Jump to content

nyrath

Members
  • Posts

    604
  • Joined

  • Last visited

Posts posted by nyrath

  1. From the time physics loads until I end the flight. Including on the pad.

    I tried using just one layer of 80MN magazines and it was stable enough to fly. I replaced two with two 400MN magazines (still just one layer total) and they induced a jitter that was quite evident in my craft's motion. I got to Kerbin orbit before I completely lost control and the craft began spinning faster and faster on its own. Two layers of magazines is immediately uncontrollable.

    All I did was unzip USAFOrionMod_v9.zip and put the subfolders in the KSP main file structure. The mod seems to be working great except the jumping behavior.

    A note: I'm currently locked in my cube, testing on a Core i5 Latitude laptop. And I've only got KSP 0.20.0 here. I don't suspect this is an issue that would change with updating to 0.20.2, but I'll spend the hour to download it and see if it helps.

    Not sure what is happening. Does the \plugins\USAFOrion.dll have a modified date of 6/4/2013?

    I made an Orion stack: a Mk 1-2 Command Pod, an Advanced S.A.S. Module Large, UKAF Nuclear Pulse Engine (with center column), two 400MN magazine, four 80MN magazines (all in a single layer), and 16 launch support struts attached to the engine right below the magazines. No jumping, launched fine, got it into orbit, but had trouble turning it because it was a bit massive for the poor SAS. I should have put a set of RCS on it but I wanted to simplify things. The magazines never jumped.

  2. I noticed everytime i put the launch clamps to low on the engine, the whole ship exploded right when the physics kicked in. Putting them near the top of the engine solver that issue.

    Very interesting! I did not realize that where the launch clamps were attached was important. In retrospect it makes sense. The engine is relatively light, the magazines are quite dense.

  3. The "jumping beans" issue isn't exactly resolved.

    They never actually let go of the craft, but the magazines were certainly trying to escape on their own.

    Arrgh.

    Did they do that on the launch pad, or only after lift-off?

    Usually when I see them jumping that much, the ship explodes shortly afterwards.

    They never ever totally let go, but while they flail about sooner or later they hit something, then Boom!

  4. It works perfect. No wobbling at all.

    Edit: i discovered a little bug. It seems like if you attach things too low on the engine the whole thing wobbles itself to death on the pad.

    Oops. What did you attach and where? Was the Orion sitting on the launch pad with no support struts or with? If with, how many? Engine with the support column on top or engine without?

    This thing has more problems with wobble than it should.

  5. I'm currently working on hoojiwana's restart problem.

    I have made it so activating the Orion does not automatically spit out a bomb. This should fix the bug with unexpected bomb detonations upon quicksaving and switching between vessels. The autospit was originally to allow the user to launch an Orion held with stablizer struts, before the throttle was implemented.

    Bombs can no longer destroy Roche. Unless the user sneakily decides to edit the part.cfg file. There is a maximum mass parameter, any part with more than that mass is immune to bombs.

    Bombs now do not just cause impact to other ships, it also heats them up.

    Struts are no longer required with individual bomb magazines. Before, if you failed to add struts, the magazines would dance around like Mexican jumping beans and explode, while the vessel was on the launch pad. This is because the attachments are shaky if one part has more than x10 the mass of an attached part. Since the large magazines have about x100 times the mass of the engine, this was a problem. Now, upon activation, all the magazines sort of weld themselves solid to the engine. This will probably prevent jettisoning magazines via decouplers, but I'm fixing one bug at a time.

    I'll get a new version up once I've fixed hoojiwana's restart problem.

  6. The engine also seems to fire, playing the animations, using a bomb, applying no thrust, but still destroying anything nearby. I just had it blow up Roche on an attempted docking with a quickload.

    Wait, the engine fired off a nuke which destroyed Roche??!!?

    Oopsie. I didn't know Roche was considered a "vessel" by the KSP code. I'll have to fix that.

    edit

    Roche has a mass of 597 tons, I'm going to arbitrarily set it so any part caught in a bomb blast will be spared if that particular part has a mass over 100 tons.

    This means if a second Orion ship is too close, it will all be destroyed except that the 15 kiloton magazines will come flying out of the explosion.

    (of course the 100 tons is defined in the nuke magazine part.cfg file, so that users can fiddle with it if they want)

  7. When I disable all nuke feeds to the engine, and then enabled NTRs on my rocket to do fine manuevering, and then disable the NTRs and re-enable bomb feeding, the first one must be launched manually.

    That's a new bug, and I will do my best to fix it.

    The G meter just doesn't seem to display G force from the 400MN nukes. I don't know if it's overflowing and can't show all the digits or whatever, but it displays 0G instead.

    That's a bug I'm trying to find a solution for. You'll also see this in MechJeb. You see, the G meter and MechJeb look in your vessel for parts that are ModuleEngines. They show data in the G meter and in MechJeb.

    Unfortunately the Orion is not a ModuleEngine, nor can it be one. It does not use liquid fuel, have constant thrust, have an exhaust plume, and quite a few other things. I had to program an entire new propulsion system in the plugin to handle pulsed propulsion.

    I'm looking into somehow spoofing the G meter and Mechjeb into thinking that Orion is a ModuleEngines.

    I've also just discovered another bug. If I have the nuke feed disabled, quick save, and then quickload, the nuke feed re-enables itself. The engine also seems to fire, playing the animations, using a bomb, applying no thrust, but still destroying anything nearby. I just had it blow up Roche on an attempted docking with a quickload.

    The feeds re-enabling themselves is a bug, that should be easy to fix. The feed state is not being stored and loaded.

    The engine unexpectedly firing has already been fixed, it will be in the next version (hopefully soon).

  8. I have noticed that after switching between engines the Orion doesn't launch with the throttle. I have to manually launch the first nuke before the throttling kicks in again.

    The G meter on end of mission and F3 seems to enjoy breaking as well.

    Glad you are having fun with the big block-buster nukes.

    The "switching engine" problem, do you mean when you switch from Orion to NTR then back to Orion? Or when switching between different types of Orion nukes? I just need the info so I can try to fix it.

    And can you please describe in a bit more detail how the G meter breaks?

  9. triOrion08.jpg

    Ah, I forgot to explain about throttling.

    With conventional engines, throttling alters the thrust. Thrust is defined as how many kiloNewtons in one second.

    The Orion bombs have a fixed explosive size. Each bomb gives a fixed amount of impulse.

    So to vary the thrust, I vary the delay between bomb detonations.

    The USAF 10 meter Orion that I am modeling had pulse delay times that varied from 0.8 seconds to 1.5 seconds between detonations.

    So I set it up such that throttle setting between 10% and 30% made the pulse delay 1.5 seconds, 30% to 60% made the delay 1 second, and 60% to 100% made the delay 0.8 seconds.

    This means that if you are using the 3500 kN bombs, 1.5 second delay gives an effective thrust of 2300 kN, 1 second gives 3500 kN, and 0.8 seconds gives 4400 kN.

  10. Easy!

    On the launchpad, when you do OnStart, do the following check:


    if(part.Modules.Contains("ModuleEngines"))
    part.Modules.Remove(part.Modules["ModuleEngines"]);

    Bam! ModuleEngines is gone on the launchpad, but still exists in the VAB.

    Ummmm, but during the mission will my module still show up when MechJeb goes looking for ModuleEngines?

    If so, that will be a most elegant solution.

  11. That "Rendezvous with Roche" challenge sounds an almost suspiciously good way to show off this mod. I'd try it if I could reliably get into kerbin orbit. Not that an orion loaded with nukes really cares about these "orbit" things in the first place.

    Side note: Nyrath, your website is awesome and you should feel awesome. Been using it as reference long enough that I can't remember when I first found it.

    yes, the Orion would be good for the Rendezvous with Roche scenario. I fear it would be so good that it would take all the fun out of the scenario. Especially if you used the super-deluxe sized 400,000 kN nukes.

    Thank you for your kind words. I'm glad you liked my website!

    But since I am the "atomic rocket" guy, I couldn't resist making this mod. Even if I had to teach myself C# and Unity3D in order to make it.

  12. Stupid newbie here again with another silly question.

    Plugins like MechJeb can calculate mission trajectories because it can discover the maxThrust and the specific impulse from atmosphereCurve from the vessel's parts that include ModuleEngines. The MechJeb FuelFlowSimulation.cs module does this in its FuelNode constructor.

    MechJeb and other plugins will not work with my Orion Drive mod, because it is not a ModuleEngines. The rocket fuel system used in KSP does not work very well with the nuclear bomb "fuel" used by Orion. Therefore MechJeb will examine an Orion equipped vessel and fail to find any ModuleEngines.

    But I do not want to add to the Orion's part.cfg file a fake entry like

    MODULE
    {
    name = ModuleEngines
    minThrust = 0
    maxThrust = 60
    atmosphereCurve
    {
    key = 0 800
    key = 1 220
    }

    }

    just to feed the proper values to MechJeb. As far as I know, if you include the module, all of its code comes along with it. I'm sure it will interfere with the Orion code.

    So my question is: how can I spoof MechJeb and other similar plugins?

    The only way I can see is to include the ModuleEngines module, stuff my Orion values into its variables, and somehow permanently disable the ModuleEngines code. But I have no idea how to do that.

  13. This takes too much time to stack magazines, and symmetry doesn't work, how can I make it faster?

    I know, I know. I'm not sure what is wrong with symmetry, maybe it is because it is because they are vertical stack not radial attach.

    I will say that with the larger magazines, you do not need to stack them very high, they are concentrated thrust.

    I'll look into it, but there are other more serious problems.

  14. No, tommygun, no it does not.

    Awesome! Nuclear detonation engine that mechjeb can control! There is no way this can go wrong!

    It does have a throttle for MechJeb, but I think there is still a problem with KSP not understanding how much thrust appears with each throttle setting. I'm not sure how to fix that, but it is on my bug list.

    Thanks for all the effort you put into this. This is by far one of the best mods for KSP.

    Thanks! That makes is all worth while.

    Nyrath, you have done well, can't wait to play with it

    Thank you, but be aware there probably still are bugs in it. Keep me posted if you find anything.

    There ain't no stealth in space, or is there? Just kidding of course:) But i found this graphical glitch:

    I'm not sure if its a problem of your mod or something else but i just post it here.

    Do you mean the beads of light along the edge of the ship? I'm not sure what is causing that.

  15. The fact is : you can't combine efficient ISP AND efficient thrust.

    Not with stock parts, at any rate.

    This holds true In Real Life as well, there seems to be an inverse relationship between Isp and Thrust.

    There are a couple of theoretical propulsion systems that have both high Isp and high thrust, but they are kind of extreme and none of them have even come close to being prototyped. They include the Orion project nuclear bomb drive, beam-core antimatter, and Zubrin's infamous nuclear salt water rocket.

  16. New Version Available for Testing

    see first post for download link.

    I did my best to ensure this version will not break your existing designs. My apologies if it does.

    However, with any luck, s20dan's double magazine stack will suddenly start working.

    The readme file explains how to use the multiple magazines. Be careful with the larger magazines, they are full of dynamite.

    I am still working on implementing the throttling. It is taking too long, so I wanted to get this version out in the meantime.

  17. That's beautiful, nyrath. I've been lurking this thread pretty hard, and when you get this version with individual magazines ready, I'm eager to try it out. Will the center column/feed stack be of adjustable height?

    Regarding the joints, I read somewhere that KSP mostly uses what Unity calls "RigidJoint" and that it's (obviously) not perfectly rigid. Apparently if you use some other kind of join (flexible joint, maybe it was called?), with the values set to 0, it's actually much more rigid.

    Thanks for your kind words. Alas I cannot make a variable height center column until I can make more solid connections. I will look into RigidJoints. I will also experiment with the method used by Quantum Joints.

×
×
  • Create New...