A107 Dekatron

[threeneurons]

Went to the TRW swapmeet last Saturday, where I traded with Westdave, to get an A107 dekatron. These are in the same glass, with the same structure as the A108 (and A109), but have a different gas mix, different operation parameters, and the glow is barely visible, form the top. Here's a short video:

http://www.dailymotion.com/video/x146fbc_a107-dekatron-selector-demo_tech

I rebuilt the selector demo circuit, with 10 LEDs in a circle. One attached to each cathode, to show it in operation, because I had been forewarned that the glow would be almost invisible.

This time I left out the shunting resistors across the LEDs. In the A108 circuit I left them in, so that only the cathode that was selected would turn ON. But in reality, the two adjacent cathodes, also conduct some current. In this circuit, that is shown.

Also, here I use a neon relax oscillator with a discrete toggle flip-flop, to step the dekatron.

The parametric differences here, are: Supply voltage is 510V and anode current of 1.5mA, as opposed to 300V & 0.7mA, for the A108.The commutating step is a 120V, that's biased at 60V, so it steps down to -60V before resting at 60V. 

Enjoy !

[Dekatron42]

Nice!

 

Have you tried to adjust the anode current to get the glow to stay on one pin and not leak so much between pins?

 

I just tested a large bunch of A-101, A-102, A-103, A-106 and earlier I have tested a large bunch of A-107, A-108 & A-109 and I needed to adjust the anode current on every single one of these to get them to spin up to the maximum frequency in the datasheet, and also to get a minimum of leakage between cathodes. Some also needed to have the guide bias voltages adjusted within the range in the datasheet to run properly at high speed. The anode current on all of those I checked needed to be adjusted to a value close to the one in the datasheet, not one would run perfectly without this adjustment. The glow was also more uniform and even at low speeds when the adjustment had been done to get them to run properly at high speed, adjusting the anode current at low speed won't do as the high speed requirement is different and you'll need to find the sweetspot for each dekatron.

 

The high speed (A-106, A-107 & A-109) ones also changed how the glow looked when they were adjusted correctly, with the wrong anode current the glow was much more "on", or closer to the pins like in a GC10B dekatron, but with the correct anode current the glow was "between" the pins, further away from the pins - the glow extends a lot more from the pin itself with the correct anode current than it does with the wrong anode current, then it looks like there is no glow "on" the pin at all.

 

I was able to press four of the A-106 up to 150KHz with ease and two of these four specimens ran at 225KHz without a hickup, these two had a more blueish glow than the others which had a more purplish glow, the two sets were manufactured different years so that could make up for the difference in gas mixture (or it might just be old age as more than 50% of them were duds, even though some contained gas they would not strike nor count).

 

With the correct anode current there was less current leaking between the electrodes on both types. Tweaking the cathode resistors also resulted in less leakage, but some needed a higher value and some a lower value so there was no way of telling what would work best.

 

It seems like the Russian dekatrons (at least all I have) need more tweaking compared to the European/American types to me, I don't own any other types like those from Asia (Japan) so I can't say of those are like the Russian or the European/American when it comes to tweaking.

 

/Martin

[threeneurons]

I've seen the trimpots on the anode circuit of Russian high speeds. I figured it was primarily to adjust for maximum speed. hadn't thought about them for adjacent leakage. I added a 200K trimpot, and have managed to get the glow to cover only two at a time, but very consistently.

[threeneurons]

[John Rehwinkel]

> http://www.dailymotion.com/video/x146fbc_a107-dekatron-selector-demo_tech

I saw another video titled "Abba Zabba Inside a Pile of Tubes", and am trying to figure out what the large neon? tube is that seems to have the glow darken on alternate sides, mounted on a circular base divided into quadrants.

http://www.dailymotion.com/video/x13sg67_abba-zabba-inside-a-pile-of-tubes_tech

- John

[robin bussell]

I'm guessing the top view of a nixie with a sequence running that
illuminates the digit stack in that fashion.

Cheers,
Robin.

[threeneurons]

That's a WE346 with the paint removed from the glass. Its a gas trigger tube:

http://www.westernelectric.com/spec_sheets/346C.pdf

Here's the internal structure:

Its driven by a simplier circuit, that later used to run this neon flower:

http://www.dailymotion.com/video/x2fz6o_neon-flowers_creation

Basically, I change the duty cycle on each half of the tube. The WE346 has that anode rod, with a half circle cathode, and trigger electrode is identical to the cathode. So, I treat it like an additional cathode. On the flower, there's only two electrodes, so I drive that from an H-bridge. Each electrode alternates being an anode, then cathode.  The effect happens by sweeping the duty cycle of each electrode (in its cathode period) between 10% to 90%, over a period of ~2 seconds. A small AVR tiny12 uC is involved. Same with the WE346, but there I don't need an H-bridge, since I have the dedicated anode. Just two NPNs. One for each cathode.

[threeneurons]

Here's a better video of A107 circuit:

http://www.dailymotion.com/video/x148ak5_a107-selector-demo-2_tech

I did add a 200K resistor in the anode leg, and adjusted it so only 2 cathodes pass current, at any time. 

[Dekatron42]

Nice that you could adjust it to pass current on two cathodes.

 

To be able to build a nice clock with these, where each cathode drive a transistor which drives the Nixie, only one cathode should be struck at the same time. I did not do any tests on whether this was possible with the A-107 or A-109. I could only get the A-108 to strike one cathode at the same time with my quick tests. Maybe some other parameter needs to be adjusted for that to work, like guide bias voltage or other cathode circuit load parameters.

 

/Martin

[threeneurons]

You mean something like this:

http://www.dailymotion.com/threeneurons#video=x148ak5

To get rid of two adjacent transistors being on at the same time, make sure that only an "odd" or "even" cathode is selected alternately. After a lot of convoluted pondering, I went back to the something that looks like the original Russian drive circuit:

Get rid of all the isolating diodes, and connect the odd cathodes to one half, of the flip flop, and the even cathodes, to the other half. Insert the nixie driving transistors in each cathode leg. LEDs are also inserted to show the current paths.

[threeneurons]

Oops, Actual video:
http://www.dailymotion.com/video/x14ftg4_a107-nixie-driver_tech

[Dekatron42]

Ahh, wondered what you meant with the first video.

 

In my experiments I only got the A-108's to work with only one cathode on at each moment, the A-107's & A-109's would always strike two cathodes no matter how I adjusted the circuit.

 

I'll definately have to try your circuit and see if my A-107's and A-109's behave the same. I had an almost identical circuit (as per the datasheet) but with a different flip flop and without the LEDs and also designed for higher speeds with a faster switching flip-flop. I might have done something wrong in my circuit. I didn't use a darlington pair of the MPSA42 so that might have affected the whole circuit.

 

Thanks for showing your circuit!

 

/Martin

[Dekatron42]

You are a bit ahead of your time too, on the circuit it says that you designed it in September 2103! ;)

 

/Martin

[threeneurons]

Have to get my transporter fixed. There's a recall on this model, since it seems to also teleport you thru time, as well as space. At least it isn't stuck to sending me to New Mexico, in 1947, anymore !

[threeneurons]

Added a reset circuit, so it will count from 0 to 5, then reset to 0:

http://www.dailymotion.com/video/x14htr2_a107-divide-by-6-nixie-counter-driver_tech

Well, actually it transiently hit 6, but then quickly resets to 0, so you never see the 6

The reset is done by raising all cathodes well above ground, while the zero cathode remains the only one at ground, and therefore forcing the glow to it. This circuit only resets from even cathodes. 6 in this case. Then the odd cathodes are already off. A single MPSA42 disconnects cathodes 2 thru 8. If this was a real divide-by-6, 8 would not be connected, at all. K6 turns off that MPSA42, but not directly, even though its in the proper state. It has to actually drop to at least the value of the low flip-flop output. To ensure that this happens a non-inverting buffer made of two low voltage transistors (PNP, 2N3906 & NPN, 2N3904) are used. Resistor values are chosen so when K6 drops below 20V, the 2N3904, goes to its saturation level (~0.5V max). This will pull the base of the added MPSA42, well below its emitter, guaranteeing that its turned off, disconnecting all the even cathodes, except 0.

I Think the 12 hour, hours section can be done only needing one dekatron, and the addition of a flip-flop. So 5 A107 dekatrons (or 108s, or 109s) are needed to do a 6-digit clock. This circuit will be very similar to Mike Harrison's circuit, but with the dekatron flip-flop groups replacing the 4017 CMOS chips:

http://www.electricstuff.co.uk/nixclock.html

[Tidak Ada]

Hi, could it be possible to add an short delay in the reset to lighten the 6th cathode?

 

eric   

 

---

From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of threeneurons
Sent: dinsdag 10 september 2013 11:14
To: neoni...@googlegroups.com
Subject: [neonixie-l] Re: A107 Dekatron

--
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 an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/3f3b07d0-7ab6-4c94-86c9-262872a81383%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Dekatron42]

This was indeed a nice way of solving the reset/set circuit.

 

If I don't misunderstand how the circuit works you can add one transistor between each pair of cathodes on both even and odd branches, and if you also add the possibility to reset/set the flip/flop, it would be possible to set the dekatron to any number as well as have it reset to any number you wish. You could then have it count from any number up to any number you wish.

 

It will be an awful lot of transistors to do this so the need for a dekatron might soon go away if you compare with a transistorized counter, but it is more fun to tinker with dekatrons than transistors in my opinion!

 

/Martin

[threeneurons]

Yes, though for the purpose I described, its working just as it should. This is how the asynchronous reset works on most modern logic chips. If you want a 4017 to count from 0 to 5 (Q0 to Q5), you connect the reset to output "Q6".  But if you stick a small cap across that 10K resistor, that's across the BE junction of the transistor generating K6, it should glow momentarily. 

[threeneurons]

This thing can easily become quite convoluted. This dekatron came out, when transistors were readily available, so it was made to be used with them. But with 10 common SCRs, and a few descrete components, you could make a functional, divide-by-10 nixie driving ring counter. I think there even is such a schematic, in the old GE SCR or Transistor Manual. I have both books. And with a ring counter, the count length is equal to the number of stage. Though using this little dekatron is much fun.

[Dekatron42]

I agree that dekatrons are fun, and these dekatrons glows so nice, especially these ones with a little different gas mixture that makes them glow purple/blueish.

 

There are many nice ring counters, the strangest one I have come across uses ten (10) pin diodes in series with nothing else between the diodes, but with a small driving circuit around them making them work as a ring counter, the number of diodes in series determines the number of counting steps (the diodes might have been called something else, can't check right now as the book they were in is packed down).

 

/Martin