Stereodude
Not really a
So who here has a Nixie clock?
I realize I'm way late to the party, but I kind of want one.
I realize I'm way late to the party, but I kind of want one.
Nixie Clocks
- Thread starter Stereodude
- Start date
ddrueding
Fixture
Way late? I had to Google it. And it looks like a nightmare to clean.
Stereodude
Not really a
I'm not sure I follow you on the cleaning part. You'd have to dust it like anything else.
jtr1962
Storage? I am Storage!
I've known about them for ages. I easily have the capability to design and make one, including programming the microcontroller. Right now it's on my bucket list. I would probably go all out and make one with either a surplus rubidium oscillator or a surplus OXCO adjusted using my rubidium oscillator as the timebase. The latter makes more sense as the bulb in the rubidium standards has a limited life. While not as accurate as a rubidium oscillator, a good OCXO will hold its frequency to within a few parts per billion. In terms of time, that means gaining or losing under 1/10th of a second per year. Good enough for me.
Stereodude
Not really a
A lot of them seem to use an Arduino in which case it seems like it shouldn't be too hard to connect a ESP8266 and simply have it use NTP (assuming there's enough unused flash and RAM left in the Atmel chip).
The Nixie tubes are rather expensive now too. A NOS set of 6 IN-18's are about $300.
Last edited:
I don't have one but I've followed various website describing projects with them. There is a person who was working on an alternative due to the availability and cost issues of genuine Nixie tubes called Lixie. LIXIE - an LED alternative to the Nixie Tube!
I've not used either so I can't speak to how well they may or may not work but it's something to consider.
I've not used either so I can't speak to how well they may or may not work but it's something to consider.
I don't know if this would work for your project but the Adafruit HUZZAH ESP8266 Breakout has a WiFi built in and maybe you could set it up to do simple NTP?
Stereodude
Not really a
I saw those yesterday also. They're still a little pricey and they don't quite have the same retro vibe to them as Nixie tubes.
Oh I agree; the Lixie displays are not an exact replacement for the Nixie tubes but they're interesting to play with on their own. I haven't seen much that can replicate the same visual experience as a Nixie.
Stereodude
Not really a
These nixie tubes are amazing and look like finely crafted pieced of art, but it's almost $1k for a set of 6. If you're going in, you might as well go all in right? :bomb:
Their completed clock is a piece of art, but at $2.2k I won't be buying one.
Stereodude
Not really a
I think a nice geeky clock could be made with RGB LED matrix panels also. Like this one. I should note they're much cheaper to buy on AliExpress. Like this 2mm pitch 128x64 one. There's code on GitHub for driving them from the RPi. There's code to render text to them, playing videos, and displaying images. I'd want to have some nice animated transitions / effects on the characters (numbers) and that's the part that's missing (as far as I can tell) short of playing a pre-rendered 24 hour video.
I bought one of these Adafruit NeoPixel NeoMatrix - 64 RGBW - Natural - ~4500K to play with but haven't had the time to mess with it yet. No where near as dense of LEDs like the one you linked to but I wasn't looking to make a clock with it.
Stereodude
Not really a
Those are cool. I didn't realize there were RGBW neopixels now.
On the clock front it wouldn't have to be limited to being a clock since it's a dot matrix display. It could have weather or other information on it too. More than one line or alternating screens. It's sort of hard for me to visualize in my head what a 10"x5" 128x64 LED panel looks like across the room in terms of graphical fidelity, readability, how bright it will be a night, etc. I guess I have to buy the parts, put it together, and see. I wonder if it would be possible to render from Qt (or similar) to the buffer the hzeller is sending to the display.
I had one in the early 70s. The main issue is that the numbers are increasingly blocked further to the rear of the tube and at angles away from the perpendicular.
7-segment displays were a huge advancement in that they are much easier to read from any angle.
My favorite clock ever had a 6 digit neon display. The segments were so bright and clean compared to LEDs. The clock chips of the day (NMOS?) were power hogs and they did not last. IIRC I replaced the clock chip after about 10 years and then the second one failed later.
jtr1962
Storage? I am Storage!
In the last decade the brightness of LED displays has improved enormously, to the point they rival those neon displays if you drive them 10 or 20 mA. I usually drive at less if it's a battery-powered device to save power. In many cases I can drive the segments at well under 1 mA and still have them very readable. The white, true green, and blue ones tend to be the brightest, although red and yellow aren't far behind. Back in the 1970s red was the only available color. Eventually you had yellow, orange, and yellow-green (565 nm). In the last few years true green (525 nm), blue, and white are readily available. Some people even make DIY 7-segment displays using LED filaments which resemble those old neon tube displays.I had one in the early 70s. The main issue is that the numbers are increasingly blocked further to the rear of the tube and at angles away from the perpendicular.
7-segment displays were a huge advancement in that they are much easier to read from any angle.
My favorite clock ever had a 6 digit neon display. The segments were so bright and clean compared to LEDs. The clock chips of the day (NMOS?) were power hogs and they did not last. IIRC I replaced the clock chip after about 10 years and then the second one failed later.
LED filament clock
The brightness is not so much of an issue as the contrast and containing the illumination to only the desired segments. Everything else should be pitch black.
Many LEDs, especially those in the early days, suffered light leak that the neon did not.
Stereodude
Not really a
I used today's eBay 15% off coupon to buy a set of 6 matched NOS IN-18 Nixie tubes and an assembled PCB to put them in.
Stereodude
Not really a
I finally got the nixie tubes yesterday. I had to pick them up from the post office since they needed a signature and I wasn't home when they tried to deliver. Air Mail shipping from the Ukraine is certainly not fast. I have it working with the stock Arduino firmware that came preflashed on the board. The order (random) that I put the tubes in initially had the last digit with a "9" that would not illuminate. I swapped the last two tubes and now they both will illuminate all digits.
The clock is cool, but I plan to modify the Arduino code running on it to make what I feel will be improvements.
1) The LEDs under the tubes don't quite achieve the visual effect I want. You can pick the color (by stopping it as it rolls through the colors) but the colors are maximum intensity. I want to get a orange glow that's basically the same color as the Nixie tubes numbers themselves, but I want it to be a bit more subtle than maximum intensity. I've already looked through the code and this should be a pretty easy tweak to fix the color and intensity (but will lose the ability to roll through the color).
2) The flashing columns between the hour and second digits don't flash on and off at what I think is the "correct" cadence. I would expect them to flash on and off once per second. They turn off and on a 2 second cycle. This looks super easy to adjust in the code.
3) On some periodic basis the clock will display the date for a few seconds and return to the time. It does a scroll through the numbers transition getting to the date and back from the date. My understanding is that scrolling is necessary to preserve tube life (anti-poisoning). I haven't quite timed it out, but it's not sequenced to the clock. It's something like every 65 seconds rather than happening say every minute at 40 seconds. Adjusting the interval in the code looks trivial. Controlling exactly when it happens looks a bit harder as the code currently doesn't use that as a condition.
The code is not well commented. It's barely commented at all. It's not too hard to follow because the variable names aren't overly generic and the code isn't all that complicated, but it could definitely be better.
My first step is to build the code from GitHub with no changes and make sure it runs correctly on the board I bought. The eBay listing says it's a NCM107 board, but the silk screen on the PCB says it's a NCM109 V1.1 board and Gra & Afch have way too many versions on GitHub.
The clock is cool, but I plan to modify the Arduino code running on it to make what I feel will be improvements.
1) The LEDs under the tubes don't quite achieve the visual effect I want. You can pick the color (by stopping it as it rolls through the colors) but the colors are maximum intensity. I want to get a orange glow that's basically the same color as the Nixie tubes numbers themselves, but I want it to be a bit more subtle than maximum intensity. I've already looked through the code and this should be a pretty easy tweak to fix the color and intensity (but will lose the ability to roll through the color).
2) The flashing columns between the hour and second digits don't flash on and off at what I think is the "correct" cadence. I would expect them to flash on and off once per second. They turn off and on a 2 second cycle. This looks super easy to adjust in the code.
3) On some periodic basis the clock will display the date for a few seconds and return to the time. It does a scroll through the numbers transition getting to the date and back from the date. My understanding is that scrolling is necessary to preserve tube life (anti-poisoning). I haven't quite timed it out, but it's not sequenced to the clock. It's something like every 65 seconds rather than happening say every minute at 40 seconds. Adjusting the interval in the code looks trivial. Controlling exactly when it happens looks a bit harder as the code currently doesn't use that as a condition.
The code is not well commented. It's barely commented at all. It's not too hard to follow because the variable names aren't overly generic and the code isn't all that complicated, but it could definitely be better.
My first step is to build the code from GitHub with no changes and make sure it runs correctly on the board I bought. The eBay listing says it's a NCM107 board, but the silk screen on the PCB says it's a NCM109 V1.1 board and Gra & Afch have way too many versions on GitHub.
Newtun
Storage is nice, especially if it doesn't rotate
Merry belated Christmas! Now you've got time on your hands. :wink:
Stereodude
Not really a
So, I identified the correct code project on GitHut for the board I have and I fixed #1, #2, and have #3 partially fixed. #3 now happens on 60 second intervals instead of 65 second intervals. I haven't attempted to make it happen at a certain number of seconds yet.