Troubleshooting 3-guide dekatron

[Andrew ODonnell]

Hi

I'm playing around and familiarising myself with Dekatrons. Have an arduino-interface PCB so running them off Arduino. OG-4 and A101 running nicely but i'm struggling with the Argon 3-guide tubes.

Here's a video of what it's doing - it looks like it's driving correctly but it's failing usually at the same pin everytime. But sometimes it's circling nicely.

https://photos.app.goo.gl/tznRK8o7DN14wGbX7 

Is this a hardware fault ('sticky' pin on the dekatron) or is this a fault in my software to drive it?

I know some Dekatrons can be a bit tricky and sticky, if this is just a sticky pin, what's the best way to try and resolve it? If the answer is 'run it for hours' - is it best to run it super fast or slow?

Many Thanks

[gregebert]

First thing to check is your timing. I use a symmetric 6-phase approach where you have overlap on the 3 groups of cathodes.

Secondly, you might need a bit more voltage-drop at the cathode resistors (ie, make them slightly higher value).

I had this problem with my homebrew A101 spinner, and once I had about 30 volts across the cathode resistors it continues to run flawlessly after nearly 8 years of continuous operation.

[Andrew ODonnell]

Thanks for the tips

I have an A101 and an OG4 running great on the same circuit - just having issues getting the 3-guide OG3 running (i'm aware the OG3 needs more amperage and a different drive circuit and it's got that). Anode to main cathode voltage is about 225 and Anode to Guides is about 180 so yeah about 45 volts drop for the guides compared to the main cathode. 

What makes you suggest a slightly higher voltage drop at the cathode resistors - something about the appearance?

As for drive, i've taken a fairly simple 4-step approach (I can reverse the direction and it doesn't stick so i'm happy the drive is working unless i'm missing anything?)

It's been running for about 2 hours now and it's improving constantly - much less sticking each time, so I suspect it's mostly the Dekatron itself being sticky. I'll leave it overnight and see if this helps. it's probably also worth noting it only sticks clockwise - if I run it anticlockwise it spins fine.

[gregebert]

The reason I suggested boosting the cathode voltage was based on my own experience with the A101; apparently I misread the datasheet and thought it needed about 15V, and it didn't work smoothly. After boosting it, it ran much better. At 45 volts, your design has plenty of margin for the A101 and the OG4 based on the datasheets I found online. If the current is correct (400uA), the only other factor you can control is the timing.

Once you have it working smoothly, try reducing your anode voltage to make sure there is some margin in the design, so that as the tube ages it will still work.

[Jon]

Difficult to say for sure without seeing the circuit, but if it spins fine one way but not the other, most likely you've just got a bit of a sticky tube. Particularly if the performance is improving with use - it's a very common observation when waking up dekatrons after a few decades asleep, though some models are more prone to it than others. I generally find that a few hours use clears up most of these issues and it doesn't seem to matter too much how fast you spin it while doing that. I'm drawn to the idea that a slower step should be better - give a little bit more time for the glow to really heat up the cathode and burn any crud off it. But I've no objective evidence to support that view - difficult to do controlled experiments. If you get a really recalcitrant cathode, then sometimes letting the glow rest there for a couple of minutes while running the tube at a current a little above the maximum rated value can be useful. What current are you operating the tube at?

Having said that, can you tell which cathode seems to be sticky? If it's K0 or the guide pins immediately either side of that, then you might have a tube which has been heavily used in an application where it's been stuck on one cathode (usually K0) for many hours. That can cause erosion of K0 by sputtering and deposition of material on the faces of the neighbouring guide cathodes - in both cases the surface properties of the cathodes are changed which can cause stickiness. And in bad cases that can be incurable. It's more of a issue on the neon-filled tubes though than the helium/hydrogen ones like your OG3 (there is no argon in there BTW, that's an oft-repeated error).

Jon.

[Andrew ODonnell]

@Jon

Really helpful thanks. This is new old stock OG-3 so certainly been sat quietly for a few decades. The max current (measured) on that circuit is 740uA or so which is just about right for the OG-3. (I have selectable 485/750uA circuit for neon/argon dekas)

Messing about with it and experimenting further - if I run a simple spinner at about 50mS per step it sticks (or rather fails to step and jumps back to k0). Running the same spinner faster at 20mS per step and it runs almost 100% of the time so there may be something to your ideas that it is struggling to heat up or needs more time to really glow. The sticky guide is about halfway around from k0 so not too worried about that.

I'll let it run for a day or three and slowly crank the speed down to try and keep easing it into service. Certainly just an hour or two of running (sticky jumping pins all over the place) and it's gone from being almost unusable to 99% fine. If this doesn't work i'll rig something up to pause the glow on the affected pin if I can and see if that burns it off.

This is all really helpful thanks - I have some other Argon (normal double pulse) Dekatrons which have been all over the place and this one is reassuring me it's probably the Dekatron and not my circuits. I'll give them another go once this one is hopefully cleaned up,

Cheers

[Jon]

Great, let us know how it goes. Always happy to hear of someone else having fun with dekatrons!

BTW, to your question about OG9 in the other thread - yes, OG9 are essentially the same as OG4, just with more main cathodes brought out individually to pins on the octal base (K0, K1, K5 and K9 instead of just K0). It's what ETL call a 'computing tube' - partway between a counter and a selector. ETL's analogous tubes are GC10/4B and GC10B.

Jon.