3 min read

19. Engineering: How To Burn Your Sensitive Electronics

It's easy, really. Let me show you...
19. Engineering: How To Burn Your Sensitive Electronics
Photo by Nicolas Thomas / Unsplash

I love engineering. I love learning, creating and seeing my work function. By and large, engineering is enigmatic. We expect it to work readily, only to find out our assumptions were all skewed. I did that this evening. I was working on a methane sensor for my winning NASA/CSA Canacompost project. This sensor is specific in that it cannot be soldered: remember that as it will become important later. To mount my sensor, I created three different prototypes of a 3D printed casing, illustrations below.

This sensor is treacherous. On the left is the casing I designed and 3D printed, and on the right is the sensor.
This sensor is tiny and mean.

The final prototype had the right well depth for the spring, the tolerances were just perfect and the sensor fit snuggly. The sensor has three protruding legs, you push it into the casing, twist it, and release it, the spring pushing it against electrical contacts. And here lay the problem. After creating the driver (see diagram below), I could not get any signal out of this thing. For those who don’t know, the signal is the analogue information the sensor spits out, from which you can derive whatever it is you are measuring...in this instance, methane gas.

A Wheatsone bridge circuit to drive the sensor.
First meandering breadboard prototype. No signal from the sensor.
Second soldered through-hole prototype. Still no signal from the sensor. This is getting fun.

Eventually, I realized that I had a short within the sensor housing. I fumbled for hours on it. I have actually been working for tens of hours on this sensor between understanding the datasheet, buying all the parts and prototyping the 3D printed housing. The sensor itself costs $60 bucks. My time is worth more than that. And hence, you have probably guessed by now, I decided to hell with it and soldered the sensor onto a PCB. Yep, I did exactly what the datasheet said not to do. Mind you, my iron was at 450ºC and I was in contact with the pin for less than 0.2s. There was no way to create a heat sink to protect the sensor as the legs are so short. I did cross my fingers, so that probably helped.

My heart rate when I realized I had burned the hell out of my sensor, just kidding, that's my sensor's heart rate, or rather, its lack of signal.

I then went on to connect the newly soldered sensor to its driving circuit. Still no luck on the oscilloscope, just a bland flat line. I ended up moving the driving circuit from my breadboard, where nothing is soldered, to a properly soldered protoboard. I thought, maybe I am just unlucky with shorts. Well, I did create a new short. I am actually not sure when it happened, but I burned my sensor. I know that legs 1 to 2 and 1 to 3 used to show continuity on the multimeter. One of them must have melted because continuity was now gone.

And so I go on, fumbling around with electronics because isn’t engineering fun? Ok, 12 more different sensors to go. Hopefully, someone actually qualified within my team can take on some of them.

For curious people, the sensor is a MP7227-DA.