Kepler Channel Noise Problem

Really BAD news...
Kepler, NASA's mission to search for planets around other stars, will not be able to spot an Earth-sized planet until 2011, according to the mission's team. The delays are caused by noisy amplifiers in the telescope's electronics.

"We're not going to be able to find Earth-size planets in the habitable zone — or it's going to be very difficult — until that work gets done," says Kepler principal investigator William Borucki.

Borucki says that the noise will hinder searches for a rarer scenario: Earth-sized planets that orbit more quickly around dimmer, cooler stars — where the habitable zone is closer in. These planets could transit every few months.


http://www.nature.com/news/2009/091030/full/news.2009.1051.html

Apparently the software fix for this known before launch issue will take until 2011 or thereabouts.

Now hang on a sec, what's with the "Look! HAT-P-7b's transit was measured so well that we have the precision to detect Earth-sized planets?"

I guess there's the issue of what constitutes a convincing detection to bear in mind here.

This kind of thing just goes to show that this kind of detection is very difficult! Unfortunate, but there you go.

Sirius_Alpha wrote:

Now hang on a sec, what's with the "Look! HAT-P-7b's transit was measured so well that we have the precision to detect Earth-sized planets?"

I would shout "HOAX" loud

Sirius_Alpha wrote:

Now hang on a sec, what's with the "Look! HAT-P-7b's transit was measured so well that we have the precision to detect Earth-sized planets?"

http://kepler.nasa.gov/ed/ppt/BoruckiPPT/Borucki2009IAUpresentation.ppt

Quote :

Kepler photometer is working well
All 84 channels are operating & producing useful data
Photometric precision is about 1.5 times design values

http://www.nature.com/news/2009/091030/full/news.2009.1051.html

Quote :

Three of the amplifiers are creating noise that compromises Kepler's view. The noise affects only a small portion of the data, Borucki says, but the team has to fix the software — it would be "too cumbersome" to remove the bad data manually — so that it accounts for the noise automatically. He says that the fix should be in place by 2011.

3/84=3.5% data in big noise?

Sirius_Alpha wrote:

Now hang on a sec, what's with the "Look! HAT-P-7b's transit was measured so well that we have the precision to detect Earth-sized planets?"

My reaction too. My surmise is that they manually filtered that data to get the result. Something they can't do for 100,000+ stars observed N times where N is a LARGE #. So they need the new software 'filter'. But still, the hype seems a bit disengenuous in retrospect given that they all knew about the amplifier problem BEFORE the launch.

To give the Kepler people the benefit of the doubt: I'll point out that they said they could find Earth-sized planets. This may well still be the case, provided they can get enough transits. Habitable-zone planets in longer period orbits - this bad news talks about planets around solar-type stars, or planets in orbits of several months - don't get so many transits, so presumably would be more affected. (It also helps that with a hot Jupiter you know the orbital period from the primary transit so you know how to phase-fold the data!)

Incredible.

Very disappointing.

I guess I was right to be disturbed by the high S/N ratio they published even after the manual filtering.

If I get this right, they can't separate out the 3 amplifiers automatically so all the data is being degraded right now.
If they change the software the two data sets before/after will be incompatible RE: one Peter van de Kamp.
Hence the wait to 2011 when the first 3 year cycle will be complete.

TheoA wrote:

Heads need to roll for this. No doubt about it.

TheoA: I suppose you have put together scientific experiments of this kind of precision all the time, launched them into space, etc, and they all worked perfectly fine every single time.

Incidentally William Borucki, the principal investigator for the Kepler mission has posted the following at Space Disco:

Quote :

There is a mistake in the Nature article. The Kepler Mission is actually doing very well and is producing planet discoveries that will be announced early next year. Data from 3 of the 84 channels that have more noise than the others will be corrected or the data flagged to avoid being mixed in with the low noise data prior to the time an Earth twin could be discovered.

The willingness of some people in this thread to immediately jump on the idea that the Kepler team has been deliberately misleading the public is a bit depressing.

It doesn't surprise me that the Nature article was mistaken, and I was confident that the news was blowing something out of proportion.

While I can't speak for others, I can certainly say that I did not doubt the authenticity of the Kepler Team's press announcement a few months ago, I just had a hard time rationalizing how it and the Nature article went together. Was this a new problem? Was Nature mistaken? I am glad it was the later, instead of the former.

Borislav wrote:

3/84=3.5% data in big noise?

I guess

Reading carefully, the Nature article never says it was a "new problem". It said "everyone" knew about the problem pre-launch. Apparently everyone does not include the public who funded the mission and has read nothing but completely effusive articles about Kepler from the team. In what specific aspects does Dr Borucki say that the Nature article is mistaken?

I'm hopefull that the 3 bad amps don't screw the pooch, but I don't understand the nature of the software filter needed to be developed for the data.

Nature article wrote:

Planet Hunt Delayed: Nature News

The very title itself is deceptive. The hunt has not been delayed, and is underway right now.

Nature article wrote:

Kepler, NASA's mission to search for planets around other stars, will
not be able to spot an Earth-sized planet until 2011, according to the
mission's team.

Kepler can see, and perhaps has already seen transits of Earth-sized planets already, here in 2009. They probably won't announce anything until then because anything now is probably far too premature.

Nature article wrote:

The delays are caused by noisy amplifiers in the telescope's electronics.

As only 3 out of 84 channels have extra noise, we still have 81 channels running perfectly (as far as I know). As Borislav pointed out, that's ~3.6%. Surely we can detect Earth-radius planets in the other ~96.4% of the clean data. Besides, the delay until ~2011 is because they need to gather enough data to be sure and not prematurely announce anything that might need to be retracted later.

While I can see how this will slightly hinder Kepler, I'm not sure why the article states that this puts the odds of finding the first exo-Earth in the favor of Radial Velocity searches. I'm also baffled by Nature quoting the Principal Investigator as saying "We're not going to be able to find Earth-size planets in the habitable zone — or it's going to be very difficult — until that work gets done." I can't help but wonder if this is a misquote? Especially with Borucki's statement at the Space Disco.

If the problem was known pre-launch, then that means it existed while they were measuring the secondary eclipse of HAT-P-7 b. Recall that turned out nicely. (If the problem was known pre-launch, any reason why they didn't fix it then? pressure to meet the launch date?)

It appears what has happened is that the article has conflated the problems with the amplifiers with the issue of needing to get 3 transits. You wouldn't get Earth-analogues detected before 2011 because of the orbital period of the planet. Obviously having data problems like this is going to make things more difficult - you might miss some of the transits, or have less conclusive data. This is not a fatal problem though. The HAT-P-7 result indicates that with the other channels they are getting very high precision results indeed. Don't start accusing the Kepler team of scientific fraud unless you are very well prepared to back up your accusations.

The article states that the Kepler team would be unable to deal with the bad data until the software fix. Borucki states that it is possible to flag or correct this data. As for why the problem was not fixed pre-launch, apparently this is because it was detected after the parts had been put together - at this point the team was forced to make a judgement call between launching with faulty components or risking further damage (and also risking postponing the launch) by disassembling the probe to remove the noisy amplifiers.

Also it is a false equivalence to suggest this is anything like the Van de Kamp telescope adjustments - new software is not going to magically fix the hardware. You still have the raw data available from before and after the software adjustment. Let's be clear here, this does not seem to be as major a setback as, say, Galileo's high-gain antenna failure, and by all accounts Galileo was a successful mission.

Lazarus,

I understand your reluctance to hammer the crew on this.

But note the casual tone of the principal investigator here.

"..Data from 3 of the 84 channels that have more noise than the others
will be corrected or the data flagged to avoid being mixed in with the
low noise data prior to the time an Earth twin could be discovered.."

This means this data is not being filtered or flagged presently. Hence it is all being mixed together. Hence they are collecting degraded data.

Keep in mind that the due to the quarterly roll the planet field affected is actually (4) times 3 or 12/84 ~ 15% once you cut through the spin.

What more are they not telling us. This batch of data will intrinsically have to be treated differently from the batch after the fix.

If they knew this was a problem prior to launch they should have released it to the public. Also there is still no full disclosure. No one knows which amplifiers these are, why they are malfunctioning, could more go bad?, etc

We accept that these things happen. I remember the heart stopping moments with the explosive bolts on the mars rovers. No one said anything and everyone held their breath as there had been full disclosure.

TheoA wrote:

Hence it is all being mixed together.

Borucki wrote:

will be corrected or the data flagged to avoid being mixed in with the
low noise data

(emphasis mine)

TheoA wrote:

Keep in mind that the due to the quarterly roll the planet field affected is actually (4) times 3 or 12/84 ~ 15% once you cut through the spin.

I'm not sure I understand. Why would there be more noisy channels after the roll?
Edit: Okay I see what you're saying. What exactly are the channels? Surely, they aren't those observation rectangles that Kepler is looking at, I counted them recently, and there's far less than 84.

TheoA wrote:

I remember the heart stopping moments with the explosive bolts on the mars rovers.

O_o what? I don't remember anything of the sort but then again, it's been a while.

I'm willing to bet that there has been disclosure of this issue, and that Nature just picked it up recently. But I don't know for sure.

Quite informative as to the nature of the data channels. Looks like TheoA's thing about the rotation of the spacecraft compromising more data fields was well-founded. The total data should still only be 3.5% noisier than usual though, as not every field is under the same channel all the time.
http://infox.ru/science/universe/2009/11/02/NoisyCCDamplifiersdelayKeplerplanetsearch.phtml (In Russian)

Sirius_Alpha wrote:

http://infox.ru/science/universe/2009/11/02/NoisyCCDamplifiersdelayKeplerplanetsearch.phtml (In Russian)

http://translate.google.ru/translate?u=http%3A%2F%2Finfox.ru%2Fscience%2Funiverse%2F2009%2F11%2F02%2FNoisyCCDamplifiersdelayKeplerplanetsearch.phtml&sl=ru&tl=en&hl=ru&ie=UTF-8

The sense comprehensible?
Everything is worse than previously thought

In general total - a loss 1-2 years in collection high precision data and 15% data after correction.

Sigh! Why does the only article with some accurate insight have to be Russian.

The present algorithm adjusts the photon count received by comparing the light from one star to all the others.
The essential concept behind differential photometry.

So its the light from all 100% of the data that is being compromised 3.84% before being sent to Earth.

Because the full data set is not sent to Earth this not a fixable with processing on Earth type issue.

Keep in mind that Kepler was designed to detect planets as small as Mars.
A Habitable zone Earth was originally a very easy catch.
The degradation is quite dramatic.

On the bright side this should not affect the ability to detect Jupiter.

Get ready for a tonnes of hot Jupiters and more spin on how valuable these are.

yaay. All together now. yaay.

TheoA wrote:

yaay. All together now. yaay.

Agree. We hope to fix the problem next year. All is not lost.

I wonder how accurate uhudshililas already collected over 6 months of data? At 2, 5 or 10 times worse?

Also note the space craft has only enough consumables for a 5 year mission, esp. to de-spin the gyroscopes.

The loss of 3 years is likely to be fatal to finding Earths IMHO.

Thanx Borislav.

I think that 2.5 year extension was to work without propellant potentially. Still think it will be fatal.

The primary mission of Kepler was actually to perform a census of the planets and so give us a very good statistical understanding of whats out there.

Right now it looks like this is severely compromised.