104 Hours With Philae

19 November, 2014

For the week of the landing I was booked to participate the events from the DLR Cologne control center, leading the people from the Philae science team and providing the interface for them towards the lander operations team at DLR Cologne and CNES Toulouse, the lander project management who was located at ESOC, and the ESA mission control center in Darmstadt. I was in a privileged situation since I could hear and see almost all the information about the lander, the landing and the scientific measurements from the start to the end. Thanks to this interfacing role, in several occasions I may have had better, earlier and more comprehensive and detailed information about what happened than anybody else.

The Philae landing operations started on monday, 10 November 2014, around 18LT. The spacecraft was started with booting the computers. This had to be done twice since somewhat unexpectedly it did not work successfully for the two data processing units in the first round, although this procedure was exercised many times before, for instance last on saturday before the landing week. This hick-up start left a kind of uncertain feeling with me. Nevertheless, the further start-up sequence of the lander worked well and it looked like that operations had stabilized in the course of the following day.

The confident situation changed drastically with two preparation steps for the landing sequence, i.e. the conditioning of the primary battery and the opening of the tank for the active descent system ADS. The battery conditioning stopped onboard after just 1 minute into the sequence and the tank opening failed shortly thereafter. The lander was in a crisis since by that time lander management – and many of us - felt to be closer to a 'no-go' for the landing than to a 'go' decision. Upon a suggestion by me, people from the ops and software teams at DLR Cologne rerun the battery conditioning sequence at the software simulator including various software patches, while in-flight the two failing sequences were commanded to be repeated again. In flight the ADS problem remained, but it was considered not a final show stopper for the descent and landing. For the battery conditioning a solution was found such that after the ground and in-flight testing one could conclude that according to all existing knowledge the remaining sequence for the separation, descent and landing could be continued. This conclusion was also recommended to the lander management, and so late during the night of 11 to 12 November 2014 the lander project manager gave the 'go' for the release of the lander which was followed by the 'go' for the whole landing sequence by the Rosetta project manager.

The Rosetta spacecraft left its parking orbit soon thereafter entering the hyperbolic trajectory towards the release point for the lander. On 12 November 2014 at 08:35UT the lander separated from the orbiter at a relative velocity of 19 cm/s. It separated via the spindle drive unit, the nominal system foreseen for separation. The spring mechanism, foreseen for the immediate emergency release, was not important anymore for the now starting Philae descent towards the surface. The close to perfect performance of the release mechanism was the first big relief for the MPS engineers, since these devices were designed, built and commanded by them.

During the first two hours after separation no direct communication contact between the two spacecraft was available, meaning no house keeping information on the lander status was transmitted to Earth. This link break was on purpose, since radio communication was considered risky for the antenna receivers. Moreover, the orbiter had to turn for the subsequent escape maneuver to reach a safe distance from the nucleus. However, one instrument onboard the orbiter could at least tell that the lander was alive: the Concert instrument. From before release to most of the descent until shortly before expected touch-down, the instrument part onboard the orbiter and the one onboard the lander were in contact with each other for joint measurements of the direct (between the two parts) and the indirect (involving reflections at the nucleus surface) instrument signal. So, we knew that the lander was working and did not have a serious problem. After two hours plus the signal traveling time from the comet to Earth (about 30 minutes) the normal link between the two spacecraft was established and all engineering and first science data from the Philae spacecraft were transmitted to Earth.

As a positive news – not only for the lander team, but in particular for the MPS engineers who are responsible for this subsystem of the lander - the initial housekeeping data from the lander showed that the landing gear was released and the Concert antennas were unfolded. Farewell images were taken by the Civa camera and also other scientific instruments onboard (Romap, Concert, Sesame, Mupus) worked well. Everything was on 'green' for a good descent and happy landing.

At DLR the excitement of the engineers and scientists in the control room and in the lander instrument support area LISA next to it was increasing. More and more people gathered in the LISA to see and hear what was happening in the control room, at the LISA display screens of the engineering information from the lander and on the voice loop between the control centers of ESA, DLR and CNES.

The Philae touch-down happened at 15:34:06UT, and it was reported by the project manager via voice loop as soon as the signal arrived on Earth. The triggering event for the touch-down came from the landing gear indicating a position change due to the contact with the surface at a speed of about 1 m/s. At ESOC the messages from the touch-down and subsequently from the anchor firing appeared as clear indicators for a successful landing. They triggered jubilation and a huge applause in the Esa control room and also in the public area where all the VIPS and press, radio and TV were following the event.

Due to an about 10 minutes delay at DLR Cologne in the arrival of the house keeping data from the touch-down, the engineers and scientists there could not follow the relief of excitement at ESOC with the same enthusiams. Trusting is good, but controling is better – and soon after the data had arrived at DLR it was confirmed that the lander had touched ground, but the anchoring harpoons were not shot and the ADS was not fired. Still the people were happy and congratulations were received since the lander seemed to be alive continuing to send housekeeping and science data after the touch-down at the cometary surface. From the short delay (only a few seconds) between predicted and measured touch-down time it was concluded that the surface contact of the lander happened very close to the nominal target position at landing site J.

Soon thereafter surprising measurements were seen by the Romap and Mupus teams whose instruments were kept switched on during and after landing to continue into the First Science Sequence of the Philae lander after a safe landing. They saw semi-periodic variabilities in the magnetic field orientation and in the measurements of the surface temperature sensor, respectively. Also the solar cell currents displayed short-term variability. Unfortunately, the Civa landscape image arrived on Earth rather incomplete. However, the existing evidence allowed only one conclusion: Philae had bounced at the comet and was still moving across the surface, most likely rotating and tumbling.

After more than 2 hours the Romap team detected another feature in the instrument data that resembled to be caused by a touch-down of the spacecraft. It was followed by another period of signal variability that then lasted less than 10 minutes. Thereafter, the Romap signal appeared as if was collectd with a stable position of the lander. This was supported by the Mupus data that showed a constant temperature after the third touch-down noted in the Romap data. And also the solar cell currents stopped varying such that the people at the DLR lander center started to believe that Philae had finally landed and remained at the surface of the comet.

The 2nd and 3rd touch-down happened well into the FSS operations of the lander which started shortly after the 1st ground contact of the spacecraft. Further science data from before, during and after the 1st touch-down were collected from other Philae instruments and arrived in the meanwhile on Earth, i.e. from Sesame, Concert, and Rolis. The latter ones were most interesting since it showed the approaching surface area of the first touch-down point from a close and decreasing distance to the surface. Sesame also detected the 1st touch-down of Philae through its sensors in the lander feet (the subsequent contacts were not recorded by Sesame since the instrument was switched to other measurement units for the FSS period). Around 18 UT the link break between Philae and Rosetta happened more or less as scheduled; for about 5 to 6 hours no engineering and science data would arrive from the lander.

Again Concert had direct signal exchange between both spacecraft, once more providing evidence that the lander was doing well. Very late in the night (around 4LT) the Concert team informed me that the instrument measurements are best compatible with a final landing site location that is about 1 km away from the nominal J site. However, an approximate position was not yet derived, i.e. we knew the Philae had landed, we did not know where it had landed, but we had confidence that the spacecraft was working where ever it was located at the nucleus.

During the morning hours of 13 November 2014 the next link period of the lander was scheduled. It was not clear when it would happen, and in fact it was delayed by about a bit more than one hour. The lander made contact with the orbiter, first with some interruptions most likely because of the low altitude of Rosetta as seen from Philae. With time the signal got more and more stable and the house keeping and science data from the period of the link break were downloaded. Romap, Concert, Mupus and Rolis were active during the first science block period of the lander. Most interesting was the Rolis image that showed indeed firm surface structures in the field of view.

Nonetheless, how to go on with the science program was no clear. The FSS plan foresaw to initiate the science block involving surface exploration activities via SD2 drilling. However, given that too few information on the landing site was available, we decided to implement the 'Philae safe block' instead that involved measurements from instruments like Romap, Mupus, Sesame. These instruments do not depend on instant information of the landing site (like for instance the local terrain around and underneath of the lander). To start with, a Civa panorama image of the landing site was taken und immediately downloaded to Earth. The exposure series took the chance of Sun illumination of the landing site and showed parts of the lander feet and Concert antennas as well as the surroundings of the site for the first time. Further analysis of the Civa and Rolis images was needed in order to prepare surface activities of lander instruments.

So, the command for executing the 'Philae safe block' went up to the spacecraft and the lander started the respective sequence stored onboard. This block lasted about 8 hours and was followed by lander stand-by until the next link period with the orbiter started. In the meanwhile it was discussed and decided at the DLR lander control center to prepare as next lander activity the first surface exploration measurements of lander instruments. First choice was the Mupus hammering and temperature measurements and the APXS release for the elemental composition analysis of the surface material. These activities were to be complemented by Sesame measurements with acoustic sensors and also dust counters as well as atmospheric sniffings of the Ptolemy and Cosac instruments. In order to support the identification of the landing site Concert ranging measurements preceeded the science activities of the other instruments. It was important to get the Concert data down to Earth during the time window of the on-going radio link and to follow at least the start-up of the Mupus activities on the surface before the radio link stopped. Due to variable signal strength between orbiter and lander at the beginning of the communication interval the commanding of the Concert measurements had to be uplinked twice which has caused a time delay in the execution of the subsequent activities. Unfortunately, the 2nd exposures series of the landing site landscape foreseen to be done by CIVA just after the Concert activities, ended up in darkness at the landing site. However, the rest of the instrument and subsystem activities continued and run well until the subsequent radio frequency link started after a communication break of some 5 hours with the lander.

Around 9 UT on 14 November 2014, the next radio link with the lander started, and it was not clear whether it will be the last communication window with Philae before its batteries were running out of engery. In the meanwhile it was found that the landing site allows only about 2 hours of power collection through the solar panels of the lander and the temperature in the compartment of the lander decreased towards 0 C. In preparation of this critical period the science team had discussed what to do next with the lander instruments, now also under the assumption that the results may not come to Earth quickly, but remain stored at the lander mass memory onboard. The goal was to perform a unique scientific measurement, i.e. addressing the organics content of the ground, and to let also SD2 have a chance, since it was the only lander instrument that was not yet in action.

So, the still on-going Mupus-APXS measurements were stopped and a sample activity involving SD2 and Cosac was initiated on-board. It was clear that only one sample could be drawn, but the lander has three instruments to analyse it, i.e. Cosac, Civa and Ptolemy. The decision which instruments should get the sample, was taken by the lander science team as a whole and it was also accepted by the teams who did not succeed. Again, Concert ranging opened the lander work at the surface and Cosac and Ptolemy had another chance to sniff the cometary atmosphere while the SD2 drill went down to the surface.

The radio link with the lander ended and the engineers and scientists waited with increasing excitement for the next and certainly last link about 5 hours later. It was not sure that it would happen at all. Nonehteless, one has to be prepared for the good as for the less good situation. And for the optimistic case that there is another radio link at 22:30UT in the night: What can and needs to be done given that the power budget of the lander batteries will not allow long lasting operation activities. Most important was to take the last measure to improve the chances for the lander survival, in particular the chance for collecting more power for a later switch-on when the comet gets closer to the Sun. In the meanwhile, the engineers had figured out that a rotation of the lander compartment around the z axis would place solar panel 1 into favourable illumination geometry. This solar panel has a larger light collecting area and thus promised to provide more solar power. A final Rolis image of the ground using the instrument illumination unit should follow in order to verify the lander rotation and extend the surface coverage underneath the lander. And Ptolemy had to get a chance to operate and measure the CASE oven where cometary gas was collected over the whole duration of the descent, landing and surface stay of the spacecraft. Last, but not least another Concert ranging measurement was foreseen letting it run until the signal from the lander part would fade away due to lack of power.

The last radio link between orbiter and lander started and the new command sequence was uploaded while the science data from the previous measurements during link break were downloaded. No one knew how long this would run. The applause by the people in the lander control center, acting and watching the events, got louder and louder with each step of the operations sequence that was completed by the lander. And indeed, the spacecraft managed with its remaining little energy to run to the end of what it was told to do and it transmitted all measurement data successfully to Earth (via the Rosetta orbiter). Around 2:15LT the last signals from the lander part of the Concert instrument were received. Thereafter, the battery voltage dropped below the critical limit and the onboard computers started to reboot, but due to lack of energy the boot process did not finish and was repeated automaticall a few times. The lander battery were exhausted and the spacecraft switched off for a longer hibernation period.

About the Upcoming Philae Separation, Descent and Landing

10 November, 2014

Philae is a so-called passive lander. This means it does not have a navigational and thruster system on board that can be used to guide it actively to the desired landing site. Or to put it in somewhat simple words: Philae is thrown down by the orbiter to the surface, however, with the aim to reach a predefined landing site on the comet. It is a bit like throwing a dart to the red center of a dart board. However, Philae has some additional requirements a dart player usually does not have to follow: it has to be thrown from a moving orbiter, with predefined speed such that it will land in a vertical orientation at a speed of about 1m/s or less at the targeted landing site. The Rosetta orbiter is the dart player running orbiting around the comet. The release velocity of the lander is fixed by hardware and commanding to 19 cm/s and the range of the touch-down speed and the trajectory orientation are fixed as well. So, the orbiter has to fly a very specific orbit for the lander release and the release itself has to happen at a well-defined point in time and space and with a well-defined orientation. This is indeed a very delicate task for the orbiter to perform and requires a special preparation phase on the spacecraft side.
 
Before release, the fly wheel of the lander is switched on since it is needed during the whole descent in order to keep the orientation of the main body axis of the lander constant in space. During the descent until touch-down the orientation of the main body axis of the lander is meant to be perpendicular to the surface of the nominal landing point. Slight rotation of Philae around this axis may happen during descent.

When the release point is reached, the lander will be ejected from the orbiter. The lander separation works through a spindle drive system. In case the spindle mechanism is not performing as expected, a spring eject mechanism will be used. Both mechanisms will be activated automatically within less than 30 sec such that Philae will escape from the orbiter within the right time window at a speed of 19 cm/s. However, first a so-called wax actuator has to to be opened since it holds the whole lander attached to the orbiter. In order not to release the lander too early after being freed via the wax actuator, the spindle drive is run in pull mode to press the lander softly against the orbiter interface until release time. The release also pulls off the umbilical cable between orbiter and lander which is used for power, command and data exchange between both spacecraft. Already some time before the release, the lander runs on internal battery power and all necessary 'life' systems are up and running. The separation phase will be a very critical time period for the MPS engineers since it is they who are responsible for the proper functioning of the lander release hardware.

Before release, the parts of the CONSERT instrument on the orbiter and on the lander will have already communicated with each other and are meant to keep contact for almost the whole 7 hours long descent period until shortly before touch-down. CONSERT intends to collect science data during descent that will be also useful to determine the separation distance of the two spacecraft and thus the descent orbit of the lander.

Very shortly after lander separation, the orbiter will slew in space and perform escape maneuvers from the release orbit, however, in such way that it can keep radio link contact with the descending lander. Philae separates further from the orbiter, and in a safe distance the landing gear is released from its storage position which it had over the long time period of the journey to 67P. This means unfolding the landing legs and 'lifting' the touch-down drive to maximum capacity. It will be initiated by so called 'hot knifes' that are heated and 'cut' through the fixation of the preloaded unfolding mechanism. Moreover, also the CONSERT antenna will unfold and the ROMAP arm will be put in optimum measurement position.

Two hours after release, lander and orbiter spacecraft are ready to start a continuous data and command link again which got interrupted due to the special orientation and maneuvering of the orbiter for the lander release. As of this time the ground control team should be able to follow closely how the lander performs during its descent to the surface – of course with a delay of about half an hour for the signal travel time from the comet to Earth. Before that time, only the CONSERT instrument signal will tell indirectly that the lander is 'alive'.

The next four hours will see lander descent operations together with some scientific measurements performed by CONSERT, ROLIS, and ROMAP (see earlier blog entry) as well as calibration measurements by MUPUS and SESAME. The latter are only possible to be collected during descent since it requires open view to dark empty space for MUPUS and the exact geometry of the unfolded landing tripod for SESAME.

During the last hour of the descent, the lander prepares for the landing. First of all, Philae needs to be able to sense the touch-down on the surface. This is done via detection of a sudden deceleration of the lander motion, not an easy task and in fact the touch-down trigger is expected to come from a motion of the landing gear motor – now operated as a current generator - when it is pushed back into the lander body as soon as the landing tripod gets surface contact. The lander carries also two accelerometers that were installed for exactly the task to sense the touch-down at the surface. However, these devices cannot be switched on, since the fly wheel of the lander 'makes so much noise' that the accelerometers would trigger the touch-down sequence (see below) already during descent and long before the surface is reached. This is bad luck and was certainly not intended, but it became only noticeable during the cruise phase of the Rosetta mission, and after analysis it was clear that nothing can be done about it apart from disabling the accelerometers from their function during descent.

Very likely surface contact will be first by one foot of the landing tripod, then a second one and then the third one. The surface contacts of the tripod will also drill the ice screws of the landing feet (one per foot) into the ground. These screws act like a drill, driven by pure mechanical energy from the touch-down motion of the lander. They help fix the lander to the ground and also prevent lateral motion of the lander during touch-down.

Due to the inertia of motion the body of the lander will move further towards the surface while the landing tripod has already stopped. As a consequence, the landing gear motor is moved further and this way motion energy of the lander is extracted and transferred into electrical energy. Through this process is will be possible to absorb most of the kinetic energy of the descent motion during the touch-down of the surface (btw, the electric energy produced by the landing gear motor will be turned into thermal energy and then conducted and radiated to the cometary nucleus and into space).

The touch-down signal by the landing gear triggers the anchoring of the lander. Anchoring is needed since the cometary gravity is so weak that the lander might otherwise bounce back into space. Moreover, any future surface operation like for instance drilling for subsurface samples would uplift the lander from the surface. So, the anchoring process starts with the 'firing' of the active descent system, a thruster that is embedded in the top surface of the lander body. This thrust pushes the lander against the surface. After a few seconds the two harpoons are ignited and shot into the ground. After they have stopped, the anchoring ropes which are connected to the harpoon heads are pulled back such that the lander spacecraft is tightened to the surface. The thruster firing has to continue well beyond the anchor shooting since the latter pushes the lander back from the surface due to the reaction force from the explosion of the harpoon firing. After less than a minute the anchoring process is finished and Philae has safely landed on comet 67P/Churyumov-Gerasimenko. The touch-down and anchoring period of the lander again will be an exciting moment for the MPS engineers in the control room since they are responsible for the proper functioning and operations of the landing gear and anchoring harpoons.

The last phase before and during touch-down sees scientific activities of some lander instruments, i.e. ROLIS imaging the surface granularity from a few meters distance until touch-down, ROMAP trying to measure local magnetic fields, MUPUS measuring the deceleration of the harpoons while they penetrate into the cometary ground and SESAME determining mechanical properties of the surface soil while the lander feet touch the ground. Immediately after landing CIVA will take first panorama images of the landscape of the landing site. They will be used together with ROLIS descent images of the landing site and terrain models obtained from the OSIRIS and NAVCAM images onboard of the orbiter to identify the landing location and its exact coordinates on the nucleus. Moreover, and also supplemented by ROMAP science measurements and voltage measurements of the solar arrays, also the lander orientation will be derived. This requires, however, that the respective instrument and housekeeping data are quickly uplinked from the lander to the orbiter and from there to Earth. Immediately thereafter the first automated instrument sequences of the First Science Sequence at the cometary surface are started.

All what is described in this blog section is performed through predefined command sequences prepared by the lander operations team and uplinked and stored in the onboard computer for execution. Due to the long delay of command and data signals between lander and Earth everything is done automatically and without intended interaction from the control center. The engineers and scientist can and will follow the sequence of events performed by the two spacecraft, the Rosetta orbiter and the Philae lander, during the period of separation, descent and landing remotely, and certainly with great excitement.

Despite of the very careful preparation and testing of all the command sequences and the tedious selection of the landing site on the cometary surface, a successful Philae landing requires still a good piece of luck and is not for granted. Regardless how it will all work out, I am sure those who can participate closely in the control rooms or close by – and maybe even people watching the landing remotely – will never forget the landing day of the Philae spacecraft on comet 67P.

Landing Site Selection - Round 3: Go for Site J

16 October, 2014

Week 41 was meant to end with the final round of the landing site selection by the Philae team and ESA representatives. Actually, this selection round was not a free choice, but had to follow a simple rule: A judgment had to be made on engineering grounds only, i.e. whether (1) the Philae spacecraft can be delivered by the orbiter such that (2) the lander can separate, descend and land safely at the primary landing site J and (3) that it can operate at the site to achieve its science mission on the surface. The expected answer was 'YES, site J can do'. If not, it meant automatically to check these criteria for the back-up site C and, if positive, to give the 'NoGo' for site J and the 'GO' for site C. A 'NoGo' for both sites was not really foreseen and would have a major impact on the Rosetta mission as a whole.

The meeting of the selection board was originally scheduled for Sunday, 12 October 2014. However, it turned out that the information and analyses results were available earlier such that the meeting was rescheduled for Friday, 10 October 2014, in the afternoon – and it was transferred to be held via internet connection since the board member could not travel on such short notice to meet in person.

An open point after selection round 2 was the coverage of landing sites J and C by boulders of different sizes. Landing on a boulder or even close to it (meaning that the lander tripod or even the lander touches the boulder) is not the goal and bares high risks for the lander: it might fall over or be damaged during touch-down. Even boulders well below 1 meter in size are considered dangerous.

The OSIRIS team delivered information on the boulder statistics of the two landing sites through close imaging of the respective surface regions and the SONC flight dynamics team performed the risk assessment due to boulders. Given the dimensions of the expected landing uncertainty on the surface, a statistical assessment of the probability to land at or near a boulder such that a safe touch-down would be at risk was the goal of this exercise - not to avoid a specific boulder piece at the landing site. The OSIRIS images have revealed really an awful lot (thousands) of boulders on both sites. Nonetheless, this large numbers resulted in a probability of a few 10 percent for a dangerous landing. And this is considered not to be a show stopper for the Philae landing, although one needs to have in mind that the risk for an unsuccessful landing is not zero.

The other assessment analyses focused again on the Rosetta delivery orbit for the lander release and subsequent descent orbits, the landing terrain properties and safe touch-down success, the power and the orbiter-lander communication situation after landing and the impact for executing the science sequences during descend and the FSS period (2 ½ days) after landing. In some sense it was a repetition of the analyses done during selection round 2, now covering the full error ellipse of the landing. A major task led by engineers from SONC CNES Toulouse in the first place and LCC DLR Cologne!

Indeed, the conditions for the landing, the operation at the surface and for the communication between orbiter and lander are not the same for all locations within the landing ellipse of an individual site. For instance, we expect a lander hibernation phase of a few weeks when landing at one end of the ellipse or less favorable communication windows at another one. Again, the evaluation of the results from these analyses is a compromise and its main aim is to exclude an unsuccessful lander mission at site J or C.
In conclusion, the landing site selection board agreed to propose to ESA to land at site J.

Some thoughts on 67P/Churyumov-Gerasimenko

29 September, 2014

The overall shape of the Rosetta target comet 67P/Churyumov-Gerasimenko (abbreviated 67P) looks funny. It was nicknamed by OSIRIS camera people as 'the rubber duck' since it reminded them of a plastic toy used by children to play with. The smaller component of the nucleus is the 'head' of 'the rubber duck', the larger one is the 'body' and the narrow part in between is the 'neck'. This shape came as a complete surprise considering that the nucleus shape models of 67P derived from Earth-based observations over the past years appeared more regular. They reminded one of the potato-shape nuclei seen on images of earlier comet flybys. The actual shape of 67P looks like a contact binary that might have been created by the two main components. The 'neck' is the gluing material that holds the two major components of the nucleus together. In this sense, the nucleus gives the impression of representing a direct relic from the early days of the solar system, and has already received attributes like primordial or cometesimal – the latter is an artificial word indicating its original character of the components at the time before the comet was formed. So, this scenario suggests that cometary nuclei are composed of cometesimals of a range of sizes represented by the two components of the nucleus of 67P, let's say typically km size bodies or maybe smaller. In conclusion, the formation process of the solar system would have to produce such cometesimals and to provide the procedure to put and glue them together to cometary nuclei.

Another view of the nucleus comes to mind, if one considers that 67P as a comet of the Jupiter family has a high probability of coming from the Kuiper belt, i.e. from a distance similar to that of  Pluto or even further out. (In numbers that means: from about 30 to 48 Astronomical Units. 1 Astronomical Unit is the mean distance of Earth from the Sun equal to 149.6 million km). This does not mean that 67P has formed at that distance, but at least since shortly after its formation it stayed in the Kuiper Belt until some 10 million years or so ago when it was scattered inward by Neptune. Thereafter, it cascaded inward by repeated gravitational interaction with the gas giants in the outer solar system towards Jupiter distance. The latter captured it as a Jupiter family comet. We do not know whether, when and how in detail this migration of 67P towards the inner planetary system might have happened. However, the transport process from the Kuiper belt is dynamically an efficient way to get Jupiter family comets recruited from the outer solar system. It is not the only one and it is usually not possible to prove it for individual comets. It explains a good part of the existence of the family of Jupiter comets and in particular its relatively high concentration towards the orbital plane of the planets.

During the long period in the Kuiper belt, objects are exposed to impacts, and indeed it is concluded from the size distribution of Kuiper belt objects that the ones smaller than about 50 to 100 km should represent relics from impact events rather than bodies of original size from the formation era. With a typical size of a few km, 67P is well in the range of collision fragments, so – if indeed coming from the Kuiper belt – the nucleus might represent a debris product from the still on-going down-grinding and clean-up process of the original formation material of the planetary system. 67P might have been part of one or even more larger bodies. The question how large the parent body(ies) may have been, remains unclear, but the Kuiper belt contains Pluto-size objects some 2000 km in size. Larger Kuiper belt objects are expected to have a differentiated interior, i.e. the original material from the formation era was modified possibly in composition and constitution. Fragments of such a body are not expected to be 'original' either.

An alternative scenario for the binary shape of the nucleus of 67P is to consider it as a pre-cursor of a comet that will split in the future. The splitting might happen in the neck region which may appear weakened due to its narrower cross section that could further be reduced by on-going gas and dust release from that region. The fragmentation of cometary nuclei is not a rare event. Some 50 comets are known to have split and produced fragments of different size and of different lifetime (days to decades). Since such events were observed up to now from Earth only, most likely only the largest pieces are detected and many smaller ones escaped unnoticed. Nucleus splitting might happen at any point along the orbit of a short-periodic comet, although there is a bias for its detection when it is closer to the Sun. On the average a cometary nucleus splits with a likelihood of a few percent per century. For 67P it could mean that it splits roughly every 50th revolution around the Sun. How close this nucleus splitting is, remains unknown and so far no signature is identified that would indicate a break-up soon.

A number of phenomena are known to accompany splitting events of cometary nuclei like increase of the coma brightness by factors of 100 and more, so called 'coma wings', an intense dust trail and last but not least the appearance of fragments. None of them is really considered an early warning signature, they are all post-event indications and only the latter phenomenon guarantees that nucleus splitting has happened. A prominent example for a split comet is 73P/Schwassmann-Wachmann 3, the very first target of the Rosetta mission. This comet split in 1995 and then again during subsequent revolutions around the Sun. The primary nucleus, component C, of the more than 20 larger pieces detected in the meanwhile, is still fully intact exhibiting a significantly higher activity level close to perihelion than known from the time before the splitting.

Certainly, from a scientific point of view it would be an exceptional extra for the Rosetta mission, if it could follow the nucleus splitting of 67P during its current visit. On the other hand, it might be frightening and a severe challenge for the spacecraft to survive. I am not sure, whether the engineers are prepared for such event to happen while they sail the spacecraft in a few ten km around the nucleus of 67P. So, I keep my fingers crossed that 67P will not split while Rosetta is visiting. Its normal activity will provide an interesting show anyway.

A last brief subject is about the surface terrains seen at 67P. 'Head' and 'body' have flat pool-like terrains surrounded by steep cliffs, while the 'neck' region appears smoother with a finer substructure. On first sight, some of these 'pools' resemble craters and look as if they were produced by impacts. The surface of 67P has certainly received many impacts over the billion years when it was still orbiting the Sun in the Kuiper belt. Nonetheless, on closer inspection the 'pools' don’t really resemble impact craters. So, it might be more likely that they are produced by the surface activity of the nucleus. Details of the processes that created these terrains are currently unknown and it is a task of the Rosetta mission to explore and explain these surface features, how they are created and how they evolve. The 'neck' region might be more strongly exposed to resurfacing by nucleus activity than the neighbouring surface areas.

Another Harpoon Shooting

22 September, 2014

On 11 September 2014, we had scheduled another shooting of the harpoons of the Philae anchoring system -  not in flight, but on Earth. This experiment took place at DLR Oberpfaffenhofen. It was prepared and exercised by MPS engineers and scientists with the help of colleagues from the MPI for Extraterrestrial Physics in Garching, the Institute for Space Research in Graz/Austria, the DLR Oberpfaffenhofen and a colleague of Pyro Globe company.

An earlier test of the harpoons had taken place at the end of July. For that test, the harpoon was shot into a long wooden container filled with loose grainy material. This material was meant to represent the cometary surface. Actually, it would be best to test the anchors for a medium that is identical or at least similar in its structure and mechanical behaviour to the one of the cometary surface. The problem is that no one on Earth knows what the surface structures and composition of a comet is like and even less what the mechanical properties are. Learning more about these properties is one of the goals of Rosetta’s landing mission. So, for the anchor test in the lab we could only apply a material that mimics cometary soil as much as we know for its bulk density (around 0.5 g/cm3). And it should be a loose granular mixture of irregular silicate grains of different sizes from micron to sub-millimeter and to a few millimeters.

The shots worked well and the harpoons flew into the cometary analogue medium at a high speed of about 80 m/s. Somewhat unexpectedly, they got stuck in the wooden walls after having traveled about 1 m in the 'cometary medium' of the experiment. This was due to a torque momentum imposed on the harpoon by the explosive firing. As a consequence the anchors missed the target entry point of the medium by a small amount and they also entered the medium with an inclined angle (expected was perpendicular entry geometry).

The key task of the second harpoon shooting was to verify the proper working of the acceleration sensors of the MUPUS experiment that are located in the harpoon heads and to allow a free traveling of the harpoons without hitting the wooden walls of the containers that hold the cometary soil analogue material. So, the harpoon experiment electronics was complemented by the MUPUS accelerometer read-out equipment. Moreover, a much larger containment for the soil analogue material was set up. As a consequence, it became a challenge to procure the analogue material (more than a ton) for delivery in time. But finally all containment equipment arrived at DLR premises two days before the harpoon shooting was scheduled to take place - just in time for preparing the experiment set-up the day before 11 September 2014. Once again, two harpoon heads plus electronics were prepared by Christian Rohe from MPE Garching and Reinhard Roll, Henning Fischer and Wolfgang Kühne from MPS Göttingen, the MUPUS electronics and sensor was set-up by Günter Kargl and Norbert Kömle from IWF Graz. The DLR facility was arranged by Stefan Völk and again Dr. Lell from Pyro Globe prepared the harpoon explosives and provided a high speed camera for the movie protocol of the experiment.

We met at DLR on 11 September at 9LT. Soil containment, the harpoon and MUPUS electronics as well as the ramp for the attachment of the harpoon had been prepared the day before. It still required the installation of the high speed camera and finally of the armed harpoon, one at a time. By about 11:30LT everything was ready for the first shot. The camera was running and Henning Fischer triggered ignition of the harpoon. The explosive generated an acceleration of several thousand Gs (Earth gravity acceleration) on the harpoon head. The projectile was accelerated to a speed of several 10 m/s and entered the soil material. After a few milliseconds it got stuck in the soil material after having traveled about 1.5 m, this time not hitting the wall of the containment. The MUPUS sensor recorded the deceleration of the harpoon head in the soil analogue material as we were aiming for. Since this time the rewind drive was not connected (on purpose), the harpoon was pulled back in the soil by human force, however using a gauge for measuring the applied force. The pull-back is meant to open the flaps at the harpoon head resulting in a large cross section and thus high resistance against the pulling force. So, our test projectile could only be pulled back by a few cm in the soil containment, then the flaps were apparently full open and it was not possible to retrieve the harpoon any further by human force. But, of course, we could still get the harpoon head out of the ground by means of shuffling away the soil material.

The second shot was foreseen for the afternoon after a late lunch break in the DLR canteen. The IWF colleagues were discussing whether to modify the experiment by including a plate into the soil material as a kind of stiff layering. This mimics a bit the layering of the cometary soils (which is still unknown, but suspected from theoretical calculations and lab experimenting) and at the same time it provides a fiducial point on the time axis of the MUPUS recording of the anchor shots. It was decided to introduce a rather soft multi-layer paper material among the soil analogue at a distance of some 50 cm behind the entrance surface of the containment. Then the second shot was prepared and successfully ignited and measured – even getting the fiducial mark in the MUPUS experiment recording. Before the end of the day the whole experiment set-up was dismantled and the test facility was cleaned up from any relics of the anchor shooting. By then, I was already on my way to Toulouse where the landing site selection process took place on 13 and 14 September 2014.

Landing Site Selection - 2nd Round

16 September, 2014

The second round of the landing site selection took place at CNES Toulouse on 13 and 14 September 2014. Again, some 60 people met for these two days in order to select the primary landing site for the Philae mission as well as a back-up landing site option. The sites had to be selected from the set of five sites that were pre-selected during the 1st round of the landing site selection which happened on 23 and 24 August 2014 at CNES Toulouse. The five sites from round 1 were labeled A, B, C, I and J (the missing letters in this alphabet series are associated with five more sites that were not pre-selected in late August).
 
Actually, I arrived in Toulouse already on Thursday evening coming directly from the anchor shooting of the MPS team at DLR Oberpfaffenhofen. On Friday, 12 September 2014, I wanted to prepare for the selection task by familiarizing with the information and analyses performed by the engineering teams for the landing site selection and also with the new and latest results from the orbiter instruments on the nucleus and in particular on the preselected landing sites. All these results became available only during the last days. For the afternoon we had called for a meeting of the lander instrument PIs at CNES Toulouse in order to brief the colleagues on the actual situation.

The engineering studies of the five pre-selected landing sites resulted in some surprises, good and less good ones. To start with the negative news: the ESOC flight dynamics team found site A too risky to be landed on under any circumstances and site C should only be approached with lander release strategy O2 meaning one has to rely completely on the primary eject mechanism and does not count on the emergency release option since it will not succeed to land on the surface – a situation which we wanted to avoid (not exclude) as much as possible.

The positive news came from the science front since organics was found to be present on the nucleus and the activity level on the surface seem to be compatible with the lander experiment expectations for successful detections of gaseous species. Other news was related to thermal properties of the surface which appeared to be interesting as well. The CONSERT PI introduced us to the success aspects for his instrument in respect to the landing sites. Due to the shape of the nucleus and the orbit trajectory that was preselected by the investigators of the orbiter instruments for the time after landing, the success chances of the CONSERT instrument differ quite a bit for the potential landing sites with B and J being best, I acceptable and C really disadvantageous.

Then, on Saturday morning, 13 September 2014, 9:30LT, round 2 of the landing site selection started. The orbiter instruments OSIRIS, VIRTIS, MIRO, ROSINA and GIADA presented their analyses of the nucleus and also of the specific landing site areas. Apart from the news mentioned above, we received information on the distributions of boulders on site, on nearby activity, on a scenario for the thermal profile in the soil and on the H2O activity of the sites. The latest images of the landing sites shown by the OSIRIS PI and taken from some 50 km distance above the surface and with pixel resolutions in the meter range offered awesome views. Of particular value for the engineering studies are the digital terrain models of the landing sites provided by the OSIRIS team from the NAC images and by the ESOC flight dynamics team from the NAVCAM images.

Next came the analysis of the engineering aspects of the landing sites. Site by site the analyses by the various engineering teams were presented, i.e. on the flight dynamics, the descent trajectory and durations, the slope and terrain properties important for the landing, the probability for a successful touch-down, the site illumination at the time of landing and during the months thereafter, the power, thermal and operational aspects of the lander system for the descent, the landing and the weeks thereafter, the chances and durations of communication between lander and orbiter during about five months after landing and consequences of all this for the performance the science sequences after landing.

This took almost the whole afternoon, but it illustrated aspects, criteria and individual site properties that were essential for a qualified evaluation of the landing sites. For instance, the success rate for lander touch-down at site I was found to be by far below that of the other sites and not convincingly high to keep that nucleus region in the list of potential landing sites. Site B was found to be not very advantageous for the long-term science program, since it has a short and low illumination pattern leading to a low power budget and a 1-2 months long hibernation period without lander science activities starting about a week after touch-down. Moreover, in the new images this site appeared to be rich in boulders. Site C was best from the power point of view, but had, apart from the disadvantages already mentioned above, also not so favorable radio link possibilities. Site J was the only site that had good to acceptable properties in all engineering and science aspects, not outstanding in a positive sense, but also being far above the discarding limits. So, it has a safe delivery trajectory, it allows the safe release scenario of the lander (nominal and emergency release with the same velocity), provides a descent to the surface in a little more than 7 hours with the second highest touch-down probability. It allows to perform all experiments during descent and most of the scientific experiments of the first science sequence right after touch-down using battery power only, and even more when considering additional energy collected by the solar panels of the lander. It has a very good success profile for the CONSERT measurements and allows regular battery recharging, scientific experimenting and radio links between orbiter and lander during the weeks and months after the landing.

This was the end of day 1, and in order to prepare the comparative discussion for the subsequent day 2 I agreed with my lander lead scientist colleague to wrap up all the pros and cons of the sites in a table file. This turned out to be very useful in order to arrive at an own opinion on how to proceed towards the site selection and priorization. ESA expected us to tell which site to be primary and which one to be back-up with a delivery date of this information by the end of Sunday, 14 September 2014.

Sunday morning started with a meeting of the lander and orbiter PIs. We came rather straight to the point, and I proposed to take site J as the primary landing site and C as back-up site. Alternative scenarios were also addressed during the discussion that followed, i.e. taking site B as back-up and how to overcome the less advantageous release scenario for site C. During the PI discussion the proposal J being 1st and C being 2nd received a majority support from the lander and orbiter PIs although not by all of them. It was certainly acceptable for all of them.

At 9:45LT on Sunday 14 September 2014 the decisive meeting on the landing site selection begun. First, agreement was achieved that sites A and I are out of the game. Once more, during a 1 1/2h discussion the proposal sites J+C was introduced and evaluated against alternatives like J+B. At the end consensus was achieved with the leading project persons, the engineers and scientists to go with the J+C proposal and to recommend it to the landing site selection board. It looked like it was the best possible under the given scientific, engineering and operational constraints.

The selection board met right after the end of the plenary discussion and endorsed this proposal. So, by the end of Sunday ESA received the proposal for site J as primary landing site and of site C as a back-up. By about 12:30LT the landing site selection round 2 was ended, but the people attending still had lots to discuss, most likely being satisfied that this important step for the lander project was achieved per se, some of them 'cum grano salis'. I admit that absolute happiness with the selection may not be with all of us, maybe not with anybody. I think this is, because we could not say: yes, this site is by far the best and absolutely outstanding and it has no risk for the landing and all science will be done for sure. At least we had the feeling that we have a good chance to achieve all this although not at a 100 percent level. And we will have to see how the future will judge about the result of the landing site selection process, mostly based on an implementation success or a landing failure.

The First Science Sequence of PHILAE after landing

10 September, 2014

The First Science Sequence (we call it FSS in our jargon) of the lander starts immediately after touch-down. There is no doubt that the FSS will have highest priority and importance, and it may be the phase that may receive a lot of attention by the scientific community and the public, although the subsequent phase of the long-term science exploration (LTS in our jargon) is for us equally if not more important. Since the FSS will last only 2-3 days, it will provide a kind of spotlight view of the comet and what is going on at the surface and underneath - a bit like the cometary flybys done in the past. The duration of the FSS is determined by the available power resources, i.e. mostly how long the primary battery of the lander will last. Unfortunately, this cannot be predicted in advance, and the battery will just stop to provide power when it is empty – without notification in beforehand. And of course the expected available total energy for FSS is not generous, and it depends on the duration and power consumption during SDL as well (a long descent will cost more power). Our engineers estimated that the lander energy for FSS could be about 1400Wh that is just sufficient to run an average desktop computer for half a day.

The FSS provides 'a first taste of the comet' in a way impossible to achieve through a flyby or orbiting spacecraft. Nonetheless, not all is fresh data: Some of the scientific data collected during SDL are still in the lander mass memory, waiting to be uplinked during FSS to the orbiter relay to Earth – this was already briefly mentioned in the blog on the SDL science.

The FSS is performed in blocks and each block is meant to empty the mass memory of the lander where the data of the block measurements are usually stored and to uplink it to the orbiter. So, the FSS consists of 3 blocks with different science activities and with different operating instruments. However, the intention is to enable at least one period of scientific measurements per instrument during FSS.

A key task of the first FSS block of bit more than half a day duration is to collect scientific measurements that enable the lander engineers and scientists to understand where and how the PHILAE has landed. In the first place, panorama images of the landscape of the landing site will be taken by the CIVA instrument and the plasma instrument of ROMAP will measure properties of the charged particles around the lander that may be useful for attitude determination of the spacecraft. The PHILAE location and attitude will also follow – apart from the CIVA images – from the solar power produced by the solar panels of the lander; this is an engineering assessment to be compared with results from lander instruments – and with images of the landed spacecraft taken by the orbiter camera system OSIRIS. So, we hope to get redundant information on the landing site and lander orientation through different measurements.

What else will be done during the first FSS block? CONCERT will start its first sensing of the cometary interior and the upper surface layers. This measurements needs to be phased properly, since the instrument units onboard the orbiter and onboard the lander need to have the same time reference and, of course, it is required that the nucleus is 'placed' in between the two instrument units. In practice, it means that ROSETTA has to fly an orbit that puts the nucleus in between orbiter and lander. And this has to happen when the orbiter tries to escape from the lander delivery orbit towards a safer distance from the nucleus; so, it is very demanding for the orbit and spacecraft engineering.

For the lander it means that it will lose the orbiter out of view after the time synchronization of the CONCERT units is done in order to get the cometary nucleus in place for the instrument sensing. The aim of the CONCERT measurements during FSS is to develop a first glance of the interior, i.e. is it homogeneous or more heterogeneous, nothing more. Greater details will only be possible through more measurement orbits of the instrument during the LTS phase.

The two gas analyzers onboard PHILAE, COSAC and PTOLEMY, will 'sniff' the local atmosphere around the lander for the C, N and O bearing compounds (PTOLEMY) and for evaporating organic material (COSAC). Naturally, they will also 'see' the evaporating water gas from the nucleus. ROMAP will follow the time evolution of the melange of solar wind and cometary plasma around the landing site through magnetic field and charged particle measurements and the MUPUS sensors in the heads of the Philae anchoring harpoons (shot into the ground for anchoring the spacecraft at touch-down on the surface) will measure the temperature evolution underneath the landing site surface for almost a full comet rotation. How deep the MUPUS sensors will be located is unpredictable – anything between a few cm to 2 ½ m - and will depend on the strength of the cometary surface which is currently unknown.

The end of FSS block 1 is the start of second FSS block. First thing to do, is to uplink and secure the data to Earth. But most likely, also to send new commands to the lander. At the beginning of the science operations of this block the lander may have to be rotated around its z axis (the one perpendicular to the cometary surface) in order to optimize the sun illumination of the solar panels of the lander. This is a must for the long-term survival of the spacecraft. The science of this 2nd block is mostly attributed to explore the composition of the comet using COSAC, PTOLEMY and the SD2 drill. SD2 shall take two samples from underneath the surface and deliver one each for analysis by the mass spectrometers and gas chromatographs of COSAC and PTOLEMY. The samples may come from a few 10 cm depths and after delivery they are heated to higher temperatures (several 100 C) which should release less volatile gases from the grain solids than can sublimate through solar heating. The latter, sublimated gases only can be measured, if at all, by instruments like ROSINA onboard the orbiter. The light elements and their isotopic ratios as well as complex, possibly organic molecules are the desired findings of the two lander experiments. Apart from that, also SESAME-DIM will be switched on to measure the dust mass and velocity at the landing site. This 2nd science block of the lander will last for about 15 hours.

Again data uplink and commanding for the 3rd FSS block will follow. This 3rd block is planned to be the longest one of the FSS and may last about 24 hours, i.e. almost two comet nucleus rotations. However, its execution may suffer from an earlier end of the battery power supply although there are chances that it may last long enough to complete all science operations including subsequent data uplink to the orbiter. Three scientific instruments will measure during this FSS block. MUPUS will release its separate probe from the lander and hammer into the ground next to the spacecraft. A 12 hour measurement sequence of surface and sub-surface temperatures will follow. In parallel to the MUPUS hammering and also thereafter SESAME-CASSE will measure the acoustic wave propagation through the cometary surface in order to determine mechanical properties of the cometary soil and SESAME-DIM will collect and measure dust grains at the landing site once more. Half way through the MUPUS measurement period, APXS will be released to the ground to perform its measurements of the elemental composition of the cometary surface. Towards the end of this block SESAME-PP will sense the dielectric properties of the ground underneath the lander in order to determine how deep the water ice could be located. Data uplink shall finish this FSS block.

Although it is not very likely to become reality there is a 4th FSS block prepared, meant to take another subsurface sample by the SD2 drill and to deliver it to a heating oven of the COSAC instrument. This time the sample, about 1 mm3 in volume will also be imaged and analysed by the CIVA internal camera and spectrometer. This should happen before the heating for COSAC is done. Well, whether this block will be started and executed at all very much depends on the capacity of the lander batteries and how much energy is left after the other FSS blocks are performed and how much Sun energy can be collected at the landing site.

It is assumed that the FSS activities will finish at any time during execution. So, the lander and active scientific instruments will be switched off by lack of energy and most likely not in a well- defined state. This will have to be recovered at the beginning of the next lander mission phase (LTS).

Landing Site Selection - 1st round

4 September, 2014

In my agenda, week 34 was blocked for participation in the first round of the Philae landing site selection process. It started on Monday, August 18th,  with the first inspection of the latest shape model provided by the OSIRIS instrument team and the overall surface illumination conditions for the expected time of landing (11 November 2014) to March 2014 when the comet will pass two Earth distances to the Sun. Despite the somewhat unusual shape of the nucleus of 67P - that is far from the one modelled from Earth-based observations of the comet over the past revolutions around the sun and that was for sure not expected by any expert in cometary science – it did not look too bad for finding possible landing sites on the surface. This was already noted during the two weeks before when the first shape model from OSIRIS became available to the lander team. In the meanwhile also a nucleus shape model was provided by ESOC Darmstadt that was derived from the images of the navigation camera NAVCAM onbord the ROSETTA orbiter.

On Tuesday, August 19th,  I traveled to Toulouse where the SONC team of CNES prepared the landing site selection. Namely, the flight dynamics team was working hard to find landing sites for a safe landing on the nucleus under conditions required for the Philae spacecraft and its science mission after touch-down on the surface. Apart from finding efficient trajectories for a safe landing they calculated the illumination conditions of the nucleus for six months after landing and the windows for the radio link between orbiter and lander for the possible landing sites. The former task, the trajectories and landing site identification, is the most demanding and difficult one. Apart from the shape model – that consisted of several hundred thousands of surface facets – one needs to satisfy hardware constraints for the lander separation from the orbiter (like separation velocity and direction) and touch-down conditions on the surface like landing during morning hours under Sun illumination and at a speed of below 1.2 m/s and in close to vertical direction to the surface.

Possible landing areas were identified by the flight dynamics team on 19 August such that the day after representatives of the Philae team (colleagues from DLR Cologne, CNES Toulouse and myself, in total about ten persons) could pre-select up to ten possible landing sites for further evaluation during the rest of the week. From the landing regions identified by the CNES flight dynamics team the group of pre-selectors compiled a list of 16 possible landing site coordinates. One should imagine that at this point one has to assume a landing uncertainty of up to about 500 meter radius around the nominal landing coordinates, so the actual landing sites consisted of areas of about 0.8 km2 on the surface of the nucleus. Then, a second iteration step of our group tried to isolate and possibly merge site areas that were overlapping and to discard others that appear to be less favorable in a closer look for instance since the NAVCAM images showed a rather rough terrain on the surface or the sites are expected to be in continuous sunlight for months after landing which may cause high temperatures in the lander compartment and due to that malfunctioning of electronic hardware. Luckily, this iteration step resulted in ten possible landing areas and thus the team had fulfilled the expected task (to pre-select ten landing sites) without the need of further down-selection and lengthy discussion
s.
21 and 22 of August saw the arrival of physical surface maps of the nucleus at SONC like for albedo and spectral slopes as well as for the averaged local surface temperature. This information allowed me to prepare a first assessment of the ten pre-selected sites for a brief meeting with the lander science team that was scheduled for Friday late afternoon. At that time it was still very difficult to discriminate pro and cons for the various sites and we were hoping for more information of this kind from the orbiter instrument participants who attended the planned meeting on Saturday and Sunday.
 
The highlight of the week was the meeting for the landing site selection process round 1. This round's goal was to select five sites from the ten pre-selected ones for a deeper analysis by the various engineering and science teams and using further observations of the orbiter instruments at greater detail and closer distance (to be implemented as of 25 August 2014).

About 60 people, scientists and engineers from the lander science, engineering and operations teams, from orbiter instrument teams, from ESOC Darmstadt and from the Rosetta management of ESA and NASA, gathered at CNES Toulouse for two days (23 and 24 August 2014) to perform this task. The ten pre-selected sites were introduced one by one and all possible aspects for the landing were presented and discussed. New contributions were calculations on the energy budget for the lander after landing at the individual sites, also whether the CONSERT instrument can expect to perform its science program the quality of which depends very much on the orbiter trajectories and the PHILAE landing site location (since this instrument has parts on both spacecraft that need to work together in a coordinated way after landing).

As expected, the meeting participants got also the latest first-hand information by various orbiter instrument teams on the surface temperature, surface reflectivity and gas activity. And the OSIRIS team showed images of the ten pre-selected sites taken during the past days and with unprecedented resolution down to meter level. Finally, I tried to make - based upon my general knowledge of comets and the most recent results on 67P that surfaced over the past days and even during the first meeting day - an ad-hoc summary of important scientific aspects of the nuclear surface and the pre-selected landing sites.
 
So during mid-afternoon all known facts and opinions on the ten pre-selected landing sites were put on the table and the difficult part was started. Of course, a major discussion point were the terrain properties of the sites, where the engineers tended to prefer flat terrains since it provides better prospect for a safe landing, while the scientists were more in favor of a terrace terrain since the prospect for getting access to surface activity appeared to be better. An important aspect was also the chance to perform the Long Term Science (LTS) phase of the lander and whether the lander will have to experience a longer hibernation period after landing since the solar illumination will not be sufficient for recharging the lander battery efficiently. From the ten pre-selected sites there was one that had almost all criteria positively fulfilled and in particular it appeared to be flat and wide enough to cover the expected landing uncertainty ellipse, but it had very bad prospect to perform the LTS phase, for sure not within the first weeks after landing and with the risk of a long lasting hibernation period for the lander and even the risk not to wake up again. The other nine sites are not considered extended and flat enough in the opinion of the engineers to allow a really safe landing although a final quantification of the risk of landing failure was not available. This was the point were trading started among the meeting participants exchanging arguments in favor or against keeping an individual site in the list. These discussions went across the participants partially in parallel, reminding more of negotiations at an ancient European market and did not give a very conclusive perspective. Hence, I proposed to approach the site selection by qualifying each pre-selected landing site in a three level scheme: very good, acceptable, not acceptable. This qualification was meant to be performed by the lander flight dynamics team, the lander operations team, the lander science team, by the CONSERT team. Also the orbiter instrument representatives were invited to provide their opinion about the preselected sites and how far they can support the instrument and Rosetta mission goals (for instance by combined or complementary observations or analysis to be performed in the future weeks and months). After some discussion, this proposal was accepted and the meeting resumed for this day. However, the various teams who were expected to perform the above mentioned evaluation met in the evening to take up and finish its task.

The next day, the meeting started with a plenary session where the evaluation results of the ten pre-selected landing sites were presented by the various teams. And by feeding this information into a table it became clear that four sites were not really good for one reason or the other and should be discarded from further considerations. So, we were left with six sites of which one had to be dropped since the goal was to provide ESA with coordinates of five sites, and this goal is taken very strictly. So, another iteration step on the quality of the remaining sites happened and at the end one of the six locations was identified as being less good than the other five. The disqualification criteria were now a set of less favorable properties like a relatively difficult terrain with no good conditions to measure with the CONSERT instruments. Moreover, it was the goal to have the sites about equally distributed on the two main components of the nucleus (the head and the body of the 'duck').

Science during separation, descent and landing

10 July, 2014

Separation, descent and landing will be the time, when the first real lander science is performed. In our lander jargon, it is called 'the SDL sequence'.

The descent will provide a few hours for scientific measurements when the 100 kilogram spacecraft Philae descents from about 2.5 kilometers above the surface to the surface itself. Right after separation and after the lander has unfolded its legs and antennae, the camera CIVA intends to take fare-well images of the orbiter. Well, these are not science images per se, but we hope that they will contribute to the illustration of the landing event and they are expected to provide the first and the last view of the Rosetta orbiter in full extent in space. Two instruments, MUPUS and the PP unit of SESAME, will take the unique opportunity to collect some calibration measurements in free space which were neither possible while the lander was attached to the orbiter, nor can they be performed in a proper way after PHILAE has landed. For exactly these reasons, these instrument activities have high priority.

Real scientific measurements are planned for the ROLIS camera, which is meant to take images of the surface area that is covered during the descent. Just before the expected touch-down of the lander at the surface, ROLIS will switch optics for close-field observations from a few meters distance which will provide high spatial resolution images of the touch-down area on the ground resolving the surface by far better than would be possible with the orbiter cameras. The ROLIS images during descent and before touch-down serve also an operational task: they are used to track the descent trajectory through landmarks on the ground and to identify the landing area for instance through the terrain model compiled before from the OSIRIS orbiter camera images. For this reason, the ROLIS exposures are foreseen to be uplinked to the orbiter and from there to Earth during descent or right after touch-down.

The radio sounder CONSERT, a two-part instrument with components onboard the orbiter and others onboard the lander, will sound the surface layers during descent through direct reflection of the CONSERT signal from the orbiter at the nucleus. There is also a direct signal between the two CONSERT units without reflection at the surface. This communication between the orbiter and lander units of CONSERT will provide information on the distance between ROSETTA and PHILAE, in this way providing an independent means to track the descent trajectory of the lander. Since the orbiter unit of the CONSERT instrument emits and receives the signals directly, results can be transmitted to Earth without further delay.

The ROMAP magnetometer will be switched on during descent in order to monitor the magnetic field in the cometary plasma and, when getting close to the surface, to sense any magnetic disturbances from the nucleus itself. Nothing is known on intrinsic magnetic fields of cometary nuclei. The spacecraft GIOTTO detected a magnetic cavity around the nucleus of comet Halley during its flyby in 1986. This showed that for high density of nucleus activity material the solar magnetic field cannot penetrate deep into the cometary atmosphere and does not reach the surface of the nucleus. These conditions, however, may not apply for the PHILAE descent and the magnetic field from the solar wind may be measureable with ROMAP all the way during the descent. Nonetheless, ROMAP may be exposed to small magnetic disturbances close to the surface, which could indicate cometary magnetism. Such a detection would be with consequences for our understanding of cometary formation: given a positive detection by ROMAP, magnetic forces would have played a role.

At touch-down, two harpoons will be shot into the cometary surface in order to anchor the spacecraft to the nucleus. This is required since the gravity of the comet is so small (example: PHILAE weighs about 100 kilograms on Earth, but only a few grams on the comet). The harpoons contain accelerometer sensors that report to MUPUS the deceleration of the shot devices during penetration and anchoring in the ground. This in turn, allows to conclude the strength of the cometary material. The harpoons also contain temperature sensors which allow measurements of the subsurface temperature up to a few meters below the cometary surface.

The scientific measurements of the lander instruments that are collected during descent and at touch-down are transmitted to the orbiter after landing, if they are not sent already during the descent. Of course, descent and landing is a risky undertaking – more about that during a later blog contribution. Moreover, it is performed under tight resource constraints, one of which is electric power. During SDL, the electric power is drawn from batteries and whatever is used during SDL will not be available for the next mission phase of the lander, which is scientifically even more important than all that was done earlier by the PHILAE instruments.

The current status of the SDL sequence preparation is such that its basic planning is completed and now the details of the commanding are compiled. Of course, an open parameter is the landing site that is currently unknown and that affects the duration and some other parameters of the descent. When completed, a realistic reference of the SDL sequence will be tested at the Ground Reference Model GRM at DLR Cologne. However, in-flight testing of the whole SDL sequence is not foreseen anymore, mostly because as long as the lander is attached to the orbiter most of the SDL operations are impossible or meaningless. Imagine, for example, trying to simulate the anchor shooting while still attached to the orbiter! If something went wrong, the harpoons would be lost  and might even damage the orbiter.

Landing site selection process – an exercise, 2nd and 3rd rounds

6 June, 2014

The second round for the test of the landing site selection process happened on 23 May 2014 in Toulouse. This time, I could participate via webex internet connection and telecon. The task of the 2nd round was to identify the primary and back-up landing sites from the 5 sites selected during the 1st round. Primary landing site means: this site is scientifically most interesting and technically reachable for the lander with safe trajectory for the orbiter for the descent injection. This selection step marks the last decision point where scientific arguments are considered.

A lot of engineering information was prepared for this decision meeting, i.e. trajectories for the orbiter to deliver the lander to the release point for the descent in time for a landing on 11 November 2014, trajectories and durations for the lander descent to the respective 5 landing sites including error ellipses for the landing, an assessment of the illumination conditions of the sites, the power and energy budget for the descent and within about a month thereafter, an assessment what consequences may exist for the performance of the science program for each site. Of course, it is important to apply a kind of figure of merit estimation for each site considering the various scientific, engineering, and operational aspects. This helps to separate good sites from less advantageous ones. Note, that for this test the landing sites considered were chosen in beforehand and they are not identical with the ones selected during the first round.

At the end, the selection team managed to identify 2 sites which were the best ones and which had prospect for a successful lander mission. One site was even reachable with a release strategy that is considered safest, and for this reason it was selected as primary landing site. One needs to know that the lander release uses two different ejection mechanisms one after the other, the so called nominal ejection which is adjustable up to speeds of 50 cm/s and the emergency release which has a fixed ejection speed of 18 cm/s. If the first option works, the second one is without effect for the lander descent; if the first one for whatever reason does not work, the second one makes sure that the lander is released in any case. However, different ejection speeds result in different descent trajectories, and it is not for sure that both end in a safe landing on the cometary surface. So, if the nominal and emergency ejection velocity are the same, the descent trajectory is basically identical regardless which ejection is going to work at the end.

The second landing site chosen would be reachable only for the nominal ejection velocity of 50 cm/s, but would not result in a landing on the nucleus if the emergency ejection would become effective during the release.

The third and last round of the test of the landing site selection procedure was performed on 6 June 2016. The main task is now to confirm the primary landing site as the best one and say 'Go for it'. In case of a No-Go for the first site, automatically the back-up site will become the site to target for the Philae landing. This time – and apart from refined assessments as during the 2nd round of the selection process – also information on the surface roughness is included in the evaluation process.

Surface roughness is another word for the importance of boulders and an uneven surface terrain at the landing sites. This information becomes available only through further and closer investigations – mainly via the cameras onboard the orbiter – of the terrain at the previously chosen primary and back-up landing sites. The boulder statistics and distribution at the chosen two landing sites are an important factor for the risk assessment of a safe Philae landing.

The outcome of the 3rd round of this test was a 'Go' for the primary landing site selected during round 2. The back-up site would have been also ok from the landing risk point-of view, but it relies on the perfect performance of the nominal ejection velocity only – see the discussion above.

Landing site selection process – an exercise, 1st round

23 May, 2014

On 16 May 2014 the first round for the test of the landing site selection process took place. It is an exercise that is meant to show that it is possible to select landing sites under conditions as realistic as possible. Of course, the necessary information on the comet is not yet available since Rosetta is still too far away. The necessary observations characterizing the nucleus and its physical and chemical properties cannot be done yet. So, one has to use a simulated target with assumed properties. On the other side, it is needed to determine, as much as possible in real time, the parameters for the landing and to perform also the selection of possible landing sites based upon the available information.

ESA had defined a target and provided its input information as needed for the exercise. This information is complemented by some – also simulated – information on the target provided from orbiter instruments. Both is done as much as possible using the original analysis tools and their results and output that are meant to be used for the real comet data.

The task of the first round of the landing site selection process was to select up to 5 possible landing sites that would be accessible for the lander – meaning it could land on any of them safely – and that would be of scientific interest for the instrument measurements during a Philae science mission at the simulated target. The process is done in three basic steps: (1) Find areas on the target that can be reached safely by Philae after separation from the orbiter. (2) Identify the landing areas that will allow to operate the lander during its science mission that consist of the descend phase, the first science sequence immediately after landing and the long-term science phase which is meant to last for about at least 3 to 4 months thereafter. (3) Identify the landing areas that promis to provide the scientifically most interesting prospect in order to achieve the science goals of the lander mission.

For each of the three tasks requirements and constraints exist. Task (1) and (2) have qualifiable and quantifiable outputs. Task (3) is only partially of the same character and it allows also preferences; for instance, a certain area is very good for measurements of one instrument, but not so much for those of other experiments. Of course, the goal is to satisfy all three steps consistently.

The practical implementation requires quite some effort for steps (1) and (2) which was done by engineers of CNES Toulouse and DLR Cologne under the leadership of the project management during the last few days. Today the results were presented to the lander team and namely to representatives of the scientists. The scientists were expected to provide their landing site selections based upon the engineering results, to iterate them with the engineering team and to destill up to 5 possible landing sites which are again evaluated by the lander project for their principle feasibility. Luckily and after some 6 hours of meetings, presentations, discussion, and separate work, the team present in Toulouse achieved this goal and 5 sites were chosen. It is not the end of the story, and more to come in a week's time – the second round of the landing site selection process.

Lander Commissioning after Hibernation - Part III

25 April, 2014

The 3rd block of the lander post-hibernation commissioning ran from 20 April to 23 April 2014 with most activities during the days in the middle of this period. This 3rd block focused on the continuation and completion of the commissioning of the lander instruments and on further interference tests between lander units.

The latter can be a testing activity with surprises, not so much for its execution, but in the results. Interference tests are performed between several units, typically two instruments, and involve the appropriate lander subsystems, that are meant to operate together. Usually, operational, commanding, power, and data transmission issues are not the point, it is rather whether the quality of the data collected by the instrument units are affected by the parallel operations.

This can happen due to electromagnetic disturbances produced by one unit and either radiated or conducted into the other unit. Of course, electromagnetic shielding principles and rules are applied during the unit designs and manufacturing; but since some instrument measurements make use of only a few electrons to identify a signal, the sensitivity to electromagnetic compatibility can be high for individual instruments. Testing this on Earth before launch is certainly important and it is done; the difficulty here is to achieve exactly the same test environment as later in flight and also to run the relevant operations sequences as foreseen for future scientific measurements. (The interference tests on Earth ran years before the details of the scientific measurements at the comet were defined). On the other side, major electromagnetic incompatibilities between spacecraft units have to be found before launch, since later on the means to resolve them are very limited. So, both testing on Earth and in-flight is important. What happened during the post-hibernation commissioning of the lander is the latter.

And indeed interferences between lander instruments were found: some expected ones at a minor level making the measurements of an instrument not useless, but degrading the quality, and unfortunately also one case with a really significant disturbance.

The knowledge of interference occurences is very useful for the future planning of instrument operations. In the first place, one can try to avoid parallel operations of units sensitive to interferences or refine the scheduling such that they can operate in parallel with the interfering operation of the originating unit occuring at a time when the receiving unit is insensitive to it. The latter uses the fact that the interference sensitivity of an instrument is localized in one specific electronic device which may not always be operated during the measurement sequence (for instance the read-out amplifier of the CCD electronics which works during read-out, but not during exposure integration). The latter solution requires, of course, a very detailed and accurate timing of the instrument operations.

So, altogether also block 3 of the post-hibernation commissioning was executed successfully and all well in time; and even an extra-test proposed by the Civa camera team during the commissioning period could be included. It was a period of day and night work and of the excitement since the lander and its instruments were operated when Rosetta has almost arrived at the comet. Next time it will count.

Lander Commissioning after Hibernation - Part II

23 April, 2014

The lander commissioning continued with part II on 14 April 2014. An important pre-requisite for this continuation was achieved already a few days before. Part I of the post-hibernation commissioning ended with an unexpected failure (see the blog entry from 13 April, 2014), and the lander was switched off in the state for radio transmission instead of using the ambilical connection to the orbiter. This status was luckily resolved on short notice by a lander command sequence introduced upon proposal by the European Space Operations Center ESOC in Darmstadt. A 2-hours shift in the night 11 to 12 April 2014 put the lander connection back to ambilical, and so the next commissioning could start in the proper start configuration of the lander.

Part II of the commissioning focusses on checking, testing and verifying mostly the proper function of the scientific instruments. Most of the data retrieved from these tests contain measurements without 'real targets' since the lander is not yet at the comet. For instance, the mass spectrometers and gas chromatographs detect gas which comes from the orbiter and lander environment itself, i.e. it is of terrestrial origin, or they can use some calibration gas from internal tanks of the instruments. There is however one Philae instrument which had a target in its field of view: The Civa panorama camera system showed the solar panels of Rosetta in two field of views of the camera – the other 5 cameras showed only diffuse light reflections from the spacecraft environment or from empty space. The solar panels are viewed from the back side, still at a weak photon level except for some bright reflections of the illuminating sunlight. There is even a dark shadow silhouette of the orbiter spacecraft noticeable in the foreground of the image scene. The end of the first day of hibernation part II saw the 'go' from all instruments (Consert, Cosac, Mupus, Ptolemy and SD2) expected to give the OK for the next commissioning activities.

During this week I follow the lander commissioning from the lander science operations center SONC at CNES Toulouse. This is where the instrument teams show up for important instrument operations or for periods where online decisions have to be taken for immediate implementation. Apart from following the activities of individual instruments it is also a good opportunity to stay in contact and discuss with the instrument responsibles. At the first day colleagues from the Civa camera team – also my lead scientist colleague Jean-Pierre Bibring - were present; in the 2nd day Consert and Romap instrument investigators joined.

The instrument teams have direct and immediate access to all measurement and housekeeping data of their instruments as soon as it arrives with SONC through the ESA ground network. When there is a direct link with the spacecraft through a ground station, it takes currently a bit longer than 30 min after the data was acquired onboard and sent through the antenna to Earth. Upon reception the scientists and engineers analyze the data and compare it with expected results from ground testings or from previous activities. Positive agreement is usually the confirmation that the command execution for the instrument onboard went well and the test was successful.
Everything is ok and finished or the next planned instrument activity can be initiated through the 'go' to the operations people.

Disagreement requires critical considerations before a 'go' is given and in some cases one prefers a 'nogo' for the next activity since either the pre-requisites for the next steps are not fulfilled or one needs to understand the actual instrument performance and the situation after the last test in greater detail before making a conscious decision on subsequent operations requests. Under these circumstances time is getting short since the time slots for lander operations are still rare; thus, the work pressure can get quite high and one still has to make the right decision. In the first place, the principle investigator of the respective instrument is the person who is meant to decide – possibly after consulting with the experts (scientists, engineers and technicians) in the team.
 
Also the 2nd day of the lander post-hibernation commissioning part II saw successful test executions onboard of the lander which resulted in the expected 'go's for the next steps. The same applies for the 3rd day although some data packets from a mass memory test were missing which at the moment is considered uncritical. A check sum error of an instrument software onboard made it necessary to upload the software package again before execution of the next activities. So, all in all also the 2nd part of the lander post-hibernation commissioning went without major issues and problems.

The lander was switched off for a short Easter break on 17 April 2014, 17:45UTC. Next and last part of the Philae post-hibernation commissioning is scheduled to start on 20 April 2014 21:00UTC.

Lander Commissioning after Hibernation - Part I

13 April, 2014

During calender week 15, the first part of the post-hibernation commissioning of the Philae lander was executed. The aim of these onboard activities is to check, test, and verify the proper functioning of the electronic, electric, mechanical, and instrument units as well as of the sensor and computer systems onboard and to identify and hopefully remedy malfunctionings and off-nominal behaviour of the onboard hardware and software. ESA has granted various periods between 28 March and 23 April 2014 for the post-hibernation commissioning of the lander. The respective lander activities are compiled in 5 blocks of a few days duration; in between the blocks the lander is switched-off.

Before its wake-up on 28 March 2014, the lander was switched off for more than 3 years, by far the longest time interval where ground control was out of contact with the spacecraft. Moreover, during this period it was exposed to environmental conditions never experienced before. So, before starting the science mission at the comet in late July to early August this year one would like to know and be sure that the lander units and instruments perform as they are supposed to do according to expectations of the engineers and scientists. This is not for granted, in particular after more than 10 years in space and with more than 3 years out of operations. That's why the post-hibernation commissioning is a major milestone in the Rosetta mission plan.

The lander sequence of commissioning events in space is actually prepared months in advance. The required operation requests are collected from the unit and instrument responsibles, are compiled in a commissioning activity plan and broken down into individual command sequences. Usually, various commissioning blocks are foreseen like switch on of lander, switch-on of units, unit status check, software update, individual unit tests, combined unit tests, preparation for subsequent switch-off phase.

The various commissioning steps require careful preparations and checking of results before the subsequent step gets the 'go' (or sometimes also the 'nogo') by the responible engineer or scientist. The 'go/nogos' have to be given before pre-defined deadlines, usually 15 min before the respective command sequence has to be uplinked to the spacecraft via the antenna ground station. Time frames between receptions of performance information of the onboard testing and check-outs and the 'go/nogo' decisions on the ground is sometimes long – several hours to days –, sometimes shorter – 1-2 hours. Depending on the importance of the on-board tests the respective unit and instrument responsibles follow the event at the lander control center at DLR Cologne-Porz and perform the result analysis on-site. Apart from the unit and instrument responsibles the key players for the Philae commissioning are the engineers of the lander operations team. They have prepared the commissioning sequence of the lander, and they run and control its execution at the spacecraft.

The whole commissioning activities are tested beforehand at the lander ground reference model which is situated at DLR premises in Cologne-Porz. It is done without short-cuts or simulations (unless hardwarewise required since the situation of the spacecraft cannot be represented one-to-one) and it takes at least as long as later on during the post-hibernation commissioning of the lander.

Part I of the commissioning focused on the lander units and on preparing the instruments for testing during Part II of the commissioning before and after Easter. For the unit commissioning a few firsts and highlights were accomplished. For instance a major new software version for the lander control system was uploaded. This software version is developed to support in a much more efficient and better way the autonomous lander operations during the long-term science phase after landing. The lander primary battery was conditioned, this way demonstrating that the battery is alive. This was not known since during the whole cruise phase the primary battery, that is a core element for the descend and landing as well as for the first days of the lander science operations on the surface, was not touched for safety reasons. Drive systems of the separation mechanisms were tested as well for the first time after launch in 2004. At that time the testing was aborted because of control problems and the subsequent risk analysis has shown that another attempt has the non-negligible chance to separate the lander unintentionally and long before the spacecraft has arrived at the comet. So, it was postponed and in the meanwhile a new risk-free test procedure was developed and verified by ground reference hardware set-ups. Drive motors of the landing gear were also turned the first time. This was critical since identical motors, that were kept for years under vacuum and cold conditions in a vacuum chamber at our institute, failed during their test sequence after more than 8 years not being used as is the case for the on-bord drives as well.

So, all went very well with the lander units and instruments commissioning during part I except until the very end. One of the last items to be tested was the radio frequency link between orbiter and lander and one of the tested configurations caused troubles such that the final preparation settings for the lander switch-off could not be transferred to Philae properly before it was switched off. This came as a complete surprise of an otherwise satisfying lander commissioning week and will require extra work to solve it or at least to get a work-around in place. For the moment the second part of the lander post-hibernation commissioning is meant to start on 14 April 2014, in 3 days from now.

After the Wake-up of the Philae Lander

3. April, 2014

The Philae lander woke up on 28 March 2014. It was a short event of a few hours with the goal
to show that the lander is alive. What happened? The lander got warmed up by heaters such that the compartment with most of the subsystems is in the proper temperature range for operations. Thereafter, the two computers onboard were booted, a new software compilation was uploaded, some housekeeping data of the lander were collected, and the execution information on the whole process plus some status data on lander subsystems was sent back to Earth. Thereafter, the lander was switched off again. The real post-hibernation commissioning of the lander spacecraft and its scientific instruments is still going to come in April.

The result of the first switch-on was as expected: the warm-up of the lander worked although it appeared a bit – not much – colder than predicted. The software upload was successful and the lander status information was ok as well. So, it looks like that the actual post-hibernation commissioning of Philae can start as planned. I very much hope that the outcome of it will fall along the lines of the lander wake-up. In a month’s time, things will be finished and it will be known, how the lander is prepared for the comet, the landing, and its science mission on the surface.

Around the lander wake-up a press event was organized for 28 March 2014, 14:30 to about 16:00LT, at DLR premises in Cologne-Porz. It was streamed into internet and can be seen here . I participated in it remotely from the lander control center at DLR grounds providing some answers to questions by the event moderator.

More interesting for me was the progress and outcome of the lander operations sequence performed onboard. Therefore, I joined the Philae operations people in the control room and tried to follow the activities from there. This has the beauty of being as close to the actions onboard as one can be on Earth. There is nothing foreseen to be done for a lead scientist in the control room during switch-on of the lander. It is all in the hands of the operations people. The operations sequence went really smoothly and the ops colleagues were quite relaxed such that they could explain to me the details of the data screening and of the online results. And at the end (or almost at the end of the ops sequence) the ops leader had the pleasure to announce to the officials, colleagues, journalists, and audience of the press event at the DLR Casino that the Philae lander woke-up successfully, in-time and as expected.

Pre-landing phase: planning in detail

23. March, 2014
What does the science operations planning (in short the sciops planning) for the pre-landing science imply from the lander side? It started already six years ago when the scientific objectives of the lander instruments were compiled and also ranked according to their scientific importance for the mission success. In total, nine scientific objectives to be achieved by the lander instruments were proposed and subsequently also ranked. The sciops planning we are in right now, however, requires more details like the exact execution sequence of the observations and the definition of the conditions of its execution. Important conditions are for instance nucleus distance, spacecraft pointing, power consumption, data volume, and also how often the measurements need to be performed.

This information comes from the experimenters. It is collected, synthesized, and compiled by an LCC engineer (Brigitte Paetz). After all, lander engineering and operations aspects (the lander platform) are important as well. The latter are usually not so much known in a greater detail to scientists. An important aspect for the lander sciops planning is to optimize the measurement sequences also with respect to minimize the number of switch-on periods for the lander.

All this is compiled in a document, iterated with all lander teams involved. At the end this document (we have a nice word for it, SOCOP) will be the backbone for the implementation of the sciops activities at the lander.

But we are not there yet. What needs to be done first is to establish the execution sequence for the overall pre-landing phase and to find proper slots for the execution of the lander science therein. That's were we are right now. The observation details as described above of all instruments that want to participate in the prelanding phase are 'thrown in' and negotiated by an ESA working group with all orbiter instruments and the lander participating. The goal is to find a suitable framework for the execution schedule and to solve conflicts between instrument requests. This is done via a series of telecons and also through in-person meetings. It can happen that not all wishes can be fulfilled and individual instruments may be assigned with less optimum conditions according to priority settings. And in very rare cases one would even have to drop measurements, if the resources are not available. These are unfortunate moments - not only for the experimenters of the instrument affected by such a decision, but also for the colleagues who are not affected.

If the framework for the sciops phase is set, the detailed scheduling is performed by ESA using computer and scheduling tools. The outcome of this scheduling process is provided to the experimenters again and, if no further conflicts occur and no further optimizations are required, it will be tried to implement them. The final outcome of this process is the sciops schedule for the pre-landing phase which is used on the lander side (by Brigitte Paetz) to finalize the SOCOP for this phase.

Next comes the command generation for the execution of the respective sequences of lander platform and instrument operations, and – very important – the testing of these sequences at the ground reference model GRM of the lander. This is required, since we want to make sure in the best possible way that the command sequences and the respective operations can be executed successfully and without mistakes and violation of constraints. Of course, usually the experimenters can test their individual instrument operations using a so-called instrument reference model. However, as a lander instrument they are not alone and always the lander platform runs in parallel to instruments and quite often two or more lander instruments operate simultaneously or combined such that a complete system test is required to get the true verification of the execution sequence done.

As lander lead scientist I am part of the whole process for the pre-landing sciops planning. Of course, scientific aspects of the lander receive my highest attention, but it usually also means to understand the operational and lander system implications, trusting the assessment of the instrument teams and of the engineers responsible for the lander platform as a whole. When the lander team started into the sciops planning, it meant to lead and guide and moderate the compilation and the ranking of the scientific objectives for the pre-landing and the other three phases of Philae science operations. Currently, we – Brigitte Paetz and me – are representing the lander community and interests for the science operations planning of the pre-landing phase at ESA level, i.e. we deal on one side with the planning working group on behalf of the lander and its instrument teams and with the lander instrument teams on the other side who want to measure during this phase. We hope that we can help to get the best and most scientific measurements scheduled and executed during the months before Philae separates from the orbiter for the landing enterprise.

Pre-landing phase: preparing for a safe touch-down

25. February, 2014

Where on 67P will Philae land? This is a question I am frequently asked in these days. The truth of the matter is: We don’t know yet. Frankly, at this point we know fairly little about the comet: only its orbital parameters and thanks to observations with Earth-based telescopes in the last years a rough estimate of shape, size, and rotation motion of the nucleus.

With our present knowledge of the comet, the selection of a safe landing site is not possible. If we would have to do it now, it would be like playing lottery with a very high risk to lose the lander, if we would dare to land. Scientifically, the most interesting information for the landing is the terrain geology, the constitution of the surface, and the activity of the landing area. For the descent and landing other aspects are important like: What is the gravity field and gas and dust flow along the descent trajectory? How is the slope and roughness of the landing terrain (both on length scales of a meter or so)? And for lander operations after landing it is important to get enough solar illumination for battery charging and to make sure that the radio link with the orbiter is possible. So, a number of measurements and parameters of the comet were defined by the lander team and by a dedicated working group among the Rosetta scientists. These measurements are meant to provide the information needed for a safe landing at a suitable landing site. Gathering this information is the prime objective of the pre-landing phase of Rosetta which is scheduled for the time between June and October 2014.

Of course, since there is only one lander, at the end only one landing site will be selected. However, one should have at least one back-up site. In fact, it is agreed with ESA to propose up to five landing sites for Philae of which possibly just the two best qualified ones will be considered for the implementation of the landing procedure.

The pre-landing phase provides the opportunity to perform the necessary measurements using orbiter instruments during the approach and rendezvous phase of Rosetta, i.e. from a distance of more than one million kilometers to close orbits some 10 to 20 kilometers above the surface.  Clearly, observations needed for the landing site selection will have priority over those to be done for purely scientific reasons. Nonetheless, ESA wants also to give purely scientific measurements of the comet a chance during the pre-landing phase. And the lander instruments want to do scientific measurements during pre-landing. It is also obvious that the measurements with priority for the landing site selection will provide scientific useful information on the comet per se, but this is not the driving aspect for its performance. 

All orbiter instruments and the lander want to participate with observations during the pre-landing phase. From the lander side five (out of ten) instruments have plans to measure cometary properties during the pre-landing science phase: the CIVA cameras want to take images of the nucleus, the two gas analyzers COSAC and PTOLEMY will be used to 'sniff' the coma in order to identifiy gaseous species therein, the magnetometer part of the ROMAP instrument shall provide reference measurements of the magnetic field in the cometary coma close to the spacecraft, and SESAME intends to study the dust and plasma environment of the coma. 

Next week, I’ll detail how the science operations are planned for the pre-landing phase. This is not an easy task, since many instruments and actions need to be coordinated. It’s a bit like defining the choreography for a complex performance with many actors, sometimes a bit like handling several 'prima donnas' : the scientists of the instruments onboard who are convinced that the own measurements are the most important ones, which per se is ok, but not always easy to satisfy…

How Philae warms up for the big game

10. February, 2014

The science planning for the lander measurements in the neighborhood and on the surface of the comet are currently more and more taking shape. Yes, you read correctly: the lander wants to measure the comet and its environment already while Rosetta is approaching the nucleus and while the lander is still safely attached to the orbiter. This is, of course, not the key science phase of the lander instruments; the best and most valuable scientific results from the lander will come from the surface of the nucleus. But this mission is so unique, that every bit of data is valuable.

The lander will investigate the comet in four phases, i.e. the pre-landing phase, the separation, descent and landing phase (called SDL), the first science sequence after landing (called FSS), and the long-term science phase (called LTS). Although not producing immediate scientific results, but equally important is the post-hibernation commissioning of the lander and its instruments. After all, the lander was switched off and its equipment was not used, monitored, and maintained for a period of more than two years. Imagine doing that with your car….

The post-hibernation commissioning is the next step ahead. It’s a bit like the warm-up before an important soccer game. We want to make sure, that the lander’s hardware and software are in good shape for what’s to come.  

This commissioning is already fully planned and it is scheduled to be performed between the end of March and during April 2014. It includes all subsystems and instruments onboard the lander and will be executed in phases.

Apart from health checks and testing of execution sequences of the instruments and subsystems, a major software upload for the onboard computer and data management system CDMS is forseen. This software upload shall enable a much better autonomous performance of the lander during the so called long-term science phase on the comet. The advances are in the handling of the power subsystem and of the lander-orbiter communication; both represent key elements and tight resources for the LTS phase that will now be handled by the CDMS with only very little expected interventions from the ground control center.

And last but not least, a number of so called interference tests and calibration sequences are scheduled for the lander instruments. The latter will ensure that all scientific measurements performed on the comet will be properly supported. The former are important to identify instruments that might get in each other’s way when operated simultaneously. This will be necessary throughout the science program of the lander. Interference tests of the lander instruments are done not for the first time and only a few are left to be done, before we’re ready.  

So who does all the work? The lander commissioning is prepared by the lander operations team at the lander control center LCC in Cologne. Its schedule needs to be coordinated with the overall post-hibernation commissioning schedule of the Rosetta spacecraft which started right after the wake-up and which is implemented by the European Space Operations Center ESOC Darmstadt. The role of the engineers and scientists of the lander is to propose and prepare the details of the commissioning operations of the respective subsystems and instruments they are responsible for. The system engineer and project management needs to emphasize the global context for the commissioning, in particular that all that is needed for the coming science operations is operational, properly tested and functioning and that deviations and malfunctions are understood and solved. My role as lander lead scientist is more in the background, mostly trying to encourage that this commissioning should target the science of the lander and to understand science implications, if things do not perform as expected.

In any case, on the basis of the available schedule for the post-hibernation commissioning of the lander, I have identified the periods of my participation and presence (in total 9 days) at the two control centers of the lander, i.e. at LCC at DLR Cologne for the subsystems activities and at the science operations and navigation center SONC at CNES Toulouse for the instrument commissionings. The schedule will require to work in night shifts and also some Easter holidays will get lost for private and family life.

After the Rosetta wake-up

25. January, 2014

About plus 5 days after the wake-up signal of the Rosetta spacecraft arrived Earth at 20 January 2014 19:18 CET...

A first glance at the status of the in-flight machinery arrived in the meanwhile. It did not contain much information on the lander, just some temperature sensor values that were found to be close to the expected ones, e.g. depending on the location between -30 and -110 C. From that, one can take the message that the lander is on-board, it gets power from the mother spacecraft through the umbilical and its autonomous hardwired temperature control and heating work. This is not a proof that the lander is in good shape all over, but it is considered positive. Apart from the temperature control, all other operational and control subsystems as well as the instruments of the lander were switched off during hibernation and they still will be switched off for another two months.

The excitement of the wake-up event shared by the officials, project colleagues, journalists, interested people, and the public has decreased. Looking back, it was an important step and a large achievement for the mission which was, after all, designed by the engineers more than a decade ago. The autonomous wake-up procedure itself was implemented operationally about 3 years ago. It was not the first time that an ESA comet mission was in hibernation, but it was the first example that it was part of the original mission plan foreseen from the beginning. And it was the first time that the spacecraft did not need a wake-up call from the parent engineers on Earth: an internal alarm clock did the job. The first ESA comet mission Giotto hibernated as well. It was set asleep in 1986 after the fly-by at comet Halley and was woken up again in 1992 for the fly-by at comet Grigg-Skjellerup.

By the time the wake-up signal of Rosetta arrived we (colleagues from my home institute, my wife and me) had already left the wake-up event at ESOC Darmstadt and we were waiting for the train at the main station. We had not given up on Rosetta, we just had to leave ESOC to get home the same day. My wife received the news about the wake-up signal reception via cellular phone from Jochen Kissel, former Principal Investigator of the COSIMA instrument and a good friend of us. Although we could not participate in the celebration party of the Rosetta success at ESOC, we drank a toast to the brave spacecraft at the train and later at home.

Of course the wake-up signal is just the beginning of the things to come. The mission hopefully holds further excitements - for sure positive and negative ones - as well as expected and also unexpected scientific highlights. The year is full of important progress steps until the lander has anchored at comet 67P/Churyumov-Gerasimenko (in short 67P) and performs its scientific measurements on the surface of the nucleus.