Picture of simple user-adjustable DIY Nixie Clock

As first instructable of the year I managed to finally complete my age-long nixie clock project.
Nixies are neon valve tubes, where ten cathodes have shape of digits and are lighted up by plasma when high voltage flows through them. I love these old era displays, which have been employed in last century before I was born.
In last year I've been slowly collecting components and knowledge to build some nixie clocks as Max Pierson's beautiful creation, I like the old style, the roundness of glass tubes, the rough wood case, the simplicity of the design. That clock has definitely inspired my project. Even though I really love vertical digits arrangement I keep that original feature for my next clock.

Therefore this first born is a six digits horizontal wood desk clock, with six big round Russian IN-4 nixie tubes, no dots, no visible buttons, no LED illumination, only a big massive rosewood block and the power of plasma ;-)
I have to explain you what the title means:

simple because it can be entirely built with common tools and from common components, you only have to order six IN-4 nixies and one nixie driver

user-adjustable because it's predisposed for many external sensors and additional features (as neon dots between digits, alarm, etc.)

DIY since you neither have to buy external shields or to pay for pcb manufacturing, just follow my instructable ;-)

WARNING: this circuit raises the voltage to deadly 300V so you must avoid to touch contacts while working, I'm not kidding, please BE CAREFUL!

Remove these adsRemove these ads by Signing Up
1-40 of 82Next »
jgluch14 days ago

Great writeup! I'm looking into making a clock and trying to start sourcing components. I was just wondering what the dimensions are on your PCBs. I think I'd like to try the photoresist method as this is my first time etching a PCB. Any other advice you have would be great. Thanks!

andrea biffi (author)  jgluch14 days ago

photoresist would probably work, but this is a very difficult pcb as first try..

you can read dimensions in step2 images. Good luck!

lawrence.loblevyt made it!1 month ago

Hey! Been meaning to have a go at one of these for a while, and have been doing my homework with a bit of help from your Instructables page. Honestly, its a fantastic design, simple yet functional, I like the way you have done away with the multiple drivers, even if it sacrifices a small amount of brightness.

Managed to get your schematic working as a base, and plan to add in either some diodes to combat the slight ghosting issues or some extra transistors to completely shut the signal off between the multiplexing signals. That and I will add in a modifiable relay circuit to attach a motion sensor to, or a simple button to display the time for a pre-determined time.

One issue i'm running into before i do that however is that the time every now and again resets itself to default with no interruption in the display. I've been puzzling over this all week and have narrowed it down to what i can think are the two issues, either the crystal oscillator has the wrong load capacitance (i'm using 16pF for the moment until the 22pF ones come through) although this doesn't seem to be likely. The other might be the N channel mosfet, as when i touch the rearmost plate, it naturally conducts but interferes with the atmega, making the display flicker and then freezing it.

I was wondering whether you ran into these issues, or if you might have an idea on what was going on?

andrea biffi (author)  lawrence.loblevyt1 month ago

looks great!

nope, I never had your issues, but it would probably fix when you put everything on a pcb and solder the components. The Atmega circuit part is a good old verified schematic, it should work well.

If you wish to add an external sensor or a pushbutton, my pcb design already provide connections for it, just on the edge near the atmega.

In my pcb I already managed to solve any ghosting problem, with no interruptions in the signal between multiplexing since there is no time to insert that with only one single driver, and now it works perfect! My next nixie clock is on the way!


I took your advice and stuck it all to the PCB. However have run into some problems again. The clock cycles fine, but only displays the digits 0 and 4 with all other blank, I've checked and there is a voltage across the 10 pins so i don't know what would be the cause of this. I suspect a burnt transistor from the soldering maybe?

andrea biffi (author)  lawrence.loblevyt18 days ago

ouch that's probably discouraging. I know that my pcb design is very compact and has very thin traces, so it's very difficult to make "at home".

usually a transistor doesn't burn for the soldering...I suggest you to look for short-circuits between adjacent traces, and also check for interrupted ones.

I went through and checked all of the traces for continuity, and to see if there where any branches. Also got the oscilloscope up and running and was able to see that the signal through the transistors was working perfectly, this leaves two options, the driver is faulty (i replaced it with another so i'm certain its not that) or the at mega has been fried. I've gone with the latter, so ordered a new one.

Running the oscilloscope was really cool and made possible to see the multiplexing signals clearly! I'll try and get a picture tomorrow to show you if your interested? Wish me luck with the new atmega!

And thanks for all the feedback, it's really refreshing to see someone with so much enthusiasm for his projects. I look forwards to seeing what new tricks you can cook up! Makes me wish i had chosen electrical engineering over mechanical.

andrea biffi (author)  lawrence.loblevyt17 days ago

It seems that changing Atmega will solve the issue.

When I saw nixie clocks I wished so much make one of those that it took two years of documenting (I'm a civil engineer, never studied electronics or IT before) to assimilate the necessary skills and design my own circuit. I also made more than 12 different versions of the pcb to reach that final one.

I'll also try replacing the atmega with a fresh one, there is a chance I've shorted it meaning the digital pin outs are not firing. I tried replacing it with another atmega which had this problem and it displays digits 0,1,4 and 5 instead of just 0 and 4. I'll report back if this solves it. I'l try to get it on an oscilloscope as well, just to be sure.

richb7728 days ago

What a great looking clock! I just wish i could program enough (anything!) to add an alarm function and night dimmer/power save.

BRILLIANT work here. Love it!

andrea biffi (author)  richb7728 days ago

I would like too... maybe I will be able soon to update the code with those features.

longpcb28 days ago


I want to diy for me one nixie clock the same this project, but i don't know how to convert code file you posted to file .hex! So, can you help me, convert the code fife of this project to file .hex and send to me please? Thank you very much!

My email:

andrea biffi (author)  longpcb28 days ago

I'm sorry I'm not very skilled in programming, not yet ;-)

I'm used to upload the code with arduino, then place the IC on my pcb. Good luck, please post the method when you'll find it, that will be useful!

So, You used file.ino or file.hex for programming Microchip ATMega168?
andrea biffi (author)  longpcb28 days ago

just connect your Arduino to your pc via USB and program the Atmega168 in that way, then remove it from Arduino and place it on the pcb socket, you don't need the ino or hex file, Arduino software will create it for you and upload to Atmega. My code is attached to step 4, you can also modify it.

EdwinasD2 months ago

hi. i have a question. i see that there is neon dot connection, but i dont understant how it works. will it blink every second?
p.s. sorry for bad English.

andrea biffi (author)  EdwinasD2 months ago

yes, since the pcb is widely ready for further upgrades, I added the possibility to connect two neon dots between the three digits groups. The dot's will not blink, it would be annoying in the night IMHO...

akirk115 months ago

Andrea, GREAT writeup! I'm planning on building this on a breadboard rather than a pcb. I just wanted to clarify a few things before I start ordering parts. I'm going to run it on 9V input only, so I can do away with C9, C10, D3, the 12-35V/9V selector switch, and the L7809? I'm not seeing on the schematic where to pipe in the 9V battery for a backup. Pin 1 on the ATMEGA, listed as PC6_(RESET, has a reset lead coming off of it in the second schematic, but doesn't go anywhere on the first. On the first schematic, pin 12 coming off of the 74141N driver shows it goes to "CND" before ground. I'm still a noob to this, so what the heck is "CND"? Do you also have the updated schematic to use the third position of the on/off/on switch to use as a sleep mode? Thanks in advance!

andrea biffi (author)  akirk115 months ago

Hi! Actually I don't suggest you to make the circuit on a breadboard, remember you have to do with high voltage, so it's very dangerous, you don't want a wire to disconnect and touch your nose when you are looking the beautiful nixie glow closely!

The battery will work good, but you have to implement a code with an automatic sleep function. Anyway, if you look my second nixie clock ( you will find the updated schematic with 3rd position as sleeping mode, and also you will find a new pcb ready for battery connections. When you connect battery the power-socket disconnects the wall-plug line.

Good luck!

I'm not terribly worried about the electricity. I've been zapped with mains power, 100,000v high power ignition coils, and others I care not to mention. I like the way breadboards look when I put them together. I found your other schematic after I had submitted my first comment. I'll be using that one to build mine. Looking forward to seeing more code implementation that you alluded to in another comment.

andrea biffi (author)  akirk113 months ago

just remember the "hand in the pocket" rule when you touch the hv circuit.

I'm sorry that I have no time to implement a new code for now... I hope that someone here on instructables can work on that.

Nighter3D akirk115 months ago

Hello. i may not be the author, but i can answer your questions ^^

The L7809 block only serves to regulate excess voltage from a big supply to a suitable 9v. if you are using a 9V supply you can safely omit it except for C10 which serves to clean the power a bit...which you probably want.

Regarding the 9v Battery. it can be attached to just the 9V line. Although a slight warning is that the 9v will be depleted in record speed as the project does not seem to be capable of intelligently turn of the Nixie without a external power-source and 9v batteries are not designed to give the kind of power the nixie needs...atleast not for long.

I think you are misreading CND for GND (ground). the reason its double is cause one is a Symbol (which connects to all others symbols on a schematic) and the other is a pin.The reset goes to one of the 2 switches in the upper-right of the schematic with the driver and power-supply. Remember if you click on a image on instructable you will get a better view...if its bigger then the browser can show you can click on it again to download a full-resolution picture! ^^

akirk11 Nighter3D5 months ago

I thought C10 might serve a cleaning function. As far as the "CND" I was looking at the high res. Still wasn't quite high enough to make it out. I figured it was a ground, but I'm still pretty green at making electrical components. Thank you for the clarification.

emil.godjaev4 months ago

Thank you very much for this instruction. I already had started soldering and noticed few missing parts. First of all its diod D5. B.O.M. doesn't have it. Then, your logo is hiding a number of resistor. I just can't read it. Could you help me? Also I wanted to ask you a bit silly question: may I change capsitor C1 1uf 350V to 1uf 400V. Here, in my city, i just can't find for 350V.

Then is it important to resistors R17 and R18 to have 1% accuracy?

I got question about programming ATmega8. May I use LPT programmator and PonyProg programm for this?

For now thats all what I wanted to know. I hope you could reply me as soon as possible. Thank you again.

andrea biffi (author)  emil.godjaev3 months ago

I hope you can find more info in my other instructable here

yes you probably can change the capacitor, and accuracy of resistors is not so important, I don't know PonyProg I'm sorry

(removed by author or community request)
andrea biffi (author)  emil.godjaev3 months ago

Emil I suggest you to make exercise with a simpler circuit. Solder this one only when you are quite sure that every simpler circuit made by you work well. It's easy find mistakes in simple pcb, and you can learn how to improve your tecnique.

Nighter3D5 months ago

Ah i love nixie clocks. Working on a couple of tube clocks myself. Kinda doing a "1 of each". Got a working Nixie prototype (Direct driven through multiple drives and a few shift-registers), Made a design for a tiny Numitron clock (They are the most beginner friendly. writing a instructable on them) and am already expecting some VFD tubes (again to be given a instructable as i work on it. Seem to be very few off)

May i make a suggestion though?

Look up the MC34063. Its a very popular "jellybean" Switching power-regulator often used to design high-efficiency step-up/down power-supplies (also very popular for Nixie supplies!). the only issue i have with your design is namely the Linear regulator used to convert power to 9v. if planning to convert a large voltage like 35v to such a low power that is going to power tubes its kinda innefficient...scratch that HORRIBLY INEFFICIENT (25-30%). Its a shame to have made a highly efficient Nixie switching supply, but to have it fed by a linear supply that is horrible inefficient... also makes for ways to make your own efficient energy-boosters/converters. Heck you can even invert power for Vacuum tubes or make a VFD power-supply.

Personally i avoid the IN-4. The IN-4 and IN-1 are quite known for lacking the Mercury within its gas mixture that most other tubes have. this shortens the life a bit compared to some other Nixies i have around. IN-1 is know to live for months at most and the IN-4 is know to last a year or so. You can see if mercury is present due to a blue hue within the Nixie glow.

andrea biffi (author)  Nighter3D5 months ago

yes, particularly the IN-1 has a very short life, but if you implement a good code to make the clock turn on only when you need it, it will last for 100 years, I'm working on it.

Yeah many tubes can live long beyond their rating if done properly. Especially russian tubes of which the lifespan ratings are quite conservative(e.g. Worst case). a tube rated for 5000h can last for years and those with mercury like the IN-12 can last decades in a well-programmed clock. I pity those planning to use the clunky IN-1 tubes.

andrea biffi (author)  Nighter3D5 months ago

Right, but with my design you are not obliged to use a 35V psu, you can use a 12V one, or even use a 9V power source and bypass the 9V regulator. Good luck with your projects, we look forward to see them as instructables!

janisalnis6 months ago

You could add to improvement list: "Use optocouplers to switch anodes instead of many transistors." Saw this workin on one Ebay board.

andrea biffi (author)  janisalnis6 months ago
Yes I know that possibility, but it's not a real improvement, it's just another way to proceed, but it's much more expensive.
If you are interested you can find the schematic with the TLP627 optocoupler here.
mm0bef7 months ago

awesome looking clock. Been going to build a nixie clock for some time but couldn't decide on style. This is it. Just need to get someone to etch the boards for me and im good to go.



andrea biffi (author)  mm0bef7 months ago

great! I have some spare pcb (new version, see my other nixie clock) so contact me if you need one. I don't have other tube shields but those are much more simple to etch (or you can simply wire the tubes)

SpectrHz10 months ago

What a gorgeous looking clock! I absolutely need to make a project like this eventually :) I really like the wood you used, the nice warm orange glow against that dark wood looks amazing.


andrea biffi (author)  SpectrHz7 months ago

thanks! :)

salexandridis10 months ago


Yesterday I finished the project and I tested it. All nixies lit up apart from the first hour nixie. After some measurements, I found out that one MPSA42 transistor was burned out, so I will replace it soon.

Now, the problem is that when I powered the circuit again this morning I noticed that all nixies would display the digit 2 and the digit they are supposed to display (have a look at the attached photo). Keep in mind that I have already tried disconnecting the K155ID1 pin that controls digit 2 (which is pin 8). Digit 2 turned off on all tubes, but when it was time to display this digit, the nixie would go crazy and a fade "red-ish shadow" would appear around each digit. Do you have any idea about what might be wrong?

Thank you in advance,


andrea biffi (author)  salexandridis10 months ago

Ouch... that's very strange. I would say that there is an high voltage across all anodes when no signal goes to transistors. I found out that there is actually an interference between hv circuit and transistor, but the effects were not so evident in my clock. I designed a new pcb where voltage is constant on all anodes, I'll post it soon here.

ps: you forgot to attach the photo

1-40 of 82Next »