Update (31/12/2022): The project has stalled after being dogged by electrical gremlins. In addition to this, the revised chassis was somehow larger than the original carboard chassis.
Therefore the complete circuit has been redesigned and rebuilt (and subsequently debugged).
The above photo demonstrates the new circuit design (bottom of photo) in comparison to the previous design. The new design is 45% smaller. Now time to begin the mechanical design.
Necessity is the mother of invention.
Plato
If only there was a way to stop the dog from barking out of the window every time something walked past.
New Invention: The Bark Blind
The first proof of concept was a success! It might have been fabricated from cardboard, but it achieved all of it’s primary objectives and the battery life exceeded expectations.
The electronics consisted of an arduino nano, L298 motor driver, geared ‘130’ hobby motor and a 2200mAh 7.4v battery. All of these parts were kindly donated by the Fighting Robot project.
Bottom: Battery and L298 driver, mic module, speed sensor
A condenser microphone module was implemented to ‘listen’ for barks; the digital output changes if there is a noise above a threshold volume. The output is configured as an interrupt allowing the system to enter a sleep state.
An LM393 speed sensor was also implemented to fulfil two roles. The first was as a limit switch by detecting notches that have been cut out of the edge of the blind. The second function of the sensor to act as a safety cut off should the blind become dislodged; thus preventing further damage from occurring.
To interface with the blind, measurements of the existing blind fittings and motor connection were taken using Vernier callipers and modelled in a CAD program. The final design was 3D printed using bog-standard PLA.
Blooper Reel
There are always teething problems with every project and this one was certainly no exception. Therefore, in the interests of openness and honesty, here is a list of the issues that were faced along the way. (Many were of my own making):
- Pretty sure I burnt out a ‘speed sensor’. It just stopped working. But it might be because…
- …The first attempt at wiring everything on the carboard chassis was a massive failure. Mostly due to my decision to make a wiring loom from CAT5e data cable with indistinguishable colours.
- The first test with the roller blind stripped out the pinion gear in the gearbox. This was due to an error in my code that immediately slammed the motor in to reverse after dropping down.
- The ‘limit’ stop notches would catch on the light sensor and cause the blind to come away. Thankfully this was resolved by placing Sellotape over the notch.
- There was considerable flex in carboard chassis which would cause the blind to become disconnected…
- …Ironically, this was useful because every reset required you to roll the blind up by hand.
Future Plans
PRODUCTION KILLS 99% OF MY PROJECTS, BUT…
I’m reminded of this project every time the dog barks out of the window. The check list for the next version is:
- Solid chassis & case.
- Reduced power consumption / increased battery life.
- Introduction of user interface.
To do this, there needs to be an improvement with every section of the current version. Hopefully writing this will help break it down and make the task more palatable.
Mechanical
- Measure and plot components in CAD
- Design chassis
- Print chassis
- Battery access lid
- Need to consider fixing points
Electrical
- Efficient motor driver (currently exploring the possibility of a transistor h-bridge)
- Create UI with push buttons. Also some sort of feed back.
- Reduce unnecessary overheads (eg: LEDs) and condense wiring.
- Introduce power switch.
Software
- Program switches to include up and down controls.
- Also program a switch to start/stop listening. The switch probably needs to be tied to an interrupt to wake it from sleep.
- Change algorithm to extend the down time of blind depending on intensity of barking.
Last updated: 31/12/2022