Build, Test, Document
Table of Contents
Build, Test, and IntegrateThe team entered subsystem testing and found issues with the design. As a part of that, steps have been taken to find and isolate issues in the current build and still keep on track. Unfortunately the project fell a couple weeks behind from where we initially planned to be. In some cases, components were still either on route or in the process of being manufactured.
The mechanical side of the build was delayed due to attempts to bring down the cost of the system to allow the electrical side more of the budget. The housing manufacturing was delayed several weeks, the mounting brackets for the hydrophone and speaker weren't finished until the week before Imagine. The dish design for the hydrophone was completed but untested. The second heat model was developed but unverified. A last ditch effort to heat sink would've been to mount a block from the power amplifier to the side of the housing for direct conduction.
After the PCB was received early on, it was quickly populated and tested. All of the board's subsystems were proven individually and then were integrated. The signals shown to be generated and amplified without much distortion. All power systems were checked for voltage ripple and noise. It was found that some of the power amplifiers were putting noise into our signals and larger caps were used to filter out the noise. In the end, the speaker was not able to operate at our desired frequencies and was then tested at lower frequencies. Even at these lower frequencies, the speaker was not able to work reliably and was therefore never integrated into the system.
From the software perspective, due to the Beaglebone Black's poor ADC capabilities, and external ADC was purchased which then also required an update for the electrical design to accommodate it. One of these ADCs was not functioning correctly and wasn't found until too late in the build so 2-way communication isn't possible without replacing this component. The code was implemented so that 2-way communication can be implemented but couldn't be tested. The architecture developed was fitted around a central control module that would handle all the other modules such that future teams could improve specific areas of the code easily. Error detection/correction and compression were implemented. Encryption was forgone because it was not a required function.
Test Plans & Test Results
Our initial plans for testing and results from early in the semester:
Our final test results and how they line up with customer needs and engineering specifications:
The mechanical assembly is designed such that the back plate is screwed into place using nuts and screws. The front is identified by having pems pressed into the sheet metal, allowing for the system to be closed. The speaker and hydrophones directly interface with this front plate. On the inside, 4 hex standoffs are used to support a mounting plate which can hold a Beaglebone and one circuit board. Velcro straps were to be used to hold the batteries in place on the inside wall of the housing. The 3 cable glands are for the Hydrophone cable out, the speaker cable in, and an ethernet cable for communicating with a computer.