Unknown Domain

Split-flap Display – Prototype V3

At last, with the help from some friends I have built the long awaited third prototype of my project. The following details some of the design changes:

Gears

I decided to go with laser cut plastic gears like last time because the only decent range of timing belts and pulleys comes from McMaster-Carr and they don’t really offer a range of metric sized parts, which is a pain, there seems to be a real lack of mechanical power transmission parts. I am going to try contacting the 3D printer community about this as the gears are always super loose.

That aside the plastic gears work okay for now, and were designed using Matthias Wendel‘s Gear Template Generator software I still want to use pulleys but I need to find a better supplier that has the range I need. Why can’t I be an expert electronics engineer, mechanical engineer, computer programmer and designer all in one?

Housing

The housing was constructed of a thinner 2mm clear Perspex, last time I used 3mm, one disadvantage is that 2mm seems to shatter easily when dropped, also coloured Perspex isn’t readily available in 2mm thickness. It is also scary to cut in the laser cutter as you can see small pockets of fire under your work which are usually hidden by the opacity of the coloured material. I bought it from Hamar Acrylic, London and used a fantastic and cheap product they recommended for gluing called Plastic Weld, it’s basically a solvent that melts the edges, bonds them and evaporates leaving the two surfaces welded but no residue.o2

Catch

The idea for the catch came from a friend who suggested inserting a sheet of rubber rather than plastic that hangs down from the top giving it a ridged flexibility that allow the flaps to push against the catch until it flaps by without putting pressure on the sprockets.

The new design is dramatically bigger, by volume probably 100-150% larger! Next up I need to work on screen printing the flaps and the electronics I was working on before, having some trouble with inverted UART to RS485 signals, anyway here it is:

 

Another email

Today I got an email I guess triggered by my last post pushing me up the Google-o-meter again? Perhaps co-incidence, each time I blog it seems to result in someone else emailing me?!?

This time more of a q&a email from someone like minded looking for help, I assure you I will be returning to my efforts this summer!

Hi there,

I came across your blog recently searching for info on Solari stuff. I’m doing a project here (New Zealand) with the Auckland City Council to install a split-flap system like this… hopefully? They’re keen and there’s some money involved but I haven’t heard anything back from Solari yet so have still been looking at other alternatives. Mine needs text, I’m planning to commission a writer to develop a speculative history about an actual train station here. It’s a very small station, but has a lot of historical baggage…

Anyway, wondering where you got to with your efforts? Did you ever get something made? Did you make it yourself? The info on your blog’s really useful, esp the link to Rebecca Baumann’s work.

Thought it was worth getting in touch,

Cheers,

Luke Wood

Signal to Noise by @lab_au

In an article by Filip (CreativeApplications.net) he writes about a project called Signal to Noise by LAb[au] which makes use of reclaimed and refurblished split-flap units from a station in Bruges.

A wonderful project, and as usual I have emailed them to let them know how much I love them, and if they know where I can get one, I so wish the UK Railways hadn’t destroyed theirs, I’d love to have a real one!

Perhaps this summer I will achieve this!?

SOURCE

Sketchup split-flap display

I thought it would be good to spend some time getting to know Google Sketchup, especially after watching Matthias Wandel’s amazingly clear video introductions on YouTube, it really helped a beginner like me move to the next level in understanding. Anyway, I made the following model of the original version one split-flap display because its simple but mainly because I still have the design, while the version two was made on the fly and I did keep the files on a purple USB stick I can’t find any more, if you see it, send me a shout!

RS485

Useful resources on RS485:

http://www.gammon.com.au/forum/?id=11428

http://real2electronics.blogspot.com/2009/09/arduino-and-rs485-english-version.html

 

Character module MCU

So I have been looking at electronics predominately today, I looked starting with the RS-422 line driver, which I think should be a Maxim MAX1483, so I will be ordering a couple to test out on a breadboard as they are available in a DIP package, unlike half the stuff out there. How are you supposed to prototype an idea or do DIY projects if there are no DIP package versions, I for one am not going to make a PCB, and I don’t fancy buying/finding a breakout board either, anyway…

Looking at what I need for a MCU to drive the character module, I would like a chip available in DIP and some kind of SMD. It needs to be Arduino compatible and there are also the following probable needs:

1x UART for RS-422 (via line driver)
2x 10bit ADC for hex Switches
2x Digital input for IR Photodetectors
1x Digital output for a status LED
2x Digital output for the stepper driver interface (1x direction & 1x step)

Thats a total of 1x UART, 2x ADC, 2x digital inputs and 3x digital outputs, the Arduino as I said in a previous post is more than capable of this however a prefabricated Arduino board, even a clone board is about £12 and that is just more than it needs to be as I don’t need USB, and even if it is cheaper I still need another PCB shield to sit on top of the Arduino that holds the other chips.

So here is a list of known Arduino compatible chips and their issues:

ATmega1280 – Used in the original Arduino Mega, and is just too expensive
ATmega2560 – Used in the current Arduino Mega, and is just too expensive

ATmega8U2 – Has no ADC
ATmega16U2 – Has no ADC
ATmega32U4 – DIP package not available (otherwise perfect).

AT90USB1286 – Too expensive

ATtiny44 – Doesn’t have UART or enough program storage
ATtiny45 – Doesn’t have UART or enough program storage
ATtiny84 – Doesn’t have UART
ATtiny85 – Doesn’t have UART

ATmega8 – Limited program storage
ATmega168 – Good
ATmega328 – Good

Looks like the best chip to power the character modules is an Arduino still but just without the big board, nice to know as it means I can just use an Arduino for now. Was that a pointless waste of 2 hours? Probably but what did I learn? What UART means, and more about the range of MCUs Atmel makes. I will choose between the ATmega8/168/328 depending on my programme memory needs vs cost as the 168 is the ideal medium, the 328 is expensive.

Electronics planning

LCD

I think it is important for there to be a interface on the master control unit so that it can be controlled without a computer when there are issues which require testing,

Go online – Listen for messages from computer
Go offline – Ignore messages from computer
Blank – Runs all modules to blank flap
Home – Returns all modules to flap zero
Identify – Blanks display and runs selected module to X (marks the spot)
Test mode – Instructs each module to perform test (output on module)
Demo – Runs random modules to random flaps
Config – Not sure what might need configuring yet

Buttons
The interface will be based on a scrolling menu of options, so there will be up, down, left, right, select, back.

MCU
I suspect as this project will only need one master control board I will just use an Arduino Mega 2560 as it has a USB interface and is hackable, I could even be useful should I want to add new features. What might make sense is to use a shield rather than just leaving cables stuck in the header pins.

Stepper
During the prototyping phase I will be using a Mercury Motor’s SM-42BYG011-25 from Sparkfun, it doesn’t look like the manufacturer still make this, however they sell a bunch of similar ones and it is a standardised stepper motor. Any bi-polar 1.8º step motor will suit.

Stepper driver
In the prototyping phase I will be using a Sparkfun EasyDriver (4.3) however as the original designer of the board explains on his website, the EasyDriver is really just a breakout board so it will be included with the MCU and other components on the character module’s PCB.

The circuit primarily consists of an Allegro A3967 which will be more than adequate for the needs of this project.

Photo detector
To detect the flaps passing I will be experimenting with a pair of photo detectors, on to sense flap zero and one to detect each flap passing, whether this will work I am not sure, but its worth trying. Right now I am prototyping with this optical detector from Sparkfun.

DIP Switch
For location programming I will probably use a pair of HEX DILs or Rotary HEX switches, however just to get an idea of how I might be able to use a voltage divider to sense values from the switch without using 16 pins per switch, I will be using this 8 position switch from Love Electronics.

MCU
Because Arduino’s are expensive it seems over the top to use one for the character module PCB and as other parts are SMD and using a EasyDriver would be expensive too, it would make sense to do a custom PCB with all the circuitry necessary to run the module. As I only know C based languages really I would probably prefer to use an ATTiny45/85/2313 if suitable as they can run proper C or the Arduino boot loader so I can keep the programming simple for my own benefit.

It will need two digital inputs, two digital outputs, serial TX/RX, and two analogue inputs for the DIPs, however it would also be convenient to have a status LED which could be used to indicate problems, or set to flash when in the identify mode, so the front would show an X for frontal recognition and the back would be flashing for back recognition when working on large panels of these modules.

 

Water/laser/plasma cut 0.8mm black anodised aluminium quote

They people I spoke to about the cutting of the black anodised aluminium quoted £1 per flap which is reasonable but then you have to buy material too and the original devices used plastic, by the looks PP 0.5mm

Stanford University Graduate School of Business

I found this amazing project using what appears to be Krone units from looking but its absolutely beautiful to watch.

Monument to Change As It Changes - Zambrano Hall - Stanford University Graduate School of Business




From the artist:

Read more: LA Times

Materials update

So after a load of looking at other materials again I am here:

Styrene – Brittle and carcinogenic when burnt

Polypropylene – Same as the original flaps

ABS – Single sided

Acrylic – Brittle

PETG – Extremely melty

Rapidform said to use 4D Model Shop as a resource for finding out what materials are available but buy else where, I have however found engineering plastics, the plastic shop and eSheet so I think I am okay on that front.

Cut Laser Cut had two new suppliers to look at Arjowiggins (Priplak range) and Bayplastics Ltd. However I think its obvious that PP is the right material, I just need to get the right grade, definitely going to harrass that VitaSheet bloke, I also got another couple of names out of Cut Laser Cut…

One last thing to look into:

Black anodised aluminium – I spoke a long time ago to a company called ML Fabcuts but I never pursued it to a quote. I will contact them, VitaSheet man and possibly Arjowiggins tomorrow.