Seven digit nixie clock

[Mitch]

Here's a short video of the clock. This is the second revision of the board, down to only four jumpers and no cut traces! This version uses separate boards for each tube. I'm using Z573M tubes now, and I have IN-14 and IN-12B tube boards on order. 

The IN-12B boards will use 90 degree header pins to stand vertically, and the RGB lighting will get to the tubes with clear, 5mm diameter optical fiber. The third revision of the board that I just ordered today, repositions some components and swaps the tube headers with the HV5530 chips, to 

move the tubes closer to the center of the board. That way the IN-12Bs won't hang from the front. The processor I'm using is the Atmel 1284P with the Bobuino/Skinny Bob bootloader. Programming is through the FTDI connector. 

David - no way I could route +5v, +12v, GND, and HV on the bottom. I spent a lot of time trying.

Software wise, the sketch is now over 2000 lines, not including libraries. Here are some of the features I included, which is everything I could think of:

RGB with several effects.

Three override (sleep) periods with independent settings for start and end times, brightness, neon effects, RGB effects and brightness. Each period can be set for weekdays, weekends, or both, and overlapping is allowed with a 1-3 hierarchy.

DS3231 RTC with GPS that monitors and resets the RTC when the difference is 2 seconds or more.

Automatic timezone setting for DST. Both Time and Timezone libraries are used. 

Scrolling date and temperature with optional RGB scrolling. Each RGB LED is individually controlled.

Ambient light sensor with brightness adjust.

Optional PIR sensor enables display when motion is sensed.

Cathode anti-poisoning routine with several options.

Four indicator LEDs for various functions, and four buttons for setting.

Supercap to keep the DS3231 going for a while, maybe days.

So far everything works, I can't find any bugs, but it's not easy to check different combinations of 53 options, so there are probably many left hidden for now. The software originally started on an Arduinix a couple years ago and it was my first try at C++ programming. I see now that there are many things that can be done better,

and I'll get to that as time goes on.

I think the third board revision will be the end for this design. I'll build a few for friends and relatives. Next up is an IN-18 version. That should be very easy, just a board redesign and some different anode resistors. The Tayloredge power supply will easily handle seven big tubes.

After that, a 15 digit Metronome version using five, HV5530 chips. Thanks, Jarek, for that idea.

Mitch

[Mitch]

I forgot to mention that as hard as I've tried, I have not been able to figure out smooth, digit cross fading. For now the clock uses a 'vibration' effect that you can see in the video, or a no effects digit change.

[David Forbes]

Mitch,

Are you willing to share your code? There are many pairs of sharp eyes here that
could help you figure that out.

[gregebert]

Regarding fading, I've seen clocks that gradually make 1 digit dimmer as the next digit becomes brighter (with overlap), and it looks more like a blurry mess. You may want to use PWM to dim the changing digit in a 100-200 msec, before you gradually turn-on the next value. It might be annoying for digits changing every second (or in your case, 1/10 second ??), but for tens-seconds and beyond it should look OK. None of the clocks I've built so far will support this, but the next one will be capable of it.

Based on the HV5530 driver, I'm assuming your clock is direct-drive, so you will have more options for effects.

Did you add a level-shifter to drive the 5530's inputs at the recommended 12V signal-levels ? I'm also curious if you cascaded the serial data-in/data out signals if both 5530's use the same clock; per the datasheet, doing so is a timing risk (data-hold time is 10nsec, and there is no min-prop delay specified hence assume 0. This implies a 10nsec timing risk).

I'm finalizing the PC board for my 14-digit IN-18 clock, and took no risks: Using a level-shifter, and separate clocks for each HV5530. I'm using an FPGA instead of a microcontroller, and I have enough room on the FPGA for a small CPU (probably Z80) if I decide to go that route instead of pure Verilog code.

[Mitch]

David, I will put everything on Github when the new boards arrive. The schematic and board were done with Diptrace.

Greg, the 1/10th second digit won't fade or use effects because it changes so quickly. The 'vibrate' effect works only on the other six digits. 

A level shifter is used, and all three 5530s use the same clock. I don't see how to implement separate clocks, and it works fine as is. I did some quick timing tests and it appears that the display can be updated no faster than 17ms, as the code stands. It is only updated at 100ms now, and 25ms when the vibrate effect is used.

The blanking line on the 5530s dims the display with PWM. I don't see a way to dim just one digit. With multiplexing, as each digit is selected the PWM rate could be set differently, but I don't see how to do it with direct drive.

A 14 digit IN-18 clock sounds interesting. What are you doing with the extra digits?

[Dave]

Mitch,
This is very nice work indeed !

Thanks for sharing.

When you post to github, will it be the code and the diptrace files?

I deeply appreciate your willingness to share your work.

[Mitch]

Thanks, Dave. Everything will be included, Diptrace schematic, pc layout, Arduino sketch, and operations manual. I have no intention of commercializing the project in any way.

This was my first experience with pc board design, and I'm hoping that someone here with good hardware experience will improve the design and pc board layout. Same with the software. I'll start reviewing and improving it in a couple weeks, but someone with extensive Arduino and C++ experience can certainly take it much farther.

I'll post more pictures in about two weeks, after the IN-14 and IN-12B boards arrive.

[gregebert]

> I don't see a way to dim just one digit. With multiplexing, as each digit is selected the PWM rate could be set >differently, but I don't see how to do it with direct drive.

Dimming individual digits is possible with hardware control, but probably not with software. The idea is to resend the serial data to the HV5530's twice every few milliseconds or so: Once to turn off the digit(s) to be dimmed, then again to turn all digits on. By adjusting the on and off times, you are doing PWM. I have 2 cascaded HV5530's (64 bits), and I clock them at 1.5Mhz, so it's possible to do PWM up to a few kHz.

 

A 14 digit IN-18 clock sounds interesting. What are you doing with the extra digits?

I went overboard, so it displays hours, minutes, seconds, date, month, year = 14 digits. My wife saw the nixie clock in the movie "Tomorrowland" and ordered me to build one for her (yeah, like I'm going to say no ???). I saw no purpose for 12 digits, so I super-sized it to 14 digits. To level-out the tube usage, I swap date and time locations hourly, then run a depoisoning routine at 4AM for 1 hour.

The clock will be in our vacation home, so the tubes will be off most of the time. It will be an art-deco theme, in a wooden case complete with 2 large toggle switches and 1950's-era incandescent panel lamps. I expect it will be 30 inches wide x 5 inches deep x 5 inches tall. I'm leery about the reliability of non-Burroughs tubes, and we already have 3 nixie clocks running 24/7 (why not, they are Burroughs tubes....) in our permanent home. 

[Mitch]

I tried doing something similar, and that's how I came up with the vibration effect, which is really just a failed attempt at cross fading. I'm not sure whether the limitation is with the hardware or software, probably both. Writing to registers rather than using digitalWrite() may speed it up enough. That can be someone else's project at this point, I don't think I'll spend more time on it for now.

The clock on the Metronome in Union Square, uses 15 digits. Time, time to midnight, and 1/100 seconds in between. I think 1/10 makes more sense because at least you can see the digits. I like your idea better, but a combination would make it very interesting.

Here's the info:

https://en.wikipedia.org/wiki/Metronome_(public_artwork)

[Quixotic Nixotic]

Perhaps a little off-topic, but I know some of you like funfairs and amusement parks, so I make little apology for posting. I dare say there will be neon and blinky lights somewhere.

Banksy has shocked us again. With the help of some top-ranking fellow artists he has taken over an abandoned seaside amusement park at Weston Super Mare, once called Tropicana, for the next six weeks. Artists include Jimmy Cauty, once part of KLF, the crew who publicly burned £2,000,000. (Jimmy is displaying a fun model village complete with 3,000 riot police in the aftermath of major civil unrest.)

Tropicana is now called Dismaland and features a number of highly dubious attractions.

"I loved the Tropicana as a kid, so getting to throw these doors open again is a real honour," Banksy said. He describes the show as a "Family theme park unsuitable for children, which offers an escape from mindless escapism."

Dismaland Bemusement Park opens tomorrow and is only on for 6 weeks, see http://www.dismaland.co.uk, with tomorrow being open only for local people to Weston Super Mare. You can enjoy it from Saturday. Limited tickets at the venue on the day, or do remember to book online. 



Highlights you can look forward to are:

1 The Galleries


Dismaland® boasts three large galleries which together comprise the finest collection of contemporary art ever assembled in a North Somerset seaside town.

2 Model Village


Jimmy Cauty’s hand crafted miniature world will delight and amaze (and potentially cause seizures in persons sensitive to strobe lighting).

3 Cinderella’s Castle


Step inside the fairytale castle and see how it feels to be a real princess in a crash situation. Souvenir photos available.

4 Guerilla Island


Features a bus-mounted museum, library, gallery of guerilla art, Comrades Advice Bureau and workshops in how to hack billboards.

5 Ben Long
has constructed a stallion entirely from used scaffolding.

6 Water Cannon Creek


An armour plated riot control vehicle built to serve on the streets of Northern Ireland. Equipped with sniper posts, grenade launchers and now – a children’s slide.

7 Mini Gulf


An oil caliphate themed crazy golf course made from the hockey pitch at Cheltenham Ladies College.

8 Giant Pin Wheel


Commissioned by Banksy to power the entire site, it turns out this ecological marvel has trouble charging more than two mobile phones at once.

9 Kids Enclosure


Strictly for the little ones – an area that combines soft play and loan shop.

10 Punch and Julie


The infamous Julie Burchill has given this seaside staple a good kicking. Now includes references to Jimmy Saville and Fifty Shades of Grey.

11 Puppets


A puppet revue show constructed entirely from the contents of Hackney skips – it’s ‘Fly Tip Theatre’ by Paul Insect and Bast. Self operated.

12 Jeffrey Archer Memorial Fire Pit


Warm yourself around an authentic real open fire ceremonially lit each day by burning one of the famed local perjurer’s novels.

13 Portrait Artist


Nettie Wakefield renders your likeness in an exquisite pencil drawing, but only the back of your head. Surprisingly revealing.

14 Picnic Area


The humble picnic table proves a rich source of artistic ammunition for American furniture twerker Michael Beitz.

15 Cinema


Truck-mounted outdoor screen playing a rolling program of short films day and night. Cushions not provided.

16

Circus Tent


A gothic sculpture park in a tiny big top tent.

17 Amusements


Win big prizes on hook-a-duck and the shooting gallery. Knock over an anvil in David Shrigley’s game of skill and cunning. Navigate the high seas with Banksy’s Mediterranean
 boat ride.

18 Big Rig Jig


A masterpiece of post-industrial assemblage art – it’s two juggernauts performing ballet by Mike Ross.

Much much more. I like the one where you can buy an Adshel bus shelter key for a fiver, so you can put your own subversive posters into bus shelters. Or perhaps the stall where kids can get an advance on their pocket money at exorbitant interest rates? Fun for all the family.

Artists exhibiting are, in alphabetical order:

Andreas Hykade (Bavaria)
Amir Schiby (Israel)
Axel Void (USA)
Banksy (UK)
Barry Reigate (UK)
Ben Long (UK)
Bill Barminski (USA)
Block9 (UK)
Brock Davis (USA)
Caitlin Cherry (USA)
Caroline McCarthy (UK)
Damien Hirst (UK)
Darren Cullen (UK)
David Shrigley (UK)
Dorcas Casey (UK)
Dietrich Wegner (USA)
Ed Hall (UK)
El Teneen (Egypt)
Escif (Spain)
Espo (USA)
Fares Cachoux (Syria)
Greg Haberny (USA)
Huda Beydoun (Saudi Arabia)
James Joyce (UK)
Jani Leinonen (Finland)
Jeff Gillette (USA)
Jenny Holzer (USA)
Jessica Harrison (UK)
Jimmy Cauty (UK)
Joanna Pollonais (Canada)
Josh Keyes (USA)
Julie Burchill (UK)
Kate MacDowell (USA)
Laura Lancaster (UK)
Lee Madgwick (UK)
Leigh Mulley (UK)
Lush (Australia)
Mana Neyestani (Iran)
Maskull Laserre (Canada)
Michael Beitz (USA)
Mike Ross (USA)
Neta Harari Navon (Israel)
Nettie Wakefield (UK)
Paco Pomet (Spain)
Paul Insect & BAST (UK/USA)
Peter Kennard & Cat Phillips(UK)
Polly Morgan (UK)
Pure Evil (UK)
Ronit Baranga (Israel)
Sami Musa (Palestine)
Scott Hove (USA)
Severija Inčirauskaitė-Kriaunevičienė (Lithuania)
Shadi Alzaqzouq (Palestine)
Suliman Mansour (Palestine)
Tammam Azzam (Syria)
The Astronauts’ Caravan (UK)
Tinsel Edwards (UK)
Wasted Rita (Portugal)
Zaria Forman (USA)

You have been warned, so do go along if you can...

John S

[Quixotic Nixotic]

In the way that art mirrors real life (or is it the other way round?) I have also been following the shenanigans of the revived Dreamland Amusement Park in Margate, UK, with a modicum of bored interest. As with so many of England's seaside resorts, when hard times hit and repair bills run high and the lure of property developers' money smells sweet, Dreamland (home to the UK's historical last wooden roller coaster and a protected structure) fell prey to a mysterious fire in 2008. It was officially found to be deliberate arson, so I think I am OK saying this publicly. I could say more, but it's probably wise not to.

With the aid of a huge cash injection from the UK's national lottery, Dreamland is being resurrected once again.

The lottery money was extracted partly upon its promise to celebrate the UK's traditional seaside amusement establishment heritage. No small part of this enticement was the inclusion of an extensive collection of pinball machines, arcade machines and a few jukes owned by my long-standing acquaintances Peter Heath and Philip Crow. Their machines were to form a working arcade for the public to actively enjoy. Peter and Phil were overjoyed and sourcing even more rare and wonderful machines from the US to add to the collection, solely for the benefit of the Margate visitors So imagine their surprise and delight when they were told their vintage arcade was no longer needed or part of any current plans, once the lottery money was conveniently in place of course.

I suspect a word has been spoken in the National Lottery's collective ear, as I now hear that Peter and Phil have been mysteriously re-invited to install their arcade in the entrance area after all. Hoorah!

John S

[threeneurons]

I see Disney attorneys sharpening their halberds and battle axes.

[Quixotic Nixotic]

On 21 Aug 2015, at 15:59, 'threeneurons' via neonixie-l wrote:

I see Disney attorneys sharpening their halberds and battle axes

Dismaland opened today. A friend of mine went around lunchtime and said that unfortunately all 1,000 free tickets for local people, today only, have gone. He was able to see some of it from outside.

The security staff have undergone special training to be surly and unhelpful.

John S

[Alex]

This is fantastic in so many ways. Also, I note the website footer states this:

"The following are strictly prohibited in the Park – spray paint, marker pens, knives and legal representatives of the Walt Disney Corporation."

[Ian Sparkes]

Hey Mitch

I did a smooth fading on a AVR, and the code is up there on GitHub already:

https://github.com/isparkes/ArdunixNix6

I do it a different way than I think you will want to, because I use an old fashioned K155,and use 6 I/O pins to control the anodes. I have an inner loop of 1000 iterations (to display each digit), and an outer loop of 6 iterations (one for each digit). I manage dimming by defining the off time for each digit (earlier off time = more dimming), and the fading by the switch point between the old and new digit. Each display impression, the switch point is progressed until the fade is done. Of course the code has to be quick enough so that all of this is done without letting flicker creep in. On my 1 x 6 multiplexed clock, this is a challenge, but with careful software design it is perfectly possible.

What exactly was the problem with the fading? Not smooth? Too hard to program? I guess is that it was not really smooth, because the process of clocking in the data was not quick enough, and that to resolve this, you'll have to be really careful about writing the clock and data lines, probably using bit level operations directly on the ports. Possibly clocking the HV5530s in parallel could help.

PV Electronics does smooth fading, but using the 20 bit HV5812, so it is possible if you are multiplexing 3 x 2.

If you post the code up, I'm sure someone will be able to add the "special sauce" to make s smooth fade. If you have some boards made, I'd like to get one and have a play around with it.

Regards

Ian

My own workings to try and understand the issue:

In my design I'm transferring 60 bits per impression. cathodes: 6 digits * 4 bits * 2 values (before and after fade). Anodes: 1 bit * 6 digits * 2 (anode on, anode off). I don't have to toggle a clock line. If I needed to, I would write the registers directly, but right now I don't need to. Standard Arduino ditigalWrite() is fast enough.

You appear to be using 3 HV5530s, I suppose two managing 3 digits each with the odd 2 bits for the colons, and the last one for the 7th digit? I suppose you are running the RGB back lighting off Atmel's PWM outputs?

An impression for you is 74 raw bits, but using bit banging on the clock line to clock the data in, so that's tripling that number. However, you only need to toggle the data line 20 times (2 for each digit, 2 for each colon point), which means about 168 bits toggled. If you don't optimize the handling of the data line, you'll end up writing 224 bits per impression.

That's nearly 3 times the amount of data transferred per impression, and all because of the clock line. I think this is where you need to optimize.

[Mitch]

Hi Ian,

Thanks for your comments.

I do use three, HV5530s to control all seven nixies and six neons. The lower right most neon is the decimal for 1/10 seconds, and the upper right neon represents degrees for the temperature display.

All four indicator LEDs and seven 2N3904s to turn each RGB LED on or off, are connected to the port expander.  The RGB LED anodes are connected to three PWM lines on the 1284P using 2N7000 FETs.

No doubt the code can be greatly improved, especially the display function, but that will have to be done by someone else if it's going to be done at all. I think I'm at the limit of what I can accomplish, here. I think I mentioned that the maximum refresh rate at the moment, is around 17ms.

The third release of the board is on order, unfortunately, because there are some extra lines available that could have been used to clock the HV5530s separately. I was hoping for no jumpers this time around, but that is an option.

Mitch

On Saturday, August 22, 2015 at 4:09:06 PM UTC-4, Ian Sparkes 

[Mitch]

No jumpers or cut traces, this time! Just one diode reversed. Unfortunately the IN-14 boards were a mess. The mess could be fixed with software, but I'd rather just have them wired the right way. Fifteen went in the garbage, but at under $4 for three boards, not a big deal.

I'm experimenting with flexible fiber optics for the IN-12 RGBs, but it would probably be best to just attach the LEDs to the tube boards and wire them down to the main PC board. Also, the neons should be farther forward with these tubes. Personally I like the Z570M series and IN-14 better, so I won't do much more with the IN-12. I did make one change to the IN-12 board for the next order - the 5mm hole is now higher up so the LED or fiber pipe will be out of the way of the tube's nipple. That should improve the brightness, especially with fiber. The National tubes that are almost the same size, should work fine with the IN-12 board, too.

On the main board, the FTDI connector is now on the left using a 90 degree header, and the reset button is near the front left corner. I don't see making any more major changes. Relatives are waiting, so it's time to build some clocks.

After fixing a major bug I found today, everything will go to Github, and the board designs will be shared on OSH Park. 

The parts that can't be substituted are of course the processor, Tayloredge power supply, audio jacks, and pushbuttons. Part numbers and sources will be included with the schematic.

That's about it. I'm working on the IN-18 version now. Hopefully that one won't take three revisions to get right.

 

 

[gregebert]

Mitch - Do you run any simulations on your design before fabbing boards ?

As a rule, I run a lot of spice simulations on the analog, particularly the power supplies. My second clock had a subtle wiring error that caused spice not to converge, and if I had not fixed the error it would have smoked at power-on. So far, I've done 3 boards with no errors; the boards were right the first time. The spice deck is partially created from the PCB netlister.

For digital stuff, I do verilog simulations, and I do have to spend a lot of time coding-up models for various ICs.

[Mitch]

Greg, that probably would have been a good idea early on. I don't know anything about simulators, and at this point it's too late.

Attached are Diptrace files, Arduino code, and documentation. It will be on Github shortly.

And I have a question - For the IN-18 version, would it be better to connect all three HV5530s to the processor using separate data, latch, and clock lines? The processor does have enough spare pins and the code changes are trivial. 

Mitch

[gregebert]

You can certainly connect the HV5530's "in-parallel" to the CPU. I would be more concerned about signal-levels, because the HV5530 is intended to run with 12V logic levels and your CPU is probably 3.3V. I'm using a level-shifter IC in my design, along with an FPGA.

BTW, I'm putting the finishing touches on my first IN-18 clock PCB; It's definitely worth the extra expense of IN-18's to get large and properly-formed digits (no funky upside-down 2's and 5's).

I decided to use socket pins soldered to the pcb, rather than actual  IN-18 sockets, because of space constraints. I hope I'm not going to regret this decision, and I'm definitely going to be very careful to form the pins to minimize stress. I also will not be swapping tubes around to even-out their usage (using de-poisoning algorithm instead).

Do you plan to run your IN-18's 24/7, or will you use a motion-sensor, etc ?

[Mitch]

The 1284P runs at 5v, and I use a CD4504 level shifter. Someone here suggested that, and even though none of the commercial clocks I'm familiar with use one they are cheap and easy to include. That will also make it easy to go to a 3.3v chip in the future, with only some adjusted resistor values. I think every other solid state component on this design is 3.3v compatible.

I'm not sure if there is any advantage to connecting the HV5530s in parallel to the processor, maybe faster updating. That's why I asked. The current design works fine as is, with three connected in series. I'm not sure if it will work for a fifteen digit clock with five HV5530s, though.

I like the IN-18, too. It is certainly the best large tube for the money, nothing else comes close price wise.

I use acrylic disks under the IN-14 and Z570M series tubes, and I'll do the same for the IN-18. Check the picture. They are 5mm think, but I went to 3mm for the reorder. Illustrator files are included for them in the zip. I'll have an IN-18 version soon, too.

The Spectrum 18 also uses acrylic disks under each tube, and individual pins. If the component layout for the IN-18 is done carefully and the pins are soldered while on the tubes, there should not be any stress on them.

I run most clocks from 6pm-11pm, which is when I'm around. This clock has three override periods that are weekday and weekend aware, so I run it 6-11 during the week, and all day weekends. I don't use the PIR now, but after I build a few I'll set one up that way.

Do you plan to share your design?

[gregebert]

Once I finish the clock (a few months from now, I hope...) and post a video, I'll probably give copies of my design report to people in this forum who request it.

Each of my display boards has two HV5530 devices that are connected in series, but I have separate clock signals in order to guarantee hold-time margin by using a 2-phase clock. When multiple display boards are used (for more than 6 tubes), they each have their own separate serial-data-in signal. All other signals are shared in-common. 

[Mitch]

Greg, I'd like to understand this so I can do the same for the IN-18 design and the next, fifteen digit clock that will use five, HV5530s. Please let me know if this is correct:

Latch, blanking, and data, are connected in parallel, on each of your boards. Each clock line is connected to a separate pin on the processor. How do you clock the data so that it moves across all HV5530s if they are not clocked at the same time? 

Would it be an advantage to connect each clock and each data line to a separate pin, to load each HV5530 separately rather than serially?

[H. Carl Ott]

On Tue, Sep 1, 2015 at 1:34 PM, Mitch <mit...@gmail.com> wrote:

And I have a question - For the IN-18 version, would it be better to connect all three HV5530s to the processor using separate data, latch, and clock lines? The processor does have enough spare pins and the code changes are trivial. 

Mitch

 If you are using level shifters, then you will need additional hardware on all the lines. 

 I would not bother,

 There is a possible advantage to just using separate data lines, while keeping the CLK/LE/BLANK lines paralleled. You could clock out the data in 1/3 the time.  

 

But, it's a trade off when you are bit-banging the serial data.

 The determining factor is how fast you can shift the data, set io ports and then toggle the clk line in software. 

I did it all in an  ISR written in AVR  assember. I clock out 96 bits of buffer data into 3 chained HV5522s. The ISR fires at 4096hz off a timer. I get very smooth cross fades.

I spent a lot of time optimizing  those inner loops, shaving cycles wherever possible . 

I doubt that I could actually do it much faster if the data lines were separate. I'd now have to de-mux the data and manipulate 3 IO lines before I could toggle the clock line,  All those extra cycles take time. 

I think you'd gain more by tweaking the code a bit, using the SPI hardware, or just by running  the AVR at a higher clock rate.  

carl
--------------------------------------------------------
Henry Carl Ott   N2RVQ    hcar...@gmail.com

 

  

 

[Mitch]

Thanks for the advice. I checked the Mod 6 and looking closely, there is no flicker whatsoever. Hopefully I can get to 4ms updates or close, by writing directly to the port, bypassing digitalWrite(), and optimizing the code. Assembly language is not an option. Last time I did that was in college, on a General Automation minicomputer.

The CD4504 has two unused level shifters so I will connect each of the three HV5530's data lines directly to the processor, clocking three bits at the same time, without adding hardware. Data for each digit is stored separately, so the software changes are easy, at least for that part. 

Hopefully fading will get done, after all.

[Ian Sparkes]

Hi Carl, Mitch,

A couple of thoughts in decreasing order of ease of implementation:

1) Software only: As a first step I'd bypass the digitalWrite() and control the port directly. I agree that assembler is probably too much for the first round of improvement. I see that your "dataPin" is a constant.

It's best not to pass it as a parameter: it can be referenced from directly within the "shiftItOut" code and the variable referencing can be skipped. I'd also do away with the number of bits variable. It's a known constant in the code which is required, and you can just call the right one straight away, skipping another comparison.

I'd also tend in the longer term to prepare a full static "data packet" to be shifted out and use the "shiftItOut" routine to just dump that to the output. This section is the innermost loop and needs the most optimisation.

2) Software only: I was mulling this over, and it occurred to me that out of each 10 cathode group, only 1 cathode is lit at a time (perhaps two if you implement a cross fade that way - but I wouldn't). This means a lot of the time you're only twiddling the clock line without toggling the data line. You're nearly doing this in "shiftItOut", but we have:

      if (val == i - 1) digitalWrite(dataPin, HIGH);
      else digitalWrite(dataPin, LOW);

and it might be better as:

      if (val == i - 1) digitalWrite(dataPin, HIGH);
      else if (val == i) digitalWrite(dataPin, LOW);

in order to save 8 writes to the data pin at the cost of 8 comparisons. I don't have any benchmarks about the speed of a digitalWrite as opposed to a comparison, but I'd be surprised if they were comparable.

3) Software only: We can do quite a bit of software optimisation in the code, without much restructuring. For example in "displayD" there are some variable declarations which we could do away with, and we tend to be shifting variable values around quite a lot. I would personally prefer something along the lines of a "display buffer" where changes are made when required. For example we only need to change the hours when they really need changing. This is the next most inner loop and needs the next most attention.

4) Software and hardware change: The most obvious improvement I can see would be to drive the 5530 clock from a PWM output in CTC mode, and use the ISR to do the shifting. As a nice side effect, it would have an easily programmable (in software) refresh rate. Unfortunately, I think you are using pin 4 as the clock pin and this is not PWM capable (unless I am looking at the wrong data sheet).

Just thoughts... Happy to discuss...

Ian

[Mitch]

Thanks again for the comments, Ian.

#1 - Yesterday I removed the pin numbers that were passed as parameters. I'll work on removing digitalWrite() by Monday or Tuesday.

#2 - Easily done at the same time. It's probably best to remove the 'vibrate' code that is more of a kluge than good programming, and do fades correctly.

#3 - Would the overhead to assemble a data buffer be worth the cycles to do it? I'm thinking about separating the data lines as Carl mentioned. Even if that's done, three bits could be located with an offset and written to all three HV5530s with one clock write. That could be done on your board with two trace cuts and four jumpers. The extra level shifters are there. What do you think about that?

#4 - That's beyond my understanding. I was thinking of using a hardware interrupt to update the display, but it would require some changes to the menu function as it runs in its own loop.And that is on the list of changes to be made.

Mitch