Jump to content
  • 55

Accelerate under Timewarp still not working in some cases (Often in Large SOI's)


jclovis3

Bug Report

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 10 Home 22H2 | CPU: AMD Ryzen 7 7800X3D | GPU: MSI Gaming GeForce RTX 4080 16GB GDDR6X PCI Express 4.0 | RAM128 GB DDR5-6000

What Happened:
While taking off from Minmus in a Northly direction, I used the time warp and noticed my Ap stopped changing. Fuel was still being consumed. The trajectory arc was also not changing.

Evidence:
I tried loading from a quick save and even restarting and this bug persists, so I'm attaching the quick save.

Video: 

 

 

Included Attachments:

BuggedSave.zip

Edited by The Space Peacock
Added video and reformatted text.
Link to comment
Share on other sites

58 answers to this question

Recommended Posts

  • 3

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 10 | CPU: 12th gen Intel Core i7-12700KF | GPU: msi RTX 3070 Ventus 2x | RAM32

when i engage timewarp with the dawn engine active, it stops accelerating my vessel

Included Attachments:

dumbsat.json

Link to comment
Share on other sites

  • 2

Just suffered this bug using Dawn ion engine. It was ok while in orbit of Kerbol, which is good since it required 2 game days of burn to intercept Jool (I think carrying enough fuel for 45k of delta V might be excessive but hey). Once I encountered Jool it's now broken. It consumes fuel and electric charge, but orbit speed and AP/PE are unchanged. I need to do a 7.5 hour burn to stay in Orbit. I'm not gonna do that !

Link to comment
Share on other sites

  • 2

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Linux fedora 6.6.8-100.fc38.x86_64 | CPU: 13th Gen Intel(R) Core(TM) i5-13500HX | GPU: NVIDIA GeForce RTX 4050 Laptop GPU | RAM16GB

I expected ion engines to produce thrust under timewarp. They do not. Electric charge and xenon are consumed, but the engine produces no thrust under timewarap. When not in timewarp, the ion engine produces thrust. Other engines(chemical, nuclear) are not affected. I am playing on Linux through proton, but the bug occurs on Windows too.

Included Attachments:

ksp_bug.mp4

Link to comment
Share on other sites

  • 2

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 10 | CPU: Ryzen 5800x | GPU: Rx6700xt | RAM32GB

Using Ion engine ("Dawn").

Attempting an ejection burn out of Kerbin's SOI:

  1. Time Warp to before periapsis
  2. Stop time warp
  3. Throttle up
  4. Engage time warp
  5. Disengage time warp
  6. Set throttle to zero.

Repeat steps 1-6 enough times and after several (5-6) orbits, thrust will no longer be produced at >1 time warp.  Curiously, throttling down allows time warp to be engaged again. Each subsequent orbit requires a lower and lower thrust in order to engage time warp. Saving & reloading does not rectify the issue. 

No mods. 

The craft is named Duna Probe I-14.

 

Included Attachments:

quicksave_1.json

Save2023-12-30.json

Link to comment
Share on other sites

  • 2

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Wine 9.0-rc3 on Linux | CPU: AMD Ryzen 7 | GPU: NVIDIA 3060 Ti | RAM32GB

(*) Note: this is a duplicate of the bug below - I'm filing a separate report to provide a reduced test case and hopefully useful observations, since I don't see a way to attach a save file directly to a post.

 

Problem: Under specific circumstances, engines under thrust produce no acceleration during on-rails time warp. This appears to be possibly due to a loss of floating-point precision, causing extremely small acceleration values to be rounded to zero.

In the attached test case, containing a simple craft with a xenon engine in a 2500km circular Kerbin orbit, acceleration on rails works only when the xenon tank has 0.219876t or less fuel. I can't find a way to display the current vehicle mass, but from calculations based on part mass shown in the VAB, this would be a total vehicle mass of 1695.876 kg. (In the file, the fuel level is set to 0.2199t, so it will cross the threshold after about 5 seconds of full throttle acceleration.)

Severity: Medium

Steps to reproduce:

  1. Load the attached save file.
  2. Press Z to set the engine throttle to 100%, and observe (in the orbital info window, or by switching to map view and pinning the Kerbin PE node) that the periapsis altitude begins to decrease.
  3. Within 3 seconds, set time warp to 2x.
  4. For 3 seconds (6 simulated seconds), observe the behavior of the periapsis altitude.
  5. Set time warp to 1x.
  6. Observe the behavior of the periapsis altitude.
  7. Set time warp to 2x again.
  8. Observe the behavior of the periapsis altitude.

Expected behavior: The periapsis altitude continues to decrease throughout all steps after throttling up.

Observed behavior: The periapsis remains constant during step 4, but decreases as expected during steps 6 and 8.

Included Attachments:

rails-accel-bug.json

Link to comment
Share on other sites

  • 1

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 11 Pro 64-bit (10.0, Build 22621) (22621.ni_release.220506-1250) | CPU: Intel(R) Core(TM) i3-9100 CPU @ 3.60GHz (4 CPUs), ~3.6GHz | GPU: NVIDIA GeForce GTX 1650 | RAM16384MB RAM

Impact: Med

Frequency: happens every time with this certain craft. Haven't tested if this happens for every ion engine, or if it's just this craft in particular.

When burning with an ion engine, increasing the timewarp stops the effect of the ion engine, even though the engine is still burning. It seems that the game just doesn't register any thrust at all.

Bug should be reproducible with the top stage of the attached craft file after releasing the fairings and firing the top decoupler.

Included Attachments:

Desktop2023_12.24-16_00_37_01.mp4

LongrangeCommsMK1.json

Link to comment
Share on other sites

  • 1

I've encountered same bug with Dawn ion engine. It affects the vessel around Kerbin, but not other with methanox one around Duna.

Edit: I found out that this behaviour happens only when pointed prograde. Once I turn about 45 degrees away from prograde, burn under time warp works fine.

Edited by T-BENZIN
New info
Link to comment
Share on other sites

  • 1

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 11 Home Edition | CPU: 12th Gen Intel Core it-12400 | GPU: RTX 3060 | RAM48 GB

After a while during a ongoing mission, engines of the vessel stop to work producing Thrust in Timewarp. This particular bug happend to me only so far, when Ion Drives have been involved. But the bug itself will kill all thrust in timewarp, even with a Nuclear Backupengine, thrust won´t be generated in timewarp (But fuel will be used up, as will electrical charge).

It has happen in differnt orbits, inclung Kerbin orbits, Duna Orbits, Jool Orbits and in this case, a Eve Orbit.

Technicly the craft is composed out of 2 seperate launches/ships, but I include the Interplanetary stage here with the Iondrive and Nuclear Backup Drive.

I included the savegame (thrust issues) not sure if you can even work with that or use anything meaningfull with that, its the save from the moment the thrust issues appeared (the start from the video file)

Thinks I tryed to workaround:

- Shutting Down Engines and Active again - Don´t work

- Quicksave/load - Don´t work

- Restart Game and Load - Don´t work

- Teleport Ship into another Orbit and try it there - Don´t work

- Undocking the second ship and try it again - Don´t work

 

Funfact: The Lander Engines still work when I undock them from the Interplanetary Stage and activate them (in regular and timewarp mode), so the bug itself seems only be situation and vessel specific.

Included Attachments:

KerbalSpaceProgram22023-12-2719-45-16.mp4

StarwayI.json

ThrustIssues.json

Link to comment
Share on other sites

  • 1

Yeah the one time you really need timewarp during a burn is while using dawn engines for an interplanetary ship.

Shame this is exactly where it breaks..

Just had to do a 3.5 hour burn in real time to enter orbit around Duna. Luckily the game is now stable enough that you don't need to babysit and you can just do something else during the burn

Link to comment
Share on other sites

  • 1

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Arch Linux (via Steam Proton Experimental) | CPU: AMD Ryzen 9 7950X | GPU: AMD Radeon RX 7900 XTX | RAM2x16 GB DDR5 @ 6000 MT/s

I launched a satellite with a single Dawn engine into Kerbin orbit and used it to slowly raise the satellite's altitude over many orbits, thrusting during timewarp throughout. Eventually, the engine stopped providing thrust during timewarp, but still worked normally outside timewarp. The engine had access to plenty of xenon and electric charge, and the issue persisted even when I turned on infinite propellant and electric charge.

I've attached a short, self-contained video of the bug occurring below (observe that the periapsis only changes during 1x timewarp), but was also recording the entire session. Here's a rough timeline with links to relevant timestamps:

I've attached a save from after I stopped recording with a two-hour manoeuvre I am unable to complete because I can no longer thrust under timewarp with the craft.

Included Attachments:

DawnBug.mp4

escapeburn.json

Link to comment
Share on other sites

  • 1

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Win 10 | CPU: 3700x | GPU: 3070ti | RAM32

I've encountered the exact same bug as mentioned in this reddit thread only in my case I was using a LV-N "Nerv" hydrogen engine.

Kind regards

Included Attachments:

NewIkeLander.json

Link to comment
Share on other sites

  • 1

Still encountering this Issue with 3 Ion Engines while trying to escape Kerbin orbit. It only happens on Ion engines though, not on others.
(also they are build on a multi engine mount, not sure if related though).

Reported Version: v0.2.0.0.30291 | Mods: none | Can replicate without mods? Yes  (tried with and without)
OS: Windows 11 Pro| CPU: AMD Ryzen 7 5800X 8-Core Processor 3700x | GPU: 3070ti | RAM32

 

Link to comment
Share on other sites

  • 1

Here to report I am seeing the same thing with ion engines as well. This will make my Jool ion probe far more cumbersome than I was hoping it would be...

The theory that small accelerations are getting neglected due to rounding error makes sense to me, this particular probe was only pulling a pitiful 0.002 G of acceleration, easily low enough to say it's 'close enough to 0'.

Link to comment
Share on other sites

  • 1

The technical side of the problem:

When warp is engaged with thrust ON, the game tries to calculate the trajectory ("VesselComponent.HandleOrbitalPhysicsUnderThrustStart" method)

The end of that method has a sanity check, turning off this thrust-on-rails if either the position or velocity calculated for the current point in time (T=0) on the new trajectory is >=0.01 units off the actual current possition of your vessel.

Now, the only way this is telegraphed - is the

Quote

Delta error between patched orbit and NBody orbit too large. Cancelling Thrust under TimeWarp.

error message in the Alt+C.

This makes it extra nasty, because the game falls back to the non-thrust trajectory but doesn't turn off the engines - they keep wasting fuel with no good way to notice this in time.

My solution for myself

For now, I've turned off this sanity check in my "SASed Warp" mod, because this bug was preventing that mod from being useful.

If the Vanila game were to turn off the thrust-on-rails - with my fix you will instead see a debug message in the Alt+C, with pos(-ition) and vel(-ocity) error deltas.

P.S. The same fix was also added to the "Community Fixes" mod starting with v0.13, so having either my mod or CF is enough to fix (hide) this bug.
But if you have both, make sure "SASed Warp" is at least v2.3, otherwise there will be a conflict with CF.

From my experience of flying in and then out of 800k Jool orbit, the position error is usually much larger for than velocity error.
But even with that, it's only enough to lose accuracy over time. There is no noticeable teleportation when warp starts.

Possible vanila solutions

I imagine a very small fraction of players will use mods. And my fix isn't really a fix - it just hides the problem to make it less annoying.
So it's still a good idea to do something official but temporary as soon as possible.

You could at least reduce thrust down to 0 and display a warning.
And maybe a settings toggle to turn off this sanity check?

As for long term - I think using the "double" type for such calculations was easy but a very bad idea:

Floating point may look like the right thing to use, given how the scale tends to change rapidly.
But when it really matters - is when you make a maneuver that goes through several very different scales.
Like LKO Jool transfer with 2 pre-planned gravity assists: off of Tylo and then Laythe.

Given that, I think a custom fixed-point numbers would be far more appropriate:

    unsafe struct Fixed256
    {
        private fixed uint data[8];

        //TODO

    }

If you properly use 64-bit carries with 32-bit data words - many things fall into place like puzzle pieces, so it's not as scarry as it might look at first.

And I don't think the speed will be an issue, since custom fixed numbers are generally faster (at least when they aren't 4 times bigger) and the "HandleOrbitalPhysicsUnderThrustStart" is already optimized to only be called once in a while.

Edited by Sun_Serega
fix the font color, it was black even for dark theme users
Link to comment
Share on other sites

  • 0

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 11 | CPU: 12700K | GPU: 3090FE | RAM64GB DDR5

 

  1. When entering Eeloo SOI ion engine was throttled up.
  2. It was observed that the orbit periapsis changed slightly whilst thrusting under 1x time, ie thrust was being produced.
  3. The periapsis does not change when time is accelerated, ie it stops producing thrust. Resources are still used as though the engine is thrusting.
  4. As soon as time warp was reset to 1x the thrust returned and the periapsis began to change again.

Loading old saves sometimes temporarily resolved the issue, but it always returned. Only when within the planets SOI.

 

 

Link to comment
Share on other sites

  • 0

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows | CPU: AMD Ryzen 9 5900X 12-Core Processor 3.70 GHz | GPU: NVIDIA GeForce RTX 3080 Ti | RAM48.0 GB

 

Severity: med, no workaround 

Frequency: Replicated 100% on save. Unable to use low thrust engine around Jool on this save. 

Description:  

When entering Jool's orbit, the craft cannot accelerate when using the xenon engine if time is being warped faster.

When activating the higher thrust engine and entering warp, the thrust is calculated correctly and as expected.

When activating the low thrust engines outside of Jool (before and after entering orbit), the thrust is calculated as expected both with and without time warp.

The issue is strictly low thrust and in Jool's orbit. 

 

Note the seemingly unrelated and separate issue regarding solar panels not generating EC, which is captured elsewhere by other threads. 

 

Edited by The Space Peacock
Link to comment
Share on other sites

  • 0
On 12/30/2023 at 5:56 PM, Kielm said:

Reported Version: v0.2.0 (latest) | Mods: none | Can replicate without mods? Yes 
OS: Windows 10 | CPU: Ryzen 5800x | GPU: Rx6700xt | RAM32GB

 

Using Ion engine ("Dawn").

Attempting an ejection burn out of Kerbin's SOI:

  1. Time Warp to before periapsis
  2. Stop time warp
  3. Throttle up
  4. Engage time warp
  5. Disengage time warp
  6. Set throttle to zero.

Repeat steps 1-6 enough times and after several (5-6) orbits, thrust will no longer be produced at >1 time warp.  Curiously, throttling down allows time warp to be engaged again. Each subsequent orbit requires a lower and lower thrust in order to engage time warp. Saving & reloading does not rectify the issue. 

No mods. 

The craft is named Duna Probe I-14.

 

 

Included Attachments:

quicksave_1.json

Save2023-12-30.json

 

 

 

Same here, at first my ion engine worked under warp at low Kerbin Ap. But after multiple passes to raise my Ap, when it reached about 2500km, the dawn engine stopped working under warp.

 

I don’t want to do a 3hour burn at x1 speed so Eve will have to wait..

Edited by Aodhan
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...