Version 1

This is my first nixie clock, based mostly on the circuit located at http://www.electricstuff.co.uk/nixclock.html.

Click here for the schematic diagram of my nixie clock.

I've modified some of the original design to solve a few problems and to suit some of my preferences. The clock was experimental until I had some experience. I first experimented with the idea of building an alternating display, using two B-5560 nixie tubes. The purpose was to display the hours, then minutes, then the cycle starts again. This idea was unsuccessful and omitted, which will be explained later.

The nixie tubes used in this clock are Burroughs B-5560 tubes, which I salvaged from a small broken HP frequency counter. Those nixie tubes are mounted upside down. Because of that, I used a separate board to hold the two nixie tubes and the switching transistors.

The original idea of the alternating display was to wire the collector ends of the 1 hour transistors to the 1 minute transistors and the 10 hour transistor (I did not use zero for the 10 hours at first) to a 10 minute transistor (only a digit 1 from the 10 hour). The emitter ends of the hour transistors were tied and separate from the tied emitter ends of the minute transistors. This meant I could feed in the signals to all the transistors as if all nixies were present, but one set of transistors' (hours or minutes) emitter ends would be grounded. Therefore, displaying only the hours or the minutes, if both sets of transistors were connected and grounded, the digits would overlap.

Anyway, I spent about a day wiring the board with the two nixies and the 26 (27 now) transistors. I used a ribbon computer cable to carry the signals from the controller board to the switching transistors. It was very difficult to wire the ribbon cable to the two boards, since most of the connections are thin and I was using a dull iron.

After the first board was completed and wired, I started on the controller board. I had problems with the hours not resetting correctly, so I used another 4017 IC instead, which will be explained later. I also used another 4017 IC for the 10 hour IC instead of the other 4013 half so I can display zero with my single digit hours (for example; 03:00, not 3:00 with the 10 hour nixie appearing dead).

I started on the power supply for both the nixies and controller board. This is the dangerous part of the overall design. A voltage doubler is used to provide HV to drive the nixies (115 - 120V version) and the HV is also reduced to a few volts for the ICs. The power supply is transformer-less, therefore some parts of the circuit can be hot. The two time set switches need insulation, such as my simple long-pole pushbutton switches salvaged from a VCR.

Now I'll explain mostly how the controller circuits work.

A half of a 4013 IC was used to buffer the 60Hz line. Four 4017 ICs were used to divide the 60Hz line signal to 1/60Hz, therefore you would have to wait 60 seconds for the fourth IC to give a signal to the 1 minute IC.

The other three 4017 ICs are used basically for what they were designed for, counting. Notice the 10 minute IC is wired differently because when it counts to the 7th output (digit 6), that output feeds a signal to the reset trigger, resetting that IC back to zero. So when the 10 minute counts to 6 (60 minutes), it resets to 0.

The 1 hour IC is quite tricky, since it does not count from 1 to 12. Ten, 11, 12 are not single digits. So after 9, the 1 hour IC can only display 0, 1, 2 then reset itself. The 10 hour IC was the other half of the 4013, and this IC controlled only the digit 1. When the 1 hour IC counts to 9 then zero, the zero gives a signal to the 4013 to trigger the 10 hour digit 1, therefore you see 10 instead of 0. As the 1 hour IC goes back to 1 after displaying zero, it continues counting with the 10 hour nixie displaying 1 for 11 and 12. Finally, when 13 is counted, the digit 3 is responsible to reset both the 1 hour and 10 hour IC. Resetting the 10 hour IC will turn off the digit 1, and the 1 hour IC will go back to 1 again and start counting. However, the tricky part is to get 13 to reset both ICs, and not 3. Unfortunately the parallel capacitor resistor in the original design didn't do their job for me here. I decided to use another 4017 IC. It was wired to do nothing on output 1, and when 3 is displayed, this IC goes to the second output, which also does nothing. Finally when three is back again (..12, 13). The IC goes to the third output, which is wired to the reset input of both hour ICs and itself. However, that IC didn't have enough signal to reset all three ICs (I'm guessing), so I used a switching transistor to solve the problem.

Now I had to build the alternating circuit. I built a solid state version using another 4017, which was unsuccessful. I used a relay with one side of the switch to ground, the closed side to the hours and the open side to minutes. Therefore when the relay is off, hours are displayed; when on, minutes are displayed. The relay version was successful, BUT I learned that after switching back and forth this would cause the time to reset. I'm not quite sure why, but I believe it's due to static or small voltage changes that caused the sensitive 4017s to reset.

It was fun experimenting with alternating displays, but it was time to switch to plan B. Old-fashioned time-viewing. I didn't want to redo the smaller board with the switching transistors and the two nixies to add two more for HH:MM. So I got another small board and added two more nixies for HH to put over the original nixies for MM. Also I unhooked the hours transistors from the minutes, and tied the emitter ends of both to ground as shown in the original design. OOPS! I unhooked the 1 minute transistors instead of the 1 hour transistors. Too late, that board is very difficult to wire, so I decided to rewire the controller board so I can feed signals for the 1 minute transistors in the 1 hour transistors, and the 1 hour signals to the 1 minute transistors on the smaller board, which is a bit easier. After fixing the "oops," I wired a new ribbon cable to the outputs of the switching transistors. Though it might seem a little unclever to feed negative high voltage through a ribbon cable, it worked. I got a few nicks in the counting, due to a few wiring shorts because of my dull iron. With the problems cleared up, I finally have a perfectly functional nixie clock. I decided this time to keep it that way without experimenting more.

Later I got a bit annoyed about the 10 hour setup, displaying only the digit 1 for 10, 11, and 12 without a 0 for the single digit hours. I threw in another 4017 in place of the 4013 and wired it practically the same. The 10 hour digit 1 was wired on the second output of the IC. After checking that the hours were counting properly, I added one more switching transistor and used the signal from the first output of that IC to display zero. The collector end of that transistor was connected to the digit zero on the 10 hour tube. That half of the 4013 was left unused.

I also added a 50K control (green) in series from the HV+ to the nixies' plates to control the nixies' brightness. Finally, my first nixie clock was completed.

When powering up, the clock will normally display incorrect time, overlapping digits, or even count incorrectly. This is normal because the 4017 ICs sometimes are in odd states. This can be fixed by holding down the fast set button for about 10 to 15 seconds until the clock functions correctly. After that, you can set the time.

Below are pictures of the version 1 nixie clock:

Version 1 Nixie Clock
Displayed Time: 05:13

Close-up Front View
Displayed Time: 12:47

Back View

Bottom View - LOTS of wiring!