That time you almost electrocuted yourself

9 Feb

The TLDR version (6 minute read)

When I almost burn, blind and deafen myself due to a battery short circuit, I change tack to home automation. Prioritising learning and iteration over detailed design and execution.

In 2013, I was on a journey to building my own electric car for fun, or at least a bike-path based prototype before scaling to road-scale. In February 2014, I made my first weld as I assembled the steel spine of the 3-wheeled recumbent trike frame. This was something I felt relatively capable of doing as a mechanical engineer and I had even sifted through the ‘Australian Design Rules’ to determine the requirements of the street-legal version. What I didn’t anticipate was that the mechanical and structural side would be the least of my problems.

In August of 2014, I took delivery of the last component that would enable me to test my concept for control of a 48V electric motor connected to a rather large battery (approx. 20Ah). I meticulously connected the many control cables between my 2kW electric motor, the power controller and my large battery, following what documentation I had been given from my Chinese suppliers. As I went to connect the final electrode to complete the battery circuit in the garage of my parent’s house, I was rocked by the loudest sound and the brightest flash of light I had ever seen or heard as hundreds of amps arced between the lead I was holding and the motor controller electrode. The copper eyelet evaporated, exploded and moments later, the battery management system caught fire. In my stupor, I raced to find something to safely smother that small flame as I began to realise that my long design and procurement activity was literally going up in smoke.

In the following hours, I poured over the electrical schematics and checked my connections and I could find nothing wrong. 7 years later, I can’t remember the precise failure or assumption made but I do remember it’s effect on me.

It was a really humbling experience and the learning was three-fold:

Always understand the fundamentals and be sceptical of a ‘black box’ component when it acts in a wider complex system
De-risk big procurement steps by building smaller and earlier, learning more with less investment
Always wear safety glasses (this habit probably saved me getting a blob of copper in my eye)

*Crimping the high-current eyelets on to the power circuit before moving to the garage for assembly*

*Re-wiring the controller cables to interface with my 3rd party 48V 2kW BLDC motor*

*The hotly awaited fuse board that should have prevented just this sort of arcing hazard. When mislabelled and without adequate questioning of the black box, it introduced it.*

1. Scepticism of ‘the black box’

At the end of the day, I had short circuited a ridiculously large and powerful battery by abrogating my responsibility to learn to a 3rd party that had a higher expectation of my knowledge. I thought I could pay the money and feel safe to just follow the instructions without learning about why the manufacturer made design choices that I accepted at face value. I had believed that trust in the black box would save me but by wiring the black box backwards, it would come back to bite me.

This experience drove me to more deeply understand power control, semiconductor behaviour and the design of battery management systems. These are things I should have done at a smaller scale and perhaps I could have avoided my expensive mistake.

What do I do 7 years later that is a manifestation of this experience? I hunt for the next biggest assumption. The next deal-breaking assumption that is just below the surface but that has been covered up by layers of confirmation bias and goodwill. This learning has been what holds me back from always taking the ‘high level’ view as a basis for substantial and naïve investment. Instead, I choose where to momentarily and cheaply go deep to test the high level assumptions that will later be the deal-breakers. This way, when I zoom out to the high level once more, I have a clear view of the investment risk and know what to look for regarding it’s performance. I am active, not a passive investor.

2. De-risk by building smaller and earlier

When pursuing a hobby project like this, I was seeking the final outcome: An electric vehicle that would get me to Uni with less operational effort than driving my car. 7 years later, I realise that the most valuable outcome from the process was the learning along the way. I had jumped from a small 5V control platform to a 48V system assuming that detailed up-front design would protect me from failure. This was a relatively complex system with enough unknowns that I should have known better but this is in retrospect. How would I have done it now? I would have acknowledged my next biggest assumption and would have devised an experiment to test it and learn. Now, I appreciate that this is what has resonated with me regarding my affection with design thinking and agility mindsets and the tools employed there to deliver early and often, testing the next biggest assumption to deliver a pleasing customer outcome.

I can actually pick the moment (May 2013) where I decided that I should up the game from a Lego based prototype to one that would involve welding and high power batteries. In retrospect, I can see that I was conflating two topics - steering control and scale where I ended up spending most of my money on the drive-train/motor rather than spending it on the thing that I wanted to de-risk in the first place, steering control. By going for the ‘big shiny’ thing, I blinded myself from my actual goal.

*Striving for the pretty things before de-risking the important things*

*My Brother in Sweden even 3D printed the badge I was going to put on the front*

*What the chassis would have looked like if my path had not changed in response to new priorities*

*When I look back at this control model that I built, I am amazed at both where I got to and also how far I’ve come in the space of microprocessor control in the practical world*

3. Always wear safety classes

If I had been holding the battery lead a little lower, I probably would have badly burnt my finger but if I got a ball of molten copper in my eye, that could have been chronic. This is one of the first times where a safety habit really saved me. By donning PPE and choosing a fire-proof area to assemble the high risk components (the garage), safety risks were well mitigated. Of course, I didn’t want or expect the battery to catch fire but the potential was there, I mitigated the hazard risk and it worked out ok at the end of the day. This is the experience I refer back to when I think of when not to compromise on safety.

What was this all about again?

Much of my engineering degree taught me that planning and detailed design is good. This works best where certainty is high, knowledge is complete and there is a pursuit of waste minimisation in the delivery process.

Much of my physics degree taught me that while we have quite a complete view of the world at a fundamental level, to venture in to the unknown, we must adopt the scientific method. When certainty is low and knowledge is incomplete, we must declare a hypothesis and not a statement of fact. We then devise methods to learn to the point that our hypothesis is supported by observable fact.

This article and reflection is about appreciating the difference of both of these mindsets and how in this case where there was a gap in knowledge around 48V power control. I chose quite an ‘engineering design’ path where in fact I should have, at least in the space of this unknown, taken more of a ‘scientific inquiry’ approach.

Changing course to home automation

In reflection, this experience, as I transitioned from university to work life, was quite formative of my belief that I should be more of a scientist than an engineer in the face of uncertainty.

With a busted battery and a realisation that I didn’t really know what a MOSFET (a key component in most power control systems) was, it was time to switch course and pursue learnings in a new and cheaper field. When I started work at General Electric, I drove down roads that didn’t trigger my need to invent an alternative mode of transport. These two elements drove me away from the electric vehicle and instead, I changed focus to the area of home automation. This was a space where I could rapidly learn about the electrical space that I had ignored to my own peril before. Components could be sourced cheaply and the internet was full of content to enable rapid development.

With the release of the ESP8266 in late 2014, a cheap WiFi enabled microprocessor, I entered the world of hardware programming of Internet of Things devices. This pursuit has taken me to a place where the knowledge is actively plugging gaps in my current workplace as WA undergoes an electricity transformation. This hobby has been worth the investment of time and the learning has been terrific.