Dekatron clock,

[Benoit Tourret]

Good morning,
I'm still with my idea of making a dekatron clock.

I found some diagrams at http://959radio.co.uk/dekatron.html

but here is my second question:
in the diagram http://959radio.co.uk/Sec_Counter.pdf there are two A-101 dekatrons to divide the frequency of 1 Hz into seconds.
each dekatron will pilot its own nixie tube to display the seconds. it's via an MPSA42 transistor. one per used cathode.
the first decatron will increment the second dekatron each time the cathode "K0" shines, to increment the tens.

here we have a vacuum triode, but I want to replace it with a transistor or perhaps a thyratron; it's a clock, not central heating...
how would you handle this?

another question, what is x (marked: To other modules)?

and while we're at it, can you just explain to me how the transition from 59 to 00 works, via I think Q11 and Q17. but how? and why is G2@DEK2 involved?
And also if there is (as I think) a small error on the DEK2 Bus, which should be marked 6 and not 10, and the MPSA42 transistors should be numbered from Q12 to Q17.
the compute module for the seconds and minutes being exactly the same, we will stop there.

P.S. If you think to have a better diagram, don't hesitate.

Regards,
Benoit.

[Dekatron42]

You can contact the designer of that clock, he answered my questions manyvyears ago.

You can also have a look here: https://threeneurons.wordpress.com/dekatron-stuff/ for dekatron driving and explanationd.

/Martin

[Mike Mitchell]

I built a dekatron clock based on https://959radio.co.uk/dekatron.html as shown here:  https://www.youtube.com/watch?v=AeS3xuuRBV4

There is a link from that video to a dropbox with photos of the build, including a photo of a hand-drawn schematic.

I used a STQ3N45K3 N-channel Mosfet as an inter-stage coupler instead of a vacuum triode.

[Benoit Tourret]

Hello Mike,

Thank you that is exactly what I wanted to do :-)

How did you made the A-101 sockets ?

[Mike Mitchell]

I just used Molex socket pins, no actual sockets.  I used wire labels to designate the pin.  You could 3D-print a holder for the socket pins, but I never did.

[Mac Doktor]

On Sep 22, 2023, at 2:44 PM, Benoit Tourret <benoit....@gmail.com> wrote:

How did you made the A-101 sockets ?

Here's a complete guide to the method. It includes photos (scroll down a couple of pages):

https://threeneurons.wordpress.com/arduino/dekatron-shield-kit/

There's a diagram that you can use to get the pattern for the holes:

I built this kit. Don't make the mistake of sliding the pins all the way in the "tubes". It was a nightmare getting the Dekatron back out after soldering the female Molex connectors in place. I had to get a friend to do it for me. I would have destroyed it.

Terry Bowman, KA4HJH
"The Mac Doctor"

https://www.astarcloseup.com

Edward R. Murrow: “Who owns the patent on this vaccine?”
Dr. Jonas Salk: “Well, the people, I would say. There is no patent. Could you patent the sun?”—See It Now, 12 April 1955

[Benoit Tourret]

Thanks Mike and Terry.

one of my friends cleaned up the plan a little, it's already easier to read.

I will see witch method I will use for connect the A-101.

it will mainly depend on witch tube / socket I will use for the nixies...

Regards,

Benoit.

[Benoit Tourret]

Trying to make something more ... modern ???

it is the left part of the schematics. without the resistors on the A-101 ... I need to work on it.

I am not sure that this project is the best to begin with kicad...

[Benoit Tourret]

I tried to better understand mike's drawing.

what is exactly the purpose of the Dekatron 1 (in red)

I see G1 and G2 receiving the AC voltage. I guess it is 60 Hz?

I just have to use the pin K5 for 50 Hz ?

K1 is to reset the clock

K0 should activate the mosfet driving the second tube at 10 Hz 

the second tube divide by 10 to send seconds. and so on until the end of the day.  do not pull the clock in water after midnight !!! Gremlins inside !!!

the columns are driven by tube 2 ?

I am using 240 V, how should I adapt the power supply part ?

I see 500V on the power part and 475V on the Dekatron anode, this the same ?

Well, I am far to be a Jedi...

Regards,

Benoit.

[Benoit Tourret]

Ah, also.... Are all the capacitors polarized ?

[Mike Mitchell]

There are eight dekatrons in the circuit.  The first (G1) divides the 60Hz signal by 6, giving a 10Hz signal.  To change it to support 50 Hz, move the emitter of the PNP transistor of the reset circuitry to the "5" cathode (K5) of the dekatron.  The furthest to the right "Witch Tube" (sic) note in the diagram refers to G1, the first divide-by-six dekatron.

The second "Witch Tube" note referencing K8, K9, and K5 are for manually resetting the seconds counter to 59.8 seconds.  K8 references G2, K9 references G3, and K5 references G4. Pressing the seconds reset button sends a 10Hz pulse to the indicated cathodes, setting them to 59.8. They'll then roll over to 00.0 after the button is released.

The high voltage target was at least 475v.  I'm using a voltage tripler with a nominal 120v RMS at 60 Hz supply, giving 120v * 1.414 * 3 = 509v assuming no diode losses.  I measure very close to 500v, due to the diode losses and the 22-ohm current limiting resistor.  It's a 5 watt fusible link resistor I had laying around.  The only polarized capacitors are in the power supply section, supplying +500v, +170v, -170v, and +40v.  The zener diode for the +40v supply is actually a 39v zener, as 40v ones are very hard to find.  A 39v supply is close enough to 40v for the dekatron.

The "reset counter 60Hz? or Witch Tube" (sic) note references G4, the tens-of-seconds counter.  When that counter reaches "6" it is reset back to zero.  The complex reset circuitry furthest to the left detects when G7 reads "3" and G8 reads "1" (hour 13), sending a reset pulse to K1 of G7 and K0 of G8, reseting them to hour 01.  The tens-of-hours dekatron spends a lot of time displaying "0", and I found that the K0 cathode would wear down.  To extend the life of the Dekatron I connected the odd cathodes to the NPN driver for the zero cathode of the Nixie tube, and all the even cathodes to the NPN driver for the one cathode of the Nixie tube.  Since all the even cathodes are connected together, sending a reset pulse moves the glow to a random even cathode.  Advancing the tens-of-hours from there goes to a random odd cathode.  Making the clock handle 24-hours instead of 12 hours is more difficult, as the simple even/odd grouping has to instead have three groups.  In that case I'd leave the K9 cathode open and make groupings of K0, K3, K6 for "zero", K1, K4, K7 for "one", and K2, K5, K8 for "two".  The reset connection to K3 on G7 would be moved to K4 on G7, the connection to K1 on G8 would move to K2 on G8, and the connection to K1 on G7 would move to K0 on G7.  That would detect "24" on the hours and reset to "00".  You'll also need another NPN driver for the "2" on the Nixie.

Converting from 120v to 240v isn't as straight forward.  You still need at least 475 volts for the Dekatrons, at least 170v for the Nixies, and a large negative voltage for the reset circuitry.  The 39 volts for the Dekatron gates can be derived with a zener from the Nixie supply.  First the voltage tripler has to be replaced with a doubler instead, giving 240v * 1.414 * 2 = 679v. The Nixie supply become 240v * 1.414 = 339v. You'll have to find higher-voltage capacitors for the supply filters and increase the anode resistors for the Dekatron and Nixies.  The transistors used are rated for 300v (MPSA42 and MPSA92) and 450v (STQ3N45K3). Note that the MOSFET part number is incorrect in the diagram, it's STQ3N45K3 and not STQ3N4K3.

All-in-all I think the easiest thing to do is use a travel transformer to change your supply from 240v to 120v, something like https://www.amazon.com/Watt-Step-Down-Converter-International/dp/B077CFPD1T/ref=sr_1_42

I did the same thing for a thyratron ring counter clock I built, the schematics were for a 240v 50Hz supply and I used a travel transformer to step up the US 120v 60Hz supply to 240v 60Hz.  I had to change a divide-by-five counter to a divide-by-six, but that was easy in comparison to modifying the circuit to use 120v instead of 240v.

Mike Mitchell

[Benoit Tourret]

Thanks a lot Mike.

sorry for my English, it's not my native language. I mixed Which and Witch... maybe a Halloween effect??? ;-)

I'm trying to imagine my card. I think I'll use 4 identical cards.

each board should have two dekatrons and two nixies, wired basically as if it were a divide by 100 board.
by changing the cathode wiring with jumper instead of resistors, I could use it for either hours, minutes, seconds, or even 50/60 Hz division.
since I'm on 50 Hz, it might be fun to double the 50 HZ to 100 HZ with a diode bridge and divide by 100, I might have something prettier to display...

all the time reset action, resetting the counter to 0 to 60 or 24 will be done on a smaller board plugged into the 4 main boards
it will be easier to set the time with an ntp clock, in the near future. 50 Hz is pretty accurate here in France, but I want the clock to at least self-adjust to winter/summer time.
an esp32 will do just fine. but I want the clock to calculate the time itself.

I have a handful of IN-4s, but I wonder if they will last... the QS30 is similar quality. the IN-12 is good, but won't match the dekatrons.
are the z560m mercury doped?
what are the top view nixie tubes available that last a long time... ???

Excluding the power supply,
for the resistors, will the classic 1/4w carbon ones do the trick?
and ceramic for the capacitors?

I should use smaller resistors and thru the board capacitors... it is just to get the idea.

but.... but ... the A-101 and the nixie tube cannot have the same position for the connectors, they don't have the same height. !!!!!

faire et défaire c'est toujours travailler !!!

Regards,

Benoit.

[Mike Mitchell]

I originally used QS30 Nixies, but soon replaced them with Z5600M.  I've been running them continuously for several years without a problem.

I used 1/4 watt through hole resistors with a 300v rating.  I used 1/2 watt for the dekatron anode resistor and the 4.7M power supply discharge resistors because I could get them with a higher voltage rating.

I used 1N4006 diodes, 1N4007 will work too.  The 47uf capacitors in the the high voltage supply are rated at 250v, the 33uf capacitors in the high voltage supply are rated at 450v.  The 47uf capacitor in the 40v supply (actually 39 volts) is rated at 63v.  The other capacitors are ceramic, my notes say 470nf @ 100v; 47nf @ 250v; 22nf @ 630v; 1.5nf @ 630v.

Mike Mitchell

[Benoit Tourret]

Great, thanks a lot Mike !!!