Jump to content

allista

Members
 View Profile  See their activity

  • Posts

    2,208

  • Joined

  • Last visited

 Content Type 

Everything posted by allista

  1. allista
    Oh, I would like it so much myself! Unfortunately, it's so much more challenging than to simply control overall thrust/torque direction. It's actually analogous to a problem of autonomous multipede robot; it would need complex ML/AI algorithms to coordinate its legs, not the "simple" optimization+control-theory problems I solve in TCA. Also, there's the issue of interacting with IR (I suppose rotatrons belong there?); cross-mod interactions are generally difficult to establish and even more so to maintain.
  2. allista
    Found it. When you're using navigation (go to, follow route), the module that handles this calculates ETA and braking time constantly to know when to start to decelerate. But braking time calculation depends on an estimation of the available braking thrust, which is the force that the ship/plane can muster in the direction opposite to the horizontal velocity. So far so good. But currently it assumes (of course) there's only one thrust provider -- engines. And VTOL engines on a plane are slow and weak; which, taken into account, amounts to very little braking thrust "available". So what happens is this: a plane using its aft engine accelerates rapidly to a considerable speed, which lowers the vertical thrust needed to counter gravity due to lift. The estimated braking thrust (small as it is) lowers further; and at some point the estimated braking time becomes greater than ETA. Autopilot decides that it needs to decelerate already, or it will skip the destination. Working on a solution to include lift+drag into these calculations...
  3. allista
    To be honest, I've also encountered this strange behaviour (with spontaneous pitching to slow down) several times; but for now I was unable to find a pattern and haven't debug it thoroughly. So thanks for that observation, I'll look into it in detail now. As for the aerodynamics: TCA really doesn't consider aerodynamic forces much, relying on its PIDs to compensate for them. Currently, the only place where lift and drag are included into calculations is the TWR (and many its functions I use internally); so, for example, when the lift is big, current TWR will be higher and engines would lower their thrust. But to use it fully TCA would have to consider things like orientation of lifting surfaces, AoA and so on; aerodynamics is huge and far from my area of expertize. So I second @Bit Fiddler: it's best to use dedicated plane autopilots for planes.
  4. allista
    @slackitude, Ok, tested it in 1.3 and found no abnormalities; so this is probably some incompatibility with the saved .craft file. Could you share the .craft that displays this behaviour?
  5. allista
    Thanks for the report. I'll check it out. And you've probably mixed up the links. No, the link is alright and I see the pictures. This seems like the adapter just shrank completely... O_O
  6. allista
    Yep. I'll set it up tonight.
  7. allista
    So, what I see here is: you have "assembly 'ThrottleControlledAvionics, Version=3.4.1.1", and "assembly '000_AT_Utils, Version=1.4.3.0", both of which are compiled against KSP-1.3 and cannot work with KSP-1.2. So it seems you've updated TCA as well as AT_Utils. You need to downgrade from SpaceDock: https://spacedock.info/mod/198/Throttle Controlled Avionics/download/3.4.1 Also be aware that AT_Utils is a library common to all my mods, so you cannot upgrade them (TCA, GC, CC, Hangar) independently, as any of them ships required version AT_Utils.
  8. allista
    @Vive_moi, @DrScarlett, here you go!
  9. allista
    Unfortunately, TCA is not suitable for plane-like horizontal flight - all its functionality is intended for VTOLs, rockets, landers and such. So what you should do is to disable TCA once the transition to horizontal flight is over. To facilitate this, TCA may be toggled along with the vertical engines using an action group. And transition itself may be handled with appropriate engines profile setup. I can make a video example when I'm home.
  10. allista
    Or you can use Configurable Containers to store material kits in any tank, stock included
  11. allista
    What are HL parts? Anyway, Cost/MassPatch options work only once: when a part is created in editor from prefab. Their only propose is to leave parts patched by MM with their original mass and cost. Any further changes made by various actions do not affect the way CC calculate calculate its share of mass/cost modifiers. Perfect! Do you want me to include it in CC distro? If so, send it)
  12. allista
    1. I'm almost sure MM is not that powerful. Parsing an arbitrary string into a set of meaningful numbers is not a trivial matter even for a full-fledged programming language, let alone a home-brewed patching framework like MM (which is still super-cool-and-absolutely-required for us modders! ). 2. True. And it may also be impossible for MM in general case. 3. Many mods, and many fuel/texture/mesh switchers in particular, run in FINAL. Beware the unpredictable ordering. 4.1 A set of MM patches like that could be easily made into a separate package and distributed as an option. There are many pure-MM-patch mods that do something like that. 4.2 In the next release (that's coming, fingers crossed, today) I also convert everything (except gas tanks) to use TankManager; and only wings have received restrictions on tank types they can use.
  13. allista
    1. The main window only displays ships that are landed. Ships in flight/orbit are omitted from the list. 2. You can always summon the workshop control UI via its part menu (right click on the workshop part that you want to use, push Construction Window button). 3. If you have a landed ship with a workshop part that you're sure should work, but it still doesn't show in main window and there's not Construction Window button, than it probably is a bug. Then, please, share your KSP_Data/output_log.txt file via Dropbox/GoogleDrive/Pastebin or whatever you use; just post the link to the file.
  14. allista
    It should translate exactly: the way you see it in VAB or SPH is the way it is spawned (using the exact same code that KSP uses to spawn on launchpad/runaway). If it doesn't, please, give some screenshots and share .craft file + list of required mods for me to test it.
  15. allista
    Cool! But I can't include a patch that affects/disables other mods into official distribution. Only as an option to download from the OP.
  16. allista
    @ShadyAct, I'm going to make a release tonight. Should I wait for your patch, or will it be going to some later release?
  17. allista
    Try to lower the Steering Gain value first. This will make the craft more sluggish, but may eliminate oscillation. Also try to change ship's design so that slow engines produce less torque (use smaller levers and more squarish engine placement for torque uniformity): less torque => less response to slight thrust oscillations. Oh! And I can provide you with the beta version of the TCA-3.5 that features new attitude control system based on a double PID cascade that linearizes much better than the single PID in use now. I need someone to test it!
  18. allista
    The first problem is currently a topic we discuss with @RoverDude. We realize the need for orbital construction. But ground construction is called and made so for a reason. As for the second, I'm currently implementing de novo kit production, which is the exact thing you want. Also, you don't ship material kits (these you're supposed to produce on site), only the equivalent of specialized parts. Kits are much lighter than the resulting vessels. In any case, you're always free to use EL alongside with GC. There's no conflict between them, only synergy.
  19. allista
    You're welcome) Be sure to check the folder structure then. Under the ThrottleControlledAvionics should be folders like Parts, Plugins, and so on. The archive tree starts from GameData.
  20. allista
    Strange! Want some screenshots or, better yet, a video of the whole thing. But if I were to guess, the first thought would be: consuming fuel moves CoM which causes imbalance that TCA compensates.
  21. allista
    So, PR to B9-PW repo is ready, as well as my patch for the wings: https://github.com/Crzyrndm/B9-PWings-Fork/pull/24 BTW, @ShadyAct, you do realize that my patches for various mods are generated automatically by a python script (which you can obtain in CC repo)? This saves me a LOT of trouble with part volume calculations and such. Feel free to use it as, at least, a starting point.
  22. allista
    @scottadges, @Fr8monkey, I would strongly recommend this tool for viewing/searching logs: http://glogg.bonnefon.org/download.html
  23. allista
    The problem with TCA is that its folder in your GameData is called "Throttle_Controlled_Avionics-3.4.1", while it should be called ThrottleControlledAvionics. Many things depend on hard-coded paths, including paths to models, textures etc. There are other problems showing in the log, not related to TCA: MechJeb and KerbetrotterTools, throws some exceptions during part loading. Probably harmless. JanitorsCloset throws a lot during its FixedUpdate (i.e. 3-4 times per frame). So it's better to remove this mod.
  24. allista
    I say, if it has resources, it can have CC tank instead, whatever it is. And yes, we can certainly do it, using ExcludeTankTypes/IncludeTankTypes options. And I do like the "kerbal way" argument!
  25. allista
    Oh my! I've never realized you do the patch for the WHOLE B9 My own patch policy is simple: only add tanks where resources are defined by original mod. Also, in CC gas containers are of type Gases, while MP is LiquidChemical (which may be wrong, but that's the way it already is). And I also do not touch stock cockpits in my patches: they're left with the monopropellent as is. Meanwhile, I'm currently testing the changes I made to B9-PW with a simple patch I added for the wingA (which is the only wing capable of storing resources). So far looks good! A question, though: for stock wings I use simple tanks (no tank type change, only liquid chemicals are allowed); so, naturally, with B9 I want to use the same. The logic is simple: wing should be light and should only carry fuel for the plane. But maybe it's too restrictive? What d'you think? And @Bit Fiddler? For example, we can limit the tank types a wing can use, instead of forbidding type change completely. *Note: the simple tank configuration actually reduces functionality compared to the B9-PW builtin fuel configurator, which allows for LFO tank inside a wing. Honestly, I dare not imagine a wing with liquid oxygen and hydrazine divided by a thin wall...
×
×
  • Create New...