Some hate them, and they can be turned off, but I like em, and they look fantastic at night against the glow of the tubes.

Here's a mod suggestion for the kit:



You can make R14 a 330 ohm resistor, OR keep it "as is" with add R14a added.

Excellent And I will add my thanks for not only working it out, but testing it to boot

However, I have a small issue with the kit I assembled tonight. The dekatron I'm using is stalled. Behavior - When I connect power the dekatron may illuminate on 1 segment and then jumps half way round to another segment and stalls. If the oscillator is turned up, there is a small glow between the segment it's stalled on, and the next segment along. The LED is not glowing. Now, if I connect and disconnect the supply in quick succession, the dekatron will step randomly onto a few segments, but none adjacent to each other, and the LED will flash briefly on occasion. I'm using a 6909, and the board is wired as for a 6802. NE2 is oscillating fine, and adjusts as expected. I also tried varying R14 as low as 5K and as high as 47K, but the behavior remains consistent.

Before I mod the circuit, I need to solve this little puzzle

P.S. I do like your tube checker - a lot!

It's the 6909. I hate hi-speed dekatrons, such as the 6909. As you may have noticed, the glow color isn't a nice neon orange, but a pinkish yellowy kind of hue. Mechanically, its identical to the 6802, just filled with a different gas (mix). They are to be operated at higher currents. First suggestion, is to make R10 (anode resistor) smaller, say 220K. Also measure the actual HV supply voltage. Because I used fixed components, and a lot of 5% parts, it may not really be 450V. These hi-speeds like higher voltage. Since the cathodes are biased up to ~60V, the supply needs to be 460V minimum. You can parallel a resistor to R8 (7.5K). Try 150K first, and see if the voltage exceeds 460V. It should, however, stay below 500V, else you'll exceed the part ratings.

More bad news. Hi-speed dekatrons have a high dud rate. The gas that's inside them, either escapes thru the glass, or reacts with the internal metal bits, until no more glow. Don't have a lock on the theory, but from empirical evidence, the hi-speeds are just trouble. Never found a standard speed (neon) type that lost its gas.

I'll have a play, but probably will end up getting a 6802 anyway...

I'll report back in due course

So...

I change a few things, and still the 6909 refused to work correctly.

Undaunted (well I was a bit fed up) I rigged up this circuit on my prototype board. I didn't have any .22 caps, so used .1uf instead.



Here are my results:

AC in from the Variac - 105vac (adjusted to keep B+ correct)

No tube inserted.

B+ at the last diode 509vdc
B+ after the 1m res 470vdc (so where it needed to be)

Tube Inserted

6909 behavior was the same

B+ at the last diode had dropped to 350vdc
B+ after the 1m res had dropped to 218vdc

My Conclusion

The tube is buggered (technical term)

I don't have any tube data to test the tube on my Eico 666, and not sure you can anyway

Have I arrived at the correct conclusion, or am I missing something? I'm assuming there shouldn't be such a voltage drop across the tube?

Data sheets:

http://frank.pocnet.net/sheets/168/6/6802.pdf
http://frank.pocnet.net/sheets/084/7/7155.pdf
http://frank.pocnet.net/sheets/168/6/6909.pdf

My own AC sync'd spinner:


It resides here:
http://threeneurons.wordpress.com/dekatron-stuff/

Another voltage multiplier:


can be found here:
http://threeneurons.wordpress.com/dekatron-stuff/auxiliary-dekatron-stuff/

For a 6909, you have to increase the anode current. A 6802 is nominally 350uA, upto 600uA max. A 6909 starts at ~500uA, and upto 2mA max. That spinner you referenced uses too small caps, and that 1M is too big. Voltage multipliers have high ripple once you put even a hint of a load on them. Increasing the caps will lower that ripple, and increase your average voltage. This is even more important since that 6909 needs more current, and that 1M needs to drop down to 330K or even as low as 220K.

I doubt any tube tester has a setting for this class of tube. They were very specialized "industrial" tubes.

On my kit, how far did the voltage drop, on both sides of the anode resistor ? Did you try my suggestions on it ?

PM me, maybe I can get you a good tube.

I replaced the anode resistor with a 220K 1w
I added in parallel with R8 a 150K

Measurements on the spinner kit

No tube connected each side of 220k 486/475 respectively

Tube connected 480/296 (so quite a drop)

From what you've posted I am assuming that current is the enemy here?

PM on the way about tubes etc If the 6909 is not 'buggered' it would be nice to get it going somehow (maybe you direct connected spinner circuit) as it does seem quite a rare thing. How long it lasts of course remains to be seen

Okay, your meter has ~10M resistance. Pretty good.



~840uA current. That's a good value for that tube.
296V - 60V = 236V, which almost exactly the nominal anode-cathode voltage given in the data sheet.



Dinner is fighting back. I'll get back to you after the Peptol Bismol kicks in.

It's Fluke



So, with the tube still stalled, apart from a random flick to another point, and assuming it gets enough current from the spinner, do I assume it's a sick old soul?

Looks like that 6909 is giving up the ghost. Its kinda a rare event actually seeing a tube in the process of dying. These are either good, or have no gas, and are bad. A member of the nixie group, in Hawaii, actually witnessed one working, when first installed, then watched it become non-operational right before his eyes. I recently picked up a few Russian OG-3 "purple" dekatrons. I tested one in the "kit", so I could verify the jumpering, for that tube. I put it away. A few weeks later, I plugged it in, and no glow. First, I suspected that I had blown my circuit, but a quick check with a meter, showed the electronics to be good. Grabbed another OG-3, and things were spinning. That original tube, just died.

There's a general consensus, of the few of us that like playing with these tubes, that the high speed versions just don't hold up well. Relating it to what the ARF people know, its like comparing electrolytic caps to mica caps. In time, its almost guaranteed that an aluminum electrolytic is bad. Mica hold up pretty good over time. Some hi-speeds can be compared to paper caps. They hold up a little longer, than E caps, but not too much longer. When looking for dekatrons, try to stick to the lower speed neon filled types. The ones that glow orange.

BTW: The dekatron is a British invention. Ericsson of the UK came up with the first of these tubes. They also had the trademark for "dekatron". That's why you see Brits use the decatron (with a "c") for generic tubes, and the "k" spelling reserved for those that are Ericsson (ETL, or Etelco) branded. Elsewhere, they are called "decade counter tubes", or "glow transfer tubes". The GC10/A is the very first production model, and pretty hard to find. The GC10B and GC10/4B quickly succeeded it, and upto a few years ago, could still be commonly found on eBay. In 2007, you could still pick on up for under $10; $5 fairly common. Also, the GC10/A, GC10B, & GC10/4B (as well as the Phillips [Mullard, Amperex, Valvo] Z303C) can plug right into a socket wired for a 6802.

Here are some dekatron links:

http://www.tube-tester.com/sites/nixie/trade10-dekatron.htm
http://www.decadecounter.com/vta/tubepage.php?item=15&user=0

On that second link (decadecounter.com) beware of the photo colors. They used a special camera that goes beyond the human color range. The result is that some colors get shifted. It makes for pretty pictures, but the colors are not accurate.

There were also some competing technologies:

Burroughs came out with a "cyclotron-ish" (some people say magnetron) type tube that worked in a vacuum. The count pulse, would create electrostatic forces, momentarily countering the magnetic field. This would step the electron beam to the next state. Since this occurred in a vacuum, there was no ionization nor de-ionization time. These tubes could count up to 2MHz. Ericsson also made tubes of this nature, and they called theirs trochotrons.

http://www.radiomuseum.org/forum/the_trochotron_beam_switch_counter.html

Phillips created their own special counter tube, working on static deflection, and specially shaped shielding windows. The tube was the E1T, which like the dekatron could both count, and work as the actual display. Ronald Dekker, here, actually works for Phillips, and got to interview some of the people who made this tube:

http://www.dos4ever.com/E1T/E1T.html
http://www.dos4ever.com/trochotron/TROCH.html

The dekatron, being the simplest, seemed to out live these others. Some of the last ones could count upto 1MHz. This, however, was too little, too late. By the late 60's a 7490 semiconductor TTL counter could count at 30MHz, and only needed 5V. Plus by ~1970, they could be had for only a buck.

"That original tube, just died."

Static damaged, no doubt. ThreeNeurons, you've gotta be more careful!

Must have a poor seal? These are gas filled without filaments. They will fail if neon gas gets out.

I read that the 6909 is filled with Argon...

I also wondered if the gas had leaked surely it would not light up at all? I can get it to glow on certain parts of the tube, but it just won't spin grrrr.

To make sure we're all on the same page, the original post referred to the 0A3 voltage regulator, but recent responses have shifted to dekatron counting tubes. The 6909, in particular, and its cousin the 6802.

Dekatrons are also cold cathode. The 6802 contains neon, but the 6909 has a different gas. Some believe the gas is argon, but others suspect its either hydrogen or helium. All these gases are at fractional atmospheric pressure. Somewhere in the ballpark of 50 torr (?), where normal sea level atmospheric pressure is 760 torr. The big puzzle is that the neon filled 6802, and like neon filled tubes, seem almost always to work. They never seem to have a seal problem. Whereas, the high-speeds, have a very high dud rate. And there's a hierarchy to it. English tubes like the GC10D and GS10D are almost always guaranteed to be bad. Tubes like the Sylvania and Raytheon 6909, 6910, and Phillips Z505S, are bad about half of the time. Russian tubes seem to have the lowest dud rate, but a discussion with other members, at the nixie group, revealed that they're showing problems too. If there's JUST a gas leak issue, it should be evenly distributed between neon and high speed types. The gas comes into play somehow. If the gas is helium, it can escape right thru the glass wall. If the working pressure is significantly different, then that can also effect life. For example, maybe the neon, is operated at 50 torr, while the hi-speed mix needs to be down at 10 torr. If there's simple structural leakage, then it will effect the lower pressure tube first.

On that note, if the pressure is going up, then strike voltage, should also rise. On that hunch, increase the guide pulse. Right now it goes from ~+60V to -60V, relative to the cathodes. If R16 (33K) is raised to 68K or even 100K, the bias would be shifted, so the negative stroke be made larger. -80V, for 68K and -100V with 100K.

If they work the same way as ordinary neon lamps, the gas doesn't "leak out" but gets occluded into the metal electrodes, so the gas pressure drops.

Perhaps an adventurous sort could attempt to heat it back out??

Don't laugh, I have a couple of friends that are some serious "coilers", as in Tesla coil. They have punched holes in the glass panes they've used for cap dielectrics. From what I hear, that's not so uncommon. I try not to stand too close to their work. I get in between electricity and where it wants to go enough, already.

According to a Russian reference:Lipkin, V. M. Dekatrony i ikh primenenie. Moscow-Leningrad, 1960

Nominal gas pressure in the range of 30 to 40 mm Hg (1mm Hg = 1 torr). Didn't specify for which gas mix.

For anyone still following this

I did the mods, and had differing results. With the 2.2uf cap the charge cycle was too slow, and the tube stepped very slowly and the adjustment with R18 was almost impossible and resulted in the tube stalled.

I tried a whole range of caps, and found that .2uf worked well, as did .15uf. I still found that R18 had a narrow range of adjustment and at either end of its range the tube stalled, either because the oscillation was too higher frequency for the tube, or it was too slow and stopped.

The brightness of the OA3 was only good in total darkness, so although it works, I wonder if I'll bother when I finally mount it all.

When it was running uber fast it looked very cool