Today's Updaily Post - Final Acknowledgements

We would like to thank Professor Steve Holt for this opportunity and for his constant feedback, guidance, and motivation throughout the program. Many thanks to Dr. Vanderlei Martins and Dominik Cieslak for their proposal and aid in this project. We would also like to thank Professor Matthew Neal, Professor Jonathan Stolk, and Professor Brian Storey for their help; and to Avinash Uttamchandani, thank you for your constant help and guidance in control theory and circuit implementation.



Documentation


General Information

  • CubeSat Projects 2010
  • MARS SW1 Detector
  • Marlow 4040 TEC

  • Project Documents

  • Final Report
  • Original Task Spreadsheet
  • PERT Chart
  • Last week PERT
  • Telecom Debrief (6/10)
  • Lunch Presentation

  • Powerpoint Slides

    Updaily Archives

    July 22, 2010

    This week we made the final push towards implementing an active lag compensator and though we were not successful, we learned a lot in the process. Tomorrow we turn our efforts to making final improvements to our system as it was earlier this week (with the gain and passive lag compensator) and delivering a useful and detailed write-up. The final major improvement includes a linearizing supporting circuit for our temperature-measuring subsystem. In other news, it's Mary's birthday this weekend- so don't forget to wish her a happy one!

    July 19, 2010

    After a grueling weekend, we now have a complete working system for proportional control (gain compensation). This update means that we have a functional setup with which we can test any compensator we want to implement. Moreover, this weekend we developed a software version of our circuitry using SPICE, which allows us to debug and test our hardware system with simulation.

    Tasks for Monday's 8am meeting include finalizing system details, such as whether to use the transistor setup we currently have or a high-current Op-Amp for the driver circuit. The transistor setup requires that we purchase an appropriate heat sink, while the second option requires the high-current op-amps. We will order parts by the end of the meeting, receiving them no later than Wednesday. Other tasks include data analysis and discussing how to successfully implement lag compensation.

    Note: The PERT chart will be updated at Monday's meeting.

    July 14, 2010

    Yesterday we attempted to perform the spectrum analysis of our TEC and learned that we do not get a system response at frequencies above 1Hz. This observation confirmed how low pass our system is and verified that our current model is an accurate representation.

    As for the progress with our compensator, we optimized and developed the corresponding circuitry for lag compensation and went to test it on the TEC this morning. We realized here that we are still missing two pieces of our system, the set point temperature and the driver. We have a rough idea of how to implement these final components, but we are meeting with our ECE guru later today to verify our designs.

    The current assignments include designing and building the driver circuit and progressing to optimizing lead-lag and comparing its simulation results with our current lag compensation. This week remains pivotal and, hopefully, we will have a working control system within the next few days.

    July 12, 2010

    We got our minds blown on Friday learning about basic control theory and Laplace transforms. This lesson gave way to a new approach that allows us to develop an optimal control system. The first control we will build is lag compensation, but we will also look into gain compensation, minor-loop compensation, and the inverted plant model.

    This new approach allows us to optimize the control system by implementing MATLAB functions to experiment with values for the poles and zeros of our system. The critical item now is to perform a spectrum analysis of our actual system. By analyze the data we obtain from the spectrum analysis, we can determine a transfer function of our actual system. In this sense, our model is an accurate virtual represenation of our actual system. We met with one of our professors to learn how to use the real-time spectrum analyzer and hope to complete the test during tomorrow's meeting.

    The revised PERT chart shows the dependence between tasks and is accompanied by a spreadsheet of due dates.

    July 8, 2010

    We head into the weekend with an ambitious list of tasks we hope to complete. We plan to meet with a recently graduated senior, who has offered to give us a lesson on controls and share ideas for our conditioning circuit. (I'm told he's an ECE gurU). Following this meeting, we will finalize the conditioning circuit design and develop a list of control algorithms to test. We can solidify a testing plan and continue evaluating controllers in software world, while continuing our work with the hardware system. We hope to complete these tasks by Monday.

    The absolute best case scenario would also include the hardware system construction (this includes designing and building the conditioning circuit, re-testing the three individual circuits independently, and interfacing them all together).

    July 7, 2010

    We received our order of relays, transistors, and capacitors today. This delivery allows us to continue developing our conditioning circuit; the first step is to test the two designs we solidified at the beginning of the week.

    In other news, we can officially start testing algorithms in software world. The software setup allows us to test controllers without the slow real-time aspect or complications of the actual TEC. We also arranged a meeting with another Oliner, known as an ECE gureau around these parts. He will guide us through controls and share design ideas for the conditioning circuit.

    Sidenote: check out our summer research project lunch presentation here.

    July 6, 2010

    This past weekend we researched transistors, relays, and capacitors, which helped us then solidify two designs for our conditioning circuit this morning. As a reminder, the conditioning circuit must function to filter and reduce the 12V power supply of the CubeSat to the optimal 7V input to the TEC. Both of our designs take advantage of an RC filter to reduce the 12V square pulses down to an apparent 7V DC supply. The differences then lie in power effieciency and the system reaction time.

    We also revisited our schedule to reflect a decision to re-prioritize the most important element of our project, the control algorithm. We will therefore not be including a PCB board in our final deliverable or tuning our controller specifically to the Marlow TEC. These changes can be seen in the most recent version of our PERT chart.

    June 29, 2010

    This week we aim to complete multiple subtasks all in the hopes of designing and ordering a PCB next week. These tasks include completing the PIC programming along with the PWM functionality, building the temperature reading and conditioning circuits, and interfacing all three components into our final system design.

    Atop these individual tasks, we must complete various tests--each circuit requires its own test as well as an initial vacuum test on the TEC to verify we are radiating enough of the heat to enable the TEC to get cold. To say the least, it is a crucial week.

    June 21, 2010

    This weekend we made a large leap by completing the interface between the model and the PID code. We also met on Sunday to detail a testing plan and break apart the tasks we want finished by the end of the week. Current assignments include designing and building the thermistor circuits, converting the PID code into PIC-C so that we can program the PIC. We are also developing software to help us find the PID constants. Lastly, we met with Matt Neil today to walk through the vacuum set up we are using.

    The BIG news: Our parts from TE Technology came in today, so we can add designing/building the conditioning circuit to our to-do list. The first step will be testing the impedance of the TEC.

    June 15, 2010

    Today we finally completed the model of a TEC (what we've been referring to as the virtual TEC) and coded a basic PID controller. After our group meeting, we also had a session with Steve and left with a deeper understanding of our system and the problem at hand. Current assignments now include building the interface between the controller and the model, finalizing the model (simply verifying the constants we used), and advancing the PID controller. Tomorrow, we will collectively detail our testing plan and continue with the current assignments.

    June 14, 2010

    This morning we received an email from Matt Neil offering another option for a vacuum chamber and, indeed, it is exactly what we need. This breakthough means we no longer have to construct our own baseplate. We also ordered parts and are waiting to hear back from our mentors about accurate-enough thermistors to prove our final deliverable. As for our PERT Chart, we updated it to reflect the feedback we received at this morning's collective meeting. We are allocating more time to the controller development and to additional experimental tests. The latest copy can be found here. The first test will be verifying the time step of our system.

    Current assignments include continuing work on the virtual TEC, building the experimental circuit, and coding a basic PID controller.

    June 13, 2010

    Our team met today to finalize part orders, discuss project management details, and plan the upcoming week. Over the weekend we received confirmation on an on-campus vacuum chamber. It requires some assembly on our part--constructing the baseplate and wiring an electrical throughput, but it will work. A special thanks goes to Professor Jon Stolk and Professor Matt Neil for their guidances.

    As for the management details, we revisited and update our Statement of Work, Work Breakdown Structure, and PERT Chart. Plans for the upcoming week include designing our final circuit, designing and constructing the software-to-hardware interface, (hopefully) receiving our parts, and starting the software coding.

    June 11, 2010

    Yesterday, we had a telecom with our mentor, Vanderlei Martins, and got many of our initial questions answered. The debrief from the telecom can be found here. As for our next steps, Mary is leading the effort to create a functional model for the TEC. Molly is squaring away the parts we need and finalizing the initial order with TE Technology. Sam is creating the website and researching PICs. Kiefer is finding out what access we have to vacuums on campus--we will need a way to test our controller on the TEC after all. The entire team is predicting tasks and putting together a final PERT chart.