Monthly Archives: June 2010

Spacecraft Update & Are Kepler Seasons Similar to American Idol's?

Posted on by .
Well, folks, it’s been a long time since I’ve written, and there’s not been a lot of Kepler news in the news, as it were.  And I know people are wondering, “What’s going on?  Are we finding planets or what?”
So in a minute, I’ll tell you what’s going on.  But before I do, let me remind you where things stood when I last put out a notice.
We had been seeing ongoing problems with the star trackers that sometimes sent us to our Safe Mode.  And we had seen one of our 21 detector modules go dark on us.
We’ve done a lot of work on both these problems, and the spacecraft has been operating really nicely for four months now.  We still don’t know what is going on with the star trackers, but we’ve put in several layers of mitigation so even if they do act up a bit, we should be able to sail through it without interrupting science data collection.  For the detector that failed, we’ve narrowed down the failure to a particular circuit that seems likely to have blown a fuse onboard.  There’s nothing we can do to fix it, but the good news is that whatever caused the fuse to blow, we’ve been able to show that it was a very rare phenomenon, so we aren’t likely to see a bunch of detectors fail.
We uploaded a software patch to the vehicle and rebooted the flight processor and that went really well.  Our ground processing software is getting better.  It now searches for planets in the data automatically, and runs a variety of tests on what it finds, tossing out most of the false positives that masquerade as planets.  We’re now working on the last really big piece of the data processing puzzle, writing the software that will stitch together all our flight data and let us look for the long-period planets that will be most likely to support life.  Right now, the software runs on three months of data at a time.  Every three months we have to rotate the spacecraft to keep the solar panels aimed at the sun, and when we do, all our data sort of “hiccups” as all the stars are moved to different parts of the detectors.  Imagine it this way…
You set up a camera on a tripod and take 100 pictures of a bowl of fruit.  Then you turn the camera 90 degrees and take another 100 pictures.  Now, if you flip through the pictures rapidly, the first 100 will look nearly identical, and unless a fly landed on your fruit, it’ll probably put you to sleep.  Then all of a sudden everything will change a bit.  It’s the same bowl, same fruit, but it’s rotated 90 degrees and suddenly you’re awake and paying attention again.  But as you flip through the last 100 pictures, it goes back to being dull again.  No matter how careful you are with the camera alignment, and even if you rotate all the last 100 pictures so they are right side up, there will be that little glitch when you go from the first hundred to the second hundred.  And that little jump wreaks havoc on our software.  We’ve always known we would have to fix this, but it’s taken us a while to get there.  So there’s still more work to do before Kepler reaches it’s full potential.
But what has been going on in the mean time?  Well, this is where I’m going to try and make an analogy to American Idol.
Kepler is like a cosmic-scale American Idol, with 156,000 stars vying to be the winner, and show us their planets.
Like Idol, the early tryouts are a real trial, with a whole lot of bad candidates.  They might be interesting, entertaining, but not what we are looking for.  During these tryouts, we identify the most promising candidates, and hand out those coveted “tickets to Hollywood”.  In Kepler‘s world, these lucky contestants get a KOI number.  They become a “Kepler Object of Interest”.  It doesn’t mean they will win it all.  In fact, most of them will be eliminated by our “panel” before they ever get to face the “vote”.
In January, we announced 5 new planets.  Five new Idols from the first 10 days of data that we gathered.  You didn’t see the details of the Hollywood portion of that competition, only the results.
So yesterday, on June 15, we were obligated to make public, nearly all the data we collected from our first Quarter of operation. That first Quarter was a short one, because we didn’t start operations until most of the Quarter was gone, but still, it’s a big step beyond that first 10 days of data.  On June 15 we released the data for 43 days of operation, and the results of the “tryouts” from that period.  What will you see?  Another 5 planets?  Nope.  This time you’ll get a bit more insight into the Hollywood portion of the competition.  You’ll find that we issued something like 700 tickets to Hollywood!  And even if most of these are eventually sent home, there are still likely hundreds of winners, hundreds of planets within this group of contestants!  In fact, by the time we’re done with the competition, we’re likely to find that in the first 43 days of operation, Kepler has doubled the number of known Exoplanets!
Four hundred contestants will continue the competition within Kepler, and a new set of winners will be announced over the next six months.
Like American Idol, each new season takes a year to unfold.  But unlike American Idol, we start a new round every three months, and all our 156,000 contestants get to re-compete each Quarter.  All this really keeps our “Judges” busy.  But every year the winners are more and more talented, and eventually, we’ll end up with the real prize of the mission, Earth-size planets in the habitable zone that might sustain life!
So stay tuned for more news coming out every couple of months before another big announcement of Season 2 winners in January!

Charlie Sobeck, Deputy Project Manager