Divergent Nixie clock designs

[John Snow]

Keeping the tubes as the only unchangeable part, I've seen clocks with:

I've seen a few implementations of microcontrollers; ATmega, ESP32 are the main two 'recent' ones I've noticed - and the entirely transistor based one was impressive.

The digits can be multiplexed or direct-driven; 74141, HV Microchips, Bipolar (BJT) Transistors

There are also radio/wifi/gps synching of time & date

What other ones have you seen?

[gregebert]

The first clock I ever built (1980) used stepper relays and incandescent edge-lit displays. Time-keeping used a 1 RPM motor with a cam. It was a noisy piece of junk.

My first nixie clock (2011) is built entirely from 4000-series CMOS logic gates, and runs off the AC line (no transformer).

Several subsequent clocks use FPGAs for the logic.

[John Snow]

Oh boy I hadn't considered FPGAs - I supposed that's beyond the scope of my current projects, just the direction I'm going in.

Added to my list!

Any others?

[gregebert]

Well, these dont get full-credit because I havn't built them yet, but I have plans to do so:

1. 3D-printed gears as the timekeeping mechanism. Most likely, I would use photosensors because mechanical switches are unreliable

2. Dynamotor for the nixie supply of a battery-powered "portable" nixie clock. It's an actual motor-generator that runs on 12VDC, and produces 220VDC. Though noisy, it sounds really neat. I salvaged mine many years ago from a "BC-733-A Radio Receiver", which at the time I thought was for voice communications, but I never was able to get it to work as such. As it turns out, I found out later it was for aircraft landing guidance. 

[Nick]

My current clock designs are using Xilinx FPGAs, coded in VHDL using Vivado.

Really not a sensible thing to do, except that I've always wanted to learn VHDL, and the only way to learn a new language (for me) is to have a decent project or two.

I started clocks on discrete logic, moved to PICs, then AVRs, then MSP430s and now FPGAs...though I suspect I'll revert to MSPs after this... FPGAs are really not ideal, but they are fun.

[gregebert]

Verilog is less-painful than VHDL for coding. In my dictionary, VHDL = Verbose Hardware Description Language....

Initially I used FPGAs to replace random logic entirely.

They also have the advantage of having flexible pinouts, so you can often save a lot of PCB routing headaches by swapping pins around.

Lately, I've needed high-speed control and a microcontroller just isn't up-to-snuff in all cases.

Where I'm at now is RaspberryPi + FPGA: It gives me all the goodness of the linux ecosystem (tons of software & apps), WiFi networking, reprogrammability at the hardware and software level, and high-speed logic. Would be nice if the RasPi could reprogram the FPGA (ugghhh...another project I want to do)

Lastly, I've been a chip designer for more than 30 years, and my current employer doesn't let me do "real design work" anymore, so that's why I design nixie clocks from scratch.

I did some evaluation of Arduino and MSP432, but stopped all work when the $10 RasPiZeroW came out.

[Nick]

Yeh - there are Verilog vs. VHDL wars just like Windows vs. Linux or C++ vs. almost anything!

If I were designing an ASIC, I would probably go down the Verilog route, but as I'm sticking to FPGAs then I like the structure of VHDL, it's better portability and tighter definition. I've no problem with verbosity - it's clean and legible - for me, that's what I like.

At the end of the day, it makes little difference as the synthesis/map/place/route phases really don't care what language you're using.

I'm using Pi Zero Ws for a few jobs - they are astonishing value for money - a complete Linux box with HDMI, USG OTG,  WiFi & BLE plus expansion for just 10 bucks... insane. (well, maybe $15 if you add in the uSD card).

Nick

[Paolo Cravero]

 Hi.

The digits can be multiplexed or direct-driven; 74141, HV Microchips, Bipolar (BJT) Transistors

Not mentioned yet but I think pretty effective: anode multiplexing with optocouplers.

Paolo

[threeneurons]

I've made a clock using a CPLD, the little brother of the FPGA:

https://threeneurons.files.wordpress.com/2018/06/mach210_clk.pdf

AMD MACH210. It was made quite a while ago. I guess MACH210 chips are vintage now.

[gregebert]

Long live PALASM !   I used CUPL a few times in the late 1980's.

Were you able to program your device at home ? Most of the tools back then were not free, and out-of-reach for hobbyists.

I remember sneaking into the lab during lunch hour to program EPROMs.

[GastonP]

Nice design, Mike

   What was the function of the microphone in the lower part of the shcematic? It goes to the connector and to a pin called X+ but I couldn't find what's after that.

Gastón

[threeneurons]

During the 80's, I had the keys to work, so I had free run of the equipment, after hours. I used that equipment on my senior project, and wrote my college papers on the company's CP/M computer in Wordstar. I think I got good grades on my reports, not for the content, but because the margins lined up on both the left and right side. The company had Tektronix development systems for working on microprocessors. An 8002 and a 8550. School had a lab full of HP 64000s, which were always occupied. Both the HP and Tektronix units cost well in the 10's of 1000s of dollars. By the mid-80s, however, emulator pods, that plugged into an IBM PC, started to become popular. Those sold in the $2K to $5K range. By ~1990, the IC makers started selling evaluation boards, for only a few 100 bucks. I picked up a Motorola HC05 EVM board for under $500, and it worked well as a development platform. Of course, now I use AVR tools, that cost less than $35.

I own a programmer than can do MACH devices, and it cost me in the ballpark of $200, back in the early 90s. I haven't used it in years. Haven't erased an EPROM, or quartz windowed uC (usually an HC05-C8) in over 20 years. Don't drop a ceramic quartz windowed 22V10. They break in two, when they hit the floor.

 

[threeneurons]

On Tuesday, June 5, 2018 at 6:14:58 AM UTC-7, GastonP wrote:

Nice design, Mike

   What was the function of the microphone in the lower part of the shcematic? It goes to the connector and to a pin called X+ but I couldn't find what's after that.

Gastón

There's a ring of LEDs around  digits, where the lights do a 4-chl chase pattern. The signals that control that are XB0 thru XB3. Those signals are connected to the LED cathodes. All the anodes (for those LEDs and NOT the digit LEDs) are connected to X+.

[Terry S]

LOVED the Mach devices! I made those things dance and sing. Complex state machines that would make a CPU blush. Controlling steppers and sonars and reading encoders and writing to dual port RAM.... 

I used DATO I/Os version of ABEL and AMD's plug-ins and I was in heaven creating those designs! Man I miss those days!

Terry

[ZethieTail]

i know this is an old pose but i have some 22v10's, what kind of stuff have you made with them?  using one as a bargraph driver or shift register woud be nice for my projects since i have some neon clock displays and uusing one to multiples it would be cool, its an old gold ceramic 22v10, but its one time programable only sadly, but i bought others that are reusable, but still no way to program them tho

[ZY]

You can also have decatrons do the counting:

https://www.youtube.com/watch?v=4bl5kSr8C3Q

I'm guessing some more parts can be replaced with tubes by replacing the rectifier with a tube rectifier.