The electronics were assembled at Microchip in June 2005, sent to Russia and transported to the International Space Station on the Progress 19 cargo ship that September. The suit was launched from the space station on Feb. 3, 2006.
Lessons learned
SuitSat-1 was a clear success, providing two weeks of transmissions before the batteries finally gave out. The suit itself finally burned up in the atmosphere after six months. However, there were a few snafus along the way.
"In the postmortem, the transmit signal was not as strong as desired," said Bible, referring to a signal 20 dB lower than expected. There were a number of theories for this, including transmitter failure, but Bible went over the schematics and could find no point of failure that would cause a reduction in output power. The coax was a possible culprit, but Bible and his team saw the antenna as the more likely problem, combined with the tumbling of the suit.
"The antenna was a hand-me-down," he said. "It wasn't designed to be used in a helmet, and it went up without a lot of testing." For SuitSat-2, he said, the antenna will be self-contained and will likely have a ground plane behind it.
That won't be the only improvement. Seeing that the original suit lasted six months in orbit, the team decided to develop a solar panel power source to extend the operating life of SuitSat-2 to possibly two months. But how do you design a converter that is efficient all the time it's out there, given that there is no control over the orientation of the suit and its panels?
That task fell to John Tray, principal applications engineer at Microchip. Tray came up with a six-panel design in which one panel will always be in full light, while two or three will be in partial light. However, with a 90-minute orbit giving 45 minutes of light and 45 minutes of darkness, Bible said, Tray incorporated a power scheme that "can run the whole system off the panels and recharge the batteries at the same time."
The converter analyzes how much power is coming out of the solar panel, "and it actually maximizes it for a peak detect on power," Tray said. "A solar panel looks like a constant current source, and often you'll hook it up to a battery directly and charge it at a constant current, but that's not taking advantage of the power coming out of the solar panel."
The resultant dc/dc Max Power Point Converter will take the maximum operating point of the solar panel and continually run at that point, even though the conditions change. "It'll be charging the battery at its maximum rate possible," Tray said. Power storage is via a supercapacitor.
"There's not a lot of info about this stuff out there," said Tray, "so we wrote an app on how to do it simply, for extremely low cost." It will be posted on the company's site, he said. "Anyone who's looking at having extended applications run off of solar power in a remote area, where you don't want to service it, that converter will run excellently."
To view the application note and read more about the new converter design, click here.
Other power innovations include two-way communication to shut down portions of the circuit if it's not meeting the current budget, as well as six individual circuits to monitor each panel separately. These are based on PIC16F690 microcontrollers.
On the radio end, SuitSat-2 will have much more capability than its predecessor. For starters, it will be software-defined. "It uses one of our dsPICs [dsPIC33F] to do the modulation and demodulation, and to control a regular audio codec that's running at 48 kHz, sampling," said Bible. The downlink is 145 MHz and the uplink, 437 MHz (amateur satellite service frequencies).
"We'll convert these to a 10.7-MHz intermediate frequency, and that will get digitized to that 48-kHz bandwidth," he said. 'And then the DSP does its processing on it." An FM signal will be stored in an SD card and sent to the dsPIC over an SPI line, then sent down.
An outstanding feature of the radio's design is the use of a quadrature sampling detector instead of a mixer. "This allows you to downconvert from 10.7 MHz to audio, with better dynamic range and performance for less power," said Bible. It's not a demanding circuit, he said, and is basically a quad switch that's switching in quadrature (I and Q). Capacitors reside at the input to these switches; the whole thing is just an integrator, he said.
"So long as you're switching in a certain order [vector], then you basically get the I and Q at output: incredibly simple," he said. Bible suspects it will increase in popularity now that it can be done digitally, rather than in analog.
Other additions to SuitSat-2 include a more-complex band plan, as more bandwidth is available, along with more temperature sensors to better understand the overall thermal environment.
Engineers being engineers
While the SuitSat project presented its share of headaches and demanded patience as the Microchip team dealt with NASA's requirements and waited for the launch (they're still waiting for a launch date for SuitSat-2), Bible said they'd do it all again--and he'd recommend it as the type of project engineering companies should encourage.
While little of what was learned from the experience is directly applicable to his job at Microchip, the intangibles are valuable, he said. For one, there's the group experience itself. "Microchip has many product groups, but they don't work together. [The SuitSat team members] have been learning a lot about each other's products and about each other," he said. Bible said he has also seen the work translate into better customer interactions.
For himself, he said, SuitSat appealed to the very core of what he's done all his life, from building bottle rockets as a kid right through his career at Microchip. It was simply an extension of who he is: an engineer.
