NASA ThinSats
Launcher and Pyrolysis CAD
Justin Kunimune, Emma Chickles, Dhara Patel, Abigail Nordwall and I developed three ThinSats in order to complete three missions for NASA. The first mission was to release at least 10 radar targets spinning at 10 rps in order to calibrate a millimeter scale radar (HUSIR). Next we needed to release 5 grams of vapor in under 1 second, twice while transmitting back the spectrograph of the vapor as it glows from colliding with the rarefied atmosphere. Finally, we needed to release an additional 4 larger radar targets without imparting any momentum to the targets. We designed mechanical, electrical and software systems to complete these missions, as well as built a series of in-depth physical simulations in order to validate assumptions and answer key design questions.
The launcher system was prototyped and tested using the Formlabs Form 2, although it's designed to be machined out of 303 Stainless (the same steel as the bearings, used to match coefficients of thermal expansion). All actuation, aside from the motor, was designed to use melt wire (kevlar melted with nicrome used as a resistive heater).
The launcher system was prototyped and tested using the Formlabs Form 2, although it's designed to be machined out of 303 Stainless (the same steel as the bearings, used to match coefficients of thermal expansion). All actuation, aside from the motor, was designed to use melt wire (kevlar melted with nicrome used as a resistive heater). We opted for geared, rotating rectangles/ellipses in order to maximize the length of object we could release (since the radar had a wavelength of 15mm, we needed to be able to contain a full-wave dipole, and the ThinSat is 12.5mm thick).
The Pyrolysis system (vapor release) was tricky to design. The most difficult part by far was containing a vapor in a vacuum for a significant period of time (we'd be in orbit for more than a week before release) while being able to release the vapor in under 1 second on command. We eventually were able to prove a vacuum seal with a dual-gasket design and a kevlar melt wire actuated system. We tested by weighing the container with liquid to be vaporized before and after being placed inside a vacuum chamber for extended periods of time. Because we weren't able to 3d print this prototype because of the vacuum requirements, longer iteration times were required to develop the system.
The electrical system was designed to be as simple as possible while completing the requirements because of the uncertainty we faced with our environment (we couldn't test anything in space). One of the primary problems we solved was odometry. We used a series of photocells around the satellites to detect orbital sunrise and sunset. We proved in simulation that we can back out the true time, and thus our position given a few assumptions about our initial launch plane and time (a real time clock was placed on board to give us a rough estimate of the true time to optimize from). Power circuitry for activating the many melt wires was primarily mosfet based.
The I2C image sensor for spectroscopy was processed by the Atmega32u4 onboard the primary computing satellite, and data from the sensor as well as additional lookup tables were stored onboard the SD card and external EEPROM.
NOTE: This is a very brief showcase of a few of the aspects worked on. Many of the calculations, simulations, and tests are not shown.