Subsystem Build & Test
Table of Contents
Team Vision for Subsystem Level Build & Test Phase
The broad-stroke team vision at the end of our Phase 1 review included completing:
- Finish fabrication of necessary subcomponents and begin assembly of subsystems
- Initiate purchase on long lead time components
- Begin carrying out test plans to satisfy fulfillment of engineering requirements
- Verify proper operation of purchased subsystems
- Continue quantifying terms of error analysis
- Correct shortcomings identified at Week 2 design review
The team's actual accomplishments at the end of this phase were:
- Fabrication & Assembly:
- Started fabrication of machined components, including baseplate, top plate, and clamp. Began assembly of entire structure.
- Drafted prototype design of thermal shroud & temperature control subsystem. Completed preliminary wiring of purchased controller
- Purchase Final Components:
- Acquired quote from Micro-Epsilon for capacitive sensor, data acquisition unit, and cabling and placed purchase order (2-3week lead time)
- Purchased final full-size electromagnet as well as permanent magnets for clamps
- Purchased both preliminary & final shroud components (insulated tubing, resistive heater tapes, electronic controller, datalogging thermometer, & thermocouples)
- Engineering Requirements Testing:
- Began testing electrical components (full information below)
- Error Analysis:
- Met with subject matter experts to formalize approach to error analysis & validate general design
- Design Review Shortcomings:
- Used Risk Management document to populate Problem Tracking, then updated with problems faced by team as phase continued
- Updated BOM to include delivery dates and lead times (turn BOM into more of a schedule document)
The team's expected accomplishments that did not get completed were:
- Testing of temperature controller could not begin until controller was wired properly. Therefore, due to time constraints, the temperature controller is fully assembled but not validated
- Lead time for components and time constraints before design review resulted in thermal shroud prototype not being assembled, although all purchased components have arrived or are in transit.
- Error analysis updated in brief (identifying most difficult contributors) but not expanded upon quantitatively
- Not all test plans were carried out: some require the full system/subsystems to be more fully assembled than they are at present
Team members' individual 3 week summaries can be found here.
Bill of Materials UpdateBased on comments from the previous design review, the project bill of materials was updated to include purchase and delivery dates for supplied components, functioning as a scheduling document as well as a bill of materials. As of this writing in week 6, all necessary hardware has been purchased and the budget information is up to date. A snapshot of the full BOM is shown below as reference only.
Mechanical Design & Build Update
Frame & Support StructureThe CTE tester's supporting structure was updated to increase the security of the top plate to the rods, and the rods to the bottom plate, as shown in CAD form below.
The support frame was machined from purchased raw stock. In addition, the carbon fiber composite rods laid up at the Harris facility before the previous deisgn review were cut to size and attached to the rest of the structure. With this, the mechanical support structure is nearly complete, and just needs to be integrated with the electrical & thermal components.
The "clip" design clamps were fabricated as well, which will be used to accomodate different sizes of sample coupons, and also features a locating pin to hold the coupon steady within the clamp during testing.
Work was also begun on assembling a thermal "shroud" to provide uniform heating to the sample coupon. To accomplish this, the outer shroud was made out of polyurethane foam insulation to retain heat around the sample. This shroud is made in two pieces to allow the sample coupon to be placed and observed, to visually confirm that the maglev system is working properly. After maglev operation is verified, the second piece of the shroud can be closed to isolate the sample. The shroud ID was selected to accomodate the flange securing the capacitive sensor, the sample coupon, and the heating element.
The heating element consists of a resistive heating tape connected to a purchased controller. With a thermocouple to provide temperature feedback, the controller will regulate the power to the heating tape to ensure a constant temperature output. The final system will consist of two thermal tapes in parallel, one for each half of the shroud, laid against the inner surface of the shroud in a "zigzag" pattern. This will provide uniform heating to the sample while still allowing the shroud to open and close.
The final shroud controller system will be integrated with our temperature data logging system and thermocouple interface, to accurately monitor shroud temperature throughout the duration of a full CTE test.
Test Results Summary
Test Plan S1: Magnetic Strength
Test Plans S2, S3 & S6: Control System
System architectureThe high level layout of the system architecture can be seen below. We have a Simulink model controlling the gate on the MOS, which allows current to the electromagnet. We’ll also use Simulink for data acquisition from both sensors.
The Simulink control portion of the system can be seen in the figure below. We are inputting the Hall Effect sensor digital value, which is the output feedback of the system, and outputting a PWM signal. In order to reduce the Hall Effect sensor biasing from the electromagnet switching we run the Hall Effect input signal through a Moving Average filter. The difference from the moving average filter and the set point is taken and feeds the PID controller. The PID controller is then tuned to have desired output characteristics.
System StabilityMotivation System stability in is a very important aspect in this project. If the system is unstable the magnetic levitation system is useless. The goal of this test is to ensure the magnetic levitation system is stable enough to support sensing of finite expansions in coupons.
Procedure This test consists of placing the clamp and coupon in the magnetic levitation system and using sensors to ensure stability. The test will also consist of changing the setpoint of the control system such that stability can be ensured. To ensure the same initial conditions every time the faro magnetic ball rests on a surface roughly 2” from the electromagnet.Results The first test ran was simply applying a setpoint to the system. From the figure below we can see that the system is stable within +/- 1 digital value, or +/- 0.0488 volts once the system settles.
Lastly, a sinusoidal signal was applied to the system. Here we can see that the system tracks the sin input reasonably well once the system settles.
Conclusion The system was stable within the +/- 0.00488 volt tolerance. The system was also able to track a sinusoidal input and a pulse input. The system could become more stable with a faster microcontroller. Further work The future test will consist of a fully integrated system. These tests were simply ran using a spherical ball which was levitated.
System Transient CharacteristicsMotivation An important aspect of control systems is rise and settle time. This test will determine the rise and settle time given a step input.
Procedure Levitate the object, then apply a step to the system. Measure the rise time, then measure the settle time.Results Applying a pulse to the system we can see the overshoot and settle time of the system. From the figure below we can measure both. The overshoot is roughly 0.009 volts, or 66%. The settle time is roughly 0.2 seconds.
Conclusion The system has a somewhat high overshoot; however, the system remains stable. The settle time is more than reasonable.
Future work This test shall be ran with the fully integrated system.
System Frequency Response
Motivation A relatively easy way to measure a frequency response of a system is to excite modes of a system with continuous sine signal with varying frequency, which is called a sine sweep. This will tell us the systems bandwidth (which is where the magnitude of the bode plot crosses -3db), resonant frequency, and where our closed loop poles are (roughly).
Procedure Levitate the object and input a sine wave to the system. Start at a low frequency and record the input sine wave and the output response. The output should follow the input. At certain frequencies, the output will be larger, smaller or equal to the input signal. Start at 0.01HZ measure the input and output, and calculate the magnitude of the amplitude of the sine waves (20*log(Vout/Vin)). Do the same for frequencies up to the Nyquist rate (half the sample rate). Once the magnitudes have been calculated for all the frequencies plot the FRF on a log scale.Results Running the sine sweep from 0.01HZ to 30Hz we can plot the FRF seen below.
Conclusion The bode plot shows that the closed loop system is under damped. We are unable to tell what the bandwidth is because the system became unstable before the -3db point, however, we know it is less than 30Hz. We can see from the bode plot that there are zeros around 1Hz and poles around 6Hz.
Future work Run test with fully integrated system.
Hall Effect Sensor Calibration
Motivation An important aspect of the system to know is the hall effect sensor reading versus object distance. This trend will tell us aid in finding the tolerance of the levitation system.
Procedure Set up the hall effect sensor. Record readings from the hall sensor with the levitated object at distances ranging from 0 to 50mm. Plot the data and record a trend line.Results Running 5 separate tests on the hall effect sensor the following curve is plotted. The best equation of the line follows a 3rd order polynomial equation seen on the plot.
Conclusion This test told us that the distance tolerance changes as a function of setpoint. The tolerance is larger at higher setpoints than at lower setpoints. We can now relate a hall sensor reading to a distance.
Future work Run tests with fully integrated system.
Risk and Problem Tracking
Risk management remained unchanged from the previous phase, as "risks" were considered more as vague, conceptual problems, not concrete issues faced by the team.
Link to the live risk assessment document here.
Previous phases' risk management was used to populate the Problem Tracking template, in order to account for the actual challenges the team phased while working towards this phases planned goals, as well as not lose sight of overarching risks to not achieving engineering requirements. Each line item in the risk assessment was added to the problem tracking, with numbers in the first column denoting which risk item corresponds to which problem. Additional problems that arose were added as well to keep all information about project-hampering occurrences in one consolidated place. These were arranged into three categories: Critical, Major, and Ordinary, based on their relative severity and ease of implementing a solution.
Plans for next phase
Error AnalysisThe error analysis document from the previous review was updated based on the subsystem tests that were carried out. This document will be revisited as the remaining tests are completed, and when the full system is integrated.
Project Planning Schedule
Major Accomplishments of Week 6
- Bulk of CTE test fixture's frame has been fabricated
- Prototype maglev system able to levitate a spherical magnet with stability
- Several electrical subsystem tests completed
- Performed update to error analysis chart
Major ObstaclesSeveral potential roadblocks that we foresee causing trouble before our Week 10 design review are:
- It is difficult to ensure that the initial conditions are consistent when testing the control system. This will resolve itself should we fully assemble the frame and begin large-scale magnetic levitation testing.
- The magnetic levitation system can levitate a ball with stability, but it is more difficult to levitate a disk. The current clamp design untilizes disk magnets, and will thus requre a more robust control system, although spherical magnets have been purchased, to be integrated with the clamp if necessary.
- The control system has certain ideal setpoint levels that are more stable than others. Full stability testing of the control system will be utilized to determine the ideal setpoints and avoid unstable ones.
- The MATLAB data acquisition software utilized on the lab computers is prone to crashing. This can be corrected by using a computer capable of real-time MATLAB simulation.
- The capacitive sensor must be integrated into the frame before a full-scale CTE test can be attempted. The long lead time associated with this component means that there may not be sufficient time in between its arrival and the Week 10 design review to perform a full-scale test.
- Temperature control unit has not been holding the thermal tape at the proper temperature and will require additional troubleshooting
- Manufacturing concerns required changes in structure design
- Many error sources are still very difficult to qauntify
Projected End State at Week 10
- Remaining subsystem tests will be complete
- New mechanical structure, control system, and temperature systems will be integrated
- Predictive ANSYS model will be prepared for comparison with full system results
- Error analysis will be updated again to reflect remaining tests and any other concerns