Orbit changes during Time Warp in Vacuum - can some PC players share some insight?

[Arch3rAc3]

Hello all, recently I created this thread in the Console Technical Support subforum, but not only Devs appear to pay zero attention to it, it also doesn't draw any noticeable attention from other KSP players

By reading some old threads I noticed players on PC also used to have the same problem; although I never saw a solution, I didn't see any newer threads about this so I assume it has been somehow fixed?

Now, this issue is very present in the Xbox version still, so I'd like to share this thread with you guys here (as I assume pretty much all of you reading are going to be PC players), and would love if you guys had any info on what might be causing the issue.

Thanks!

[Superfluous J]

I do not recal there ever being a problem with orbits while in time warp. The problems I experienced and to this day experience very rarely (though I suspect mods now) were always when the ship was not in time warp.

Note that the PC KSP developers do not work on the Xbox. They help but as far as I know fixing this kind of bug - if it is only on Xbox and not on PC - is up to the Xbox KSP developers. Who have no forum presence to my knowledge.

 

[noname_hero]

At least some of it is caused by stock game on PC, as I've seen this too yesterday in a fully stock game. Apoapsis of my aerobraking Eve ship went *up* a bit a short while after leaving Eve's atmosphere. I waited till I cleared the atmo, quicksaved, switched to map view, and watched my Apo go up

However, I wasn't time-warping during this.

This is not the first time I've encountered this problem. But the change this time was IIRC somethig like 150m on a 3Mm Apoapsis, and never anything significant, so I always thought this is probably a known small glitch I don't have to report.

 

[fulgur]

I frequently get this where, on Minmus trajectories I have carefully corrected to give me a periapsis of 30km, I time warp and end up missing it entirely...

[Fraktal]

Same here. More often than not, timewarping after doing the intercept burn towards Minmus takes off a couple hundred kilometers from my apoapse and screws up the intercept.

It started happening after updating to 1.8.1.

[luizsilveira]

I had this just yesterday. Was going to keep the quicksave but ended up overwriting it.

I had burned to intercept an Asteroid and ended up with <1km intercept. I hit quicksave and warped ahead only to find I missed the asteroid entirely. I thought I had done something wrong so I loaded the quicksave and there it was, an intercept path. Ok, warp again (10kx I think) and, again, miss it entirely. 

Ok, at this point I thought the mun could be getting on the way and pulling my craft somehow so I quick loaded again and this time I warped slower so I could keep an eye on where exactly it was deviating. I think I must have warped to 100x or so? Lo and behold, perfect <1km intercept just like it was showing from the beginning. 

As said I was going to keep the quicksave and post about it but my reflexes to hit F5 after it finally went right got in the way. 

It was not the first time I experienced this, but it was the first time I noticed - because I had the perfect quicksave - it had nothing to do with my flying and it only seems to happen in higher (10kx+) warps.

[linuxgurugamer]

Most likely floating point error.

This is why I never timewarp for an entire trip, I always stop several time to check and correct the orbit

[Arch3rAc3]

4 hours ago, linuxgurugamer said:

Most likely floating point error.

But would it cause such disparity though?

I understand it's more noticeable when the eccentricity is close to zero, but at times it completely messes an intercept orbit.

[linuxgurugamer]

4 minutes ago, Arch3rAc3 said:

But would it cause such disparity though?

I understand it's more noticeable when the eccentricity is close to zero, but at times it completely messes an intercept orbit.

Errors get multiplied over time.  For a really long time warp, it can become very significant

[noname_hero]

I've encountered this again today, at it does look like some kind of a floating point error, or something else of that kind.

Whatever it is, it sometimes makes my Apoapsis change as I'm drawing close to Jool's atmosphere for a bit of aerobraking.

[Jognt]

I had this too on pc. Sending a new creation to the mun and when I arrived, the mun wasn’t there? Wait what? What happened to my orbit?

I tried to figure it out, figuring maybe a mod caused it. Had it again and eventually just threw my hands up and quit playing KSP. 

I take it it still happens then

[Arch3rAc3]

On 1/13/2020 at 1:55 AM, Jognt said:

figuring maybe a mod caused it

Can confirm a mod is not necessary. Some may increase the intensity of it, but it's 100% possible to have this bug on a stock install.

[KerballingSmasher]

Yes, can confirm.

 

Appeared to have a Duna encounter upon departure from Kerbin only to have it shift in timewarp and need a 150 m/s burn to fix it.

[Starman4308]

Are you warping through any SOI transitions? That could cause issues if you are at high timewarp while passing into a new SOI; I'm not sure the patched conics solver handles that properly. I advise using KAC to set SOI crossing alarms and keep timewarp at a crawl for the transition.

[luizsilveira]

10 hours ago, Starman4308 said:

Are you warping through any SOI transitions? That could cause issues if you are at high timewarp while passing into a new SOI; I'm not sure the patched conics solver handles that properly. I advise using KAC to set SOI crossing alarms and keep timewarp at a crawl for the transition.

In my case, no. In fact the game slows down by itself to a 50x warp when transitioning between SOI, doesn't it?

All my cases were within Kerbin's SOI. And it seems like pressing the ">>" button to warp to next maneuver node when said maneuver is a bit far away is very risky to say the least. It immediately warps hard and one's orbit changes quite significantly. 

[Fierce Wolf]

I can confirm the bug too. Also, how to avoid it: Time warp slowly from x10, x100 then maximum. Somehow doing this preserves original orbit.

[Superfluous J]

6 minutes ago, Fierce Wolf said:

I can confirm the bug too. Also, how to avoid it: Time warp slowly from x10, x100 then maximum. Somehow doing this preserves original orbit.

That explains why I never see it. I never insta-warp anywhere and use the . key to speed up and let KAC slow me down, which is also gradual.

[MechBFP]

Another thing that helps is to warp to SOI exits/entrances and cross the threshold in normal time, before engaging warp again. 
That will also prevent your orbit from getting moved around. 

[SpacedInvader]

On 1/22/2020 at 7:17 AM, Fierce Wolf said:

I can confirm the bug too. Also, how to avoid it: Time warp slowly from x10, x100 then maximum. Somehow doing this preserves original orbit. 

 

On 1/11/2020 at 12:04 PM, linuxgurugamer said:

Most likely floating point error.

This is why I never timewarp for an entire trip, I always stop several time to check and correct the orbit

This specific problem has stopped me from playing for years. I don't have the issue while playing within the Kerbin SOI, but 100% of my attempts over the last few years to launch interplanetary missions have failed because the mid-course corrections have needed hundreds or even thousands of m/s of dV that I could never seem to get ahead of. I even tried building an extra 1.5km/s of dV into one design only to have it need nearly 2km/s at a midway course correction. If this gradual timewarp change actually does fix the issue, I might just come back to play again. 

[GluttonyReaper]

On 1/24/2020 at 12:13 AM, SpacedInvader said:

This specific problem has stopped me from playing for years. I don't have the issue while playing within the Kerbin SOI, but 100% of my attempts over the last few years to launch interplanetary missions have failed because the mid-course corrections have needed hundreds or even thousands of m/s of dV that I could never seem to get ahead of. I even tried building an extra 1.5km/s of dV into one design only to have it need nearly 2km/s at a midway course correction. If this gradual timewarp change actually does fix the issue, I might just come back to play again. 

I can't *quite* remember how long ago it was (definitely post-1.0), but they did some pretty serious work a while back changing up the orbit code to try and reduce those kind of problems somewhat. That combined with some other tweaks they've made (e.g. automatically lowering timewarp across SOI changes) has made it all pretty solid in my experience, although I admit I had no idea you could instantly jump to high warp levels without going through the lower stages first.

[SpacedInvader]

2 hours ago, GluttonyReaper said:

I can't *quite* remember how long ago it was (definitely post-1.0), but they did some pretty serious work a while back changing up the orbit code to try and reduce those kind of problems somewhat. That combined with some other tweaks they've made (e.g. automatically lowering timewarp across SOI changes) has made it all pretty solid in my experience, although I admit I had no idea you could instantly jump to high warp levels without going through the lower stages first.

The instant jump to high timewarp is the result of "warp to" commands like what you get from "warp to next morning", but also mods like KAC and KCT will instantly set max timewarp when you use certain functions. That all said, I first noticed the issue around the 1.0 release and persisted through the last time I tried playing around a year ago. I can't rule out mods causing it to be reintroduced, or at least exacerbating it, but if they did work to fix the issue, I'd definitely not seen it as of a year ago.