Four days on from my original blog post my YouTube video has been viewed over 11,000 times and details of my project have spread right around the globe, translated into German, Spanish, Portuguese and French… and others I’m still finding!

I’ve posted the code and the circuit online for anyone else who’d like to try it, but I’ve had a lots of first-time Arduino users asking for a complete “from scratch” guide via Twitter, email and in the online comments.

I guess from a standing start – without experience of coding and/or electronics – there’s quite a lot to it, so I’ll make a start for anyone who wants to have a go themselves.  Click the link to read more…

Where to start

Before opening up the alarm clock (or even buying one) I’d recommend getting familiar with Arduino and the basic circuit.  To get all that working you will need:

• An Arduino (a Nano model if you want to eventually fit it inside the clock). For regular Arduino projects I’d recommend buying an official Arduino Uno, but when space is tight pretty much any Nano clone will do. You can usually pick one up on eBay for around £10.
• A solderless breadboard – around £2 – something like this.
• Some LEDs – pennies each – any old kind… like these.
• Some 330Ω resistors – pennies each when bought as a bag – any kind… they don’t need to be as fancy as these.
• Some jumper wire – any old solid wire will do, but if you’re planning to get into it, invest in a £10 kit like this.

Getting the Arduino set up

Visit arduino.cc and download the latest version of the software, then install it.  If you’re using an Uno you’ll be able to start testing it right away, but if you’re using a Nano you’ll also need the FTDI drivers.  It’s well worth following this guide to get started.

If you’ve followed the getting started guide linked to in the last paragraph you should have the Arduino environment running and the blink example running on your Arduino board. Congratulations – you’re half way there.

Wiring it all up

To put the circuit together, simply follow the layout in the diagram below.  If you’re using a Nano, make sure to leave the bottom row of pins off the breadboard (otherwise the two sides of pins will short together).  Make sure the resistors straddle the gap in the middle of the breadboard.  You won’t yet have the little white chip at the left of the diagram, and you won’t need the ‘battery terminals’ lines coming off the bottom, but that’s fine for now…

Testing the program and circuit

At this point you can upload the Arduino sketch from my original post. If successful, you shouldn’t really see much of anything happening.

Now the fun bit – open the serial monitor in the Arduino software and type the number 3, then press return.  If everything works, 3 of your LEDs should light. Now type 1, 2, 4, 5 or 6.  Each time the relevant number of LEDs should light.

The next step is to run the Python script that will drive those numbers – 1, 2, 3, 4, 5, 6 – from your email inbox rather than directly from your keyboard. Download the Python environment and the pySerial serial port extensions and install both following the instructions.  Then simply run the Python script from my original post (you may have to change the COM6 value to another number depending on your PC, and you’ll have to enter your Gmail user and password) and, with a bit of luck, you should have the LEDs updating directly from the number of unread emails in your inbox.  It’ll probably take a bit of fiddling to get it working, but if you get stuck please feel free to ask me in the comments.

Congratulations, if you’ve reached this point you’ve got the project 90% working.  The chip we left out earlier is called an Optocoupler (aka Opto-isolator) like this and it’s job is to basically switch other stuff on and off.  It does this using light, so there’s no electrical connection between the two devices (so, in theory at least, the clock can’t blow up the Arduino or vice versa).

If you want to finish the project, just pop in this chip where shown on the diagram and hook up the switch you want to open/close to the other side.  In my case I used the wires going to the ‘demo’ button on the clock, but you could just as easily control a button/switch on almost any low voltage (battery powered) device.

Good luck, and please share your results!!!

A couple of weeks ago I was in our local Boots chemist. They were clearing out all the Christmas gifts at 75% off.  I saw this Space Invaders alarm clock reduced to around £3 ($5) and had to have it, even though I had no use for another alarm clock.  (Btw, even though Boots are now sold out, the same clock is available from Play, IWOOT, Menkind and others.)

The challenge was… what to do with it?

In the end I decided to turn it into an ambient device, connected to the internet and capable of alerting me when things happen. Right now I’ve got it firing up with a new message arrives in my Gmail inbox, but really it could do anything – flag up Twitter replies, indicate the chance of rain or even, paired with one of my previous projects, act as an additional ringer for the doorbell.  I’ve added a row of LEDs (three red, three yellow) to give it some more feedback options.

Every time I get a new email the clock moves from side to side and does the classic Space Invaders sound, then lights one of the LEDs to give me an idea of how many unread messages there are.  There’s loads more that could be done with it, but I was happy just turning it into something more useful without destroying it in the process.

If you want to make one yourself, here’s what I did…

I used an Arduino to give the clock some more logic. I didn’t want to alter the external appearance of the clock and space was incredibly tight inside, so I used an Arduino Nano clone that I had lying around.

Update: I’ve since written a post with a parts list and some setup instructions that may be useful to read in conjunction with this.

First step was to get the components working on a breadboard. I ran 6 LEDs off the Arduino’s digital pins and added an opto-isolator to another digital pin to act as a switch. I knew the Arduino wouldn’t be able to drive the clock’s motor straight off an IO pin, so my plan was to simply trigger the clock’s handy ‘demo’ button from the pin and deliver the main clock/motor power from the direct 5V pin on Arduino.

Here’s the wiring layout that I ended up with…

The original power source for the clock was 3 AA batteries producing around 4.5v. I was planning to power the clock directly over USB – I figured 5V was close enough – so the battery compartment provided the perfect space to house the additional components. Unfortunately the battery box wasn’t quite deep enough for the Nano, so my first job was to cut out the battery recess with a craft knife. I left a small lip on the bottom edge to support the additional components and to retain the original motor mounting point.

With a bit of breadboard trimming (with a hacksaw… quick and nasty!) I was able to fit it into the clock…

Next step, the Arduino sketch. The code is currently dead simple: listen to the serial input and if it’s a number, light that many LEDs; if the new number is higher than the old number (i.e. you’ve got new emails, not just cleared out old ones) then also close the ‘demo’ switch to activate the sound and movement. Here’s the finished code.

int ser = 0; // for incoming serial data
int prevCount = 0;
int timer = 100;

void setup() {
 for (int thisPin = 2; thisPin <= 13; thisPin++) {
 pinMode(thisPin, OUTPUT);
 }
 Serial.begin(9600); // opens serial port, sets data rate to 9600 bps
}

void loop() {

 if (Serial.available() > 0) {
 // read the incoming byte:
 ser = Serial.read();

if (ser >= 48 && ser <= 58) {
 int newCount = ser - 48;
 if (newCount > 0 && newCount > prevCount) {
 digitalWrite(12, HIGH);
 delay(100);
 digitalWrite(12, LOW);
 delay(1500);
 digitalWrite(12, HIGH);
 delay(500);
 digitalWrite(12, LOW);
 }
 prevCount = newCount;
 } else {
 digitalWrite(13, HIGH);
 delay(100);
 digitalWrite(13,LOW);
 }
 }

for (int thisPin = 3; thisPin <= 11; thisPin++) {
 digitalWrite(thisPin, LOW);
 }

int topPin = prevCount + 2;

for (int thisPin = 3; thisPin <= topPin; thisPin++) {
 if(thisPin <= 5) {
 digitalWrite(thisPin, HIGH);
 } else {
 digitalWrite(thisPin+3, HIGH); // Because I left a gap in the digital pins
 }
 }
}

The last step was to write a quick script to poll Gmail and get the number of unread messages.  I’ve tried lots of different languages for Arduino serial control in the past – Perl, PHP and DOS shell scripts – but every time it’s been a little bit flakey, at least under Windows. Flash Actionscript is pretty reliable using Serproxy, but that was overkill for this project. So in the end I decided to give Python a go. It was a good choice – reliable and really quick to implement.  Here’s the code…

import serial
import time
import imaplib

ser = serial.Serial('COM6', 9600)
time.sleep(2)

while True:
 obj = imaplib.IMAP4_SSL('imap.gmail.com','993')
 obj.login('GMAIL_USER_HERE','PASSWORD_HERE')
 obj.select()
 obj.search(None,'UnSeen')
 count = len(obj.search(None, 'UnSeen')[1][0].split())

if(count > 9):
 ser.write(':')
 else:
 ser.write(count)

time.sleep(120)

ser.close()

The only bit that had me scratching my head was Arduino’s trick of resetting every time the serial line is opened.  The Nano takes around 1.5 seconds to restart, meaning it would always miss the first mail notification. A wee sleep command fixed that.

And that’s about it. Three hours and around £12 of components (including the Nano clone) to turn an alarm clock into a Gmail notifier.

Future scope

There are still lots of free pins on the Nano, so plenty of scope for expansion. A light sensor could be a useful addition so it doesn’t trigger when the room lights are off, and maybe a knock sensor so it could be set to ‘snooze’ with a quick tap. Apart from that, most other mods would be software-based. I might make the three red LEDs linked to Gmail and the yellow ones linked to Twitter.

If you decide to make one of your own please let me know.  And if you make any improvements to the software, please share!

Update: The response to this hack has been phenomenal – thanks for all your emails and feedback. Here are some of the mainstream sources that have picked up on it.

• Make Magazine: Space Invaders Gmail Notifier
• Engadget: Arduino hack turns Space Invaders alarm into Gmail Notifier
• Ubergizmo: Space Invaders Gmail notifier
• Technabob: Space Invaders Alarm Clock Becomes a Gmail Notifier
• Engadget Español
• How-To Geek

On Twitter:

• The Gmail Team
• Massimo Banzi (co-father of Arduino)

In previous years I’ve just pulled together a picture, some clip art and a Christmas-y tune in Flash to make something like this: http://www.heraldandtimeslabs.com/christmas2010/

…which is OK, but not very exciting.  So this year, I decided to try something different.

Over the space of a week I took around 3,500 still photos around our Glasgow office and Cambuslang print plant.   I then stitched these photos together using a free Mac app called Time Lapse Assembler to produce a pretty smooth 30 frames/sec timelapse video.  In turn, this was edited to match an appropriate Christmas tune in iMovie.  Here’s the result…

Two cameras were used on the shoot.  Most of the pics came from my Nikon D3100 using both an 18-55mm zoom and an 8mm fish eye for the really wide angle shots.  Interval timing was handled by a great little remote shutter release I got off eBay for about £15.  It does everything the official Nikon one does for about a tenth of the price.

My backup camera was a Ricoh CX1 which I used to capture a second angle while the D3100 was in use (like when the drivers were loading their trucks and while everyone gathered for the final “Merry Christmas” shot).

In total I reckon I spent approx four hours shooting the pics (elapsed time was more like 18 hours, but most of the time I could leave the camera and get on with other work) and then about another three hours assembling the footage and editing in iMovie.  So, for less than a day’s work — or about the same as I’d have spent creating another cheesy Santa animation — I reckon we got a much better result.

Producing this video has really given me the timelapse bug, so for Christmas I’m getting a custom-built robotic tripod head which will allow me to combine real tilts and pans with my timelapse photography.  Exciting stuff!

The awards ceremony, hosted by Catriona Shearer, was held on Thursday night at Oran Mor in Glasgow. Congratulations to all the winners and finalists listed here - especially the 999 Design team who scooped three awards and Screenmedia who took the top award for the second year running.

There’s also a wee write-up of the event here.

From Algarve 2011 – the pics Tara didn’t upload, posted by Grant Gibson on 9/18/2011 (28 items)

Generated by Facebook Photo Fetcher

From Algarve Holiday Sept 2011, posted by Tara Gibson on 9/18/2011 (41 items)

Generated by Facebook Photo Fetcher

As in previous years, we ended up with thousands of photographs.  And, as in previous years, we made hundreds of those photos available to buy via our photo sales site.  But the problem we’re always left with is one of discovery: without having the names and email addresses of those pictured, how could we let people know their photo is available to buy?

So, this year we’re trying a new experiment.  As well as putting the pics up on our photo sales site we’re also uploading them all to Facebook.  Users are able to ‘tag’ themselves and their friends in our photos with prizes available to act as an incentive.  All the pics going up on Facebook have a unique reference number which allows them to be found quickly and easily on our photo sales site.

Adding this reference number was actually the trickiest part of the whole process – standard Photoshop actions, even combined with the new Photoshop Variables feature, don’t offer the flexibility needed for this.   Instead, the solution was to script a Photoshop extension in Javascript that would extract the reference number, overlay it on the image and resize the output to a Facebook-friendly res.  What… you can script Photoshop in Javascript?  Who knew?  Not me, until last week.

Anyway, it’s early days but it seems to be working quite well so far.  The technology part was a breeze and should be reusable in future with mimimal effort.

Unfortuantely the Evening Times Facebook page was only five days old at the time of the Great Scottish Run, so we were a bit short of friends to give this a big kickstart.  Hopefully we’ll be able to use this technique again in future to aid discoverability of our photo sales service.

If you took part in the run, go here to see  if you can find yourself or your friends!

What they don’t tell you anywhere on the packaging or instructions (and by implication deny) is that the Wi-Drive works with ANY web and Wi-Fi enabled gadget.

Want to use the Wi-Drive with an Android phone or tablet, Windows or Mac laptop?

Here’s how… Simply connect to the Wi-Drive’s Wi-Fi AP then point your browser to http://kingston.  As if by magic, the main window of the iOS app appears within your browser.  You can browse folders and open files just like in the app…. or perhaps more accurately, just like the internet.

I’ve tested it on a few different devices – phone, tablet and laptop – and it works nicely on all of them.  It’ll even stream movies quite happily from the web interface.

Personally, I think this transforms the Wi-Drive from a super-niche proposition (iPhone users who are stuck for space and aren’t due an upgrade) to something far more mainstream and interesting.

Hosting a personal site from home is a great idea, and something I’d strongly recommend to anyone.  As well as learning the basics of server configuration and maintenance you also get several unique benfits, like:

• You can hook up other devices like weather stations, door bells or ambient Twitter clients
• Storage is virtually unlimited
• It’s pretty cheap to run (I reckon around £20 a year in electricity)

The one problem with hosting from home is bandwidth.  A basic DSL or Cable line is fine for handling modest everyday traffic (this site gets between 300 and 1,000 visits per day), but isn’t suited to handling large spikes.  Hopefully, Google’s Page Speed Service (PSS) will be able to smooth out those spikes and give me the best of both worlds.  So far it’s looking good.

Set up

Setting up PSS is pretty straightforward. Once you’ve signed up and have been accepted, you just need to confirm ownership of your domain (by adding a TXT DNS record) then provide Google with a ‘reference domain’ on which they’ll be able to access your site. The reference domain needs to be something new, because you’ll be redirecting www to Google’s servers.  The final step is to update your CNAME record to redirect your www to Google’s page speed domain, then it’s just a case of waiting for DNS propogation.

The process was pretty painless, taking less than half an hour even though I was being super-cautious throughout.

Results

Initial results are pretty impressive.  Independent speed tests suggest an improvement of approx 50% in page delivery times and that’s while my server is otherwise quiet.   I suspect the difference will be much more noticeable when my Cable line is busy downloading the latest multi-gigabyte OSX/PS3/Windows 7/Wii/iOS update.

Issues

I had a few teething problems with specific pages on my site.  My weather station graphs showed garbage data and the lightbox effect stopped working on my photo gallery.  I’m sure it would be possible to view source and debug where it’s all going wrong, but for simplicity I added the weather page to Google’s handy page speed blacklist which stopped them trying to process that page.

For now I’m going to put up with the missing lightbox on my photo gallery.  Those galleries are only there for pre-2010 content anyway – more recently I’ve been using Facebook and a handy WordPress plug-in to import pics back onto my site.  Plus, I’m more interested in Google taking the strain of image delivery than I care about the lightbox effect.

Final thoughts

I’m really impressed so far.  Every test I’ve run on the site suggests the pages are far better optimised and download faster.  For a completely transparent, coding-free solution that’s a great result.

My only slight concern is over Google’s as-yet-undisclosed pricing for PSS.  I like that my site is faster, but it’s not something I’d be willing to pay a lot for.  Hopefully pricing will be based on usage with a free entry level for personal sites like mine.

Generated by Facebook Photo Fetcher