are contaminated in normal operation. Regular CPP process assumes that normal operation of the cathode will drive off any impurities and we all know that in the real world even at nominal currents, cathodes still get contaminated even when we think that all of the digits are being operated. 

 

In the case of a high speed CPP process, a short, higher current pulse of short duration may be able to clean a mild buildup on the cathode surface.  Again, all cathodes are contaminated by the operation of another “ON” cathode.  Your process does offer an opportunity to sequentially hit each cathode with a pulse that is different than normal operation and may provide a cumulative effect over time to actually clean cathodes without the need for long-term and sustained overdriving of a single cathode.  It potentially could be done without a significant impact on the useful life of a tube in normal operation.  Rather than thinking of it as prevention, your technique might offer a way to mildly clean and prevent.  It actually does look interesting in the video…

 

Jeff

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6c8a3859-79f1-43d2-acdc-dbbb9e143f02%40googlegroups.com.

[Spirit]

Well, my current setup does not allow to vary the drive current, as that would require a custom HV supply capable of rapidly switching between regular operation(say, 170V), and "extreme" mode, at maybe 220-260V.

By all means, try it out. It could work, but I have no predictions on the outcome besides "the tube will get hot".

[Keith Moore]

Am I the only one who thinks the slot machine effect is much nicer to see than this alternative? However, I am on-board if this actually helps longevity. 

The slot machine effect is usually what gets non-nixie folks excited about the display. It's the "blinky-lights gets the attention" effect.  :-) 

[Spirit]

I wanted to design a clock that would sit under my main monitor, and so I didn't want it to be distracting while still maintaining good tube life. I prefer my neon to be calm :)

The slot machine effect is still perfectly valid for those that don't mind having the clock flash obnoxiously every few minutes.

[Nicholas Stock]

I agree Keith, it’s eye candy so I quite like the slot machine effects and similar. You don’t need it every other minute, once every ten minutes or so seems fine to me.

Sent from my iPhone

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/f0e93070-eb34-489b-9ac8-a9c577bbf7cc%40googlegroups.com.

[Kiran Otter]

While I don't mind the slot-machine effect, I don't know how effective it is.  My tubes only had slight poisoning when I built the Spectrum 18 clock 3 years ago, and now several have developed problems.  Here's a video.  I slowed it down 50%.  The two left tubes are '82.. the others are 82-83.

I intend to get one of Marci's testers in hopes of healing them.

Kiran

[Kiran Otter]

Oh btw, the tubes had about 16K hours on them 3 years ago.  In the new clock they're on for 15 hours a day, for 3 years.. is another 16K hours, so they have about 32K hours on them.

Kiran

[gregebert]

The '4' in the unit-hours position I saw in the video should not have been poisoned during normal operation, because all 10 cathodes will get exercised.

Could this tube have been in a different position, such as the 'tens hours', and been poisoned that way ?

What I'm wondering is if this is normal 'wear-and-tear' , or a small leak, rather than classic cathode poisoning (contamination of a seldom/never-used cathode from other cathodes that are used).

[Kiran Otter]

When I built a new clock 3 years ago, I didn't keep the tubes in the same positions. However, there was one that was showing signs of poisoning and I swapped it with the left most tube at the time.  Otherwise, they haven't moved for 3 years.

Kiran