The main counter boards (8 in total) are all identical. I've included the schematic for that board. It it essentially 4 master/slave flip/flops connected together with some other gating logic functions to enable resets and modified counts. For any board that counts to 6 or less, unused chips are jumped out and bodge wires are added to bridge over any unused circuits. 

These boards take a clean 60Hz in from the power supply board and divide it down. Starting with the lowest counter board dividing by 10, then the next one up by 6, and so on. The last two boards at the top are modified to count to 23 and then reset. 

Each board has a white power LED and a group of 2, 3 or 4 blue LEDs to indicate which bits are active for each counter. The BCD outputs from each board are brought out to 0.1" headers and drive the decoder boards.

The decoder boards use NAND gates wired as inverters to buffer the incoming signal and invert each line. This allows me to use 4 input NAND gates (7420) to decode each of the 10 outputs. In turn, the outputs of these NAND gates are then inverted by more 7400 gates and then drive MPSA42 high voltage NPN transistors. These transistors drive each cathode for the respective NIXIE tube. Much like the counter boards, all the decoders are identical and any chips or circuits not required are jumpered out. There are 6 driver boards in total; one for each NIXIE tube.

The power supply is over complicated and in need of a redesign. The time setting circuits are located here and are the part of this project that still doesn't work properly. Aside from that, the clock runs on 120VAC and uses a compact switchmode power block to generate the main 5V rail for all the logic. A small transformer generated 12VDC which is used to provide isolated power to a step-up converter to provide the 170V for the tubes. Additionally, there is a circuit with two 18k power resistors driving an opto0coupler directly from the AC line. This is then cleaned up and used to generate the 60Hz square wave for the clock time base. The original time setting circuit was created with op-amps and a 7400 to create a schmitt trigger of sorts. This idea failed and does not work. I am experimenting with swapping out the entire circuit for a quad NAND schmitt trigger IC (74132). However, I am still experimenting as I'd like to keep all the parts on this build of the same vintage logic wherever possible.

Once the time setting circuits are fixed, a case will be built and this project will finally be complete!