vyznev

Okay. Craft file: https://kerbalx.com/vyznev/TurboFun Imgur album with more screenshots: https://imgur.com/a/k1bYB05

Could be many reasons, but I've had that sometimes happen with too many autostruts. For example, from an old mission: IME, autostruts are really useful, but they're a bit like a strong spice, best used sparingly. I use them, but only on specific parts that I think will need them. And I particularly try to avoid having too many parts autostrutted to the same (typically root) part, since that seems to be a common trigger for weird physics glitches. Of course, all this could also be completely irrelevant to whatever's messing up your vessel.

I believe all the engines I used (Spark, Terrier, Mainsail and Vector) are stock. So is most of the rest of the ship. I did use the 5m diameter DLC fuel tanks in the first stage asparagus cluster, but I don't think replacing those with a bunch of smaller tanks would've made much if any difference. Not sure what kind of kraken attacks you're experiencing, but if it's just the ship wobbling too much and shaking itself apart, may I recommend more struts? I have quite a few of them on my ship (many aren't visible since they're hidden between the asparagus cores and the side boosters) and I also have a bunch of parts autostrutted to either "root part" (which was supposed to be the probe core on the last stage, although I think I accidentally made the fuel tank under it the root at some point) or to "grandparent part" (handy on side boosters attached via radial decouplers). My basic guidelines when applying struts are that: 1) every side booster should be attached (by strut or by decoupler) from at least two points, one at the top and one at the bottom, 2) side boosters on upper stages should be strutted to lower stages to support their weight, and 3) if you must have a bunch of small parts sandwiched between big fuel tanks in the core stack (which tends to make that section of the stack springy and bendy), either bypass them with struts or use rigid attachment to essentially fuse those small parts together.

@farmerben Here's the video: https://www.youtube.com/watch?v=FgKEVLOf0go (already edited into my earlier post)

Nope. I checked, and this was my highest speed during the run: (It's relative to Minmus, since I flew most of the ascent with the navball in target mode, but that shouldn't matter much since the orbital velocity of Minmus around Kerbin is only 274 m/s, and it's at an almost exactly 90° angle to my flight path. Doing the math, sqrt(95062 + 2742) = 9510, so I shouldn't have been going faster than about 9510 m/s relative to Kerbin.) I'm not sure why you're getting a slower time with a higher peak velocity. Maybe I'm just doing more efficient (= shorter) acceleration / deceleration burns? Mine are definitely not optimal, but I don't think I'm more than a couple of minutes off from the best my craft can do. And even that took quite a few retries to practice. I do have a video of the entire flight, I just need to edit and upload it…

Same rocket as before, new flight, this time without premature staging. No mid-flight reloads this time, either. Time at landing 1:31:24. So close to 1:30… (Don't ask me where the heck the F3 display pulls the "highest speed achieved" value out of. My actual peak velocity at the start of the coast phase was just a hair over 9,500 m/s.) Edit: Here's the whole flight, sped up 4x:

Well, if you need a lot of thrust for cheap, the answer is surely to use SRBs, right? You just need to use enough of them. More images in imgur album: https://imgur.com/a/u93vgAf IIRC, a full S4-512 tank costs 51,200 kerbucks, so my score should be 442,632 – 51,200 = 391,432. I probably could've squeezed a little bit more off that by trimming off unnecessary parts (like do I really need that engine plate?), but meh, whatever. Craft file: https://kerbalx.com/vyznev/Fuel-Lifter BTW, one issue I noticed while building this craft was that the Clydesdale SRBs kept breaking off from the core on the launchpad. Apparently I'm not the only one who's noticed this issue. FWIW, my solution to the problem was: 1) struts, 2) autostruts, and 3) more struts.

Here's 1:37:10: More screenshots and flight details on the imgur album: https://imgur.com/a/xul1flc I actually kind of messed up this flight by dropping one of my booster stages while it was still half full, which probably cost me nearly 1000 m/s of total delta-v. With that mistake fixed and better flying in general, I think this rocket could probably do sub 1 hour 30. Craft file here: https://kerbalx.com/vyznev/Minmus-Express

I might try this later. In the mean time, some random thoughts: It looks like most (all?) attempts so far have spent some time coasting under time warp. That's wasted time that could've been spent accelerating (and later decelerating), if one just had more delta-v. Over two hours seems long enough that nukes or even ions might be worth using for their high Isp (and consequently high achievable delta-v), even with their low TWR. At the kinds of velocities and accelerations involved, both Minmus and Kerbin gravity should be basically negligible. So is the orbital motion of Minmus, too. That means it should possible to approximate the ideal flight profile with some basic high school physics. In particular, I believe an ideal flight profile (using stock KSP with no fancy autopilot / landing assist mods) should look roughly like this: Time warp until Minmus is near its ascending / descending node. Time warp further until Minmus is overhead from KSC. Set Minmus as target and burn towards it until your remaining delta-v approximately equals your velocity relative to Minmus (with a bit of safety margin). Turn around and coast (hopefully not too long) until your time to target is somewhere around half(!) your total remaining burn time. Burn retrograde to slow down. Make sure not to miss the target. (The "half your remaining burn time" estimate is assuming constant thrust and neglecting gravity. A proper suicide burn calculator should be able to give a more accurate estimate, or you could just quicksave and use trial and error to find the optimal start time for the deceleration burn.)

Uh, this counts, right? Because damn, this thing gets wobbly near the end of the landing. I guess it's my own fault, I should've included a reaction wheel. Or at least some RCS thrusters. But I wanted to show that this can be done with just one Puff engine and no other attitude control mechanisms. Which I guess I technically did. In general, I went for minimalism with this design. There's the single Puff engine (rotated and offset so that its thrust vector points straight down and lines up with the center of mass), one monoprop tank (technically about 30% more fuel than needed for an optimal landing, but a bit of extra delta-v margin is always nice), one OKTO probe core (lightest possible, no built-in reaction wheels), six batteries (I guess I could've left one out; did you know that the small batteries are cheaper and lighter per unit of charge than any of the bigger ones?), two solar panels (again, technically one more than needed), one antenna and of course one seismometer.

That's true in KSP too.* KSP plane builders just have to either manually tilt their wings by a tiny amount (which is awkward, since the rotate tool will only snap to multiples of 5°) or live with the fact that their plane must keep its nose pointed slightly above the horizon in order to maintain level flight (and the slight extra drag this produces, for most conventional plane designs). *) Easy enough to test by putting some vertical wing panels on a rocket, launching it straight up (and holding surface prograde with SAS) with aero force visualization turned on, and observing that the wings indeed produce no lift in that case. Do they, tho? I was under the impression that, while e.g. the "Big-S" wing parts indeed have variable thickness, they're still symmetric and thus have zero camber (and thus work equally well even if placed upside down). I also suspect that, even though those wing models indeed look back-to-front asymmetric (both in thickness and in overall wing shape), the stock KSP aero model probably ignores that too. That shouldn't be too hard to test: just build a plane using those wings, with nothing else attached to them, then turn them backwards (while keeping the nominal center of lift at the same place, as shown in the SPH, which may require translating the wings forward or backward) and see if the plane still flies the same way.

…anyway, I decided to try cheating a bit by moving the engines down and out (which of course would never work in real life, but in KSP connected parts don't have to touch and same-vessel collisions are optional ). With that change, it does fly. Not well, mind you, but well enough to take off and stay in the air. Trying to rotate the engines and change flight direction went about as well as I expected, though. Mind you, that's probably in part because some of the control surfaces started acting the wrong way when I changed the control point. I think I remember having that issue with the CuBoid too, and IIRC I "solved" it there basically by having a completely separate set of control surfaces for sideways flight and locking the surfaces that were not in use based on the selected heading. That might be the way to go here, too.

Anyway, that crazy airplane concept mentioned in your original post intrigued me enough that I wanted to try building a working replica of it on KSP and see how it flies. To absolute nobody's surprise, it flies like excrements. Or rather it doesn't, because it can't even take off — the elevons don't have nearly enough pitch authority to counteract the massive downward pitch moment from the top-mounted engines, and its negligible yaw stability isn't enough to keep it going straight. And the wobbly rotation servo doesn't help with that, either. About all this thing's good for really is doing donuts on the runway. I could maybe fix the yaw stability by adding some kind of deployable tail surfaces, like I used on the 8-way CuBoid plane I linked above, but I see no easy way to solve the pitch problem — at least not while leaving the result looking anything like the original concept design. (A bit ironically, turning the landing gear 90 degrees and taking off narrow-side-forward might actually work better, at least if I added some extra pitch control surfaces near the nose and the tail in that orientation, since that would put them further fore and aft of the center of mass.) Alas, all this means that I didn't even get to test how switching direction mid-flight (near Mach 1!) would've worked. I suspect it would've also ended rather poorly, though. Honestly, looking at the design, I have no idea how anyone ever thought this would fly in real life, and I suspect making it flyable in KSP isn't possible without massive cheating* either. But who knows, maybe someone better than me can pull it off. *) Like adding lots and lots of reaction wheels or abusing the aero model to make parts dragless, etc.

Not sure if this old craft qualifies, since it's uncrewed. If it does, I believe it meets the "Reuse Me", "Look Ma, No Wheels" and "Double Trouble" extra goals.

That's a good question. I can see at least three reasonable rulings: If the contract doesn't tell you to return (or to "transmit or return science" or return samples or recover a part or a kerbal, etc.) then you may not return. Better send a probe or bring plenty of snacks. Returning without being told to is allowed if you've already returned from all the locations (fly-by / orbit / surface of a given body) visited during the trip, as that means you'll get no new milestone for it. Returning is always allowed. Kerbals want to get back home to their own bed. I'm not sure which of these (other than option 3, obviously) my own playthrough actually qualifies for. I've tried to use probes for contracts that don't require returning, but looking back, I might've technically blown it with the very first progression contract, which IIRC strictly speaking only requires launching a vessel, not recovering it. In any case, if you're not playing a strict "one contract, one flight" version of this challenge, I see no problem with grabbing a second contract to return from wherever your kerbals happen to be. That said, I would probably not allow just any repair contract as an excuse to return Bill home, since you could always hire another engineer instead. (BTW, I already added one exception to my strict rules in my own playthrough: if the contract requires me to rendezvous and/or dock two vessels, then I'm allowed to launch two of them. While I'm pretty sure I could've finagled some way to complete that contract with just one launch, I felt that if I'm going to do what's basically a "re-enact Gemini 8" contract, then I should do it properly. )