kellven

I have absolutely no idea how to do anything with particles. Except shut them off. Look into a LiquidEngine .cfg file at the fx definitions, that\'s all the help I can give.

Making a completely custom engine class set derived straight from part. Can\'t seem to find a straight answer as to whether or not to factor Time.deltaTime into AddForceAtPosition force param. onActiveFixedUpdate { ... this.rigidbody.AddForceAtPosition(thrVector * thrust, this.transform.position, UnityEngine.ForceMode.Force); } I already have the thing flyable, and a simple test showed identical results with equivalent liquid/custom engines, but it\'s difficult to produce lag-on-demand(Only shows up when you don\'t want it). Anyone know definitely if Unity physics handles framerate internally, or if I need to compensate explicitly? Much appreciated if anyone knows.

Work on an actual prop engine class seems to be going well so far, though I haven\'t gotten to to meaty parts or testing yet, but I\'ve yet to find a roadblock. I also found Gimballed engines gimballing the wrong way in the archives. I noticed the helo units have a negative gimbal set, I don\'t know if the base class interprets that correctly, or even if the bug still exists, but it would explain why I couldn\'t get a puller unit to fly quite right, and also why I\'ve had problems with forward rockets in the past. I will try to make a really, really long rocket with puller props on the rear, and check where the gimbals send me when I get to testing my propengine class. ED: In regards to your swingwings public float AoA; public float deflectionLiftCoeff; public float dragCoeff; ControlSurface also has these, if your part bases from one of these, you should be able to update DLift and dragC through code.

Now that I figured out how to interface with MuMech code, I also believe I just figured out how to get rid of the engine death ray completely. I\'d better test it out first to make sure it really works this time, but I think it will, and you can thank NS if it does. Actual prop thrust? Possibly.

I finally figured out why I couldn\'t read tank class template out of MuMechLib. Could you please add public before class MuMechVariableTank : MuMechPart { And also possibly VariableEngine.cs at least, when you release next time? It will also require at least Part.cs, as subclasses can\'t upscale access in C#. I\'ve been trying to code a general purpose reactor module, but to make it able to read VariableTanks, it needs to reference MuMechLib for the templates. Without the public access modifier before the class, C# defaults to internal, and the class isn\'t usable outside it\'s own assembly. Without being able to tell my code what a VariableTank is, I can\'t even count them, much less find out what type of fuel they hold. I don\'t have any need right now to reference the engines, but that would seem to be the other main thing someone might find they need to read. Best wishes, Kellven.

Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. You forgot to put public in front of class BaceControlUnit : SASModule in BaceControlUnit.cs, and it\'s the best thing to happen to me all day! I spent much of the rest of the day trying to figure out why r4m0n\'s MuMechLib.dll absolutely refused to expose any of it\'s classes, and it wasn\'t until I recompiled your code that I realized he forgot to put public before any of his classes. It also explains why Tosh\'s cart has the same problem, but C7\'s gear doesn\'t. I can now tell my code what other peoples parts are without the compiler throwing a tantrum spiral. Did I say thank you already? Just for that, I\'m launching a BACE mission to the Mun in your honor. PS If anyone is completely confused or worried, the code is just fine as far as I can tell, it\'s purely a programmer thing. PPS Kellvonic Imperial Robotics and WarPeacemachines is still eagerly awaiting your hingy doodad thingies.

If you want autohover, you need to use long axis engines, unfortunately. MechJeb won\'t even try to do off-axis thrust with KillVert, so helo units don\'t work right. If you set translatron killvert to 0 and check kill hs, MechJeb (assuming you have enough thrust) will try to counteract anything changing your position, EXCEPT your manual input. You end up hovering in place once it stabilizes, and you can carefully position yourself with wasdqe before you land/whatever. You can however, put a rack of rockets on the VAB roof, and use a helo craft for bowling. I hope nobody minds my answering questions here, I\'ve just been playing with the MuMechLib and multi-axis thrust craft for a while.

I think it\'s either that the thrust for the puller engines is coming from 15m in front of you(center of thrust? I\'m so confused), or that the CoM of the part is 15m in front of where it attaches, but physics is one of those things public education failed to impart to me. Either way, try adding a little more mass to the back end, and see if that balances it out, or add more claw canards forward. I didn\'t have too much issue with the puller helo unit, once it was horizontal. The quad puller prop I tried was uncontrollable once it pitched horizontal, as you were describing. I did have another idea on the deathray issue, make it an RCS 'engine', but then I remembered that only uses the RCS controls and ignores throttle. Definitely some strange changes happening inside the mechjeb code from what I\'m reading now. I\'m beginning to wonder if there isn\'t some hyperspace communication network between coders I\'m just not tuned into.

I know one should never cross the streams, but I just couldn\'t help myself, so I made the rotors run on my LH2 tanks. Then I noticed a distinct absence of 3m command pods in my folder. Now, my prisoners say you plan to take your gasoline out of the Wasteland. You sent them out this morning to find a vehicle. A rig big enough to haul that fat tank of gas. What a puny plan! Look around you. Nothing can escape! The Humungus rules the Wasteland! 8 x 2m rockets wouldn\'t lift this tub, but 14 additional propfans did the trick. For some reason, the props still spin in space, I thought there was a section in the code about air pressure and prop rpm, perhaps I misread it. Or my potatoputer may have just been bugging with the 294 kerbograms of LOX I was hauling. To Damnyoujapan: I don\'t have anything you need to worry about. Keep on trucking with what you were doing. I wasn\'t referring to r4m0n\'s code before, I checked that first, and it was fine. The case of the missing thrustAxis is in the actual base class, LiquidEngine. It\'s the base game part classes that haven\'t been documented or source released, r4m0n\'s code is all there to be parsed like it should be. I may be imagining things, but I think the shifted origin trick also throws off the net CoM of the part. Not a huge issue as you made the shifted parts ultra light. The problem is this also gives them an enormous Thrust:Mass, which I ran into trying to adjust them to SE 2 ratings so I could use them semi-legit for the first 10km of reaching orbit. That\'s why I was trying to find another solution. The work fine for cruising around in airplane mode though. Of course, I still couldn\'t come up with an actual mechanical reason to use props on a launch vehicle, but they just looked cool, so I made them high ISP high T:M. Ed: Congrats on landing a helo on the VAB, I had to use a massive RCS system to get the truckodyne to set down without bowling the other rockets over.

AngryPointer->Fix(&LiquidEngine.thrustVector, Now) was reflected by undocumented dll class Kellven cancels Fix Item: Sourcode inaccessible Kellven cancels Fix Item: Throwing tantrum Seriously, over a month now, and not a single source release. There\'s only about 15 or so part types, is Squad really worried counterfeit parts will be sold in front of Walmart? Almost every single plugin made so far violates the EULA because they never gave us the source for the parts. The only way I could see to do it is to hijack liquid engine calls, shunt them to solid rocket members, and hope for the best. Too much headache to even try now.

I had stopped modding, but apparently, I seriously need to join modders annonymous. I finally got the 'death ray'/CoM/VAB offset half fixed by bringing your model above origin, and flipping it upside down and backward. Then redoing the nodes to make it attach right, and the ray comes out the top of the rotor instead of down from above, with thrustCenter adjusted. Unfortunately, negative thrust and negative thrustAxis both just make the engine do nothing but burn fuel. Someone obviously thought it would be fun to add an extra clamp0 in the code somewhere. Probably the same someone who made the node 'vector' ignore negative alignment, making this impossible to stack attach like I had it earlier. About the exploding engine thing, the death ray gives a message that it damaged the part below it, but the only way I could get an actual explosion was bumping the heatProduction to 1000, and that only exploded the engine itself unless it was surface attached. Why your rotors are producing rocket exhaust heat to begin with is a bit of a question, but the only actual damage seemed to be just the engine overheating, unrelated to the position/orientation of the deathray. I\'m now determined to find out where the major malfunction is with the vectors and thrust clamp, and I went and registered that copy of MSVC#. I\'m all out of bubblegum now, and I\'m about to put on my angry pointers. It was a line I didn\'t want to cross, but the code just wouldn\'t listen. I just wanted to say, this is all YOUR fault, Damnyoujapan!

I may be a little late to the party, and this may not be the largest helo so far, but I\'m fairly certain nobody else has used this configuration. And I\'m positive nobody would ever build a second after flying the first. I didn\'t see this yet, and it simply had to be done. Had two bugs, or perhaps design issues. That thing would consistently fire 'something' straight up in the air shortly after launch and vertical hover slide. It looked like it was shooting part of the prop very fast Y+, it would only do it once per flight, and I was never able to figure out what exactly it was losing. Those are 0.5m puller props on it. The other thing you\'re probably already aware of, but the large origin offset seems to exacerbate CoM issues. Perhaps it\'s just my imagination, but even stuff that should be balanced just won\'t stay vertical off the pad. Jeb, Bob, and Bill, along with countless other Kerbals died to get that t-chopper off the ground without a backflip. Oddly, MechJeb KillVert works perfectly for auto-hover precise positioning on the tailsitter propcopters, but the actual helicopters can\'t hover worth a thing.

I have to say I was quite upset when I first read this today. I\'m slightly mollified now that I see exactly what you two did. Moving stack nodes is something else entirely. Anyway, I think I saw someone mention something about autogyros. While this technically doesn\'t count, I clearly over-engineered this one by a factor of 2.

The equations weren\'t mine, I just did a top forum search for 'drag', and picked out the various formulas from other posts. This is the most recent one I could find describing drag FD=.5*density*velocity2*(max_drag1*m1+max_drag2*m2...) This is the most recent one I could find describing drag FD=Force Drag density=atmo density(below) velocity2=velocity^2, I believe max_drag1=max drag of part #1 m1=mass, yes mass of part #1 Atmo Pressure, I know HarvesteR posted this way back in Aug/Sep, and I think(hope) that was where I stole the equation from, but now I can\'t re-find the post. It\'s also gone through several redesigns, so I really have no idea what\'s going on with this. Patm(alt) = (Pasl * 2.71828183)^-((alt / 1000f) / scaleHeight) Patm=atmo pressure at (altitude) Pasl=Pressure at sea level alt=altitude scaleHeight=...? You can find the MaxDrag values in the part .cfg files, but almost invariably it\'s 0.2 Mass conversion is possibly the single most subjective factor for converting Earth stuff to KSP stuff. It depends entirely on how you want everything else to balance, as a straight conversion results in completely different ratios for all parts than you find in real stuff. For instance, rocket mass fractions, engine weight, thrust:weight, even ISP and fuel cost. You\'ll just have to find your own, I use 0.0000971*5.64/10 or 0.000054708 * Earth(kg) = KSPMass However, that only works for SE v2a style engines, and it\'s only a rough approximate at that. I was hoping one of the resident engineer/physics types would pop in and give up the answers, but I think I\'m done with modding for now, so I\'ll just leave this up if anyone else is working out wings.

Note to mods: This is specificaly about an addon, however I\'m asking for general game mechanics, so I stuck this here instead. May well prove pertinent for general use, at least for those who like to calculate their astrophysics instead of just Jeb\'ing it. If you feel it\'s really in the wrong section, go ahead and move it. -What I\'m asking for: Consensus formulas for how KSP deals with atmospheric effects. Specifically; Drag, DLift, ADrag, and how they relate to Mass/Area/Vol. For wings and other parts(tanks, etc) also. -Reason I\'m asking: I tried to make an updated, normalized wing set, people could use, or base parts off of. Wanted to just get all the wing data compiled, updated, fixed, and all in one place. They may not work like real wings in real air, but at least all the parameters would be used properly. I did the best I could, but after some searching, I found stuff in the archives that seems to completely contradict what I was using. Unfortunately, I couldn\'t seem to find a consensus on how it all hung together, or even if it was still applicable. -I (think I) can do the conversion to part logic myself, but I need to know what\'s what if it\'s going to be worth anything. So far I\'ve found: drag = -IAS^2 * atmDens * partDragFactor d(h) = Fg / ( v2 * kd ) Patm(alt) = (Pasl * 2.71828183)^-((alt / 1000f) / scaleHeight) Vt=45.519/sqrt(maximum_drag) And most recent: FD=.5*density*velocity2*(max_drag1*m1+max_drag2*m2...) -Assuming the last function holds true today, am I correct in assuming DLift gets multiplied by 'FD' somehow along with AoA and applied perpendicular to part global Y(Z) axis? How is DragCoef factored in, multiplied by FD and set to -Y? -Angular drag (from Unity reference) is essentially 'roll resistance' Is this following something similar to the above equation, or is it simply a denominator to input torque? -Yes, I could try testing all this stuff myself, but I\'m quite sure there are a couple aerospace engineers only too happy to answer the questions. I\'m also modding, which means I don\'t get to actually play the game, much less experiment on it. I only have so many fingers and eyes.

0.17 stuff here As this was never completely updated for 0.15, I decided it was time to update to 0.17. New name for this project is KSylon, reciprocating LED strip not included...yet. For KSP 0.17 https://dl.dropbox.com/u/74137297/KISylonV08a.zip 9m Bonus Engine: https://dl.dropbox.com/u/74137297/KISEngKSylon9m.zip for use with GQDM "Prometheus" http://kerbalspaceprogram.com/forum/showthread.php/5316-0-17-Gaby-s-Quick-And-Dirty-Miscellania-v1-1-We-like-big-LFT-s-and-we-cannot-lie For those unfamiliar with this mess of a thread, here you will find at least a crude approximation of the hybrid engines proposed for the Skylon SSTO drone/cargo/spaceplane thingy. For those who still don't know what this is, essentially it's a Jet engine welded onto a rocket, or a rocket with an air intake. -Below 30km(Kerbin), this engine gives you an air-breathing ISp of 3500 -Above 30km it functions in pure rocket mode with an ISp of 450 -This is offset by the TwR of about 0.58(see notes 1 & 2) This means it's extremely efficient at moving stuff to/from Kerbin orbit, but it can push less rocket. Highly similar to the nuke rocket with the ISp range reversed. 4+1 sizes available: 1m: Both stack and radial versions. Radial model was from Thobewill10. 2m: Stack only. 3m: Stack only. Bonus 9m: Stack only, for use with GQDM "Prometheus" 9m tanks. To further help with lag reduction when hauling fuel dumps to orbit. 1)Note on TwR: This is actually the Thrust/Mass compared to LiquidEngine 1. Spreadsheet of KSP&RL engine specs mostly updated to 0.17: https://dl.dropbox.com/u/74137297/KSP_0.17_EngineSpecs.ods 2)Note on specs: I tried keeping the specs of this as close as possible to the proposed specs for the SABRE engine. The stumbling block is the mass, which should be around 4.56 instead of 1.8. The problem is the stock "jet" engines have a TwR of 0.58, and 0.76, which should be somewhere around 0.10, meaning the Jets in KSP are ridiculously light, and I had to compensate to make these engines competitive. 3)Note on "Air": "Airbreathing" mode (3500 ISp) should be activated whenever 2 conditions are met: 1-Atmospheric Density is >= 0.003 & 2-Engine is in an atmosphere containing Oxygen. Keep these in mind when planning trips to other planets. Author is still attempting to get a plane onto Laythe for proper testing. Atm Den switch (atmCutoff) is in the part.cfg, 26km(SABRE) is 0.006. 4)Note on "Code": This version extends the LiquidFuelEngine class new to 0.17, adding atmospheric switching, and "Disable"-ability. Known Bugs: 1) The visual "exhaust" effects can get stuck "on" if you disable the engines with throttle >0, then cut throttle. Just toggle the engines on/off with 0 throttle to get rid of it. 2) The single tiny exhaust effect looks ridiculous on the 9m engines, but they're there just for lag reduction, so not a priority. 0.17 Version has nothing to do with MechJeb, and is under MIT license. 0.15 and below below .15 stuff here, for .14, look at the bottom spoiler. ->V06 Repackaged everything properly and added Copyright(left) + GNU blablabla. Also added readme with description/install/usage/configurables. https://dl.dropbox.com/u/74137297/KISBroadswordV06.zip Fries and a drink extra. You also need the latest MuMechLib(MechJeb) by r4m0n installed in Plugins directory. If you don\'t have the camera update, get latest version from here: http://svn.mumech.com/KSP/trunk/Releases/MuMechLib.dll ^Latest Version Up Top^ Recompiled against the camera MechJeb version thingy, should be working again. https://dl.dropbox.com/u/74137297/ZKISAtmoEngine.dll Too tired to repackage it, so it\'s just the plugin this time. Can\'t remember if I released the LOX/LH2 MuMech Tank parts, but those seemed to burn through LOX way too quick, have to look at it some time later. I also got my gigantic resource monster version to burn through standard fuel tanks as 'generic resource mass' when they run out of actual resources, but having problems getting the fuel bars to show on the tanks when stock engines aren\'t also burning them. Also got gimbal mostly figured out, but the animation seems to be bent 90\', so need to fix that first. Need to recreate the fxgroup from scratch once I make sure the gimbal is working right. Updated for MechJeb 1.8.4 http://dl.dropbox.com/u/74137297/KISBroadswordV05a.zip Updated for MechJeb 1.8.3 Seems I\'m going to have to recompile everything every single time MechJeb updates now. http://dl.dropbox.com/u/74137297/KISBroadswordV05.zip If you want to try out spaceplanes, but just can\'t seem to make a plane that flies well: This is a little big, but well balanced, wingtip anhedral for auto-leveling, enough wing to use as a glider, can be easily flown without (A)SAS/MechJeb, and is perfectly capable of landing on the Mun, Minimus, and back to Kerbin. Only thing it doesn\'t have is RCS, or Jeb. http://dl.dropbox.com/u/74137297/KISSpaceplaneA.zip Requires Broadsword & MechJeb. Broadsword description, and MechJeb 1.8.2 version: Thobewill10\'s model looks very nice with C7\'s new parts. Really did a nice job with that model, and it\'s not even using the new .15 graphic features. Picture here: KIS Broadsword V04a -This is a MuMechVariableEngine that combines the functions of an air-breathing ramjet, and a traditional vacume-capable rocket. -Below 30km, it functions as an air-breathing Jet engine with high fuel efficiency, and can be used for purely atmospheric flights, though it doesn\'t track the advanced parameters C7\'s engines do. -Above 30km it automatically switches to pure rocket mode, and burns fuel at a comparable rate. This is also the mode used when near an object with no atmosphere, such as the Mun. -It also utilizes the Engine Control menu so it can be turned on/off as a group at will. Very helpfull if you\'re using the DA props. -This version uses only the normal fuel found in the default tanks. LOX/LH2 versions will be coming as they become available again. -As I\'m using the default fuel for this one, I adjusted mass to match the standard rockets. !!THIS REQUIRES MECHJEB 1.8.2!! http://kerbalspaceprogram.com/forum/index.php?topic=8021.0 If you have MechJeb, or either of the updated DA packs working, you should already have the correct library in your Plugins folder. Plugin, Parts, and source here: http://dl.dropbox.com/u/74137297/KISBroadswordV04a.zip My sourcecode references MuMechLib, which is under GPL, so mine is as well. If you\'re still running .14x for some reason, you can still use the stuff in the spoiler below: For clarification, the Broadsword download is an atmospheric and vacuum hybrid engine part, and plugin. The pods are non-plugin parts (2-3m command pods). The longship is a pre-built ship, and self contained set of custom parts; it\'s a 'dual-vector thrustship' or aircraft without wings that uses engines pointing down for lift, and rear for propulsion. Broadsword: Upd 28 Apr 12 V3a to correct a small oversight in altitude check. Now correctly checks for presence of an atmosphere also. Zip files for those Mac people http://dl.dropbox.com/u/74137297/KISBroadswordV03a.zip http://dl.dropbox.com/u/74137297/KISMMTanksV02.zip http://dl.dropbox.com/u/74137297/KISBigPodsV02.zip Screen of Thobewill10\'s new model. Many thanks, Thobewill10. (Sorry, deleted original picture, this is the best one I could find right now.) Old screen of my own ugly tanks and engines. Broadsword and Tanks -Broadsword & Air ram: A combined cycle rocket for Jet level efficiency in atmosphere, and high rocket efficiency in space. Light enough to reach orbit; but like all airbreathers, ++ heavy compared to 'pure' rockets. Required You need 2 plugins for this now, so be sure you end up with both MuMechLib and KISAtmoEngine in your plugin folder. KISAtmoEngine is in Broadsword v03 download. MuMechLib is MechJeb: MechJeb 1.8 Broadsword v03a http://dl.dropbox.com/u/74137297/KISBroadswordV03a.rar Second plugin and source is included in this download. New model by Thobewill10 makes the 1m version look like an actual engine. New code also completely automates the air ram, and eliminates many problems. Note: I added a temporary simple collider untill I can get a better fit. There are also lots of hidden toys in the code, but please don\'t play with them yet, unless you really know what you\'re doing. Excluding Broadsword specifics, they aren\'t done yet, and have never been tested. LOX/LH2 tanks v02 http://dl.dropbox.com/u/74137297/KISMMTanksV02.rar contents: LOX tank-(0.5m base) LOX tank, heavy and tiny. Lots of sizes, total tank equivalencies noted. LH2 tank-(1.5m base) LH2 tank, big and light. Lots of sizes, total tank equivalencies noted. Tanks are pre-matched. You need 1 tank equivalent of LOX AND 1 tank equivalent of LH2 when above 30km. Below 30km, you just need 1 tank of LH2. Space: 1 LOX + 1 LH2 = 1 stock fuel tank Atmosphere: 1 LH2 = 1 stock fuel tank General rule of thumb that works for me is to pack 2 LH2 for every 1 LOX required to reach orbit. This reliably reaches orbit for me, and usually gives me a bit of LH2 left over for reentry. Broadsword engine was converted from SABRE specs (Skylon) and normalized to Silisko Edition 2 Alpha. Notes/bugs: The LH2 tanks are extremely light, and double as structural members in all my ships now. The LOX tanks are massively heavy, and are best kept right around where your desired CoM, as they should rightly have their own gravity well. LH2 tanks are probably unrealistically light when empty, mass was calculated paired with a LOX tank to equal a standard 'fuel' tank, but those are highly unbalanced to begin with. BigPods: 2 Big pods, for big rockets, for big Kerbals, for big people with big tanks, and bigness issues. http://dl.dropbox.com/u/74137297/KISBigPodsV02.rar Longship-VTOL style 'thrustship' No wings, just engines. Same idea as the prop pack, only on the bottom, and highly unpretty. This is old and neglected now. mediafire Semi-functional sub parts. Causes stack damage from engine thrust, project neglected for some time. mediafire Licensed under MIT. If anyone has a licensing issue, please contact me, if contact cannot be made, consider it in principle copyfree. Under GPL, too confusing otherwise, as r4m0n\'s MuMechLib is under this.

It really depends on what kind of modding you want to do. Making the actual parts requires 'no new coding', just dig into the cfg files, and try to get comfortable doing basic stuff in blender if you want new models. If you want to use one of the plugins someone else made, take a look through here, you just replace the part type with what the threads tell you to. There is plenty you can still do without going any further than that. New plugin stuff is done in C#, which is sort of a hybrid between Visual Basic and C++. I know C++, but not the '.net' school of coding, so I can\'t really give you much advice on where to start there, other than the actual language is still 'C style'. And that apparently, you import the game code from a set of classes or dll (not sure how exactly); then extend it with your own code, and finally export your code in dll form. Basically, you\'ll need to be familiar with 'C style' syntax, classes and inheritance, and using and making .dll stuff. Can\'t tell you anything about Unity itself, as I don\'t know anything about using middleware. If you want to get into the coding stuff, and already understand the Visual Basic way of doing stuff, you might want to start with C++ reference and tutorials and try some simple stuff in either a C# IDE, or you can use something like Code:Blocks and GCC for free. Just remember that if you start modding, you stop playing, so please make sure you\'re completely done with actually playing KSP before you start.

KIS DC: Pictures D/L(parts + craft file included) --The main rar uses landing legs which are not compatible with demo version.-- For Demo users only: D/L and install main pack, then D/L DemoPatch. Install it over the main pack, overwriting the old files. Simply replaced legs with leg shaped struts. mediafire DemoPatch Wings, pylons, 'elevatrons' Tanks, structural tanks, inline RCS Wing bin: BYOBBQ Boards, Deltas(Back and Forward, forward has +lift +AngDrag), 'elevatron' all purpose control surface, and- whatever this thing is <bug alert-Apparently, there may be a stray 'm' factor in the KSP drag equation, which means mass/density/area is factored twice in these. Going to wait for better data before further 're-fixing' the wings. In the mean time, I have absolutely no idea how the drag ends up, but they still deflect air> mediafire Structural bin: Various empty tanks and pylon beams to seperate parts/stacks. These pylons work off surface attach, not the erector set I tried earlier. <bug-The .5m pylon may be oriented incorrectly> mediafire Misc bin: Only thing in here right now is the 1m inline RCS thrusters, and 0.5m radial/stack ones. mediafire These thrust in only 4 directions, but you only need 4 total on a given ship. See bottom of post for VTOL help. Use for whatever you want, don\'t care anymore.