Integrated System Build & Test with Customer Demo
Table of Contents
Team Vision for System Level Demo with CustomerDuring this phase the team desired to complete the entire platform and execute the launch. Below are the plans that the team had for this phase:
- Order all last minute components.
- Finalize software for all sensor acquisition for Pi HATs.
- Finalize software for all sensor acquisition for MSP430s.
- Finalize software to communicate SPI between Raspberry Pi.
- Finalize software to communicate UART between all Pi HATs.
- Finalize software to communicate between MSP430s via SPI.
- Finalize software for controlling motor.
- Finalize software for reading motor battery voltage.
- Develop software for turning on reaction wheel and turning on cut down at specific altitudes obtained on the Pi HATs.
- Maintain the team's Trello page.
- Maintain the team's budget.
- Maintain the team's BOM with team.
- Finalize all Pi HAT software including communication to MSP over UART.
- Implement communication and command structure from COMMs team.
- Continuously improve the reaction wheel control system/model.
- Create code for the reaction wheel commands.
- Work with the entire instrumentation platform model.
- Design wiring harnesses for SL connectors and ribbon cable to cameras.
- Conformal Coat all PCBs.
For this phase, the team actually accomplished:
- Pi Hat:
- Sensor Acquisition software finished.
- UART communication software finished.
- All software acquisition finished.
- All harnesses/wiring for SL connectors created.
- Mounted inside of the platform.
- External sensors installed to outside of platform.
- Balloon Sensor Board:
- Installed inside of balloon plug with wiring to telephone cable to platform.
- Pi Camera Extender Boards:
- Installed/utilized for each Pi HAT camera inside of platform.
- Mounted inside of platform.
- All harnesses/wiring for SL connectors completed.
- All LEDs installed in the walls of the platform.
- Buzzer installed on the top of the platform.
- Battery jumper installed on the top of the platform.
- Host Board:
- Reaction Wheel control model/system/implementation complete.
- SD Card writing software implemented.
- IMU data acquisition software implemented.
- SPI communication between COMMs Pi and Host MSP430 software implemented.
- UART communication between Pi HATs and Host MSP430 software implemented.
- SPI communication between both MSP430s implemented, but not reliable.
- Software on Host MSP430 for handling all COMMs commands.
- Software on Host MSP430 for polling all Pi HATs for latest sensor data.
- Software on Host MSP430 for enabling/disabling reaction wheel.
- Software on Host MSP430 for cutdown.
- Cutdown mechanism utilizing Nichrome wire created and installed through plug.
- Mounted inside of platform.
- All harnesses/wiring completed.
- New battery obtained after one malfunctioned.
- New battery wiring created.
- Motor mounted/installed in platform.
- Motor controller installed in platform.
- All PCBs conformal coated.
- Launch Day, 4/29/17:
- Successfully finished platform, launched, achieved 83,000ft altitude, landed near Syracuse, NY and recovered the platform.
Launch DayOn April 29, 2017, we had an actual launch of the platform:
Here is the Launch Video.
Here is the timelapse of the entire flight.
Here is an hour of launch day and flight footage.
Pi HatsThe first HAB launch flew with four PiHAT boards. Each board was configured to store photos/video as log sensor data as detailed in the following table.
During pre-launch checks, it was determined that PiHATs B2 and B3 were not powering on. There was no time to debug the issues, so the mission proceeded with PiHATs B0 and B1 properly working. Post launch, it was determined that the two UART cables connected to B2 and B3 were not wired correctly. The UART Host Tx was tied to the RST input on the PiHAT and therefore disabled the PiHAT’s power supply, keeping B2 and B3 powered off. No data was recovered from either of the two boards. Also, a plastic lens cover was placed on the bottom camera of B0 to protect during HAB pre-launch checks. This cover was not removed before launch and thus the camera pictures and videos were tinted.
PiHAT B0 I2C and 1-Wire data was successfully logged. PiHAT B1 was the only completely functioning board. Post-processing the sensor data showed that the maximum mission altitude achieved was between 82416ft and 78396ft. External HAB temperatures were measured down to -58F. The mission ascent time was approximately 63 minutes. The calculated ascent rate using the maximum altitude was ~1275ft/min. The mission descent time was approximately 25 minutes, with a calculated descent rate of ~3175ft/min. The two below plots document the recorded data through the mission. Following the plots in an image of Earth and space captured by PiHAT B1.
Host Board / Reaction WheelFor the first HAB launch, only the critical functionality was implemented for the DAQCS Host board due to time constraints and the large number of features that could be implemented. On the Host MSP430FR5994 micro-controller, the UART communication between all of the Raspberry Pi Sensor Nodes, SPI communication between the Host and the COMMS Raspberry Pi Board, cutdown mechanism control, reaction wheel battery voltage sensing and simple GPIO based control of the Motor MSP430FR5994. On a software level, SPI communication between the two MSP430FR5994s was also completed but due to some bugs, the functionality was decided to be removed from the final launch platform. Regardless, the Host MSP430FR5994 was able to receive all commands from the COMMS board and either send commands to the Motor MSP430FR5994 or forward commands to a particular Raspberry Pi Sensor Node. The Host would then be able to store local copies of all incoming data from both the Raspberry Pi sensor nodes and the Motor MSP430FR5994. This data would then be sent to COMMS for transmission back to earth.
On the Motor MSP430FR5994, only limited functionality was implemented due to timing constraints and to optimized the reaction wheel control algorithm. All data from the IMU was able to be acquired over a SPI interface. An analog signal from the reaction wheel motor controller was able to be acquired and converted to received the actual speed of the motor over time. Using the built in Real-Time clock unit, all data collected was able to be timestamped down to millisecond accuracy and up to hour accuracy. The read/write speeds of data to the SD card limited the overall speed of the reaction wheel control algorithm when constantly acquiring and storing data to the SD card. In order to compensate for the slow write speeds, built in FRAM was utilized to quickly save over 2 minutes of data collection while the reaction wheel is being used. This allowed the reaction wheel controller to operate at approximately 200Hz. After 2 minutes of running the controller, the controller would be disabled and the data in FRAM would be dumped to a single text file on the SD card. When the motor controller was not operating, data would continue to be logged to the SD card approximately every 100ms, and would log data for 5 minutes at a time. Finally, using all of this data, PWM signals were generated to control the speed and direction of the motor through the motor controller. The final assembled DAQCS board can be seen below.
DAQCS MSP430 FIRMWARE : Download Here
The above plot shows the final test data of the reaction wheel and the instrumentation platform. It can be seen that the approximated 1:170 proportion between the platform and the reaction wheel was proven to be correct. However, unlike what was first thought, the controller ended up becoming a proportional-integral controller, rather than a proportional one. This left two gains to be determined; they were found through testing to be equal to approximately 120 and -0.2 for the Kp and Ki, respectively. It can be seen that after being implemented, the platform maintained its zero RPM angular velocity relatively well, with a steady state being reached within approximately 0.5 second. There was some oscillation around the zero point, which was mitigated by filtering the IMU with an exponential moving average. A boundary was then set at +/-0.2 RPM to further try to mitigate the “bouncing” of the IMU data.
Here is a video of the reaction wheel rotating the final HAB platform during pre-launch checks.
GRSSHere is the video of the GRSS working installed in the platform. The GRSS was useful in recovering the platform due to the worker in the treatment plant hearing it and being curious. The GRSS lasted the entire flight and still has plenty of battery left over.
Risk and Problem Tracking
- Latest Risk Management document is found here: Risk Management
- Latest Problem tracking document is found here: Problem Tracking.
Plans for next phaseChris Schwab's Plan: Chris's Goals
Lincoln Glauser's Plan: Lincoln's Goals
Steven Giewont's Plan: Steven's Goals
Sydney Kaminski's Plan: Sydney's Goals