Friday, 27 March 2020

One year after India's ASAT test

click diagram to enlarge

Today it is one year ago that India performed an ASAT test codenamed 'Mission Shakti'. The test consisted of the on-orbit destruction of the Microsat-R satellite (2019-006A), launched specifically to function as target for this test. The intercept occurred at 285 km altitude, but created debris pieces with apogee altitudes much higher than that. I have earlier published an extensive OSINT analysis of the test in The Diplomat of 30 April 2019.

The test generated large amounts of debris. A total of 125 larger debris pieces have been tracked and catalogued by the US tracking network. Note that these only concern larger pieces: most of the generated debris probably was too small to be tracked.

Over the past year I have periodically posted an update on the status of these larger debris pieces on this blog. Whereas the Indian DRDO claimed at the time that all debris would have been gone 45 days after the test, the reality has been quite different: 45 days after the test, 29% (less than a third) of the larger debris pieces had reentered. It took 121 days for half of the pieces to reenter, and some 200 days before 75% of the tracked debris pieces had reentered.

One year after the test, some 114 of the tracked debris pieces have reentered according to CSpOC tracking data. And two more objects for which no decay message was published by CSpOC, 2019-006AR and EA, have reentered according to my own analysis with SatEvo, bringing the total tally of reentered larger tracked pieces to 116.

Nine, or some 7%, of the original 125 larger tracked debris pieces are still on orbit.

It concerns objects 2019-006V, AJ, AX, BD, DC, DD, DE, DM and DU (red orbits in the image below: the white orbit is that of the ISS, as a comparison).  They have apogee altitudes varying from 600 to 1500 km, and perigees generally near 260 to 280 km. Six of these are expected to reenter over the next half year 9 months. And the last debris pieces may not reenter before 2022-2023.


click image to enlarge

Wednesday, 25 March 2020

SpaceX's Starlink Darksat is, indeed, darker

clic image to enlarge
The image above is a composit of stacked frames from four video sequences shot in the evening of March 22. Apart from a stray Chinese rocket booster that happened to cross the field, it shows four Starlink satellites from the 2020-001 launch: Starlink-1114 (2020-001P), Starlink-1030 (2020-001N), Starlink-1084 (2020-001B) and Starlink-1098 (2020-001D). These satellites are currently at their intended operational altitude.

Starlink 1030 is also known as DARKSAT
- it is the Starlink satellite that has been given an experimental coating to reduce its brightness.

As can be seen in the video stack, the coating indeed seems to reduce the brightness. The effect is also very apparent in the photographic imagery below, comparing Darksat to two other operational altitude Starlink satellites in the same orbital plane, Starlink-1114 and Starlink-1084 that both passed within 5 minutes of Darksat. The two regular Starlink satellites are well visible, but Starlink-1030 Darksat is very faint in the image:

click to enlarge
The video images were taken with a WATEC 902H and Canon FD 1.8/50 mm lens at 25 fps. The photographic images were taken with a Canon EOS 80D + EF 2.5/50 mm lens at 1000 ISO, 10 seconds exposure.

It is difficult to attach reliable magnitudes to the video and photographic imagery, but I'd say the magnitude difference between Darksat and the others is probably in the order of 1 to 2 magnitudes. Given their shape, the brightness which Darksat and other Starlink satellites can attain will probably be  highly depending on the viewing angle (as well as of course the phase angle at time of observation), i.e. which part of the satellite you are looking at.

Saturday, 7 March 2020

Dragon CRS-20, 23 minutes after launch, with thruster firings

click image to enlarge

SpaceX launched the Dragon CRS-20 cargoship to the ISS this morning at 4:50:31 UT. Some 23 minutes after launch from SLC-40 at Cape Canaveral in Florida, it was visible from the Netherlands around 6:13 local time (5:13 UT) in morning twilight. There were some fields of clouds in the sky, but I nevertheless got a clear view of the four objects associated to the launch, all still closely together.

The image above is a 2-second exposure at 800 ISO which I took during the pass, using a Canon EOS 80D DSLR and a SamYang 1.4/85 mm lens. The image shows the trails of  four objects, two of which are tumbling. In the annotated image below, I identify what is what:

click image to enlarge

The Dragon cargoship, the Falcon 9 upper stage and the two solar panel covers were easy naked eye objects. The Dragon and Falcon 9 upper stage were very bright and steady, while the two solar panel covers slowly flashed alongside them. These solar panel covers varied in brightness between invisible (with the naked eye) and magnitude +1.5. The Falcon 9 upper stage and Dragon were about +1.5 to +2: with the naked eye, being very close together they seemed one object, while on the photographs they are clearly two.

The image below, taken a few seconds after the previous image, shows one of the tumbling, slowly flaring solar panel covers at its brightest, rivalling the Dragon and Falcon 9 upper stage in brightness:

click to enlarge

The slow regular flashing behaviour was nice to see: the two tumbling solar panel covers were alternating, when one of the two was bright, the other was faint (clearly visible in the image above and the video below). Due to the alternatingly flashing panel covers above and below the Dragon, it looked a bit like an aircaft.

I also captured a small part of the pass on video, using the WATEC 902H with a 1.8/50 mm lens on a fixed tripod in autonomous mode (I was outside myself witha sceond tripod and the photo camera). In this video segment (below), a thruster firing is visible as a cloudy upwards moving "puff"starting at 5:13:00 UT:



Dragon CRS-20 will berth to the ISS on Monday 9 March near 11:00 UT.

This was the last flight of a Dragon 1, and the concluding flight of a contract awarded in 2008. All future Dragon supply flights will be done by an updated model, the Dragon 2 as well as the crew-rated Crew Dragon variant of the latter.

Friday, 21 February 2020

Launching cubesats from the X-37B OTV 5: lifetime modelling with GMAT

image: USAF

Last week, CSpOC issued catalogue entries for three cubesats released as part of the X-37B mission OTV 5.

It concerns USA 295 (2017-052C), USA 296 (2017-052D) and USA 297 (2017-052E). No orbital data are given, but the catalogue entry did explicitly indicate that all three are no longer on orbit.

That cubesats were released as part of this X-37B mission had been clear from a US Air Force statement made after completion of the OTV 5 mission in October last year. The wording of that statement is however ambiguous: while most analysts take it to mean the cubesats were released by OTV 5, it is also possible that they were released as ride shares by the upper stage of the Falcon 9 rocket that launched OTV 5 in 2017.

In this blog post, I will do an academic exercise aimed at guessing when, at the latest, these cubesats could have been released by OTV 5, assuming release from the latter.

OTV 5, the 5th X-37B mission, was launched from Cape Canaveral on 7 September 2017. It landed at the Kennedy Space Center Shuttle Landing Facility on 27 October 2019, after 780 days in space. Unlike previous missions that were all launched in 38-43 degree inclined orbits, this one was launched into a 54.5 degree inclined orbit. Combined with the fall launch date, this meant it took our tracking network a while to locate it on-orbit: the first positive observations were made in April 2018, half a year after launch.

From April 2018, when we started to track it, to October 2019, when it landed, OTV 5 orbitted at various orbital altitudes between 300 and 390 km altitude (see diagram below):

click diagram to enlarge

The CSpOC catalogue entry lists all three cubesats that were released as part of this mission as "no longer on orbit". Assuming they ended their orbital life by natural decay (rather than, for example, being retrieved by OTV 5 again at a later stage, which is in theory certainly possible!), the fact that they were no longer on orbit by 11 February 2020 might yield some constraints on when they could have been released.

To get some idea of the orbital lifetime of a cubesat released from OTV 5, and spurred on to do so by Jonathan McDowell, I ran several GMAT models in which I modelled a 5 kg 3U cubesat released at three altitudes: 400 km, 360 km and 325 km.

We do not know the actual orbital altitude of OTV 5 at that  moment. Nor do we know when the cubesats were released. Hence the three altitude variants. The start point of the modelling was an assumed release into the OTV 5 orbit on October 7, 2017, one month after launch of OTV 5.

For each cubesat, the models were run in two variants: one with the cubesat in minimal drag orientation (0.01 m2 cross section), and one with the cubesat in maximal drag orientation (0.03 m2 cross section). I used the MSISE90 atmosphere in the model, with historic Space Weather data for October 2017 to February 2020 and estimated solar and geomagnetic activity parameters from the 'early cycle' variant of the GMAT Schattenfile for dates past early 2020.

For the three assumed orbital altitudes and an assumed release one month after OTV 5 launch, the GMAT data produce the orbital decay plots below. In these plots, the red data are for minimal drag orientation, the blue data for maximal drag orientation. If the cubesats in question were similar to NRO's Colony II cubesats, then the red minimum drag orientation curves probably represent the orbital evolution best. If they were more like Colony I cubesats, then the blue maximal drag curves are more representative.




Taking the minimal drag variants, and under the assumption that the cubesats were 3U cubesats and not retrieved on-orbit by OTV 5 at a later stage, the suggestion is a release below 350 km. Released at higher altitudes, they would still be on-orbit.

Assuming reentry before 11 February 2020 after natural orbital decay, a minimal drag orientation and release no lower and no higher than 325 km, the latest possible moment of release would be late August 2018, give or take a month to account for the uncertainties.

It appears we can rule this out however, because we know that OTV 5 was orbiting at 380 km altitude, not 325 km altitude, at that time. So the best guess (although one under many assumptions) is a release some time before August 2018, i.e. within 1 year after the launch of OTV 5.

It is still possible that the cubesats were released at a later date, but next retrieved while still on-orbit by OTV 5. If the cubesats were smaller than a 3U cubesat, a later release than August 2018 is possible as well.

Finally, given the ambiguity in US Air Force Statements on the matter, it is also possible that the cubesats were released from the Falcon 9 upper stage on the day of launch.

For more about the X-37B, and especially the active myth-making that seems to be at play around this secretive space-plane, see my earlier post here.


OTV 5 rising in April 2018. Click image to enlarge

Tuesday, 11 February 2020

Iran's failed Zafar launch: where did it go?

Zafar-1 launch on 9 Feb 2020. image: IRNA

On 9 February 2020 at 15:48:15 UT, Iran tried to launch a new satellite, Zafar-1, on a  Simorgh (Safir-2) rocket. Video released by the Iranian government shows that lift-off was succesful, and so was first stage separation and second stage ignition around 15:50:00 UT, and fairing separation around 15:50:18 UT. The upper stage next however failed to reach the necessary speed to put the satellite into earth orbit.

The intended orbit according to Iranian sources was a 530 km altitude, 56-degree inclination orbit. Orbit insertion however failed because the Simorgh upper stage burnt out at a speed of 6.533 km/s, almost 1 km/s short of the necessary 7.4 km/s,  according to the Iranian minister of Communications and Information Technology, Mohammad Javad Azari Jahromi. The upper stage and satellite reached an apogee at 541 km altitude before making a long ballistic flight back to earth surface.

Zafar 1 on top of the Simorgh rocket at Semnan. Photo: IRNA

In order to get some idea where it's flight ended, I have modelled the failed launch in STK and GMAT.

The ascend to 541 km altitude was modelled in STK, with launch into the azimuth needed to reach a 56.0-degree orbital inclination (launch azimuth about 134.7 degrees - this was calculated with software I have written myself). I positioned apogee such as to correspond with an attempted orbit insertion about 10 minutes after launch (a typical value for launch into lower LEO). Burnout speed was put at 6.533 km/s, per Iranian sources.

The resulting State Vector was then used as input in GMAT to model the ballistic descend. I did this for two cases: for a 90 kg mass, ~0.25 m2 cross-section object corresponding to the Zafar satellite; and for a 1000 kg mass, 4.5 m2 cross-section object corresponding to the spent Simorgh upper stage. As I had no values for mass and size of the latter, I used values similar to a North Korean UNHA-3 upper stage. The MSISE90 atmosphere with current Space Weather was used in GMAT.


click map to enlarge

The result of this modelling is impact in the Indian Ocean some 25 minutes after launch and some 6400 km downrange from Semnan, at about 12 S, 88 E, for both the satellite and the Simorgh upper stage (see map above). These values should not be taken too strictly, given several uncertainties in the model input: they are ballpark figures.

As it turns out in this case, varying the mass and size have mostly minor effects on the impact position only (note: in an earlier modelling attempt posted on Twitter, the impact point came out closer to Iran, because in that initial model run I had been using a lower burnout speed).

Friday, 31 January 2020

California 30 January 12:30 UT: the "space debris" reentry that wasn't




On 30 January 2020 near 12:30 UT (10:30 pm PST), a bright, slow, spectacularly fragmenting fireball swooped over southern California. It was seen and reported by many in the San Diego-Los Angeles area. The video above was obtained by a dedicated fireball all-sky camera operated by Bob Lunsford. The fireball duration approached 20 seconds.

In the hours after the fireball, the American Meteor Society (AMS) initially suggested that this was a Space Debris reentry, i.e. the reentry of something artificial from earth orbit.

But it wasn't.

Immediately upon seeing the video, I had my doubts. Upon a further look at the video, those doubt grew. To me, the evidence pointed to a meteoritic fireball, a slow fragmenting fireball caused by a small chunk of asteroid entering our atmosphere.

A discussion ensued on Twitter, until NASA's Bill Cooke settled the issue with multistation camera triangulation data, which showed that this was an object from an Apollo/Jupiter Family comet type heliocentric orbit with a speed of 15.5 km/s. In other words: my doubts were legitimite. This was not a space debris reentry but indeed a chunk of asteroid or comet.

I've already set out my argumentation about my doubts on Twitter yesterday, but will reitterate them again below for the benefit of the readers of this blog.

My doubts started because while watching the video I felt that the fireball, while slow and of exceptionally long duration, was still a tad too fast in angular velocity in the sky, and too short in duration, for this to be space debris. In the video, it can be seen to move over a considerable part of the sky in just seconds time.

The image below shows two stills from the video 6 seconds apart in time. The fireball passes two stars, alpha Ceti and beta Orionis, that are 35 degrees apart in the sky, and it takes the fireball a time span of about 6 seconds to do this, yielding an apparent angular velocity in the sky of about 5-6 degrees per second. That is an angular velocity that is a factor two too fast for reentering space debris at this sky elevation, as I will show below.

stills from the fireball video, 6 seconds apart, with two stars indicated

Orbital speed of a satellite is determined by orbital altitude. Reentering space debris, at less than 100 km altitude, has a very well defined entry speed of 7.9 km/s. This gives a maximum angular speed in the sky of about 5 degrees/second would it pass right above you in the zenith (and only then): but gives a (much) slower speed (2-3 degrees/second) when the reentry is visible lower in the sky, such as in the fireball video.

To gain some insight in the angular velocity a reentering piece of space debris would have at the elevation of the California fireball, I created an artificial 70 x 110 km reentry orbit over southern California that would pass the same two stars as seen from San Diego.

The map below shows that simulated track, with the object (marked by the green rectangular box) at 70 km altitude and positioned 6 seconds after passing alpha Ceti (marked by the green circle):


Simulated reentry track. click to enlarge
The angular velocity in the sky for a reentering object at this sky elevation suggested by this simulation is barely half that of the fireball. During the 6 seconds it took the fireball to move over 35 degrees of sky passing alpha Ceti and beta Orionis, the simulated reentering object would have moved over only 15 degrees, i.e with an angular velocity of 2.5 degrees/second rather than the 5-6 degrees/second of the fireball.

So this suggested that the fireball was moving at a speed a factor two too high for space debris. This therefore pointed to a meteoritic fireball, not a space debris reentry.

There were other reasons to doubt a reentry too. There were no matching TIP messages on Space-Track, the web-portal of CSpOC, the US military satellite tracking network. A reentering object as bright as the fireball in the video would have to be a large piece of space debris: this bright is clearly not the "nuts and bolts" category but suggests a large object like a satellite or rocket stage. It is unlikely that CSpOC would have missed a reentry of this size.

To be certain I ran a decay prediction on the full CSpOC catalogue with SatEvo myself: no object popped up that was expected to reenter near this date either, based on fresh orbital elements.

The fragmentation in itself, one of the arguments in the AMS' initial but mistaken conclusion of a "space debris reentry", is not unique to space debris reentries. It is also a common occurence with slow, meteorite dropping asteroidal fireballs, especially when they enter on a grazing trajectory. Take the Peekskill meteorite fall from October 1992 for example:




Likewise, while a 20-second meteor is not everyday, it is not a duration that is impossible for a meteor. Such durations (and even longer ones) have been observed before. Such long durations are especially the case with meteors that enter in a grazing way, under a shallow angle.

At the same time, a 20 seconds duration would be unusually short for a satellite or rocket stage reentry. Such reentries are usually visible for minutes, not a few seconds or a few tens of seconds.

So, to summarize:

1) the angular velocity in the sky appeared to be too large for space debris;
2) the fireball duration would be unusually brief for space debris;
3) and there were no obvious reentry candidates.

On the other hand:

a) the angular velocity would match those of slow ~15 km/s meteors;
b) the 20 second duration, while long, is certainly not impossible for a meteor;
c) the fragmentation observed occurs with slow asteroidal origin meteors as well.

Combining all these arguments,  my conclusion was that this was not a space debris reentry, but an asteroidal origin, slow meteoritic fireball. This was vindicated shortly later by the multistation camera results of Bill Cooke and his group, which yielded an unambiguous speed of 15.5 km/s and as a result a heliocentric orbit, showing that this was not space debris but a slow chunk of asteroid or Jupiter Family comet.

In defense of the American Meteor Society (who do great work on fireballs): it is not easy to characterize objects this slow, certainly not from single camera images and visual eyewitness reports. Given the slow character and profuse fragmentation, it is not that strange that the AMS initially (but incorrectly) thought it concerned a space debris reentry. It does go to show that you have to be extremely careful in drawing conclusions about slow moving fireballs: not every very long duration fragmenting fireball is space debris.

Monday, 20 January 2020

Testing a new lens for GEO and HEO (SamYang 2.0/135 mm)




The past week brought some clear skies. It also brougt me a new lens, a SamYang 2.0/135 mm ED UMC.

This lens had been on my wish-list for a while, as a potential replacement for the 1979-vintage Zeiss Jena Sonnar MC 2.8/180 mm I hitherto used for imaging faint Geosynchronous (GEO) and Highly Elliptical Orbit (HEO) objects, objects which are typically in the magnitude +10 to +14 range.

The 2.0/135 mm SamYang lens has gotten raving reviews on photography websites, several of these reviews noting that the optical quality of this lens is superior to that of a Canon 2.0/135L lens. And this while it retails at only half the price of an L-lens (it retails for about 460 to 500 Euro).

While I have the version with the Canon EF fitting, the SamYang lens is also available with fittings for various other camera brands.

Focussing is very smooth and easy with this lens. Unlike a Canon-L lens, the SamYang lens is fully manual (both focus and F-stop), but for astrophotography, manually focussing is mandatory anyway. The general build of the lens is solid. It is made of a combination of metal and plastic.

While not particularly lightweight, the lens is lighter in weight than my 1979-vintage Zeiss (which is all-metal and built like a tank, in true DDR fashion). The SamYang has a somewhat larger aperture (6.75 cm) than the Zeiss (6.42 cm), meaning it can image fainter objects. It also has a notably wider field of view (9 x 7 degrees, while the Zeiss has 7 x 5 degrees).

So for me, this seemed to be the ideal lens for GEO and HEO.

And after two test nights I can confirm: this SamYang lens indeed is spectacularly sharp. The first test images, made on January 15 and 16, have truely impressed me. Even at full F2.0 aperture, it is sharp from the center all the way to the edges and corners of the image.

Here is a comparison of the image center and the upper right corner of an image, at true pixel level. There is hardly any difference in sharpness:

click to enlarge

The images below, taken with the SamYang on a Canon EOS 80D, are crops of larger images, all but one at true pixel level.

The first image is a test image from January 15, a nice clear evening. It shows two objects in HEO: a Russian piece of space debris (a Breeze-M tank), and the classified American SIGINT satellite TRUMPET 1 (1994-026A). Note how sharp the trails are (this is a crop at true pixel level):

Click image to enlarge

The next night, January 16, I imaged several geostationary objects (which at my 51 degree north latitude are low in the sky, generally (well) below 30 degrees elevation). While the sky was reasonably clear, there were lingering aircraft contrails in the sky, locally producing some haze. Geostationary objects showed up well however, better than they generally did in the Zeiss images in the past.

The image below, which is a crop of a larger image, is not true pixel size, but slightly reduced in size to fit several objects in one image. It shows the Orion Nebula, several unclassified commercial GEO-sats, the Russian military comsat KOSMOS 2538 (BLAGOVEST 14L), and the classified Italian military communications satellite SICRAL 1B (2009-020A):


Click image to enlarge


The images below are all crops at true pixel level. The first one shows the US classified SIGINT satellite PAN/NEMESIS I (2009-047A), shadowing the commercial satellite telephony satellite YAHSAT 1B. It also shows a number of other unclassified commercial GEO-sats.

PAN/NEMESIS 1 is an NSA operated satellite that eavesdrops on commercial satellite telephony (see my 2016 article in The Space Review).

Note that this image - just like the next images- was taken at very low elevation, and from a light-polluted town center.

click image to enlarge

The image below shows another US classified SIGINT satellite, Mentor 4 (2009-001A), an ADVANCED ORION satellite. It shadows the commercial satellite telephony satellite THURAYA 2 (more backgrounds on this in my 2016 article in The Space Review). At magnitude +8, it is one of the brightest geosynchronous objects in the sky (note how it is much brighter than THURAYA 2):

click to enlarge

The last image below again is a classified US military SIGINT satellite, MERCURY 2 (1996-026A). While 24 years old it is, together with its even slightly older sibling MERCURY 1 (which I also imaged but is not in this image), probably still operational:

Click image to enlarge

After these two test nights, I am very enthusiastic about the SamYang lens. It is incredibly sharp, also in the corners, easy to focus, goes deep (in terms of faint objects), and overall performs excellent. I also like the wide field of view (compared to the 180 mm Zeiss which I previously used to target GEO). Together with the equally well performing SamYang 1.4/85 mm, it might be the ideal lens for imaging GEO and HEO.

Astrometric data on the targetted satellites from these test images are here and here. The astrometric solutions on the star backgrounds in the images had a standard deviation of about 2".

Added 20 Jan 2020:

This last image (reduced in resolution to fit) was taken this evening (20 January) and shows Trumpet 1 (1994-026A) passing the Pleiades:

Click image to enlarge


Sunday, 19 January 2020

Imaging Starlink 2

click to enlarge

A new set of 60 Starlink satellites, Starlink 2 (the third launch), was launched by SpaceX early on January 7th. Over the past 10 days, all passes were in earth shadow for my 51 degree North latitude, but as of this weekend, the satellites start to make low visible passes in evening twilight.

Yesterday evening was one of the first opportunities. The Starlink satellite "train", already dispersing as their orbits are raised, would make a pass low south with a maximum elevation at 28 degrees, where they would enter earth shadow.

Conditions were dynamic, with fields of clouds moving in the sky. Initially, the part of the sky where they should be brightest was obscured by a cloud, so I pointed the camera more west and lower in the sky.

The image below is a stack of 65 images, 5 seconds exposure each with 1 second intervals, taken between 17:52:50 - 17:59:15 UT (representing a 6m 25s period), with a Canon EOS 80D and EF 2.5/50 mm Macro lens set at F2.8, 1000 ISO. There is a band of Starlink objects, diagonally from lower right to upper left crossing behind the tree. These are objects in the 'head' of the Starlink 2 main "train":

click to enlarge

When the sky near the satellite culmination point also cleared of the field of clouds, I repositioned the camera to that point and captured the last part of the main "train" tail.

The first image below is a stack of 10 images, taken between 17:59:30 - 18:00:30 UT, representing a 1-minute period. The objects can be seen entering earth shadow at left.

The second image below is a single shot image (5-second exposure) from that series, showing four Starlink objects.


click image to enlarge

click image to enlarge

Near their culmination point, the Starlink satellites were clear naked-eye objects, with a brightness of approximately mag. +2.5 tot +3.0.

The images were taken from the center of Leiden town in the Netherlands, in a twilight sky that suffers quite some light pollution.