Dekatron for Demo Display

[peter bunge]

Having finished the Amperex 8453 display I started on my Dekatron. Thanks to all for the suggestions.

I read up on Ronald Dekker's clock project https://www.dos4ever.com/decatron/decatronweb.html

 and decided against a direct drive from a 74141 because, even if it was practical, it would not look as good as using all the pins. It would look like the 8453 I just finished except without the number mask.

A few quick experiments showed that a -24v power supply was needed for the easiest implementation.. This is my design. I will let you know if it works.

A PIC drives the circuit: it starts with Q3 off to force a start at 1. A high on R3 or R4 is the same as the switches in Ron's test circuit placing -24v on the guides. As my PIC sends the BCD for the other displays and clocks the E1T it will generate the sequence to advance or retard this Dekatron..

[Jon]

Interesting approach - not seen it rigged up quite like that. Let us know how it goes!

My immediate question is whether there's a big enough potential difference between an inactive guide and a main cathode to get a reliable transfer forward from a deactivating G2 to the 'next' main cathode rather than back to the adjacent recently used G1 - 5V is much lower than the datasheet guide bias. Might be OK at slow stepping speeds with long guide pulses. Also the leading edge of your guide pulses is going to be fairly slow as Q1/2 come out of saturation and the guides are passively pulled down to the 'active' voltage. Most guide drive circuits use a NPN pull-down to the active state which creates a sharp leading edge and then a slower return to the inactive state.

Jon.

[peter bunge]

You are absolutely right. 5v was not high enough despite Ronald's and my confirmation that it would work.

I am changing to +25v and -25v and will have an optocoupler with the diode driven from the 5v PIC and the transistor at the bottom of the resistor between the two power supplies. It keeps it simple. I have to order the optocoupler because none of my old ones have higher than 30 v rating.

I will have to think about the extra two power supplies. Maybe I can't avoid them. It will be a few days until I get the parts.

Peter

--
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 view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/62979e37-ef21-46e1-9b7c-45a4c4080238n%40googlegroups.com.

[Dekatron42]

I'd raise the resistance to at least some 100k for the two bias resistors R5 & R6 in your diagram above.

I'd also use the correct bias voltage and just use an MPSA42 with its emitter to common ground for the driver to simplify the circuit as Ronald Dekker and Michael Moorrees with their dekatron circuits.

Different dekatrons need different bias and pulse voltages on the guide electrodes to count properly so accomodating for those requiremenst will remove a lot of problems and keep down the fault finding time.

/Martin

[Jon]

I'd had a similar thought about increasing the bias resistors - it's not necessary to run so much current (5mA) through that part of the circuit.

OK, so your proposed conditions are that the dekatron will see are 400V anode to main cathodes; guides swinging from +25V to -25V with respect to main cathodes. Tube current will be about 345uA.

I've replicated these on my dekatron tester using a nice NIB GS10C/S as the test subject and can confirm that they do work - the tube stepped fine up at speeds up to over 4kpps once I'd woken it up a bit.

Caveats:

1) I only tried one tube.

2) My circuit is rather different to yours so although the static voltages are the same, the pulse shapes are almost certainly different. And we didn't even talk yet about pulse durations, so I've no idea what your PIC is spitting out. But as long as you're not trying to cut things too fine, there's lots of latitude to find patterns that work. Keep things north of 100us per phase and you'll be fine unless you've got a really reluctant tube.

3) I was running at slightly lower current (300uA) so you've probably got a bit more margin for speed than I had. Recommended operating conditions are 325uA +/- 20%, so we're both inside that range. But a little more rather than less current is useful when pushing higher speeds. I'm guessing though that you're looking for a much slower stepping speed for this application though.

If you've not already done so, I'd definitely second Martin's recommendation to take a look at Michael Moorrees' dekatron work (he looks in here from time to time too). He used a couple of elegant design tricks which simplify the interfacing of dekatrons to modern electronics / microcontrollers - I've followed his approach in pretty much all the stuff I've built and it works a treat. Not to say that other approaches aren't equally useful too of course.

Jon.

[peter bunge]

Those values were left over from the 5v circuit. I was more concerned with the level shifting from the PIC at 0 to +5v.

Peter

You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/dQn3tFBYfoc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/75f258db-83ea-4df3-9d50-52dcaed1461en%40googlegroups.com.

[Dekatron42]

There is one more dimension to think about when running a dekatron and that is that over time it will degrade due to sputtering affecting the electrodes and to maintain a correct counting when it ages you should keep the voltages as described in the datasheet, especially the guide and bias voltages as those are needed when the dekatron nears its end of life - it is the same effect that is in play if you want to raise a dekatron from the dead if it has been heavily used or just stored for a very long time without use, raise the bias/pulse voltages and sometimes the anode voltage to surpass the effects of the sputtering - effectively increasing the current flow between the anode and the guides/cathodes.

So even if dekatrons work at other voltages you will see the effects of lower/incorrect voltages when they near their end of life, this has been described in older litterature where the internals of dekatrons has been disected in detail - you might just need some 10V below the glowing cathode to move the glow on a new dekatron but as it ages it will stop working at that voltage.

/Martin

[gregebert]

Martin - Do you know how long dekatrons last ? I have an A-101 running as a spinner in one of my clocks, and it's been going 60RPM  24/7 for almost 8 years with no signs of degradation. I use 30k dropping resistors at the cathodes, and the current is right at the spec value of 450uA, so that gives about 15V for "steering", and the driving waveform is 3-phase with overlap. I

[Dekatron42]

No, unfortunately not, no hard figures - maybe Jon has. I've only read in some books about dekatron construction (neon tube construction in general where dekatrons are shown as special variations) that they have similar lifespans compared to small neon lamps but that the complex design complicates the failure modes/rate and also that keeping currents/voltages within the values specified in the datasheet will guarantee that they work as long as possible.

It depends on when you say a dekatron fails, it can get a silvery finish on the glass so you almost can't see the glow but it still works and it can fail a lot earlier than that due to internal flash-overs due to sputtering onto the ceramic material which means a current can flow in this sputtered material as the distance is shorter there compared to other paths (I have a few dekatrons with that failure and of course a few with a silvery look).

I have to count my dekatrons one day to see if I have enough to build a second Harwell WITCH..... ;) :)

/Martin

[Jon]

No hard figures - it's dependent on a bunch of factors; principally manufacturing quality, how you use the tube, operating temperature and gas fill. If you keep the operating current within spec and preferably at the lower end, avoid the tube getting heated from its surroundings and above all else keep the glow moving around all of the electrodes, they will last a seriously long time. The electrical properties will change gradually - the maintaining voltage rises and the latitude around the voltages required for reliable stepping decreases. So you were absolutely right to point out the importance of taking note of these elements of the spec when designing - you can get away with a lot on a NIB tube, but progressively less with age. Of course if the application requires clear glass to see the lovely glow, then the tube may reach end of its useful life in that application as sputtering gradually obscures the view, and that may happen well before it ceases to step reliably.

I'm not surprised by gregebert's description of his A101 - his application sounds like a great recipe for a long dekatron life. The slow speed ones are tough as anything and they love to work! Just for context, I believe that the large majority of dekatrons working in the stores of the WITCH today are from the original complement of tubes the machine was built with at the start of the 1950s. Can't formally prove it of course, but the date codes are consistent with that. Also, as we've noted before on other Russian glow tubes, the guaranteed life spans on the datasheets are ridiculously conservative if the tubes are treated well. The A101 datasheet gives a 1000 hour life - his A101 has done 70x that! I have a similar experience with my IN9 clock - the prototype unit is still going strong on its original tubes after 120K hours - datasheet life is 1K hours.

Note that we're talking here about use-related life limitation - the death in storage of the high-speed dekatrons is a whole different topic.

Jon.

[peter bunge]

I tried my new circuit with the optocouplers using 270k for R5 & R6 and 1k for R3 & R4.

It works within the limits of +0v and about -22v. It is interesting that the other circuit did not work for low + bias. It must be the slower positive edge as someone mentioned. I will use about +30 and -30v, whatever the rectified transformer gives.

This is my preliminary software. It only counts up at present from 0 to 9 fast (10Hz) slowing down to 1Hz then repeating.  The delay between the steering pulses depends on the speed so at 1 Hz you can actually see every pin lit. I thought it a waste to not see them. I have not fully tried this software, only a Dekatron test version, so if there are errors please forgive me. There is a BCD version that also works. I may add a "Spinner" between 0 to 9 counts in place of the 2 second delay, 1 second forwards and 1 second backwards.

 // main loop

   while (1) // loop endlessly
   {
 for(j=1;j<=10;j++)                // sets the delays
{
     for(i=0;i<=9;i++)             // the BCD digit displayed (Nixies, etc)
     {
// generate clock and BCD
      output_high(pin_C4);        // high for 10 uS
 delay_us(10);
      output_low (pin_C4);        // end pulse
      portC = i;                        // output BCD
// Dekatron clocking
    output_high(pin_A5);    // steering ring 1
delay_ms(j*50);
    output_high(pin_A4);   // steering ring 2
delay_ms(j*50);
    output_low (pin_A5);
delay_ms(j*50);
    output_low (pin_A4);
// period of sequence
delay_ms(j*100);                  // set period
     }                                  //end for i loop

delay_ms(2000);       // wait 2 seconds after each 0 to 9 count
    }                               //end for j loop
   }                           //end of endless while loop
}                        // end of main function

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/bcaa0282-1c44-4b37-9f6f-98845208efecn%40googlegroups.com.

[Jon]

Good to hear - another successful neon spinny thing!

Jon.

[peter bunge]

I tried another Dekatron but I'm not sure what it is, and it doesn't work. The glow just flickers back and forth on adjacent elements.

This tube looks a bit like the GS10C/S but has no markings, has a metal base, and the end is domed instead of flat. It fits the same socket and appears to have guide rings.

My high voltage optocouplers don't arrive until this afternoon and I have already tried +/- 25v which is 30v above the rating of the optocoupler I am using. They have not blown, yet. 

I am trying to understand the data sheet for the GS10C/S, looking at recommended operating conditions: 

What does double pulse drive-amplitude mean?

What does Integrated pulse drive amplitude mean?

I just finished re-reading Ronald Dekker's description of how it works and his experiments.

Peter

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/3cc6ab77-f4c7-4ae9-8d65-6a60653b7895n%40googlegroups.com.

[peter bunge]

I found a reference:

http://www.r-type.org/articles/art-132.htm

that explains the difference. So I am using double pulse rather than an integrated single pulse to get the delay.

The data sheet implies I should be using -80v pulses to drive the rings. I guess I will up the drive voltage when my optocouplers arrive. I ordered 80v ones, perhaps I should have ordered the 300v ones. Oh well, another $8 shipping if I must.

Some useful parts from DigiKey:

MOC8050M-ND optocoupler 80v

SFH619A-ND optocoupler 300v

HM4682-ND Hammond 186C120 transformer to isolate my CRT clock that is live to line. It has dual primaries and I will see if I can use it backwards to get two 120v outputs isolated from line (which will only work for 120v power in Canada/USA)

497-2344-5-ND  ULN2003A 7 segment driver for Vacuum Fluorescent displays, stepper motors, relays. Observe voltage ratings. There is an 8 transistor version as well.

Peter

[Jon]

Yes, the double pulse situation is the relevant one here - the other two on the datasheet refer to specific valve circuit configurations that were used at that time to generate the guide pulses. Modern components make the job much easier in most ways. The datasheet recommended conditions call for guide pulses of 80V +/- 10V amplitude, negative going, from a baseline of +36V (the guide bias voltage). A swing from +40V to -40V will do the trick just perfectly.

As discussed earlier in the thread, there is quite some latitude around these conditions, particularly for 'new' or lightly used tubes. In my experiment a +25V to -25V swing worked fine, even up to the full spec counting speed once it had woken up a bit from being asleep for decades. But that was a NIB tube. I suspect that your other dekatron has seen more use and/or is more deeply asleep, so for that one you'll maybe have to get closer to the datasheet spec. And maybe start with longer pulses. I'll PM you on a couple of additional details.

Jon.

[Jon]

>No, unfortunately not, no hard figures - maybe Jon has.

I've done a bit more digging in the literature on this topic, and found basis for an operating life exceeding 100K hours for the ETL low speed tubes if all the various pieces of advice on how to keep the tubes happy are followed. End of life is application dependent as we've discussed - there's no cliff-edge failure, it's really when the progressive aging of the tube parameters becomes an issue for the specific circuit it's being used in, and the stepping starts to fail. A typical spinner where the tube is just whirling away at fairly low speed with long guide pulses of the specified amplitude and timing may pretty much go on indefinitely.

Jon.