Orbital Periods and Drifting Satellites

[AlamoVampire]

Ok. I am completely stumped here. I have 4 Communication Relays <RA-100's> is what should be synchronous to each other orbits. All 4 are set to roughly the same orbital height <you will see their altitudes in the images, I cannot recall off hand, even though I just took these 4 pictures lol> but, the important part is, they are in temporal LOCK STEP at a period of precisely 1 hour 4 minutes 10 seconds EACH. They were roughly equidistant making a rough + sign in orbit where each relay was at the end of each segment of the + sign. I thought I was seeing drift in them as I was doing missions, and dismissed it as me just being jumpy. Turns out, I am seeing drift, and a whole heck of a lot of it. Is my save modded? Yes. Do I think that has something to do with this? No. Why? All parts on the 4 satellites in question are 100% stock parts. The only item in active use to have any mod parts on it is my science station Station 1 and its only got 2 parts from planetary base. Every other part that is in active use on ANYTHING aside from 2 corridor ends is 100% stock KSP <DLC included>. So, can someone who is far smarter than I am at orbital maths please tell me why 4 satellites in IDENTICAL periods at roughly the same orbital height +- 15-20km at MOST on either side what was intended are drifting THIS badly?

When I finished launching this 4 satellite array up, number 4 was across from 2 and 1 across from 3, but now? well, I think the images speak for themselves. I have no idea what happened. Not a single 1 is in orbit with any kind of engine on it. All drives were dumped once all 4 were in orbit and I was happy with where they were. Yes, they have RCS, but, despite what the pictures state, they have all been shut down at the blocks themselves. Besides, I dont think RCS is capable of triggering this kind of drift w/out seriously altering the orbits themselves. Am I just glitching or is there something else at play here? 044802102025

[AlamoVampire]

Does anyone know why 4 satellites with nearly identical orbits with identical orbital periods are drifting as shown above? The only thing I can figure is altitude is a factor but im recalling someone saying as long as periods match it shouldnt matter. Please i beg help me understand. 202302102025

[Scarecrow71]

Floating point error?

[AlamoVampire]

To help facilitate my fellow forum goers help, I am going to give you a build list for my 4th Generation Communications relay as well as a picture of it on orbit. Parts List will be listed in 2 ways, first will be from the antenna down, and then from the probe core out. I hope what follows makes sense to you guys. I think the picture should help understand the build. You can launch this to orbit using what ever launcher you wish. That is, if you deign to launch this thing or a few of them to see whats going on. I will also provide orbital parameters too, but, Im not sure how much they will help.

Ok, Antenna down:

RA-100 Relay Antenna

Z-4K Rechargeable Battery Bank <stacked 4 times>

Advanced Reaction Wheel Module, Large

RC-L01 Remote Guidance Unit

Gigantor XL Solar Array in 4x Symmetry on the RC-L01 Remote Guidance Unit

Advanced Reaction Wheel  Module, Large

Z-4K Rechargeable Batter Bank <stacked 4 times>

That is the top down build, from the antenna down. To build:

Select the RC-L01 Remote Guidance Unit, then place 1 each above and 1 below the RC-L01 Remote Guidance unit 2 Advanced Reaction Wheel Module Large. Next grab and place 1 Z-4K Rechargeable Battery Bank on either of your Advanced Reaction Wheel Module, Large, the starting battery location really makes no difference, but, you want to then copy said battery onto its self until you make a stack of 4, then do the same on the other Advanced Reaction Wheel Module, Large. Now grab a Gigantor XL Solar Array and turn on 4x symmetry. You want to place them onto the flat squares on the RC-L01 Remote Guidance Unit. Finally, place the RA-100 Relay Antenna onto the top of the stack. Your final product should look like:

This is what your final product should resemble if built correctly. This particular satellite that has given us its pretty picture is the 3rd one in the set of 4 that has the satellite that has drifted. This one in particular according to the tracking station appears to be in the same slot I put it in in the set of 4. Next image will be a closer in image of the core of the satellite showing the placement location of the Gigantor XL Solar Array in 4x symmetry.

Just for clarity sake. I am sure my description of the flat squares was sufficient, but, clarity is clarity right?

Ok, up next, its orbital parameters, taken again, from THIS satellite. All 4 share IDENTICAL periods with only minor variations of altitude across the whole of each orbit.

As you see, it is in a 500.162km by 497.345km orbit, with a period of 1h, 4m, 10s.

---------------------------------------

Satellite 2  is in a 500.201km by 497.220km orbit, with a period of 1h 4m 10s.

------------------------------------

Satellite 1 is in a 500.169km by 497.230km orbit with a period of 1h 4m 10s.

-----------------------------------

Satellite  4 is in a 500.214km by 497.381km orbit with a period of 1h 4m 10s.

As you see, they are all physically identical satellites in nearly IDENTICAL orbits with EXACTLY identical orbital periods as shown by each image. So, again, I ask, what could be causing the drifting as shown in the original images above in my OP? Someone claims floating point error, but, without an explanation of how this could affect this i am still at a loss. As I said, maths on this scale are not with in my wheelhouse. I can calculate an emergency glide slope or an equal time point for a 777-300ER easily enough, but, combining coding which is what floating point error sounds like and the maths required to explain this issue may as well be in sumerian cuneiform for all the clarity I have on combining the maths affecting these 4 satellites and what ever coding black magic floating points are. 043001112025

[Poppa Wheelie]

If one of these satellites has an actual period of 1h4m9.50000000, and another has an actual period of 1h4m10.499999999 seconds, then the game would display the same 1h4m10s period, even though the periods are almost 1 second apart.  In 60 orbits they will have drifted  almost 1 minute apart. 

How long ago (game-years) were they launched?  I don't know if this could account for the entirety of what you're seeing, but this is one version of how the floating point error might be occurring.  Actually, this is more about rounding error than floating point error.

My understanding is that it is not possible to avoid drift over time in this game, just as it's not possible in real life.  Get them the way you want them initially, declare victory, then periodically cheat them back to where you want them.

Edited Tuesday at 05:54 PM by Poppa Wheelie
floating point versus rounding error

[jimmymcgoochie]

As above, slight variations in the orbital periods that are a rounding error over a single orbit, will add up over time- especially when the orbital period is relatively short. The orbital period can also change slightly due to how a vessel's position is calculated- IIRC it bases the position on the centre of the root part but will rotate around the centre of mass, so unless they're perfectly aligned (which due to the big relay dish your satellites are not) rotating the vessel can subtly change the orbital parameters.

KER (Kerbal Engineer Redux) can display orbital periods down to milliseconds, which will keep your satellites several orders of magnitude more stable. You could also try putting your relays in a much higher orbit where the combination of longer orbital periods and greater distances will reduce the impact of orbital drift.

[AlamoVampire]

5 hours ago, Poppa Wheelie said:

How long ago (game-years) were they launched?

They have been on orbit for 3 years 240 days and change. All 4 were roughly back to back launches within a day to day and a half time delay at most (their METs are displayed in the images) between number 1&4. The drift was first spotted in an earlier generation that was placed in geostationary. 172202112025 ill have to try a higher orbit later.

[Kerbart]

3 years and 240 days is  1335 days. That's 13100 hours(assuming your clock is set to Earth time). Or arounf 3300 if using Kerbin days.

Your "1h, 4m, 10s" likely isn't exactly that but anywhere in the 9.5s and 10.49s range. That'll give you potentially a drift of up to 2s per orbit, or using the lower number above, over 6600s in those 3+ years—nearly two full orbits. The drift you're showing is perfectly normal, in that sense.

[AlamoVampire]

I think the ksp clock is kerbal time. If the drift is nominal then why do 3 out of the 4 still appear to be holding their initial location in the race track with only 1 being the rogue drifter wildly out of place? But seeing as I am not currently at my computer I couldnt begin to tell you which of the 4 has gone rogue. I do know its one of the two in the OP that are close to each other. 1 was across from 3 while 2 was across from 4 at initial set up. Which ever one between the two isnt across from its partner is the rogue. I can check later if needed. 181302112025

[DennisB]

5 hours ago, AlamoVampire said:

I think the ksp clock is kerbal time. If the drift is nominal then why do 3 out of the 4 still appear to be holding their initial location in the race track with only 1 being the rogue drifter wildly out of place? But seeing as I am not currently at my computer I couldnt begin to tell you which of the 4 has gone rogue. I do know its one of the two in the OP that are close to each other. 1 was across from 3 while 2 was across from 4 at initial set up. Which ever one between the two isnt across from its partner is the rogue. I can check later if needed. 181302112025

Looking at your orbital parameters, it should be Satellite 4, because it has the longest orbital period compared to the other ones, which are more similar to each other.

[AlamoVampire]

On 2/11/2025 at 11:39 PM, DennisB said:

Looking at your orbital parameters, it should be Satellite 4, because it has the longest orbital period compared to the other ones, which are more similar to each other.

While getting infuriated over the central launch site not unlocking, I rechecked my relay net. Its actually sat 1 thats gone rogue. 2 is across from 4, as was intended, 1 WAS across from 3, but has fallen behind 4 where its not supposed to be. This means 1 of 2 things. 1. It has gone rogue. Or 2: the other 3 are in absolute lock step down to the nanosecond with each other and that their periods are somehow shorter or quicker by less than what stock shows. Which is eerie in and of itself yet more likely, because I never altered number 1's orbit once I got it stable during circularization and dumped its engine package. It was set as the key at what ever period it wound up at, which was 1h4m10s.

051002132025