Earth surface stations drift while in low lunar orbit

[periodically-makes-puns]

[KSC in the wrong place, rest of the ground stations have snapped back to starting position, but they drift as well.]

When in a flight scene in low lunar orbit, ground stations (and the KSC) will begin to change surface position when entering timewarp. RealAntennas will very occasionally correct the ground stations back to their proper location, but antenna pointing (lat/long) and the KSC location remain incorrect until scene switching back to the Tracking Station (or possibly some other conditions).

I'm not sure whether to file this as a Principia issue or a RealAntennas issue. It does seem to affect the apparent KSC location, so I'm going to file it as a Principia issue for now. If it's a RealAntennas issue (improperly setting positions) I'll move it over there.

As far as I can tell the issue starts occurring somewhere in the lunar SoI; I originally reproduced this bug while doing a crewed lunar orbit with Skopos2 active. SLA broke for half my connections and I assumed I needed a stationkeeping pass, I did a stationkeeping pass, and it immediately broke again when I switched back to the crewed craft in low lunar orbit. I then noticed that my C-band spot beams weren't routing any connections, so I tried to tweak their lat/long targeting (while still in the flight scene of the lunar craft) and noticed that the actual spot beam targets didn't match up at all with the lat/long positions. The issue (mostly) disappears once you leave the relevant craft.

Reproduction:
Install Express RP-1 and Principia (Legendre). Make a new sandbox save, and launch two vessels. One should enter a low lunar orbit (<100km sea-level altitude in my tests), and the other should enter a geostationary orbit.

I used the following procedure for setting the necessary orbits.

1. Turn on Hack Gravity.
2. Set the orbit of the vessel. For the lunar orbit, I used a 0km SMA and let the cheat menu kick it up to an acceptable SMA for orbit. For the geostationary orbit, I used an SMA of 42164172m. These don't need to be stable or accurate.
3. Turn off Hack Gravity.
   In the video, to save a few scene switches, I put the two probes on the same craft and decouple them; it doesn't particularly matter. I was able to reproduce it earlier with two separate crafts as well.

The geostationary satellite is here to demonstrate what happens to lat/lon targeting; it doesn't particularly matter where it is. Target the antenna on the geostationary satellite to any visible ground station. I used Cape Canaveral, so 28lat/-81lon.

Switch to the lunar vessel's flight scene, and enter the mapview. The KSC should have suddenly moved to the middle of the Pacific. If you start rails-warping, the ground stations will also start migrating across the surface. RealAntennas will snap them back to the correct position occasionally, but the targeting will stay at the new location (and therefore still be broken). This bug is retained for a moment when returning to the Tracking Station, but RealAntennas will eventually snap all of the stations back into the proper position.

Video of start-to-finish of the bug: https://drive.google.com/file/d/18ZbONnTK6S-yeWZ7w7IgYrM0IU-lnOVY/view?usp=sharing
[set-up ends around 5m20s into the video]

Relevant files and logs:
surface-logs.zip

[DRVeyl]

First test is going to be: can you reproduce this without Principia?

RA doesn't do anything to move the ground stations except set their Unity transforms on creation. Unity takes care of the rest. It is possible RA is doing this incorrectly somehow, but that is has never been determined. I wrote the code long ago to periodically scan all stations and correct deviations by snapping them back to where they should be.

[periodically-makes-puns]

Failed to reproduce without Principia. https://drive.google.com/file/d/1QnZoIMh66zwi7WyzW_uePTIMnPkfJzYG/view?usp=sharing

[pleroy]

Yeah, this doesn't sound like something Principia would do.

[eggrobin]

Yeah, this doesn't sound like something Principia would do.

It’s not something we’ve had problems with in a very long time, but I dimly recall vessels dancing around the surface many years ago (but in far more egregious ways which have long been fixed).

We orient the planets, and we do whatever is needed for the stuff that sits on the planets to follow. That probably involves composing rotations without the safety nets of Principia, just like what we looked at in #4452; it is possible that something is wrong when the current main body is different from the one which has things on the ground.

[DRVeyl]

https://github.com/DRVeyl/RealAntennas/blob/master/src/RealAntennasProject/Targeting/AntennaTarget.cs#L66-L71

This is the relevant code for an antenna targetted at a body lat/lon/alt. Note it is using a separate GameObject parented to the Celestial Body. So expected behavior is the target and the ground station move together, albeit independently. Which means if something is "drifting" then they both should drift together. However they are not directly linked, and the RA snapping code only moves the ground station. So once you observe this, the lat/lon/alt targeting is disconnected from the position of the ground station until it regenerates on a scene change.

[periodically-makes-puns]

Reproduced the KSC drift portion of the bug without RealAntennas. Unfortunately I don't know of any other non-vessel surface references in-game I can test with. It appears to be RSS-specific, as far as I can tell, and it definitely triggers only when the current main body is not Earth. I attempted these same steps without RSS and it worked fine.

Mods: RealSolarSystem + required dependencies only, Principia

Same reproduction steps, put a probe in orbit around the Moon. The KSC will appear to be in the Pacific, and drift around at high warp. Landed vessels on Earth appear to only be affected during warp (once you drop out of warp, they return to the correct location naturally). If you then perform a burn to eject the probe from lunar orbit, the KSC will instantly snap back to the correct position upon leaving the Moon's space-high boundary. Attempting to do the same with a probe in Mars orbit causes the same issue to occur.

2026-04-02.23-05-16.mp4

[Semenar]

The same happens to me with Sol instead of RSS.