Integrated System Build & Test
Table of Contents
Team Vision for Integrated System Build & Test Phase
In these past few weeks, we have merged the various subsystems into one single system capable of being an autonomous HABIP that can support plant life at 100,000ft for extended periods of time. Each individual subsystem has been refined. We have been working to integrate the communication between subsystems, as well as finalize the system structure and biocell.
Test Results Summary
Our test plans are contained in the following document: Test Plans of the entire system
Updated test plans for the power subsystem are linked: Power System Test Plans
Solar Panel, MPPT, and Batteries TestThe solar panel, MPPT and batteries were all tested together. The test setup is shown below. The MPPT is able to successfully charge the batteries. We were getting around 6.5W out of the solar panel. With a better light source this will be higher. This charged the batteries 50mV in 25 minutes.
This was then tested with the PMB. The PMB acts as a through connector for the solar panel into the MPPT. This is done so that the input current and solar panel voltage can be monitored. This test was also successful. Images are shown below.
PMB Data Collection TestData from the sensors on the PMB is collected over I2C. The data collected is shown below. All of the data measurements have been validated with the exception of the battery current which always reads zero. This issue will be examined further.
GPS Lock TestThe below picture shows the Communication System getting GPS lock.
Cut Down MethodThe Cutdown method Guide explains the how to setup the cutdown method.
GRSS(Ground Recovery Signaling System)
The P17105 senior design team designed an constructed a 4 layer PCB that has several LEDs and one buzzer that will assist in the recovery of the platform. The board has a 555 timer that beeps approximately once a second.
The Video of GRSS shows the system in action two years ago.
The image below shows the system in its current state that has new LEDs, and wires/connectors.
Comms, DACQ, and PMB Communications Test
The data collected from the PMB is sent over SPI to the DACQ. This is then send from the DACQ to the Comms board. This test is shown in the image below. The DACQ, Comms board, and CBOB are all also being powered by the PMB.
BioCellIn this phase, several issues were encountered with the I2C library and the FreeRTOS environment: success was met with immediate failure and consistency could not be reached. Eventually, through testing, it was determined that there was a difference in execution depending on the machine the code was compiled on. The MSP432 performed noticeably different when compiled on MacOS and on Windows. On MacOS, the start bit condition was not generated leading to a failed transaction.
Despite moving to a Linux platform and achieving more consistent transactions, several issues still do exist. When debugging, the code does hang at the start condition and the start condition will not be generated; however it has been shown that the drivers do work when transactions are successful and the data received is valid. Below is an image of the full BioCell with LEDs enabled.
Additional issues include a phenomenon in which 5V was seen on the SDA line. 5V does not exist in this system and this can only be explained by external noise; albeit the associated power is concerning. This is shown in the image below.
Unfortunately, due to the issues with I2C, no time was spent developing the ArduCam. There were also issues working with SPI between the CBOB and BioCell, and it is expected that there may be similar issues with the ArduCam.
Lastly, the CO2 sensor was tested. After extensive testing with failed results, it was determined that power was not properly delivered to the sensor and the net was not connected in the Eagle design.
Structure Build and Assembly
All of the components for the structure were manufactured during this phase. To assemble the structure the top sheet, supports and bottom sheet were riveted together to save weight. In addition, the board stack, reaction wheel board and reaction wheel were mounted to the bottom sheet to validate hole location and to define necessary wire length for all of the components. Finally, a mounting location was selected for the resistor and board needed for the nichcrome wire to cut down the balloon. Additional testing needs to be done to confirm nichrome wire gets hot enough to cut balloon. The next phase will focus on cutting the insulation panels and confirming all wire routing.
Risk and Problem Tracking
A link to the Problem Tracking document can be found here: Problem Tracking
A link to the Power System Problem Tracking document can be found here: Power Problem Tracking
An updated Risk Document can be found here: Risk Document
As of 4/1/2019, we have $318 dollars left. This should be enough to cover any additional costs with the project.
Learnings From this Phase
Engineering is hard. However, in times like this it's important to realize what we've learned and reflect on that. The following is a list of lessons learned by our team so far.
We have learned:
- How to work with a Pocket Beagle.
- Lots about Linux development.
- Python for embedded and how powerful the libraries are.
- Some important things about layout and PCBs.
- Much more time is needed to work with the hardware after they are built.
- Always have a backup plan for critical functions.
- Designing PCBs is hard work and not as glamorous as advertised.
- Compiled code can be machine dependent.
- Do not necessarily trust provided libraries.
- A lot of time should be left for debugging.
- Account for potential delays when manufacturing parts.
- Some processors may have unreliable serial interfaces so expect a lot of debugging.
- There is a lot of things that have to go right in order to get a large system working.
- In order to be confident that a system is reliable, a lot of validation must be done.
Schedule for Next Phase
A link to our schedule for the next phase is here:
Three-week plans for individual team members: