≡ Menu

Interstellar Communications: The Pointing Problem

Some topics just take off on their own. Several days ago, I began working on a piece about Europa Clipper’s latest news, the installation of the reaction wheels that orient the craft for data return to Earth and science studies at target. But data return is one thing for spacecraft working at radio frequencies within the Solar System, and another for much more distant craft, perhaps in interstellar space, using laser methods.

So spacecraft orientation in the Solar System triggered my recent interest in the problem of laser pointing beyond the heliosphere, which is acute for long-haul spacecraft like Interstellar Probe, a concept we’ve recently examined. Because unlike radio methods, laser communications involve an extremely tight, focused beam. Get far enough from the Sun and that beam will have to be exquisitely precise in its placement.

So let’s take a quick look at Europa Clipper’s methods for orienting itself in space, and Voyager’s as well, and then move on to how Interstellar Probe intends to get its signal back to Earth. NASA has just announced that engineers have installed four reaction wheels aboard Europa Clipper, to provide orientation for the transmission of data and the operation of its instruments as it studies the Jovian moon. The wheels are slow to have their effect, with 90 minutes being needed to rotate Europa Clipper 180 degrees, but they run usefully on electrical power from the spacecraft’s solar arrays rather than relying on fuel that would have to be carried for its thrusters.

Image: All four of the reaction wheels installed onto NASA’s Europa Clipper are visible in this photo, which was shot from underneath the main body of the spacecraft while it is being assembled at the agency’s Jet Propulsion Laboratory in Southern California. The spacecraft is set to launch in October 2024 and will head toward Jupiter’s moon Europa, where it will collect science observations while flying by the icy moon dozens of times. During its journey through deep space and its flybys of Europa, the spacecraft’s reaction wheels rotate the orbiter so its antennas can communicate with Earth and so its science instruments, including cameras, can stay oriented. Two feet wide and made of steel, aluminum, and titanium, the wheels spin rapidly to create a force that causes the orbiter to rotate in the opposite direction. The wheels will run on electricity provided by the spacecraft’s vast solar arrays. NASA/JPL-Caltech.

Interstellar Pointing Accuracy

How do reaction wheels fit into missions much further out? In our recent look at Interstellar Probe, the NASA design study out of the Johns Hopkins University Applied Physics Laboratory (JHU/APL), I mentioned problems with pointing accuracy when it came to a hypothetical laser communications system aboard. The team working on Interstellar Probe (IP) chose not to go with a laser comms system, opting instead for X-band communications (or conceivably Ka-band), because as principal investigator Ralph McNutt told me, several problems arose when trying to point such a tight communications signal at Earth from the ultimate mission target: 1000 AU.

IP, remember, has 1000 AU as a design specification – the idea is to produce a craft that, upon reaching this distance, would still be able to transmit its findings back to Earth, but whether this distance can be achieved within the cited 50 year time frame is another matter. Wherever the distance of the craft is 50 years after launch, though, the design calls for it to be able to communicate with Earth. We can still talk to the Voyagers, but that brings up the issue of the best method to make the connection.

Both Voyagers are a long way from home, but nothing like 1000 AU, with Voyager 1 at 158 AU and Voyager 2 at 131 AU from the Sun. The craft are equipped with six sets of thrusters to control pitch, yaw and roll, allowing the orientation with Earth needed for radio communications (Voyager transmits at either 2.3 GHz or 8.4 GHz). But what about those reaction wheels we just looked at with Europa Clipper, which allow three-axis attitude control without using attitude control thrusters or other external sources of torque? Here we run into a technology with a history that is problematic for going beyond the Solar System or, indeed, extending a mission closer to home. Just how problematic we learned all too clearly with the Kepler mission.

For reaction wheels are all too prone to failure over time. The hugely successful exoplanet observatory found itself derailed in May of 2013, when the second of its reaction wheels failed (the first had given out the previous July). Operating something like a gyroscope, the reaction wheels were designed to spin up in one direction so as to move the spacecraft in the other, thus allowing data return from the rich star field Kepler was studying. Kepler had four reaction wheels and needed three to function properly. With only two wheels operational, the spacecraft quickly went into safe mode.

The problem, likely the result of something as mundane as issues with ball bearings, is hardly confined to a single mission, and although the Kepler team was able to mount a successful K2 extended mission, the larger question extends to any long-term mission relying on this technology. Reaction wheels were a problem on NASA’s Far Ultraviolet Spectroscopic Explorer in 2001 and complicated the Japanese Hayabusa mission in 2004 and 2005. The DAWN mission had two reaction wheel failures during the course of its operations. A NASA mission called Thermosphere, Ionosphere, Mesosphere Energetics and Dynamics (TIMED) suffered a reaction wheel failure in 2007.

So by the time Kepler was close to launch, the question of reaction wheels was much in the air. We should keep in mind that the reaction wheel failures occurred despite extensive precautions taken by the mission controllers, who sent the Kepler reaction wheels back to the manufacturer, Ithaco Space Systems in Ithaca, NY, removing them from the spacecraft in 2008 and replacing the ball bearings before the 2009 launch. It became clear with the reaction wheel failures Kepler sustained that the technology was vulnerable, although it did function up to the end of the spacecraft’s primary mission.

Based on experience, the technology shows a shelf-life on the order of a decade, which is why the Interstellar Probe team had to reject the reaction wheel concept for laser pointing. Remember that IP is envisioned as a fully operational spacecraft for 50 years, able to return data from well beyond the heliosphere at that time. As McNutt pointed out in an email, the usable laser beam size at the Earth, based on a 2003 NIAC study, was approximately Earth’s own diameter. Let me quote Dr. McNutt on this:

“With a downlink per week from 1000 au that lasted ~8 hours for that concept, one would have to point the beam ahead, so that the Earth would be “under it” when the laser train of light signals arrived. It also meant that we needed an onboard clock good to a few minutes after 50 years at worst and a good ephemeris on board to tell where to point in the first place. These start at least heading toward some of the performance of Gravity Probe B… but one needs these accuracies to hold for ~50 years.”

This gets complicated indeed. From a 2002 paper on optical and microwave communications for an interstellar explorer craft operating as far as 1000 AU (McNutt was a co-author here, working on a study that fed directly into the current Interstellar Probe design), note the possible errors that must be foreseen:

These include trajectory knowledge derived from an onboard clock and ephemerides to track the receiving station and downlink platform so that the spacecraft-to-earth line-of-sight orientation is known sufficiently accurately within the total spacecraft pointing error budget. In order to maintain the transmitter boresight accurately a high-precision star tracker is also needed, which must be aligned very accurately with respect to the laser antenna. Alignment errors between the transmitter and star tracker can be minimized by using the same optical system for the star tracker and laser transmitter and compensating any residual dynamic errors in real-time. This must be accomplished subject to various spacecraft perturbations, such as propellant bursts, or solar radiation induced moments. To also avoid significant beam loss when coupling into the receiver near Earth, the beam shape should be controlled, i.e., be a diffraction-limited single mode beam as well.

X-band radio communications, as considered by the Interstellar Probe team at JHU/APL, thus emerges as the better option considering that a mission coming out of the upcoming heliophysics decadal would be launching in the 2030s, with the recent analysis from Pontus Brandt et al. noting that “Although, optical laser communication offers high data rates, it imposes an unrealistic pointing requirement on the mission architectures under study.”

What to do? From the Brandt et al. paper (my additions are in italics):

The conclusion following significant analysis was that the implementation with the largest practical monolithic HGA [High Gain Antenna] with the corresponding lower transmission frequency to deal with a larger pointing dead-band. This corresponds to a 5-m diameter HGA at X-band for Options 1 and 2 and a smaller, 2-m HGA at Ka-band for Option 3 [here the options refer to the mass of the spacecraft]. The corresponding guidance and control system is based upon thrusters and must provide the required HGA pointing as commensurate with spacecraft science needs.

I checked in with Ralph McNutt again while working on this post on the question of how IP would orient the spacecraft. He confirmed that attitude control thrusters would be the method, and went on to note that, at flight-tested status (TRL 9), control authority of ~0.25° with thrusters is possible; we also have much experience with the technology.

Dr. McNutt passed our discussion along to JHU/APL’s Gabe Rogers, who has extensive experience on the matter not only with the Interstellar Probe concept but through flight experience with NASA’s Van Allen Probes. Dr. Rogers likened IP’s attitude control to Pioneer 10 and 11 more than Voyager, saying that IP would be primarily spin-stabilized rather than, like Voyager, 3-axis stabilized. The Pioneers carried six hydrazine thrusters, two of which maintained the spin rate, while two controlled forward thrust and two controlled attitude.

As to reaction wheels, they turn out to be both a lifetime and a power issue, ruling them out. Both scientists added that surviving launch vibration and acceleration is a factor, as are changes in moments of inertia as fuel is burned for guidance and control.

Says McNutt:

“One way of dealing with this (looks good on paper) is actively moving masses around to compensate for pointing issues – but then one has to worry about the lifetime of mechanisms. Galileo actually had motors to control the boom deployments of its two RTGs to control the moments of inertia of the spinning section (a different “issue”). Of course, Galileo is also the poster child of what can happen if deployment mechanisms fail on a $1B + spacecraft – in that case the HGA deployment. The LECP [Low-Energy Charged Particle] stepper motors on Voyager have gone through over 7 million steps – but that was not the “plan” or “design.”

What counts is the result. Will engineers fifty years after launch be able to download meaningful scientific data from a craft like Interstellar Probe? The question frames the entire discussion as we move toward interstellar space. Rogers adds:

“We can always mitigate risk, but we have to think very carefully about the best, most reliable way to recover the science data requested. Sometimes simpler is better. The key is to get the most bits down to the ground. I would rather have a 1000 bit per second data rate that would work 8 hours per day than a 3000 bps data rate that worked 2 hours per day. X-band is also less susceptible to rain in Spain falling mainly on the plains.”

Indeed, and with RF as opposed to laser, we have less concern about where the clouds are. So the current thinking about using X-band resolves issues beyond pointing accuracy. Bear in mind that we are talking about a spacecraft deliberately crafted to be operational for 50 years or more, a seemingly daunting challenge in what McNutt calls ‘longevity by design,’ but every indication is that longevity can be achieved, as the Voyagers remind us despite their not being built for the task.

And while I had never heard of the Oxford Electric Bell before this correspondence, I’ve learned in these discussions that it was set up in 1840 and has evidently run ever since its construction. So we’ve been producing long-lived technologies for some time. Now we incorporate them intentionally into our spacecraft to move beyond the heliosphere.

As to Europa Clipper’s reaction wheels, they fit the timeframe of the mission, considering we have a decade to work with, from 2024 launch to end of operations (presumed in 2034). But aware of the previous problems posed by reaction wheels, Europa Clipper’s engineers have installed four rather than three to provide a backup, and we can hope that knowledge hard-gained through missions like Kepler will afford an even longer lifetime for the steel, aluminum, and titanium wheels aboard Clipper.

Image: Engineers install 2-foot-wide reaction wheels onto the main body of NASA’s Europa Clipper spacecraft at the agency’s Jet Propulsion Laboratory. The orbiter is in its assembly, test, and launch operations phase in preparation for a 2024 launch. Credits: NASA/JPL-Caltech.

Many thanks to Ralph McNutt and Gabe Rogers for their help with this article. The study on optical communications I referenced above is Boone et al., “Optical and microwave communications system conceptual design for a realistic interstellar explorer,” Proc. SPIE 4821, Free-Space Laser Communication and Laser Imaging II, (9 December 2002). Abstract. The Brandt paper on IP is “Interstellar Probe: Humanity’s exploration of the Galaxy Begins,” Acta Astronautica Volume 199 (October 2022), pages 364-373 (full text). For broader context, be aware as well of Rogers et al., “Dynamic Challenges of Long Flexible Booms on a Spinning Outer Heliospheric Spacecraft,” published in 2021 IEEE Aerospace Conference (full text).

{ 23 comments… add one }
  • Alex Tolley December 7, 2022, 14:37

    The pointing problems apply to an SGL telescope. I have naively assumed the pointing issues of communicating with Earth are trivial compared to alignments with orbiting exoplanets around moving stars, but clearly not. X/Ka-band bandwidth is orders of magnitude lower than with a laser, which might not be possible.

    What about interstellar communication? Laser seems more suited in some ways, but if pointing accuracy is problematic, even with the broader beam at those distances? What does that mean for optical SETI?

    Thinking back to Robert Forward’s multi-stage solar starships, a laser array rather than a collimated beam from the sun might have similar pointing issues, especially critical for the deceleration and return to Earth stages.

    For interstellar travel times, the star charts to determine position not only have to be updated based on movements of the stars but those movements must be known in 3-dimensions for the vehicle and the differences between their positions as seen from the spacecraft and from Earth known due to the distances and time differentials. I trust using pulsars that these stars are sufficiently distant that apparent movement is very limited. It would be tragic if some unexpected change made the craft’s celestial positioning go awry. [Asimove wrote a short story where a new supernova doomed a person in a hyperlight ship as the computers could no longer match the starfield to the ones in the computer.]

    All of this makes me wonder if physical data transport might be a better solution, even if much slower. If the data package arrives before the very low bandwidth beam with data finishes its transmission, then physical transport might be advantageous. A package from ETI loitering in our solar system might be a better means of communication than an optical/radio transmission with no certainty of receipt by the target audience. Similarly, with METI. [I think this was James Benford’s argument for the probability of success when looking for “Lurkers” or artifacts in our system, compared with SETI.]

    Is it any wonder why there is a Fermi Question given these difficulties? Even ETI’s advanced technology may come up against limits.

    • charlie December 7, 2022, 20:25

      [Asimove wrote a short story where a new supernova doomed a person in a hyperlight ship as the computers could no longer match the starfield to the ones in the computer.] – name ?

  • Dave Moore December 7, 2022, 17:01

    A couple of comments:
    The first is, why in the hell is NASA still using reaction wheels with ball bearings? Magnetic bearings are already in use in similar applications.

    And the second is, with respect to pointing accuracy vs data transmission, why not split the difference and use, say, long wavelength IR. This will have the spread to make pointing easier and the wavelength for a good rate of data transmission. The data would be beamed to a telescope in orbit around Earth or at its L2 point.

    • ljk December 9, 2022, 11:10

      Probably for the same reason NASA often does not (or did not) use the latest computers for their spacecraft. They make sure the ones they use have been thoroughly tested in rocket launch and space conditions first. That often means they are not the latest machines due to how fast computer development evolves.

      The Space Shuttle was using computers made in the 1970s for the longest time, although the astronauts did have laptops aboard to make up for this.

      I am not aware if NASA has been working on magnetic bearings in spacecraft development, which is probably why they are still using reaction wheels and ball bearings. I am sure this will change just as it has with computer systems.

  • Jens Fridthjof Knudsen December 7, 2022, 19:02

    Is there any chance the technology being used for LISA, the ESA testbed for building a space based gravitational wave observatory, could be of any use in building an ultra precise pointing system for an antenna on a very distant spacecraft? It occurs to me the same micronewton thrusters used for station keeping could see use for getting ultra precise pointing accuracy for a space probe mounted communication antenna?

  • Stan Clark December 7, 2022, 19:35

    The life expectancy of bearing appear to be the main problem with using reaction wheels. Have magnet bearing been considered?

  • charlie December 7, 2022, 20:23

    “Oxford Electric Bell” , not surprised its still running; better craftmanship then.

    • Alex Tolley December 9, 2022, 13:01

      Much simpler technology too. Oxford Electric Bell. There are no requirements for bell tone or ringing accuracy either. ;-)

  • Michael Fidler December 7, 2022, 22:13

    What about superconductors?
    A propellant-free superconducting solenoid thruster driven by geomagnetic field.

    Novel superconducting magnet thrusters to be tested out on space station.

    • Alex Tolley December 9, 2022, 12:58

      How does this help in deep space for orientation or propulsion?

      Interacting with the Earth’s magnetic field for “propulsion” has been suggested for decades. David Brin even wrote a short story using the idea.
      I haven’t seen it suggested for other bodies with magnetic fields, like the sun and Jupiter, but clearly, it is an option. [Interacting with charged particles with a magsail or electric sail is another option.]

      But for deep space missions, you need those magnetic fields for the propulsion outlined in the paper, and they are not really available far from the sun.

      • Michael Fidler December 10, 2022, 10:41

        That’s a good question, deep space is near absolute zero which means superconductors would work with zero resistance without coolant. They are able to store energy for ever and energize things like torque rods which can save the energy back to the superconductor when finished
        “A magnetorquer or magnetic torquer (also known as a torque rod) is a satellite system for attitude control, detumbling, and stabilization built from electromagnetic coils”

        The ambient magnetic field in deep space may be strong enough for orientation since time periods could be long. Maybe the Voyager spacecrafts data could answer that question.

        There may be other ways to use a ring type superconductor and torque rods to move the spacecraft orientation…

  • Henry Cordova December 8, 2022, 9:50

    Its comforting to hear that the difficulties of aiming precise com beams across vast distances are finally being addressed. Its essentially the same problem as aiming thrust beams to starsail craft. I’ve always felt that this was always a major obstacle to using lasers etc to push solar sail missions along their way, especially since this would involve two-way communications. Everything, both platforms and targets, is in constant complex motion, and feedback to the transmitter is essential to help overcome minor, but accumulating pointing errors. And of course, there is the problem of lightspeed delay…

    Another thing, the interstellar medium is very tenuous, but it is still not a perfect vacuum. Refraction and adsorption at some frequencies will be a problem, it won’t take much of a bending or scintillation to degrade the performance of a laser or radio beam.

  • Alan Mimms December 8, 2022, 15:10

    Why use reaction wheels? Why not put these offending ball bearings into races and drive them directly with magnetics to do the reaction? Why have a shaft and a motor at all? A fluid based magnetic mass would be even better (no idea if such a thing can be made suitable at temperatures, etc.) if has very low viscosity and it’s in a sealed very low friction tube and magnetically actuated as above. The tube inner diameter can be very tiny indeed if it’s surrounding the probe circumferentially (large overall diameter). If it’s 100% filled with no gaps the magnetics can be attached at any convenient point or even added to multiple points for redundancy. I suspect the “holding current” could be near zero and only significant power needed is during movement. Even that would be small given the tiny magnitude of mass/motion required.

  • Michael December 9, 2022, 7:29

    Just wondering if resonators could be used, a change in direction could have a very small change in the light moving around the ring but enough to give usable signal of change of direction. Its is all solid state tech so no moving parts.


  • Javier Garcia December 10, 2022, 9:38

    I suppose that Kepler location was carefully studied and there are important reasons not to be different.

    I wish if, from a more long term perspective, should not have been better to locate it in a more easy to access location.

    The telescope can not fix itself. No machine can do it, so far. Consequently, the advantage of having a human hand close to it means the posibility of sustain the mission for longer time.

    At the end of the day it is all about economics. But it is hard to imagine that Kepler mission could not have been more efficient by having a human fixing feature nearby.

    I am thinking more about a long range mission such as one to the outer planets or to another star system. Under does circunstances, the help of a flexible, prepared crew, able to confront the unimaginable, will be key for the success of the mission.

  • Alex Tolley December 10, 2022, 19:35

    @Alan Mimms

    Why not put these offending ball bearings into races and drive them directly with magnetics to do the reaction? Why have a shaft and a motor at all? A fluid based magnetic mass would be even better (…)

    The mass and velocity of the ball bearings or low-viscocity liquid (mercury?) would have to provide the same torque as the reaction wheels. While magnetic bearings make sense, I don’t see the value of using lots of bearings on a race[s] to provide that torque.

    This issues of reaction wheels and torque to reorient the vessel seques to one of my annoyances of sci-fi spaceships with rotating sections to provide artificial gravity. At spin-up, the main hull must be counter-rotating until spin-up is achieved, and then a countervailing force is needed to eliminate the hull’s rotation. This needs reaction engines (RCS). The design could apply the torque to a flywheel instead of the hull, but that would add useless mass and reduce the acceleration and deceleration of the ship for the same fuel load. Any friction between the rotating hab area and the hull due to operating spinning access rather like an airlock between the bridge in the hull and the rotating habitat will require firing of the RCS to prevent the hull from starting to rotate. (c.f. 2001: ASO – access to bridge from carousel section.)

    One solution is to have 2 contra-rotating habitat areas, but I recall having noticed that design.

  • Afterthought December 11, 2022, 13:09

    Use a series trailing relay probes, one per 100 AU. The relay probes can also do useful work of their own.

    • Michael December 12, 2022, 10:19

      They would offer a very long base for parallax indeed, possibly even as a gravity wave detector. Anyway if it takes 50 years to get into position we could have extended our receiving base considerably.

  • John I Davies December 20, 2022, 8:38

    My experience may be terminally rusty (1970) but control moment gyros (CMGs) were a serious altenative to reaction wheels. Since they spin at constant speed they may be more reliable long term but their gimbals need to be active to achieve the desired attitude adjustments. I guess this means that static friction might be the long term problem?

Leave a Comment