We have previously mentioned that we plan to connect our sensors up to Adafruit multiplexer using pluggable Pimoroni Breakout Garden Extender connectors. This exposes jumper pins for us so we can simply wire it up to the multiplexer pins using dupont jumper wires (no soldering necessary).
There is omne downside to this solution... robustness. Our extensive (cough) research has taught us that anything connected via jumper wire is liable to lose connection at some point during the rigours of competition day. So, we decided to mitigate this risk by housing our sensors in a casing designed to hold the sensor, breakout garden extender and dupont wire connectors together. At the same time, you might recall that we wanted to be able to use larger bolts for securing the sensors to the carpenters blocks.
We decided to reach our to Kerry Kidd (twitter handle @raspikidd), the Dundee Raspberry Jam organiser who we know also has links to the Dundee Makespace. We asked her if she could fashion some cases for us baased on some images of our prototype sensor mounts. We also gave her some samples to take accurate measurements from.
Below is a diagram showing what we had in mind:
We soon discovered a flaw in our design when we spotted a tweet from a fellow competitor, Paul Fretwell (twitter handle @drfootleg).
He reported having problems getting his distance sensors to properly sense the walls of the straight line speed test course during a robot testing day run by Brian Corteil (twitter handle @cannonfodder) at the Cambridge Makespace. He haad figured out that his sensors had been set too high.
We ventured to ask what height his sensors were mounted at and found out that this was similar to our planned for sensor height. So it was back to the drawing board to rethink our design.
In very short order we eventually came up with a new casing design which allowed for both variable height attachment and variable orientation. The image below shows this new design.
We are now waiting with baited breath to see what Kerry delivers. This is going to be an extreme test of accuracy and completeness of requirements specification because we do not have time to iterate this solution. It's simply gotta be right first time around (and when has that ever happened in the history of the universe?).