THE SECRET SPIES IN THE SKY - Imagery, Data Analysis, and Discussions relating to Military Space
SatTrackCam Leiden (Cospar 4353) is a satellite tracking station located at Leiden, the Netherlands. The tracking focus is on classified objects - i.e. "spy satellites". With a camera, accurate positional measurements on satellites of interest are obtained in order to determine their orbits. Orbital behaviour is analysed.
This blog analyses Missile tests too.
In the early morning of February 19, 2025, near 3:45 UTC (4:45 CET) scores of people in the Netherlands, UK, Germany and Denmark saw a slow, fragmenting object move through the sky. The event garnered much attention.
Above is a video of the event obtained by the camera of @dutchspace.bsky.social in the Netherlands. Another stunning video, by the meteor camera of André Knoeffel in Brandenburg, Germany, is here.
The event was caused by the reentry of a SpaceX Falcon 9 upper stage, object 2025-022Y (cat nr 62878) . This upper stage, from a Starlink launch on 1 February 2025 (Starlink group 11-4), failed to deorbit in a controlled way after inserting the payloads into orbit. It stayed on orbit for almost 3 weeks, and then had an uncontrolled deorbit over NW Europe on February 19.
The general character of the event (slow, fragmenting) as well as time and location fit well with the Falcon 9 upper stage reentry. The CSpOCfinal TIP for this reentry is 19 February 3:43 UTC ± 1 min, near 52.8 N, 5.6 W. The reentry trajectory runs over the British Isles, the northern Netherlands, northern Germany and Poland.
Below is the ground-track for the last orbit up to reentry, and a detail of the approximate final reentry track over Europe:
click map to enlarge
click map to enlarge
A possible piece of debris has been found at Komorniki near Poznan in Poland, media reports. The object on the photograph looks like a Composite Overwrap Pressure Vessel, a component of the Falcon 9 upper stage that has survived reentry before, e.g. in March 2021 when a Falcon 9 upper stage reentered over Washington. The location of the find matches well with the reentry trajectory, as can be seen in the map above. [update: reportedly two more of these composite-wrapped pressure tanks were subsequently found in Poland]
The event caused a lot of media attention: I did several media interviews on the reentry, amonst others with two Dutch national tv channels.
In two contributions to The Space Review (part 1 here and part 2 here) for the first instance, as well as two follow-up blog posts for later manoeuvers (here, and here), I analyzed three periodic orbit raising manoeuvers by the North Korean military reconnaissance satellite Malligyong-1 (2023-179A) done in February, June and September 2024. They followed a similar pattern: an orbit raise (orbit maintenance manoeuver) in five daily increments, all taking place on late evening passes (13 - 14h UTC = 22 - 23h local time) passes over North Korea.
Now Malligyong-1 has manoeuvered for a fourth time. And while there are some similarities, this time the followed pattern was different.
The orbit raising manoeuver was done between 2025 January 16 and 18, a month later than I had anticipated. However, as can be seen in the diagram above, it was this time done in three daily increments, not five as was the case in previous orbit raising events.
Compared to the previous orbit raises, the raise during each daily increment was larger, some 2.5 km per increment rather than 1.2 km as in the previous cases, for a total orbital altitude raise of about 7.5 km, which is also somewhat larger than the previous orbit raises (which was 4 km in February 2024 and 6 km in June and September 2024). As during the previous orbital altitude maintenance raises, it brought back the orbital altitude to the initial value from November 2023, when the satellite was launched. See the diagrams below:
click diagram to enlarge
click diagram to enlarge
The three sequential manoeuvers between January 16 and 18 raised both apogee and perigee. The daily rate of RAAN precession is still very close to the ideal sun-synchronous value.
click diagram to enlarge
As was the case for all previous orbit raises, the times I reconstruct for the three incremental orbit raises correspond to the orbital plane of Malligyong-1 passing over or near North Korea (for the method used, see here);
# DATE UTC LAT LON ORBITS USED FOR ANALYSIS RAISE M1 16-01-2025 13:14 28.6 N 53.4 W 25016.48110101 25017.33556354 2.50 km M2 17-01-2025 12:52 19.4 N 46.3 W 25017.33556354 25018.12465384 2.51 km M3 18-01-2025 12:33 20.5 N 41.7 W 25018.45348685 25019.24307946 2.52 km
However, there is a clear difference: all nominal positions do not plot near North Korea this time, but over the mid-Atlantic. Nominal manoeuver times were about half an hour before passing over/near North Korea.
The map below plots the nominal manoeuver positions I reconstruct, as well as a part of the ground trajectory from 10 minutes before to ten minutes after the nominal reconstructed manoeuver time.
click map to enlarge
The red circle in the map is the area where the satellite would be above
the horizon as seen from Pyongyang. Clearly - and unlike previous
occasions - the manoeuver points do not coincide with this area,
although the satellite would pass through the area about 30 minutes
after the reconstructed manoeuver moments (for one of the manoeuver
moments, I depicted a longer part of the ground trajectory as well with
markers each 5 minutes of flight time).
The manoeuvers not conciding with the satellite being over the horizon as seen from the Pyongyang General Satellite Control Center (PGSC), is something new and intriguing. The nominal manoeuver points being over the mid-Atlantic is interesting.
So how where these manoeuvers initiated? Assuming my reconstruction of the manoeuver points is correct, here are three options, all having their own implications:
(1) use of a pre-programmed, automated orbit raising burn;
(2) an orbit raising burn command sent through a (Russian? Chinese?) relay satellite in GEO;
(3) an orbit raising burn command sent from a groundstation or ship near/around the mid-Atlantic.
The white area depicted in the map is from where a command from a ground station or ship should have been sent in the case of option (3), possibly a location in Brazil or the mid-Atlantic.
It might be interesting if someone better versed in that than me, could check the presence of North Korean vessels (and Russian and Chinese space tracking vessels) in the mid-Atlantic between January 16 and 18, 2025.
All three nominal positions correspond to a manoeuver just after passing
through the Ascending Node, which is often a standard practise with
orbit raising manoeuvers when smaller or larger alterations to the
orbital inclination are required. However, no such alterations to the
orbital inclination are apparent:
click diagram to enlarge
This was the fourth orbit rasing manoeuver episode since Malligyong-1 was launched on 21 November 2023. Here are they all in a table:
# period incr raise before afterinterval 1 2024 19-23 Feb 5x 4.0 km 498 km 502 km 90 days 2 2024 03-07 Jun 5x 5.7 km 497 km 503 km 105 days 3 2024 06-10 Sep 5x 5.9 km 498 km 504 km 95 days 4 2025 16-18 Jan 3x 7.5 km 496 km 504 km 132 days
A next raise might occur in the period April to June 2025. It will be interesting to see where those manoeuver locations will end up geographically and whether at some point the orbital inclination is adjusted or not.
the metal ring found near Mukuku in Kenia. Image: Kenia Space Agency
the metal ring found near Mukuku in Kenia. Image: Kenia Space Agency
the metal ring found near Mukuku in Kenia. Image: Kenia Space Agency
On 30 December 2024, reportedly near 12:00 UTC, an odd object is believed to have fallen from the sky near the village of Mukuku in Kenia (approximately 1.58 S, 37.61 E, some 100 km from the Kenian capital Nairobi). It is metal ring of about 2.5 meter in size and reportedly 500 kg mass, although that mass could be an estimate only.
The Kenya Space Agency is investigating, believing it to be Space Debris. Apart from the metal ring in the pictures, other fragments looking consistent with space debris, for example what looks like carbon wrap and isolation foil, were found several kilometers away from it (see video below):
It is still not entirely clear if the object is space debris (although it looks likely), and if so, which object from what launch. There are two reentry candidates for this date, only one of which looks viable as a candidate (see also Jonathan McDowell's summary here).
That viable candidate is object 33155 (2008-034C), an Ariane SYLDA adapter from flight V184, the launch of ProtoStar 1 and BADR 6 to geosynchonous orbit on 7 July 2008. This SYLDA adapter was left in a 1.6 degree inclined GTO following the launch and had its reentry on or near December 30.
As I will investigate below, using a reentry simulation, both the location where the ring was found and the reported fall time are realistic for it to be this object.
CSpOC, the US military tracking network, last recorded 2008-034C in a
1923 x 146 km orbit on December 23, i.e. a week before the Kenia impact.
As this is a very low inclination orbit (1.56 degrees), it belongs to a
class of objects that is ill-tracked due to a lack of tracking stations
close to the equator. This explains the 1-week gap between the last
available orbit and the reentry.
As a note: what is a SYLDA? A SYLDA ("SYstème de Lancement Double Ariane") is a kind of hollow shell put over the first payload, in order that a second payload can be mounted above it.The conical upper part of the SYLDA has a smallest diameter near 2.6
meter, i.e. similar to the size of the ring found in Kenia, which then
could be an upper Payload Adapter Fitting (PAF).
A SYLDA (black) as part of a stage, satellite and fairing stack (image: ESA)
An Ariane SYLDA (image: ESA)
CSpOC issued a reentry TIP for this SYLDA for 30 December 2024, 21:38 UTC +- 59 minutes. That is the date of the Kenia event, but not the correct time, as the Kenia event reportedly happened near 12:00 UTC, nine hours earlier. However, the quoted uncertainty of 59 minutes from this TIP is not realistic, if based on the last available orbit (a week old at the time!). A more realistic uncertainty estimate would be +- 1.5 days.
Ignoring the CSpOC TIP time, I did an independent impact prediction, using the development version of the open source TU Delft Astrodynamics Toolbox (TUDAT).
I used the last available orbit (epoch 24358.42010446) and the nrlmsise00 model atmosphere to run a reentry prediction, using a trial-and-error approach to see whether I could tinker with the drag area such that it would reenter near 1.58 S, 37.61 E near 12:00 UTC on December 30.
From @DutchSpace on twitter, who is very knowledgeable on Ariane hardware, I got a mass of 505 kg for the SYLDA in question. The dimensions for the SYLDA on flight V184 should have been about 4.5 x 6.4 meter (there are different versions of SYLDA with different mass and sizes).
After some trial-and-error, I can make the object reenter at 1.57 S, 37.61 E on 30 December 2024 at 11:49 UTC, close to the reported location and time, if I use an average drag area of ~18.24 m2. That is a value which is about 63% of the maximum drag area of this SYLDA (roughly 28.8 m2). This is a reasonable value: during earlier reentry analysis for elongated objects like rocket stages (or in this case, a hollow elongated adapter), I found that a drag area of about 60% - 62% of the maximum area is usually a good approximation to account for the variability in drag due to tumbling .
Below is what the approach trajectory from this simulation would be:
click map to enlarge
While my TUDAT simulation does not prove that the object is debris from 2008-034C (SYLDA), it does show that it is feasible for the reported time and location.
How about that 'other' candidate? That was an Atlas Centaur booster, 2004-034B, for which CSpOC gives a TIP of 30 December 2024 21:30 +- 1 m UTC. However, the orbital plane of this candidate did not pass over Kenia at the reported time (12:00 UTC), and moreover, this object was still detected on-orbit several hours after the reported time of the Kenia event (see also Jonathan McDowell's analysis here): the last reported orbit is for epoch 30 December 2024 15:50 UTC (but it is always possible that a part came off earlier). For these reasons, it is not that likely that the Kenia event was due to a part of this object.
@DutchSpace on twitter, who as mentioned is very knowledgeable on Ariane hardware, so far has trouble positively identifying the ring as a SYLDA part (and that worries me). If the reported mass of 500 kg is correct, that is too heavy for it to be part of this SYLDA too. I have some suspicion however that the reported mass is an overestimate.
For now the verdict is: possibly the reentry of parts of the Ariane SYLDA 2008-034C, but not proven beyond doubt yet.
Here is the final output [revised after running both a TUDAT and TUDAT script update] of my TUDAT reentry model (I had it stop at 50 km altitude, as at that altitude the object should have completely fragmented and decelerated, with fragments falling down basically vertically):
mass: 505 kg drag area: 18.236375 m^2 altitude limit: 50000.0 meter
propagation start: 2024-12-23 10:04:57.030000 UTC propagation end: 2024-12-30 11:49:25.029545 UTC final altitude: 49.879
reentry after 7.072 days
REENTRY AT: 2024-12-30 11:49:25.029545 UTC lat: -1.57 lon: 37.61
Values in the last three lines are nominal only, the error margins over a 7-day integration period are large. Also ignore the superfluous digits. As a reminder: I tinkered with the drag area untill I got a value that made it reenter as close to 1.58 S, 37.61 E and 12:00 UTC as possible, and the above output gives the relevent drag area and the resulting modelled reentry time and location.
The TUDAT script used can be downloaded here (note: you have to use this script with the 'development version' of TUDAT, as the current non-development release of TUDAT has a bug where the epoch of a TLE is incorrectly read). The development version of TUDAT and installation instructions can be found here.
UPDATE 9 Jan 2025:
In a statement to Gaël Lombart of Le Parisien, Arianespace engineers have cast doubt on the identification of the crashed object as a SYLDA part, indicating that the size of the ring does not fit and stating that "this part does not belong to an element of a European launcher operated by
Arianespace". So the mystery remains as to what this object's origin is.
X-37B OTV 7 near apogee imaged by the author on 3 October 2024. Click to enlarge
It looks like the time on orbit is about to end for mission OTV 7 of the enigmatic US Space Force X-37B spaceplane (2023-210A). Launched on 29 December 2023, it went into an unusual Highly Elliptical Orbit with apogee near 38 600 km and perigee near 300 km and an orbital inclination of 59 degrees (see various earlier blogposts).
On October 10, the US Space Force announced that OTV 7 "will begin executing a series of novel maneuvers, called aerobraking, to
change its orbit around Earth and safely dispose of its service module
components in accordance with recognized standards for space debris
mitigation"
I already wrote earlier, e.g. in this blogpost from February, that the mission likely would end by using aerobraking in perigee to lower apogee, circularize in a Low Earth Orbit, and then land.
Aerobraking is a technique where, by a manoeuver in apogee, the perigee altitude of the orbit is lowered such that it is in the top of the atmosphere: not low enough to make it reenter, but enough to significantly slow it down. When the spacecraft goes through perigee in that situation, it experiences enhanced drag, that will result in drastically lowering the apogee of the orbit, certainly after a few of such perigee passages.
This will bring the orbit down and eventually change the Highly Elliptical Orbit character into a Low Earth Orbit. Orbital velocity near perigee (over 10 km/s while in a Highly Elliptical orbit with apogee near 39 000 km) will be drastically reduced (to 6.8 km/s) by this, allowing the vehicle to reenter the atmosphere and land without experiencing too excessive forces during reentry.
It looks like the process of lowering perigee might already have started around October 4, when for the first time perigee (while earlier just above 300 km) seems to drop below 300 km:
OTV 7 apogee and perigee altitudes over time. NOTE: logarithmic Y-axis! Click to enlarge
This is difficult to say for certain, as frequent larger and smaller manoeuvers by OTV 7 (it seems to have manoeuvered daily, as it never was on the ephemerids during a next observation) combined with a sketchy observational coverage (most of the observations from the last two months have been done by me, with some by Tomi Simola), means that orbit determinations are not always that easy and it is not clear how real the minor variations in perigee altitude from orbit determination to orbit determination are.
The wording of the US Space Force news item is such, that it seems to suggest that after apogee lowering and orbit circularization through aerobraking, OTV 7 might for a while continue its mission in a lower (Low Earth) orbit, as they write:
"Once the aerobrake maneuver is complete, the X-37B will resume its test
and experimentation objectives until they are accomplished, at which
time the vehicle will de-orbit and execute a safe return as it has
during its six previous missions".
So rather than land directly after the aerobraking sequence is finished, it might stay on orbit for days, weeks or months, in an orbit that is more like those of previous X-37B missions.
Over the past two months, perigee has been kept on the equator (argument of perigee kept near 180 degrees). That is a situation where during a perigee pass, there is the possibility to change the orbital inclination. So it is possible that near the end of the aerobraking sequence, the orbital inclination (currently 59 degrees) will be changed to a lower value, e.g. around 40 degrees as with previous X-37B missions in LEO.
As an interesting aside, the US Space Force bulltein also mentions that mission OTV 7 in its unusual HEO orbit "has conducted radiation effect experiments and has been testing Space Domain Awareness technologies in a Highly Elliptical Orbit".
In a surprise move, China conducted a (for them) unusual ICBM test launch on September 25, 2024. They launched an ICBM at full range, targetting an RV splash-down area in the central Pacific.
According to China the missile, with a "dummy warhead", was launched at 00:44 UTC (Sept 25) "to the high seas in the Pacific Ocean". Several countries (including the US, Japan) were reportedly informed before the test, and Navigational Warnings were issued for the RV impact area and missile stage splashdown areas two days before the test.
The Navigational Warnings and NOTAM's indicate that the missile was launched from the northern part of Hainan Island, with RV splashdown near 10.4 S, 146.5 W near French Polynesia, some 700 km west of Nuku Hiva and 875 km northeast of Bora Bora. The indicated rangeflown by the ICBM was about 11 700 km.
I have plotted the relevant hazard areas from the Navigational Warnings (HYDROPAC 3118/24, HYDROPAC 3121/24) and a NOTAM (A3054/24) with a matching reconstructed ballistic flight path on the map below, while the two illustrations in top of this post show the approximate trajectory in 3D (assuming apogee at 1200 km and launch on the Hainan coast).
click map to enlarge
Given the launch from Hainan, it was likely a road-mobile ICBM (perhaps a DF-31 or DF-41 [UPDATE: it was a road-mobile DF-31A or AG, see update at bottom of post]) launched from a TEL. The exact launch location is still unclear at the moment. The location of the hazard areas (especially that from NOTAM A3054/24) seem to rule out launch from the Wenchang Space Launch Complex, as they rather point to slightly more north on Hainan.
China usually test launches its ICBM's over land, on lofted trajectories (e.g. see this 2019 blog post). The last time they launched one at full range on a non-lofted trajectory into the Pacific was 44 years ago, in 1980. So this launch is far from a standard test.
In addition to being an ICBM non-lofted test, the test could perhaps also have been for the purpose of testing China's TJS/Huoyan early warning satellites in geosynchronous orbit.
231141Z SEP 24 HYDROPAC 3118/24(91,93). PHILIPPINE SEA. PHILIPPINES. DNC 23. 1. HAZARDOUS OPERATIONS, SPACE DEBRIS 242200Z TO 250400Z SEP IN AREAS BOUND BY: A. 19-46.00N 118-15.00E, 19-36.00N 119-48.00E, 18-33.00N 119-41.00E, 18-44.00N 118-08.00E. B. 19-06.00N 124-41.00E, 18-57.00N 125-42.00E, 18-11.00N 125-36.00E, 18-21.00N 124-34.00E. 2. CANCEL THIS MSG 250500Z SEP 24.
231521Z SEP 24 HYDROPAC 3121/24(83). SOUTH PACIFIC. DNC 06. 1. HAZARDOUS OPERATIONS, SPACE DEBRIS 242200Z TO 250400Z SEP IN AREA BOUND BY 09-37.00S 147-14.00W, 10-18.00S 145-31.00W, 11-08.00S 145-44.00W, 10-27.00S 147-29.00W. 2. CANCEL THIS MSG 250500Z SEP 24.
A3054/24 NOTAMN Q) ZGZU/QRDCA/IV/BO/W/000/999/1952N11145E018 A) ZGZU B) 2409250020 C) 2409250230 E) A TEMPORARY DANGER AREA ESTABLISHED BOUNDED BY: N200247E1113156-N200222E1120118-N194011E1120042-N194142E1113027,BAC K TO START. VERTICAL LIMITS:SFC-UNL. F) SFC G) UNL
There are three more NOTAM's that correspond to the areas from the Navigational Warnings. (HT to Cosmic Penguin on Twitter for the NOTAM).
LUCH (OLYMP) 2 position change. Click image to enlarge
Luch (OLYMP) 2 (2023-031A), the second Russian OLYMP-K/LUCH 5X SIGINT satellite in geosynchronous orbit, has changed position several times since its launch in 2023. Each time, it was placed near a commercial communications satellite. I have written about it before on this blog (e.g. here), and two (here and here) 2023 Space Review article by Bart Hendrickx provides more background on the OLYMP-K program.
And now LUCH (OLYMP) 2 has moved again. From its previous position stalking THOR 7 (2015-022A) at longitude 0.54 W, where it arrived on July 1 2024, it has now made a small hop to the other side of the THOR + INTELSAT grouplet, to 0.92 W, taking a position inbetween THOR 6 (2009-058B) and INTELSAT 1002 (2004-022A).
The move started on 16 September 2024 near 22 UTC and was completed on September 18.
LUCH 2 positions over time. Click diagram to enlarge
detail of the lastest move (top). Click diagram to enlarge
The photographs in top of this blogpost shows the change in position by
LUCH (OLYMP) 2 over the past week: basically moving from one end of the grouplet to the
other.
(as soon as the moon is less of a nuisance, I will attempt to get a
better picture of LUCH (OLYMP) 2 at its new position, with a larger
phase angle).
This is not the first "small hop" of LUCH (OLYMP) 2 to the other side of a visited satellite grouplet. In December 2023, it also made a small hop, from 3.2 E to 2.6 E, moving from EUTELSAT 3B to EUTELSAT KONNECT VHTS.
Meanwhile, it is not the first time either that a LUCH (OLYMP-K) satellite is checking out INTELSAT 1002. The latter has been visited by an earlier LUCH (OLYMP) satellite, LUCH (OLYMP) 1 (2014-058A) twice before.
The relocations of LUCH (OLYMP) 2 so far come at intervals of roughly 3 months.
It is still a bit mysterious why exactly these LUCH (OLYMP) satellites are stalking commercial satellites. The roles of their victims are somewhat diverse, although most of the stalking targets in one way or another have to do with data transmissions and TV broadcasts (but there appear to be no relations to recent Russian satellite TV hacks). They could perhaps be mapping contact networks, tapping data streams, analysing frequency hopping patterns, or even analyse weak energy field transmissions within their target satellites. Or they are just there to feed paranoia and provoke counterspace methods.
At the end of the first week of September, the North Korean military reconnaissance satellite Malligyong-1 (2023-179A) made an orbit raising manoeuver again, as I predicted in June.
The raise can be seen as sudden altitude jumps at the righthand side in the diagrams below, that plot the change in the satelllite's apogee and perigee altitude over time. The first diagram shows Malligyong's orbital evolution since launch in 2023, with three orbit raising events, one in February 2024, one in June 2024 and now one in September 2024 (the gradual sinusoid trends are due to natural orbit decay and
periodical evolution of the orbital eccentricity: the sudden stepped
"jumps" are manoeuvers). The second diagram is a detail and shows the current September orbit raise, in five distinch daily steps:
click diagram to enlarge
click diagram to enlarge
As was the case for the orbit raising manoeuvers in February and June 2024, the raise was performed in five incremental steps, one per day. The first manoever was on September 6, the last on September 10. It raised the average orbit by 5.9 km, similar to the altitude raise in June, to 504 km (see diagram below), slightly above the initial orbit insertion altitude from the launch in November 2023. While the June manoeuvers raised both perigee (slightly) and apogee, this time only the apogee was raised, from 499 km to 511 km, a raise of 12 km (see diagrams above).
click diagram to enlarge
As a result of the manoeuver, the value for the daily RAAN precession is now slightly under, but still very close to, the ideal sun-synchronous value, with the match improving over time (se diagram above)
From the pre- and post-manoeuver orbital data, I reconstruct these five sequential nominal manoeuver times (for an explanation of how these times were determined, see my earlier analysis of the February manoevres in The Space Review of 8 April 2024):
# DATE UTC LAT LON ORBITS USED FOR ANALYSIS RAISE M1 06 Sep 2024 13:12 36.9 N 137.4 E 24250.58978379 24251.57588172 1.17 km M2 07 Sep 2024 14:24 48.6 N 122.4 E 24251.57588172 24252.49645891 1.18 km M3 08 Sep 2024 14:01 60.6 N 133.0 E 24252.49645891 24253.21993764 1.19 km M4 09 Sep 2024 13:43 55.0 N 134.8 E 24253.54879868 24254.20668986 1.20 km M5 10 sep 2024 13:32 26.0 N 130.6 E 24254.20668986 24254.93046907 1.19 km
These nominal positions correspond to the crosses in the map below, with the lines showing the trajectory from 10 minutes before to 10 minutes after the nominal manoeuver time:
click map to enlarge
As can be seen, and as was the case in February and June, all manoeuver times correspond to passes within direct line-of-sight range of the Pyongyang General Satellite Control Center (PGSC) in North Korea (the red oval in the map is the geographical area where the satellite will be above the horizon as seen from Pyongyang). And as was the case in February and June, all manoevers were done on late evening passes, between roughly 13 -14 UTC (10-11 pm local time in Pyongyang).
The manoeuvers started three days earlier than I had predicted. In June, I had predicted the next manoeuver to start either Sept 9, Sept 16 or Sept 23. That was based on the time between previous manoeuvers, and the fact that these were initiated on Mondays. This time however, the series of manoeuvers started on a Friday.
We can expect the next orbit raising manoeuver to happen mid-December 2024, most likely somewhere around December 13-16, on local late evening passes (13-14 UTC) within direct range of Pyongyang.
