Jump to content

thyriel

Members
  • Posts

    158
  • Joined

  • Last visited

Posts posted by thyriel

  1. I've been trying to design a Mars lander using RO, however the terminal velocity appears completely out of whack. Because I must bring massive amounts of rocket propellants. Things don't match up at all with what I calculate that they should be. I managed to get a rather silly terminal velocity of 8000 m/s at sea level with a giantic 50 meter pancake tank.

    This actually causes problems just for landing capsules on Earth now that parachutes can't be deployed at super sonic speeds. The mk1-2 capsule got a 200 m/s a terminal velocity at sea level, which is at least twice as fast as it should be. As a result I have on a couple of occations not been able to pop the parachute, because I was still super sonic when I smashed into the ground :/

    I'm not sure what you exactly mean with all this. Is the 8000 m/s at sea level on mars or on earth ? Is it really terminal velocity (if yes displayed by which mod ?) or actual velocity ?

    But generally: Lower your speed before you enter atmosphere (=lower orbit). With RO you can't enter atmospheres like in stock game, you need always to enter a low orbit before starting to land or your speed will be too high and atmosphere won't brake you enough.

    Well I'm using the 8192 DDS but sure I will go with 4096 and see how it will go.

    Since RSS uses almost double the RAM then it did in 0.25 now you almost can't run 8k textures with 32 bit client now. Even 4k i didn't managed yet to run stable with a lot other mods.

  2. I think Ralathon is correct with his measurements of RAM usage. I repeated them with my actual texture folder (had to reduce it to 160MB meanwhile so KSP doesn't crash every 30 mins). The interesting thing is, with my 160MB texture folder RSS uses only ~150MB more in 0.90 then 0.25, if i replace that with 8k textures its about the same as Ralathons RAM usage. There just isn't any other explanation that RSS does something strange it doesn't do in 0.25 when 0.90 KSP+RSS with 2k textures uses only 1,3 GB (slightly more then 0.25) but with 8k textures it needs 3 GB while in 0.25 it was only 2/3rd of that. - The bigger the textures the more RAM it uses compared to 0.25, and it's always about double the size of the textures currently used.

    On the other side, there seems to be some kind of memory leak, but i can't say yet it if it's in stock game or any of the mods, but even if i start with low res textures and an inital RAM usage of 2 GB (with all other mods), it gets more and more over time and somewhen crashes. I could therefor imagine that a bug in stock game prevents RSS textures to be unloaded when they should.

  3. 8

    Hey. I am loving this pack. However, the size of kerbin makes it really hard for me to play as I haven't mastered the normal solar system yet. Is there a way to maybe tone down the massive solar system to kerbal size? I tried removing the Solar system mod, but then the game wouldn't start.

    Oh also, The game seem to run out of memory really quickly unless I add the -force-opengl parameter to steam. However, this makes the edges ragged. Is there a better way to reduce the amount of RAM the MODs use? (I tried using 64 bit KSP, but that would not even start ;( )

    Thanks,

    Hoaz.

    http://forum.kerbalspaceprogram.com/threads/55145-0-90-Real-Solar-System-v8-4-Dec-21-14?p=1599467&viewfull=1#post1599467

    That's a kerbin sized config for real solar system, but you will need to install other mods you like with that yourself. It isn't compatible with realism overhaul (your rockets would be too big and engines too powerfull) - but it removes the "need" for other mods and is playable with only stock parts

  4. That was meant for orbital inclination changes. Sure a launch into the correct inclination is the cheapest with max 176 dV. (if correctly done) But already in orbit it's a huge difference. For example a 90° change at LKO costs 1300 dV more then the same change when Ap is 2880km high

    http://forum.kerbalspaceprogram.com/threads/103279-Trajectories-right-after-launch-why-is-that/page3

  5. Not always. 90 degrees is certainly vertical, but it could be north or south depending on where you are and how the orbit is oriented. The target orbit's inclination doesn't change as you rotate with the planet underneath the orbit, but what direction the orbit's going in relative to you does.

    Forgot to mention that ;) Ascending node launch it's north, descending node launch it's south.

    Another important thing i forgot: Switch to orbital navball to see where orbital prograde is currently and make small corrections with that. In the end you want the orbital prograde to be at the desired degrees not the surface prograde

  6. Not hard to calculate yourself. Open the +note in contract window, launch when KSC is in map over ascending or descending node. Take the inclination value from the contract window. 0° inc is 90° on the navball, 90° inc is north on the navball. So if you want to start into a 20° inclination that would be 70° right of north.

  7. The fact that the world rotates beneath the satellite doesn't seem to be of any concern to kerbal scientists

    With it's zero inclination, zero axial tilt and equatorial KSC there is just no need to bother much about that ;)
    If you are short on deltaV go with Laie's approach.

    However, if you have a surplus of deltaV and the target orbit is not heavily inclined, simply launch into an equatorial orbit (west or east), create a maneuver node right below the AN or DN of the target orbit and adjust the inclination to match the target. It's easy to eyeball it. After burning this node you will have a nearly identical LON of AN and Inclination. This allows to launch any time you want and in the "easiest" direction (unless it's a retrograde orbit) but costs more fuel due to plane change burn.

    To save deltaV don't correct inclination as soon as possible. Make it right after the hohmann transfer to target orbits apoapsis (so Apo is still wrong inclined). The lowest dV for inclination changes is at high apoapsis and low periapsis.

  8. Ralathon: working as designed. The original textures are loaded unreadable to save RAM, then when I need a texture for a vertex color map I have to reload it as readable, create the map, then unload the new copy of the texture.

    As i somewhat too have the feeling that it uses a lot more RAM now (couldn't get more then three 8k textures to run stable now and the rest is only 2k textures which was 4k before mostly). Is it normal that those use more RAM then harddrive diskspace ? Looks like it at least. My KSP uses 1.8GB when the first loading is finished (about 200 more then with 0.25), and ends with 2,9GB usage after RSS is done, but the RSS+Textures folder only has 450MB. (Folder 150MB smaller then 0.25, end RAM usage about 800 more MB)

  9. I look at the original post at the Active Texture Management thread and read that it is actually expected to cut from a few hundreds of megabytes to a couple of gigabytes of memory usage and read from various sources that using a different graphic mode usually cuts it pretty much in half again. But when I try this with my game installation the effect is no bigger than the natural oscillations. The same thing happens with both the stock and the modded game just the textures are really compressed and blurry but the memory usage does not want to decrease. The only difference with mods is that it hits the limit quickly and issues start to arise.

    I think active texture management not doing much on top is because ATM compresses to DXT5. DXT5 is a DirectX thing and supported by openGL only via extension. Plus, if you convert all to DDS and use DDSloader there shouldn't be much textures left that can be converted.

  10. After i worked the last days on the "kerbin sized rss config" (real solar system) i thought gameplay wise it would be fun making a config that basically has the same planets / moons like stock game, but moved to a lot harder to reach positions (aka fantasy solar system).

    Things like planets orbiting in the opposite direction, extreme eccentricity, probably could even work raising mountains and adding an extreme thick atmosphere so you can only land on mountain tops... maybe a different starting condition on Kerbin then 1G and normal atmosphere...

    So what do you think ? Would that interest someone beside me playing a crazy solar system for fun ? What are the kind of planets, moons, specific orbits, or even sun parameters you always thought that would be cool to try out ?

  11. bei den Schwierigkeitsgradoptionen gibts auch noch den Punkt 'hire crew members automatically' - ohne dem flag wird immer nur die minimale crew genommen (also ein kerbal oder wenn probe vorhanden dann niemand). Machts leichter bei solchen Sachen nicht darauf zu vergessen, aber wenn man mal mehrere losschicken will sollte man halt nicht vergessen die auch einzuladen ;)

  12. I agree about the outsourced R&D problem, but I don't understand why people are saying contracts give too much science. For me, the only ones that give very much are the ones that are making me collect science anyway, like getting a bunch of seismic readings from other planets/moons. Repeating experiments in multiple biomes is still the easiest and second-most-broken feeling (after outsourced R&D) way of collecting tons of science.

    Some contracts do give a very high amount of science. Especially the bad paying "test x landed at kerbin" or "in flight over kerbin" give an incredible amount of science sometimes. Easy to farm a couple hundred science with only equipment testing on launchpad in some minutes.

    The same with a couple other contracts. My first mun visit in 0.90 (orbit only) brought 1.5 Million cash, some hundred reputation and 1500 science back to kerbin. Only with the basic mun contracts and i think 7 or 8 "perform visual survey" contracts at once around mun. (Nothing better then upgrading contract building to maximum before anything else)

  13. Regarding Mars and Venus: Duna and Eve already have PQSMod_VertexColorMapBlend so there's no reason to add it, it should be changed. Besides, RSSTextures.cfg is an MM patch, it modifies RealSolarSystem.cfg's REALSOLARSYSTEM node. I will verify in MM cache it's working properly. I guess it's possible they changed the stock PQSMods in .90...

    The workaround with adding that "@Add" in RSSTextures.cfg worked for me too. Only thing i noticed now, reduced body size for 1/10th size stretches the ground textures. Mars Northpole is like a quarter of the whole planet. (only near ground, texture from space is ok - seems like it's for all planets and moons like that i just didn't noticed before. But that's not happening with normal RSS size, only with kerbal sized). Not sure if that's another new thing with 0.90 or if it was like this before.

    edit: I will play around a bit with my deformity parameters, i think the 'polar-stretching' could have something to do with wrong scaling / deformity in my cfg.

    edit2: Didn't work, regardless if i set terrain scaling 1:1, 1:4 or 1:10 always the same polar-stretching.

  14. Checked Venus now on my 1/10th too and it's the same. Entering Atmosphere looks like i would crash into some violet water (could be intended it's a soupy atmosphere) and when i'm through that it's mostly white texture fading to some red on the edge:

    http://steamcommunity.com/sharedfiles/filedetails/?id=360484063

    I will check now every planet / moon (wanted to do that anyway to ensure terrain is at it should be with my config file) and edit results in here.

    edited results:

    Perfectly working:

    Mercury, Moon, Ganymede, Callisto, Titan, Pluto

    (sidenote: Pluto's biomes don't work. CB shows the correct map and correct list + colors, plus biome debug (alt+f12) shows the CB map, but everything is considered "poles" biome somehow)

    Visual Problems only:

    Venus: Surface Texture missing - landing works

    Mars: Texture fading out at around 110km until its completely red, landing works,Atmosphere from ground during day is black (no stars visible like during night)

    Uranus: Atmosphere not visible during day (black sky like on Mars)

    Landing problems:

    Io: No visual problems but exploding - terrain attitude jumps from +260 to -440 the moment it explodes (crashed into pol), speed was at 3m/s only

    Europa: Falling through ground, landed at -100m but can't move camera there, no visual problems.

    Phobos & Deimos: exploding on touchdown

    Would be great if someone could confirm me if Io and Europa are working in RSS or having the same issue. Not sure if that's a general problem or my config file.

    No problems but inaccurate behavior i think:

    Titan: Atmosphere seems to big with 600km titan vs 130km earth - wikipedia saying it should be 1.19 times as massive as earths. (would be rather strange if a moon thats only 3 times bigger then earths moon had an atmosphere almost as big as saturn or jupiter)

    Jupiter / Saturn: flying inside the atmosphere (descending) on Saturn works, but getting under ~700km on Jupiter just makes *poof*. I guess they should behave more or less the same.

    (all done now..Phobos & Deimos i tested now with them being normal RSS size (except semiMajorAxis reduced, both exploding on touchdown with surface)

    edit 10 or so:

    Engineer Redux showing Infinity Mm instead suicide burn was my own fault :/ Can't work when i use fuel debugger to check the planets ^^

×
×
  • Create New...