Lt_Duckweed

For fun bonus frustration, docking ports also inherited the robotic drift bug. If a craft has no docking ports or robotics, and it gets bent via some sort of stress, you can quicksave and quickload to restore it to an unbent state. With robotics parts, anything that is on the child tree of the robotics part instead become permanently bent when you quicksave and quickload. With a single docking port on the craft, EVERY SINGLE PART on the entire craft will become permanently bent if you quicksave and quickload while under stress. This nearly ruined my recent mission, which happened to span across 1.11 and 1.12. It launched in 1.11 with a docked lander and lander transfer tank, and was nice and rigid, when the whole craft (main craft, transfer tank, and lander) landed on Laythe, the now floppy connections let the lander and transfer tank severely droop, and unknowingly at the time, saving on the surface of Laythe made the droop permanent. This nearly ruined the whole mission because the nerv engine on the lander pulled double duty as the only vaccum engine for the whole mission. And for the ascent up from Laythe and the transfer home it was off center, but fortunately my reaction wheels were just barely strong enough to handle it and I had enough margins to correct all the burns that were thrown off by the off center engine. The really important thing to note about this is that any craft sitting on a surface is under stress, and will bend a tiny amount even with autostruts. What this means is that in 1.12, if you craft has a docking port, the little bends will become permanent every time you save, meaning no matter how well you design your craft, you have a limited number of loads before it slowly but surely bends to unusability.

I just finished my first ever Jool 5. With the added twist that: 1. The craft travels from Kerbin to Tylo orbit as a single stage, then to laythe and back home as a single stage. 2. All parts are recovered on the runway at the end of the mission. 3. No isru, aero exploits, or abusive part clipping are used. 4. The transfer to and from Jool is direct, no assists. 5. And finally, it only uses 2 engines, one Rapier, and one Nerv. This would qualify for Level 1, and I dont think it qualifies for any special accolades, as it is too heavy for low mass, too costly for low cost, and too partly for low part count.

Today (really the last couple of weeks off and on, you know how it goes lol) I did my first ever Jool 5. With the added twist that: 1. The craft travels from Kerbin to Tylo orbit as a single stage, then to laythe and back home as a single stage. 2. All parts are recovered on the runway at the end of the mission. 3. No isru, aero exploits, or abusive part clipping are used. 4. The transfer to and from Jool is direct, no assists. 5. And finally, it only uses 2 engines, one Rapier, and one Nerv.

Realistically, you are going to be looking at 6000-8000 if you want to have a reasonable fraction of your craft be payload, and also want a sane twr (0.1-0.2, bellow this and direct burns to Eeloo or Jool become a real pain in the ass even with periapsis kicks, you will end up spending extra dv both on the transfer burn itself, and on correcting for the inaccuracy).

1600

As long as the exterior nodes are occupied by parts that are as large, or larger, in cross sectional area as the cargo bays, there is no need, or indeed any benefit, to occupying the internal nodes. The rear node will subtract area from the rear face of the drag cube, and the front 2 nodes will both subtract area from the frontal face of the heat shield, and likewise the heat shield will subtract area from whichever face the attached parts nodes are tied to.

Exposed surface area of a part's drag cube causes drag. Node attachment has a side effect of reducing this exposed area. Thus reducing drag. See my earlier post in the thread. It's not about the node being occupied/unoccupied, it is about the relative sizes of the parts that are atached at the nodes (NOT node size, it is based on the actual visual cross section of the part)

Not quite. Ksp calculates drag based on exposed part surface area. There are a couple ways to reduce the exposed surface area. 1. Place the part in a service bay, cargo bay, or fairing. 2. Node attach parts. When you node attach 2 parts, the two node atached faces have the area of the attached face subtracted from their exposed surface area. Thus, the smaller face has its area set to 0, and the larger face has the area of the smaller face subtracted from it. There is an additional "streamlining" factor applied to each face of a part (each part has 6 faces) at play, the "pointier" and more aerodynamic a face is, the lower the factor is. (The front face of a nosecone has a low streamlining factor. The sides, being less pointy, have a high factor). Infinite streamlining has a factor of 0, a perfectly flat plate has a factor of 1. Rapiers have a very flat back end, with a factor that is nearly 1. Adding a nosecone removes some of the exposed area on the rapier and replaces it with the much more streamlined area of a nosecone.

Ksp 1 actually simulates based off of what are called "drag cubes". Each part is simulated to have 6 faces, each face has an area, and a streamlining factor. Each attachment node on a part is tied to a specific face. If you attach two parts via nodes, the area of the drag cube faces associated with those 2 nodes gets mutually subtracted from each other. After the exposed area is calculated, each of the 6 faces has 3 types of drag applied based on a variety of factors. The three types are: 1. Frontal drag (a component of form drag) 2. Side drag (analogous to real world skin drag) 3. Backface drag (the other component of form drag) These get calculated based on the angle of each of the 6 faces to the airflow. A full writeup exists here: https://github.com/Ren0k/Project-Atmospheric-Drag

Unfortunately the largest/heaviest thing kerbals can move/assemble in freefall is the aerospike.

I have found that a pure prop + Nerv design is unable to get sufficient altitude to provide reasonable performance at Nerv ignition. There has been a recent Renaissance in Eve sstos, and they tend to cluster around a family of values and ratios. 85-111 tons per vector engine Vector:Nerv ratio between 1:1 and 1:2 Engine ignition between 15-17km High twr, high nerv designs turn off vectors at about 2450 m/s Low twr, low nerv count designs turn off vectors at about 2800m/s As for drag reduction, if a fairing is the root part of the vessel it will have much lower drag than it should. So the minmax drag reduction strategy for an eve ssto would be to use a small root fairing and smash everything inside of it.

In real life most wings have built in angle of incidence, and the majority of lift is produced via the wing's angle of attack. Wing camber does produce lift, but it is much less important than the wing having an AoA to the airflow. After all, planes can fly inverted, and many acrobatic/stunt planes actually have symmetric airfoils with no camber at all, and get all their lift from AoA. That's really all there is to it. Wings in ksp have no camber, thus they must have AoA to produce lift. And with the lego like piece by piece construction that wings have in ksp, giving the model for each individual wing part camber would ruin the ability to tile them. With the procedural wings, if they give the ability to add camber, then they could implement it such that wings will be able to produce lift at Zero wing AoA. Of course then we will have to deal with "why my lift negative" questions from people who put the wings on upside down or used negative camber lol. To add to this, most airliners when fully loaded at cruising speed and altitude actually fly at a few degrees of positive fuselage AoA, as they are generally designed so that they never have to fly at negative AoA when light on fuel/cargo. The true issue with KSPs aero model is not that the wings have no camber, but that body drag is so excessively punishing, and wing lift to drag ratios so low.

Incidentally, these numbers make for a perfect transfer vehicle made from a cluster of 7 nervs, and 2 mk3 long tanks. Dry mass of the tanks and engines is 35.28 tons. If we assume a further 4.72 tons of structural, power, and control mass, this gives a nice round 140 ton wet, 40 ton dry mass. This gives us the following dv benchmarks for this very simple craft, in terms of attainable acceleration and payload at these dv numbers: 2000 m/s: 0.95 m/sec2, 304 tons 3000 m/s: 1.34 m/sec2, 174 tons 4000 m/s: 1.68 m/sec2, 110 tons 5000 m/s: 1.98 m/sec2, 72 tons 6000 m/s: 2.25 m/sec2, 47 tons 7000 m/s: 2.49 m/sec2, 29 tons 8000 m/s: 2.69 m/sec2, 16 tons 9000 m/s: 2.88 m/sec2, 6 tons

I had done more extensive math and a longer comment, but I ended up closing the browser out not realizing I hadn't posted it (oops lol). The tldr is that 5000 m/s is enough to get from low Kerbin orbit to low orbit of any other body with no gravity assists. If you assume a starting twr of .2 (with an initial kick or series of kicks worth 800m/s to an elliptical orbit this is about the lowest you can go and still get a reasonably efficient burn for a direct transfer to Eeloo), then Nervs with mk3s will get you about 80% more payload per overall starting mass than wolfhounds will. Using mk0s will extend that to about 85%, so only worth it for extreme missions, or smaller scale crafts (with just one nerv you are looking at about 58 mk0 tanks, vs using just one mk3 short + one mk1).

That is well into diminishing returns though, for a more realistic case, 100 mk0s and a single nerv gives you a starting twr of .2, and a delta v of 13,438 m/s Doing the same thing with mk3 tanks (ie max dv at 0.2 twr) gets you 12,204 m/s so there is a pretty nice upside to using the mk0s More realistically, lets ask, "What cargo fraction could you carry with a single stage craft having at least 2 m/s^2 of acceleration, and at least 5,000m/s dv (enough to, assuming plane changes are bundled into capture/ejection, capture into orbit of any body on a direct transfer from Kerbin, and vice versa). At 2 m/s^2, the ejection from an elliptical staging orbit to Moho and Eeloo would take 11min, Jool would take 9 min, Dres would take 6.5 min, and Duna and Eve would take about 2 min For Nervs with Mk3 tanks: 10% engine mass, 6.7% tankage mass, 47.3% propellant mass, 35.9% payload mass. For Nervs with Mk0 tanks: 10% engine mass, 4.7% tankage mass, 47.3% propellant mass, 37.9% payload mass. (A payload to payload improvement of 5.57%) For the Wolfhound: 1.76% engine mass, 9.23% tankage mass, 73.87% propellant mass, 15.14% payload mass. (A payload to payload decrease vs Mk3 Nerv of 42.17%)

1000 Mk0's plus a single nerv would do the trick lol.

Small nitpick here. Mk1 liquid fuel tanks have the same mass ratio as rocket fuel tanks: 9. And Mk0 tanks are actually the best in the game, with a mass ratio of 11, which gives a maximum possible delta v of 18,812 m/s

This reddit thread has a couple of techniques described in it:

And wing incidence helps there too. Both of the sstos in these videos use 3 degrees angle of incidence to greatly reduce body drag Let me emphasize again, if you build and fly a craft correctly, some incidence will always be better than none. The only question is how much.

This craft has 2.49 kn of body drag out of 25.815 kn of drag total. This is a body drag of 9.6% Wing area is 6 strakes with Angle of Incidence of 5 degrees, it has a mass at altitude of 30.7 tons and cruise speed of 1700m/s, and an estimated runtime and endurance of something like 20 circumnavigations in ~14 hours of flight.

Ah, I see where the disconnect is. When you use wing incidence you have to use a different design philosophy. You do NOT pick an altitude AND speed to fly at. You fly at 0 Angle of Attack, and put exactly enough wing on the craft so that it achieves 0 angle of attack at the intended altitude and airspeed. Your test craft has FAR too much wing for it's mass and is flying far too low for it's speed in the first 2 tests. A single pair of Big-S wings will happily hold up at least 50 tons of plane. At MINIMUM you should be using a wing loading of 5 tons mass per 1 wing area. My standard ratio is 1 rapier and 1 nerv per 36 tons of takeoff mass, and 1 wing area per ~5 tons. This typically results in a flight ceiling of about 20 km at 1650 m/s, after which I kick on the nervs and cruise to orbit. With this style of craft, 8km/s in LKO, or alternately, 50% payload fraction, is trivial to obtain. When built and flown correctly, wing incidence will always improve efficiency to orbit, via an enormous reduction to body drag by flying 0 AoA through almost the entire flight. Also forgot to mention, but Mk2 parts are really, really bad. They generate both regular body drag, AND wing drag. So they have extremely high drag at 0 AoA, and it only gets worse as you add AoA.

A couple things: 1. It is possible to take off from the runway at 100 m/s at about 1/3 the wing loading you are using, the overwing is also cutting into your l/d ratio by requiring a non ideal AoA. Move CoL and landing gear closer to CoM and add canards to greatly ease the takeoff flare, and remove one set of big-s and replace with fuel tanks. 2. Something is generating an absolutely horrendous level of drag in those side stacks. Go into the Alt+f12 menu and go to the aero tab, turn on "aero data in part action windows" and open the action windows for all parts on the craft one by one to find the culprit. Do this on a run where you do not use the props at all, keep them closed in the bays all the way from launch (there is an issue where closing the bay does not update the aero gui and part action windows, so it makes diagnosing drag issues harder), I suspect the issue is likely something attached to the wrong node by mistake, resulting in a large flat face presented to the airflow. Double check the attachments around the prop bays to make sure everything is actually attached how you would expect. 3. Get the mods "CorrectCoL", and "RCSBuildAidContinued", they are vital for efficient spaceplanes, they streamline the build process immensely. Between the two of them, you can see your wet and dry CoM at the same time, see the TRUE CoL (taking ALL parts into account), and generate a stability graph for a selected speed and altitude. Between these you will be able to accurately diagnose the flipping issue, and better tune the craft wet/dry CoM excrements and the CoL-CoM relationship. 4. Once properly tuned, you should expect to see a lift to drag ratio of at least 2 at Mach 1.15, and at least 3.5 at Mach 2 (Maximal values for a non clipped craft are somewhere near 2.9-3 at Mach 1.15, and 4.8-4.9 at Mach 2).

Not angling your wings is a horrifically bad decision. Proper wing incidence + the consequent reduction in needed wing area can more than double a craft's lift to drag ratio, allowing ascent with half as many engines.

I don't have any of them uploaded, but Brad uses a very similar ratio of rapiers and nervs in this video: In terms of how it is to fly.... really damn hard. You accelerate really, really slow, and need quite a high takeoff speed, and once you get in the air, you spend a while with very little excess thrust.