Amazing facts to impress your friends....

If you want to really impress someone just tell them this:

“If you want to use MQTT with PubSubClient and talk to an Azure IoT Hub over secure sockets you should use Version 2.4.2 of the LOLIN Wemos framework or your program will crash when it runs out of memory during socket open.”

Then again, it might just be me that would be impressed by this.

Building an Air Quality Top Hat and Letting the Smoke Out

I’m building an Air Quality Measuring Top Hat for my Red Nose Day Lecture in Rhyme next week.

I think I’m the first person in the world to do this. No idea why.

Anyhoo. I’m going to equip the hat with a bunch of coloured leds to indicate the quality of the air around the wearer. Yesterday I started building up the hardware. I was driving all the leds with one of my trusty Wemos 8266 devices. Half way through my first test of the leds I noticed smoke rising from the device. Now, I don’t know much about electronics (obviously) but I do know that all electronic devices are actually powered by “magic smoke”. I know this for one simple reason.

If you let the magic smoke out of a device, it stops working.

It turns out that I hadn’t turned on my main power supply for the leds, so all the power was being drawn through the Wemos device; specifically the blackened and unhappy looking component above. Which got very hot, let out its magic smoke and stopped working.

Oh well.

The good news is that the Wemos devices only cost around two pounds each. I’ve ordered ten more…

Sponsor Me for Red Nose Day Madness

I’ve just sponsored myself for my lecture in Rhyme next week. I’d love it if you did too. Click on the image above to go straight to my donation page.

If you can make it to the lecture I’d love to see you. It should be fun and there will be FREE DONUTS as well as lots of the usual fun and games.

You might even learn something about air quality, which is what I’ll be rhyming about.

The lecture is at the University of Hull on 15th of March (Red Nose Day) at 1:00pm in Lecture Theatre A in the Robert Blackburn Building. David from c4di-Eagles Labs is bringing the donuts. I’m bringing air quality sensors and maybe even a light up top hat. As you do.

You can find out all about my illustrious Red Nose Day history at my special red nose day page.

Connected Humber MonitAIR Circuit Board

There comes a point in a project where things start to look real. I think we’ve hit that in our Air Quality monitoring project. At the Connected Humber hardware group this evening Brian gave out some populated versions of the sensor circuit board he’s designed and had made. Rather than making all the individual connections by hand we now have a beautiful little board that takes our processor, some power supply and switching circuitry and sensor connections.

It’s very pretty. The next step is to marry it up with our housing designs and start putting sensors into the hands of our waiting test users.

Exciting times.

Sensor Base Design the Hard Way

If I ever join you climbing a mountain, I’ll probably be the one wearing roller skates and carrying a piano.

Sometimes it seems to me that I have a stubborn determination to do things the hard way. For example, today’s task was to create some designs for an outdoor air quality sensor. I wanted a circular fitting to go in the base of a pipe and a support that would attach to the fitting and hold the processor and air quality sensor.

Some people (I think I call these people the “sensible ones”) would reach for their CAD tool and begin designing. Not me though. I reached for my Python interpeter and the ezdxf library that lets you write Autocad files from your code.

I wanted to use the “nut holder” approach to fitting where the laser cut plastic (which is what I want to make the stand out of) holds nuts that a bolt engages with to grip it. I’ve used these on the Hull Pixelbot quite successfully. I was quite proud of the “nut holder generator” that I’d written for the Pixelbot, it even makes each change of direction into a tiny curve, which apparently reduces the chances of the plastic cracking after it has been cut. However, my original code only drew the holders in one direction, whereas for the holder I wanted to put them all around the edges of a circle, as shown above.

Cue lots of bits of paper covered in drawings of triangles, sines and cosines and cursing over code. Eventually I managed it though, and the next step will be to get the plans cut into plastic so that we can make some sensors.

DDD North was super awesome

The best conferences are the ones where you go and learn a bunch of useful stuff and also like to think that you’ve told a few folks useful things that they didn’t know.

DDD North yesterday was one of the very best. I was blown away by the quality of the sessions, the enthusiasm of the audiences and the sheer good humour of the whole event. I learnt a whole slew of new stuff; from IoT development tools that you can run from you laptop through Mob based development techniques, some nifty .NET library tricks and a lovely take on how to use generics in C#.

I even presented a session of my own which I hope taught people a few new tricks. I was so engrossed in my bits and bobs that I totally failed to take any pictures (which is most unlike me and a measure of the quality of the occasion).

And to make it even better, it was in based in my favourite university and old stomping grounds in Hull, so I just had to get up, grab breakfast and tootle down the road to take part.

Hats off to the organisers for making it all work, the sponsors for paying for great food and a lovely setting and the speakers for taking their time out to spread knowledge. You could pay an awful lot of money in conference fees for an experience nowhere near as good as this one.

If you ever get the chance to go to one of these in the future, you really should.

DDD North is tomorrow and I'm presenting

I’m getting rather excited. It’s DDD North tomorrow. Four tracks of excellent computing content crammed into the lecture theatres in Hull University. A whole day of splendid sessions.

And one from me.

I’m on at 2:30 pm in Lecture Theatre A in the Robert Blackburn Building talking about Azure Functions, Air Quality and a bit of LoRa. I’ve just finished the slide deck and the demos. Should be fun. If you happen to be on campus at Hull you really should come along. I’d love to see you.

Azure Functions with Azure IoT Hub

I really like Azure Functions. I love the idea that you can write pure code that sits in the cloud and just runs. Sometimes all you want is a thing that takes in some data, does something with it, and spits it out somewhere else. You don’t want to build and manage a server, set programs to run on startup, bind them to events and so on and so on. You just want a piece of code with a name and an address.

I’ve been using Azure Functions to capture MQTT messages and store their payload data. This turns out to be extremely easy to do, until you deploy your function in the cloud. Then it stops working. And you scratch your head and wonder what stupid thing you’ve done. You scratch your head, turn on lots of tracing and eventually find that the function is failing on load.

It turns out that one of the configuration items, the connection string to the IoT Hub that the program is using, is not passed over to the settings in the cloud app. You have to manually create the setting for the application and then set it to the required value. Then it all works. Wonderful.

Hololens 2 and Kinect for Azure

Microsoft have been busy. And yesterday night they announce what they had been working on, with a new Hololens device and a cloud powered reprise of an old favourite.

The new HoloLens looks lovely. It is apparently more comfortable with the processing power (which is completely contained in the device itself) at the back of the device to improve balance. It also has a neat “flip up” mechanic that makes it easy to get out of the augmented reality it gives you. They’ve also announced a major improvement in the single, principle issue I had with the original device. The augmented display (the computer generated bit that is laid over your view of the world) is now around twice the size of the original, vastly improving your field. of view. Which makes the virtual elements much more impressive. And, as a final flourish, the Hololens now also includes eye-tracking, so the device can work out where you are looking, which should greatly ease interaction and reduce the amount of hand waving you have to do to attract its attention. It looks like another wonderful device that I can’t afford, having a business oriented price tag of around 3,000 dollars.

However, with a bit of luck it might encourage lots of people upgrade, leading to a flood of cut price Hololens 1 devices going onto the market. I wouldn’t mind that at all…

The second big announcement was of the return of the Microsoft Kinect sensor, at a slightly less eye watering (but still clearly business focused) 399 dollars. I loved the original Kinect senor so much I wrote a book about it. For the second sensor I created the awesome Carbonite machine that used to sensor to make a 3D printable object of you embedded in carbonite. The first Kinect was, at the time of release, the fastest selling peripheral ever. Microsoft had high hopes for its successor, in fact for a while it was impossible to buy an Xbox One without getting a Kinect as well. Unfortunately the device never really lived up to is promise on the home front, and the Kinect sensor was quietly retired a while back.

But industry has always found the KInect very useful, and so Microsoft has brought it back. It has the same depth sensor as the Xbox One version, which is fine by me, but that is paired with a much higher resolution video camera and a bunch more microphones. It’s a great way for generating spacial data that can then be processed in the cloud. I’m really pleased to see it back again, and might even see if I can afford to get one to play with.

Presenting with Surface Go

If you look very carefully at the audience picture for my AI Frenzy talk on Thursday you’ll notice that in the bottom right hand corner you can see a tiny slice of my Surface Go. I’ve used it for presentations in the past and I thought I would again. This time I was running PowerPoint, Visual Studio and the Machine Learning app that I was demonstrating. It worked just fine. I was very keen to show people that you can use such a lowly machine to deliver presentations and it worked just fine. I love this little machine. It is properly portable and properly useful.

AI Frenzy at c4di

We had a Barclays AI Frenzy with Codepen Hulll tonight. There was plenty of frenzy in my session, where I built a complete Machine Learning application in around 12 minutes. There were also a bunch of other great sessions from Sherin Matthew who gave a splendid overview of the field, Aparna Garg who showed a lovely example of an advanced AI application and Neil Gordon who talked about research at Hull University and how companies can work with it. 

If you want to see my sample application and the presentation you can find it on Girhub here.

Red Nose Day 2019 is live

II’m doing another Red Nose Day lecture. This time I’m talking all about Air Quality. In Rhyme. Just because it’s important doesn’t mean that we can’t have the occasional (and I do mean occasional) laugh when talking about it.

I’m going to talk about good air and bad air, and then describe how you can get involved and build your own air quality monitors and use them as part of the awesome Connected Humber Monitair project. It might even be interesting and useful, as well as raising some proper cash for great causes.

And David from Eagle Lab Hull is going to bring along some donuts.

As usual, entry to the lecture will be free, but you may have to pay to get out. If you can’t make it to the event then you can still contribute here. And there will be a video. Oh yes. And if I hit my funding target I will wear a tutu. Again.

You can read more about my illustrious Red Nose Day career here.