Rage against the machine

I was expecting a delivery of a mattress today from a large retail chain.

When I ordered from the store I requested a specific day for the delivery and noted the time in my calendars. I kept telling my family that it would be delivered today and even prepared the space by taking out the old furniture to be ready. Lots of emails from the retailer including earlier today an invoice, and I noted that the money had moved from my bank account today as well. I was prepped and ready to receive!

Cue waiting at my window at the appointed time. With a sore back from an earlier exertion I had my son on hand to help with the lifting and shifting. Towards the end of the delivery window a large truck pirouetted around on the street and pulled just past my driveway. Woo hoo! son alerted, I opened the front door and put on my best delivery man smile … only to watch as they took out an mattress and walked around the corner and into someone else’s driveway! What, I thought, perhaps they have the wrong address – and my anger was confounded when they appeared with plastic in hand from unwrapping the mattress somewhere else!

I watched them drive off, any slower and I’d have been over the road accosting them and asking to see the delivery invoice. Back into the house and searching Google for things like “can I keep an item if it is misdelivered to me?”. Ready to stalk around the corner and ask a neighbour “have you got my stuff?” On the phone to the store asking what was happening, spouting off to my son and beginning to feel really passive-aggressive! Finally got through to the store and gave them my order number.

My delivery wasn’t today.

Apparently in the 10 or so emails which they’d sent through they’d moved it back 4 days. I’d skim read all of them including the three or so PDFs and missed the one where 14 ==> 18, at the same delivery time.

I was the one who was wrong, but I had plenty of ammunition to prove I was right!

It shows me how often incidents are caused by self-righteous people who think they are right, and perhaps even are right on occasion, but for whom that ‘rightness’ justifies all types of reactions. I’m glad I didn’t tell the delivery guys they were idiots, or bang on the door of my neighbour (although I might have met a new one that way!). I’m glad that I didn’t prove to be a fool to a greater extent than merely watching them and then calling the store, but I wonder how many other people around me today are encountering misunderstandings which escalate into angry words or deeds rather than being clear up?


Creality Ender 3 bed leveling

I’ve brought a Creality Ender 3. This is one of the cheapest 3D printers available and as many people were giving it good reviews on YouTube, decided to take the plunge at £165 on eBay. It was shipped locally from the UK and arrived quickly.

First impressions on unpacking were very favourable. All the parts came neatly labelled and I especially liked the small plastic bags with extra spare parts for screws, washers and other attachments. They also supplied a full set of tools including side cutters, hex spanners, zip ties and even a platen scraper to free models from the bed. This is a quality product and is very complete for the price.

But the bed levelling is a disaster.  Full disclosure – I may not know the slightest what I am doing, and all this could be a noob’s series of mistakes.

As I understand it, bed levelling is either done automatically with presence or induction sensors on higher-end machines, and manually on lower-end machines such as the Ender 3. While there is an after-market 3rd add-on available, I only have the stock machine. The whole idea of levelling is to make the plane of the print head x-y axis’s parallel to the plane of the printed material. This is done by adjusting the bed to within a paper’s width of the print head, by moving the print head to the four corners and moving the bed up or down.

I did this, and could not ever get it level. There was never enough play length on the bed screws to adjust the amount needed. The issue seems to be that the side z-stop sensor is too high – even the installation instructions mention that it should be “~32mm” from the bracket bottom to the underside of the machine – however there was a lug on the printed bracket which would not let it descend to that level and held it at 35mm height.

  1. I removed the lug and set the z-stop to the suggested 32mm above base level. The bed was now too high and the auto-home feature would have broken the glass platen if I had not switched off the machine. IMG_20180809_110853
  2. Still not enough play, so move the z-stop higher. I can adjust the front thumb-screw enough and I am hopeful (- and this would work with the stock epoxy bed cover, not the glass one which is thicker). Still not enough play in the back thumb screws, especially the left rear. Heated bed electrics are in way. That spring is at the tightest depression while the nozzle is touching the glass. InkedIMG_20180809_110940_LI (2)
  3. So I removed the rear heated bed bracket, remounted without the bracket and now I get a lot more travel in the height at the rear. That leaves the heated bed electric exposed but I can live with that for a while.sdr

Much better. Now on to levelling the bed with the paper shim and manually moving the print head to the four corners.

I can either build a new support for the 24V electric line using Sugru or more likely, once I come up to speed with TinkerCAD and Cura or whatever software I start using, I can 3D print a new one.

Just One More Year

Financial Independence, Retire Early (or ‘FIRE’) is the domestic monetary policy that says when you have enough, you stop working. After all retirement is just earning money by other means.

I’ve gone the FIRE route and are stopping regular, salaried work. While I’m not sure that this will mean stopping all paid work (‘retirement’ itself being a concept that is changing), it does mean stopping any notion of a ‘career’ and simply getting money through anything that works – barring illegal activities of course!

A corollary of taking early retirement is the “just one more year” syndrome where you feel that you would like to stop working but think that one more year of a regular salary can help you pay off the mortgage, get the children through university, create a nest egg, or whatever. I encountered a quite few colleagues who felt this applied to them. Once I started planning and saw that my outgoings could be controlled and that I could achieve what I felt was a comfortable existence the idea of trying to make extra money seemed ridiculous. When would enough be enough?

You’ll never get as much income as when you are working and if you really cannot make your outgoings be less than your incomings, then you will never be able to stop work. As someone noted usually the event that gets you to stop thinking this way is something like a big workplace bonus, a buyout or redundancy offer, a friend’s health crisis or death, or a personal health crisis. For me it was a mix of redundancy offers which seemed to come around every couple of years (I personally sat through about five over the last ten years) combined with some maths that showed me I could survive outside the system. Once the next offer came around, I jumped at the opportunity and was the first to apply.

But it took guts to stop the job I had continuously for 25 years. I still have a very small mortgage (which will be paid off from my redundancy lump sum) and my grown up children still haven’t fledged yet. I walked back from posting my papers with tears in my eyes, and are still washing those decades of work out of my hair.

But the sheer energy and happiness I have now is amazing. I’ve helped with a week-long children holiday club, helped Syrian refugees in Europe, and gone on a missions trip with my local church. I’m looking at helping build houses with Habitat for Humanity, and exploring other things where my skills fit with a need. Every day I look forward with expectation as to what I can do and create without any feeling that I have to do tasks.

I’ll close with a quote from one of the early retirement forums:

“It’d be especially sad to work longer for more money while missing an opportunity to do the tremendously satisfying things that don’t happen to require much money at all.”


As I work more in my garden shed some of the conveniences of the house become apparent: heating, power, internet connection, cups of tea, and doorbells. I realised this last one after missing a couple of deliveries which while they were expected, weren’t expected at exactly the moment I stepped out my back door.

The problem isn’t the doorbell, the problem is that having a doorbell convinces you that you’ve answered the question of knowing when someone is at your door.

Nest are now supplying their WiFi doorbell in the UK and it looks a very capable device, and Amazon also do the Ring video doorbell.  While the price is high there is a lot of function provided and they variously do facial recognition, storage of video for review (with a subscription), presence detection, and letting you know when someone is at the door. That last one is strange: I run design thinking workshops and one of the exercises involves asking people to re-imagine the door bell. First exercise begins with: “draw me a doorbell”, while the second is exactly the same but begins with “draw me something to tell me that someone is at the door”. Needless to say the second elicits a wider range of devices including dogs, pressure mats, windows, and pull bells. But doorbells these devices are not: their primary functions are to enhance the door arrival experience so that we can either speak to the person there, let them in remotely, or see unwanted visitors before they see us. So why am I leery of wifi doorbells?


Power is the enemy of remote devices. At least one of the current crop of internet connected doorbells needs a wired connection and this smacks to me of a fudge. My bottom line is that I might as well wire in a press button and a Raspberry Pi with video camera via power-over-ethernet as load such a big device on my doorstep. This aversion stems from my earlier work with power budgets on embedded devices and how it compared to other devices running in the ISM band, some of which operated for up to 2 years on a coin battery! Astonishing to consider that SBCs like the Raspberry Pi only operate for a day or so on a massive 13Ah battery pack, and even lower powered devices such as the Pycom LoRa struggle to get micro-amp-hour range while sleeping.

433MHz/868MHz devices, on the other hand, can exist for much longer on a smaller power budget – my Friedland front door chime runs for years on a coin cell battery. While I’ve played with ISM band devices before, their OOK or FSM modulation is difficult to understand unless you have a spec sheet. Long ago I had a set of devices from a now-defunct small electronic company which used an embedded protocol called LLAP and a cyclic wake mode which worked perfectly for my variety of monitors. Sadly, I am resurrecting the whole setup using more functional, but more power-hungry LoRa devices.

While there are other attempts such as DoorPi and some convoluted GitHub repos that write everything in Python, including event handling and streaming of video, I want to simply use my NodeRed installation to slave off the door press and everything else can happen from there. From my central NodeRed running on an Odroid SBC I can stream video, send a text message, speak back to the caller, switch on lights, and so on.

Question: how do I slave off the wireless ISM Friedland door chime?

Commercial Alternatives


The Ring doorbell by Amazon is an interesting alternative. It has the advantage that it is wire-free, but some of the reviews warn that it only trickle charges while connected through a wired extension and that ‘excessive’ number of people walking past can drain this – for which read 6-7 people max! That means a busy footpath like our road would very quickly drain it, so a non-starter. Also, lots of reports about delayed triggering of the phone application and


I like the Nest ecosystem, it handles well and rarely gives any problems. I have the Nest thermostat and their fire alarms and while the new Nest doorbell is currently retailing in the UK for around £250 and looks good, the base requirements specify a wired chime unit with exposed lugs on which to wire the leech wires from the door press. That makes it a no-starter for me as the initial cost of installing another wired door chime would be higher still. Perhaps if they do a wire-free option


While there are others out there and some pretty good CCTV combined monitors and door releases, I wanted to attempt my own because, well – because! I have plenty of experience of using radio frequency ‘stuff’ and while I’m not an embedded microprocessor engineer, I will shortly have more time on my hands to complete projects.

I’ve started by using some Arduino code on a Wemos (- now called ‘Lolin’) D1 mini, which is an ESP8266 encapsulated on a board to expose the GPIO and powered by micro-USB. This can be programmed using the Arduino IDE and I’ve copied some code from Jamie Jones’ doorbell project. My sounder or chime is wirefree and links from the Friedland door press by using the ISM 433MHz radio band. Getting this part working was quite simple and a matter of adapting the code to local environment, and putting it into an enclosure – which is still to be done.

Next is to work out how to sound the chimes triggered by an MQTT message on a specific topic.

Extending WiFi to the garden shed

I’ve purchased a pair of Ubiquiti Litebeam M5 network dishes to extend my house wifi to the garden shed.


Normally you’d only need one of these either to send or receive the broadcast from the main house, as my shed however is made of metal it tends to act as a Faraday cage and shields most radio frequencies, hence the pair of CPE devices. I have these set up so they point directly at one another and connect via the proprietary AirMax protocol which speeds things along as it is a layer 2 link (which means that broadcast and other packets normally inspected at the bridge boundary are passed as-is). While Ubiquiti do provide an excellent management server (UNMS) running in a Docker container, this is supplied only in a x86 image and won’t run on my always-on ARM SBCs.

These M5 are nice units because they are cheap, simple to assemble and operate, and gain up to 23db over a simple omnidirectional antenna.


Careful positioning is needed, and when they are talking to each other directly the signal is taken into my shed using normal ethernet Cat6 cable – although for professional and hardier situations there are toughened cables that you may use. This signal is fed to a Vigor 800 access point and spreads around the inside of my shed. Ethernet connection to the AP is also possible which gives really good latency response for playing games. I’ve measured 18ms response on a speed test which is pretty good considering how many bits are in the way, and the cable/radio/cable arrangement back to the main modem. I now have a number of managed switches and ADSL devices spread around my house, which is pretty good considering I don’t run a lot of servers.

Interestingly, the M5 are PoE meaning that they only have one ethernet cable into the actual radio ‘horn’, and further back you supply power and a connection to a PoE box – passive in this case – which feds 24 volts along the ethernet cable to the assembly. I also use this for my LoRa antenna in the loft as it means you don’t have to have power so close to the installation. They work well and my multi-protocol house is rapidly gaining an assembly of antennas!

(images taken from Ubiquiti web page)

Life Skills

As the world progresses the necessary ‘life skills’ needed in each era seem to change. The simple ones like being able to drive a car or horse, being able to read or use social media and understanding the difference between a cursor and and screen pointer! But even those get very old, very quickly.

Take, for example, the analog clock: the clock face with 12-hour divisions came from the ancient Egyptians with the the duodecimal system based on finger joints or lunar cycles. The skill based on being able to interpret the clock face is diminishing, and no-one needs it except those taught many decades ago. “Teach children to spell” yell the older generation while fumbling around on a keyboard and ignoring social media. “Make the remote control easier” grumble the same folks, while most younger people use their mobile phones to control their home lights, order meals, go shopping or find the next public transport. Skills change far more rapidly than a codified education system can keep up.

Take typing. If the only time you extensively have to use a pen is during an exam, surely you are training people for … the wrong things? Post university most people type things versus using a pen at almost a 10:1 ratio, yet we take many years to teach writing first. Teach them to drive a car just when autonomous cars appear and you’ve done the same thing as encouraging weaving just before the Industrial Revolution, or spelling just as spell checkers arrive.

We need to urgently identify those skills that are universal such as relationship management, ability to learn, adapting to change and fundamental things like self-learning and being curious. To paraphrase Pink Floyd, “we don’t need no prescriptive education”.

Life is a Journey, not a Destination

I’ve always had an itch to try out new ways to get somewhere. I guess I got that from my father; he’d be admonished by mother for wanted to try ‘shortcuts’ when driving which often resulted in back-tracking or otherwise ending up lost – and when lost means the middle of the Australian bush that is very lost indeed. His knowledge of remote country roads was only matched by his seeming disregard for transition from single man into family patriarch and the responsibility that entailed. I don’t want to paint a picture of someone who was frivolous or ignored safety, but at his heart there lurked an element of wildness.

So too for me. Before GPS, Google mapping every road in the world, and mobile phones I would study maps and work out that there may, just may be another way to get from A to B. I really enjoy the path less trod, understanding a new contraption from more than just the instructions, and trying new experiences.

Stoke left, stroke right

I have for many years enjoyed a personal computer from Apple. Having used hundreds, if not thousands of computers during my time in IT, I fully appreciate the fundamental advantage of a Macbook: it just works.

I’ve now passed this older laptop to a member of the family and are now using a newer work machine, but need to keep both work and leisure separate – so I run two different browser windows with the ‘left’ one covered in employment related tabs, and the ‘right’ one open with personal items such as my home email. The method seems to work although I too often find myself at the wrong end of the spectrum, so to speak.


It reminded me of some of the dating apps that use the ‘swipe left, swipe right’ idea of approval – only my swiping is replaced by the MacOS idea of four-fingered strokes left or right on the enormous Macbook Pro touchpad. And oh – the Pro is a very enjoyable type of machine.

Home broadband

I’m really impressed by my internet provider, Andrews and Arnold. I’ve used the internet since before it was named, and Fidonet, 1200-baud modems, bonded twin ISDN lines and the whole shebang. Today’s FTTC and ADSL are fast and simple in comparison.

However the last physical leg is often the source of issues. Trees rubbing on telephone lines, storms, and cable modems losing sync when it rains are all part of that ultimate physical connection to my home. Even mobile data via 4G gets rain fade in a heavy rainstorm. A wet piece of string will still transmit a signal, but it will be pretty weak.

When I speak to my sister in remote Australia it seems that they are lucky to get even a mobile phone signal when standing in a particular corner of one room of their house. There is no internet via land-line, and often reasonable internet connection is only available from her office in the town. Netflix and streaming movies, Facebook and the like are all distant, unreasonable realities and for many people in developing nations. So my little worries and 80Mb per second seems like luxury. My quota is 1Tb per month, but I rarely reach that. I am also served by cable and could use that but the latency is too great for good gaming.

So I’m pretty happy with my internet connections!

ctypes and uctypes

Having obtained the raw protocol description from the vendor, I confirmed many of my earlier observations and now have a working system which can read the information. Like many programmers I approach the problem in a step-wise manner by getting each element working in turn and refining my solution. These are the steps I took so far:

  1. Obtained and assembled hardware, including a XY-017 TTL converter in a USB stick. This is used to read the incoming RS485 protocol.
  2. Attempted to read anything from the USB interface. I developed a program (‘disp.py’) to do this and the results are in the Git library.
  3. Cycle on #3 above until satisfied.
  4. Started work on a simple sending program on a Pycom LoPy which transmits to my LoRa antenna in my loft. I’d previously got the whole LoRa thing working hardware assembly and work on connecting to The Things Network.
  5. Vacillated for a few months. Hey, this is a hobby!
  6. Created a program to read the incoming message from MQTT as The Things Network publishes all incoming messages on a MQTT broker, and by subscribing using the appropriate keys you can read the raw messages.
  7. Decomposed the messages by using ctypes library on the CPython implementation, and uctypes on the MicroPython implementation.

This latter stage wasn’t so easy, as having muddled through to gain an understanding of how ctypes works and mapping the fields, I then had to learn a new way of mapping bit structures since MicroPython tries to be helpful by removing all the challenges of endian architecture in their implementation and it works different to CPython. It took a good chunk of time to get the normal Python working, and then realising that I had to grok something that looks completely different.

I’ve written some small-n-dirty test programs to make sure I understand this new aspect of MicroPython, and then I will implement it in my actual program running on the LoPy. I also need to create an accumulation as not all data will be sent immediately, and also turn off WiFi radio and set the IoT device into low power state. Lots more fun along the road.