46

Gaia is an astrometry spacecraft that's currently operating around the Sun-Earth L2 Lagrangian point. Question: why here? Why not the Sun-Neptune L2 Lagrangian point? By orbiting the Sun at a larger distance, it should be able to get more accurate parallax measurements.

Only reason I can think of is cost. I'm not familiar with estimating how expensive space probes cost, but Wikipedia says Gaia cost ~\$1 billion and this is comparable to the cost of the Voyager program, which also cost about ~\$1 billion. Of course Gaia's instruments should be more sophisticated than Voyager's, but there were also two Voyager probes, not one.

Allure
  • 4,534
  • 16
  • 34

5 Answers5

71

Well, you thought about the spatial aspect of a parallax measurement, but not about the temporal one.

Gaia's intention is to measure 3D positions as well as 3D velocities. For the distance, you need accurate parallactic measurement, which come in with your orbital period.
For a typical Gaia-star with several measurement per year, you'll get 5 values of the parallax after 5 years of time, which you then average. If you'd send Gaia towards Neptune (besides the fact that no one has ever sent an orbiter, to say nothing of a L2 mission that far out) that has a period of 168 years, then after 5 years you'd get... 5/168 th of one paralactic measurement.

It simply couldn't achieve its science goals if put around the L2 behind Neptune. Also no one on this planet has any experience in putting something into a outer system L2 point. This is different than putting it into Earth's L2, because reaching the L2 around one of the giants has vast and very precise $\Delta v$ requirements. This would be a massive technological leap, and things don't work that way in space. Small, incremental technological steps are required in an anyways unfriendly environment, to make sure everything works properly and no millions of dollars have been wasted.
Compare that to Gaia's predecessor, the Hipparcos satellite, which was parked in geostationary orbit.

Now you could still say, why not use Jupiter hypothetically anyways. Well, the orbital period there is still 11 years, and Jupiter's L2 still suffers from the intense radiation environment that is provided by Jupiter's magnetosphere. This would lead to rapid degradation of the CCDs used for scanning across the sky.

MJ713
  • 103
  • 3
AtmosphericPrisonEscape
  • 9,840
  • 1
  • 31
  • 42
  • 39
    Also, good luck getting power from Gaia's solar panels out by Neptune. – notovny Nov 25 '19 at 02:41
  • 2
    There seems to be some confusion here that a Parallax measurement has to combine observations from opposite sides of an orbit. It doesn't. – ProfRob Nov 25 '19 at 07:40
  • 5
    @RobJeffries It doesn't, but it gains benefit from large orbital velocities. Neptune is moving slower, as simple as that - you need longer to get the same parallax. Of course, that's not the only consideration - the size of the orbit affects the maximal precision you can get from your measurements; measuring parallax from a Moon orbit would be much faster, but also gives you less precision than waiting for the Earth-Moon system as a whole to reach the opposite of its orbit. – Luaan Nov 25 '19 at 09:54
  • @Luaan perhaps read my answer. – ProfRob Nov 25 '19 at 09:55
  • @RobJeffries Yup, your answer is spot on, and this answer is wrong. I was just addressing your comment, not this answer or your answer. – Luaan Nov 25 '19 at 09:57
  • 3
    It isn't wrong, so much as not telling the whole story. – ProfRob Nov 25 '19 at 09:58
  • @Luaan: Where is this wrong? You want a large arc on the sky, and this arc has to be $2\pi$ ideally on the sky. With only a small arc of the full ellipse traced, you get massive fitting errors when determining your parallax. Rob Jeffries answer doesn't contradict me there in any way. It's not the orbital speed, its the size of your parallactic ellipse that you want to have. An arc doesn't help. – AtmosphericPrisonEscape Nov 25 '19 at 13:13
  • 2
    @AtmosphericPrisonEscape You're mangling together two things that aren't all that related. Larger distance means more accurate measurements. Less time to reach the same distance means you get the same accuracy in less time. There's very little difference between having the full ellipse and having an arc of roughly the same curvature and distance (as a flat projection relative to the object you're observing). – Luaan Nov 25 '19 at 13:20
  • @Luaan: Well, then remind me what the angle is that I take for the parallax, if I have only a small arc of an ellipse to measure this angle in the first place? Without the full arc I don't have my $\pi$ that I need to compute the distance from. – AtmosphericPrisonEscape Nov 25 '19 at 14:14
  • 3
    @AtmosphericPrisonEscape Ah, I was hoping we were just misunderstanding each other, but apparently not. Are you actually really trying to say that parallax measurement is impossible if you don't have measurements from two opposite ends of an ellipse? Surely if you know the shape of an ellipse, you can calculate the "expected" angle for the full ellipse regardless of where the two measurements are - of course the more distant the projected points, the more accurate the parallax measurement. – Luaan Nov 25 '19 at 14:28
  • @Luaan: That's my whole point. The ellipse solutions will be all over the place for a few closely-spaced points, while the same number of points spread apart will give a family of ellipse solutions with very similar parameters, i.e. small errors. If you want to learn more about how this process works, I suggest you check out the graphics on https://en.wikipedia.org/wiki/Hipparcos . – AtmosphericPrisonEscape Nov 25 '19 at 14:48
  • 16
    @AtmosphericPrisonEscape You don't need an ellipse at all to measure a parallax angle. All you need are two separate points. It doesn't matter how you get from point A to point B, whether its a full 180 degree arc, a small part of an arc, a straight line, or a detour through hyperspace. Triangulation has been used for centuries in construction and engineering, and I guarantee you that those engineers didn't walk in a half-circle just to get to their next measuring point. – HiddenWindshield Nov 25 '19 at 16:09
  • @HiddenWindshield: This is not about whether it can be done, but how precisely it can be done. See my previous comment. – AtmosphericPrisonEscape Nov 25 '19 at 17:53
  • 1
    In addition to the excellent points here, a probe in the Neptune/Sun L2 point would have a much harder time getting data back to Earth: the antennas' aims need to be more precise, the transmission much stronger (and/or the receiver much more sensitive while still dealing with background noise), and there are a lot more things that would block line-of-sight from Earth to the probe (so, the probe would need more data storage to queue up reports and would need an even more robust re-transmission protocol, plus more data runs the risk of being lost due to limits in the probe's hardware). – minnmass Nov 25 '19 at 20:05
  • 1
    Side issue, if you're going to put something in the outer system, why L2 rather than a Trojan point? – Anton Sherwood Nov 28 '19 at 05:25
  • I am no expert, but my understanding is that having many measurements from opposite sides of the ellipse, is what allows you to factor out the relative motions of the target star and our sun. – jwdonahue Feb 11 '20 at 20:54
  • @jwdonahue Indeed it does. Distinguishing proper motion from a partial orbit is difficult unless your parallax baseline makes a significant angle with the proper motion. – ProfRob Feb 16 '20 at 19:32
  • Jupiter's magnetosphere is almost absent at the Jupiter-Sun L2. Magnetosheath is less then 10 millions of km [https://en.wikipedia.org/wiki/Magnetosphere_of_Jupiter#Size_and_shape ]. From L2 to Jupiter is about 50 millions of km. – Imyaf Jan 10 '23 at 14:48
54

I think it is to do with (a) orbital speed and (b) telemetry and (c) power.

In order to measure Parallax you need to measure the position of the star from different locations in the solar system. The Parallax becomes more precise the greater the separation between those positions.

At Earth-Sun L2 you get a difference of about 2 au in 6 months. i.e. the spacecraft has a baseline that changes at 4 au/yr. In a 5-year mission, you essentially get 10 samples of the full baseline, that enables you to beat down the errors by $\sqrt{10}$, equivalent to an effective baseline of 6.3 au. At the same time, because the spacecraft has executed complete orbits, all of the sky has been sampled with a similar baseline (imagine viewing a line tracing out the spacecraft orbit from a distance - it will have a similar length when viewed from any direction).

If you calculate how long it takes a satellite in orbit at Neptune to make a baseline (defined by the chord of a circular orbit) of 6.3 au, it is only 5.5 years.

However, that would only be for part of the sky - the part at right angles to the spacecraft motion. Large portions of the sky would have barely any baseline at all because the spacecraft motion is essentially straight towards it. Solving for Parallax and proper motion (the relative tangential velocity of the stars) would also be difficult if the proper motion was then parallel to the satellite motion. At Earth-Sun L2, this problem goes away because every 6 months the Parallax motion reverses, but the proper motion doesn't. Around Neptune you would have to wait 84 years for that to happen.

Of course you would also get the observational baseline between where the spacecraft journey began (the Earth) and Neptune, which is potentially 30 au. However, this doesn't solve the problem of all-sky coverage and also wouldn't solve the issues discussed below.

The other issues are practical and I suppose potentially soluble if you throw enough money at them.

Gaia has a limited telemetry bandwidth. At the moment there is significant autonomous decision making and processing before a subset of the data is sent back to Earth. These problems become many orders of magnitude harder when you are 30 au, rather than at the Earth-Sun L2 point which is a mere 1.5 million km away.

Gaia also needs power and it uses solar panels. You get about 900 times less power per unit area at Neptune, meaning 900 times bigger solar panels or some alternative (nuclear) power source.

Lastly, it's much harder/more costly to send the same spacecraft to Neptune rather than the L2 point.

ProfRob
  • 151,483
  • 9
  • 359
  • 566
  • 10
    One could imagine a future mission that sent four identical spacecraft out to Jupiter and then used Jupiter's gravity to put them on solar escape trajectories in (as far as possible) four directions spaced tetrahedrally (like the bonds in methane). Over time, they would give a set of steadily increasing baselines in all directions. This would be massively more costly and technically challenging that Gaia though, and that wasn't easy. – Steve Linton Nov 25 '19 at 08:56
  • @SteveLinton Yes, that would be a way to do it. – ProfRob Nov 25 '19 at 09:59
  • 2
    nicely written and clear, and still with all the numbers! props. – Spike0xff Nov 25 '19 at 20:15
  • It's worth pointing out that one might propose an alternative power source (EG thermal nuclear), similar to those used in robot rovers. So whilst distance applies with solar panels, it might not be necessarily relevant with an alternative power source. – SE Does Not Like Dissent Nov 26 '19 at 10:25
  • I don't know how exactly Is position of the Gaia itself measured (you must know baseline for triangulation) but maybe measuring the spacecraft position at Neptune orbit would be less precise? – Leos Ondra Jan 10 '23 at 12:05
  • 1
    @LeosOndra I think spacecraft positions are known very, very precisely. – ProfRob Jan 10 '23 at 12:54
  • @SteveLinton So… you want to calibrate four imaging instruments, instead of one. Four instruments, with large-format detector arrays, two each, in two paths. Sounds like someone who’s never calibrated a camera to astronomy standards. There’s also the issue of building four cameras instead of one, but by four you’ve crossed the threshold of economies of scale. – caInstrument Jan 10 '23 at 17:09
5

3 problems.

1) Time. As previous answers say, to make use of the larger diameter around the sun at Neptune's L2 point, you need to wait for a full rotation which takes 168+ years.

2) Energy. Solar panels provide significantly less energy, potentially not enough.

3) Distance. Data from a probe around Neptune take a good 4h10min to earth on average, which limits the data rate you can transmit, just like New Horizons from Pluto.

Camille Goudeseune
  • 317
  • 1
  • 12
eagle275
  • 311
  • 1
  • 4
  • 1
    Nice, I forgot about the data rate issue, which is already nearly critical for Gaia. Missions with higher data rates and/or further away would have serious trouble with that. – AtmosphericPrisonEscape Nov 25 '19 at 14:51
  • 1
    No, you don't have to wait for a full rotation. Same mistake as AtmosphericPrisonEscape's answer. The other two points are already in my answer. – ProfRob Nov 25 '19 at 18:19
  • @RobJefferies - Sure, for one parallax measurement. But for the whole sky, a full rotation is preferable, as implied in your answer. – IronEagle Nov 26 '19 at 01:46
  • OK - not a full rotation - half rotation works - but its harder to "guestimate" the correct position of the satellite in relation to the star to make best use of the large axis... its definitely easier for a satellite around earth where the view is very much the same as from earths telescopes – eagle275 Nov 26 '19 at 07:28
  • 3
    I am not sure that (3) is sound either. Essentially you are saying that a larger latency implies a lower bandwidth, which is false. Data can still flow at the same speed. – Federico Poloni Nov 26 '19 at 20:44
  • @FedericoPoloni True, it is not the latency that will limit the bandwidth. But the strength of signal will be considerably lower at that distance, which will limit the bandwidth. Depends on how literally you take the answer, but it could be clearer for sure. – Katie Kilian Nov 26 '19 at 21:14
  • 2
    I'm undecided if I should upvote this or downvote it. On the one hand, you're right that distance means a lower data rate, but on the other, you're wrong about why. – Mark Nov 27 '19 at 06:12
4

One could certainly send a Gaia-like spacecraft into deep space, and take parallax measures at all times along its orbit. This is unattractive for several reasons, however. In short the large baseline may get you only a factor of 10 accuracy at the cost of several expensive modifications. The money would be better spent making a more powerful telescope for use in the Earth environment. Some issues:

  1. Deep space missions require an RTG power supply, careful thermal control, and so on. Solar power is much simpler and the thermal control simpler.

  2. Telemetry, data transfer, commanding, etc. become much, much more difficult with increasing distances, requiring large dish antennae, powerful transmitters, complex downlink schedules, and so on. The squared-distance ratio between a even a few AU and local Earth space is immense.

  3. Injection of the spacecraft into is trajectory is considerably more expensive.

  4. The angular/temporal profile of when what parallax accuracy is available is extremely anisotropic. Stars along the trajectory will still show small parallaxes. Good parallaxes to stars perpendicular to the trajectory will only be available after several years.

Tod R. Lauer
  • 216
  • 1
  • 3
-1

Lots of interesting answers. Lots of unnecessary arguing.

1) Neptune sweeps out a parallax of nine billion km every 84 years, Earth sweeps out a parallax of 300 million km every six months so if you want max parallax, then Neptune is a 30 times better place to be. However, Neptune only sweeps out 53 million km in the same six months so you can get better results there but you will have to wait six times longer before you start to get them.

2) Earths L2 point moves with Earth so only requires getting a satellite to a speed of zero (relative to Earth) in the right place 1.5 million km away. remember that when you launch something away from Earth, the Earths gravity will start slowing it down as soon as the engines stop burning so all you have to do is run out of fuel at exactly the right place. Difficult and expensive to say the least but yes, we can manage that. Neptune is 4.5 billion km away, give or take 300 million km depending on the time of year. The good news is that if we could get a satellite out there, we would just require a stable orbit, we would not need to have a zero velocity (relative to Neptune) at Neptune's L2 point. The bad news is we do not know how to put a satellite into any sort of orbit around Neptune at a price we can afford.

Paul Smith
  • 135
  • 4
  • Your first point is not the full story. While you are waiting "six times as long" you get a $\sqrt{6}$ improvement in the Parallax precision from Earth L2 by repeated measurement. That is how Gaia works. Secondly, it is no good waiting 6 times as long (say 30 years), because you won't get accurate parallaxes across the whole sky. – ProfRob Nov 28 '19 at 07:40
  • @RobJeffries - Did you read your own answer? The improvement from repeated measurements on Earth which as you described give a 'effectively' longer baseline, are replaced in Neptune with an actually longer baseline. After 3 years you will have mapped out the same baseline as Earth. After 5 years you will have mapped out a baseline longer then the improvements possible from repeated measurements, after 168 years you will have mapped the whole sky with a baseline 30 times longer then possible from Earth, and 20 times longer then the 'effective' baseline possible with repeated measurements. – Paul Smith Nov 28 '19 at 12:32
  • And you can see from my answer that you would get the same precision as the 5 yr Gaia mission in about 9 years at Neptune, not 30 years as implied in your answer. You misunderstand how averaging many position measurements increases Parallax precision. Indeed, (diminishing) improvements will continue as Gaia extends its mission. There would be some maximum precision reached asymptotically, but we aren't at, and won't get to, that point. – ProfRob Nov 28 '19 at 12:57
  • @RobJeffries - I never mentioned 30 years. I don't know who you are arguing with, but I don't think it is me. – Paul Smith Nov 28 '19 at 16:46
  • Then what do you mean by "you will have to wait six times longer"? The original Gaia mission is 5 years. – ProfRob Nov 28 '19 at 17:03
  • "Earth sweeps out ... 300 million km every six months ... Neptune only sweeps out 53 million km in the same six months so ... you will have to wait [300/53]six times longer before you start to get them [them = the results of a longer baseline]" Any clearer? – Paul Smith Dec 02 '19 at 16:26
  • Gaia operates (in the first instance) for 5 years. If it gets a parallax to a precision $\Delta \pi$ from one 6 month 2 au baseline, then the final mission precision is $\Delta \pi /\sqrt{10}$, the equivalent of a 6.3 au baseline. In $x$ years (where $x < 80$ years), Neptune gives a baseline (length of chord on assumed circular orbit) of $\sim 60 \sin (0.019x)$ au. To get a chord baseline of 6.3 au in Neptune orbit takes just 5.5 years. Not 30 years. The Gaia mission does not last 6 months. – ProfRob Dec 02 '19 at 18:59
  • Again with the 30 years? Please reread what I actually wrote, it does not say what you think it does. The only reference to the number thirty is to point out that Neptune's orbit is 30 times wider then Earths. – Paul Smith Dec 03 '19 at 11:43
  • And again. The Gaia mission is 5 years long. So what do you mean by "six times longer" if not 30 years? If you are simply trying to argue that it takes 3 years to do what you can do with Gaia in 6 months, then that is also incorrect. – ProfRob Dec 03 '19 at 14:04
  • It really would help if you read what was written instead of making things up and then arguing that they are wrong. For the third and final time, I made no mention of or reference to a time period of thirty years. Do you accept that Earth moves through space faster then Neptune? And that Neptune has a wider orbit then Earth? If so, then you must accept that you have to wait longer on Neptune to get the same results as Earth, and if you wait long enough, you will get better results then are possible from Earth. But, and this is the important bit of my answer, Neptune is much harder to get to. – Paul Smith Dec 04 '19 at 19:32