Difference between revisions of "User:Quizzical"

Revision as of 17:07, 5 September 2020

My user page is no longer a red link.

I created this account mainly to post my game guide, which is available at Tutorial:How_to_Play. Rather than all of the theory, some people just want a vehicle that will work for a given purpose. As such, I'll explain the main vehicles I used for a run of recovering all science on the moderate difficulty level. I had an artificial rule for myself of no mining, and that does affect the vehicles I created. Furthermore, when recovering science, I wanted to deliver three copies to Kerbin, while keeping one copy on a Mobile Processing Lab MPL-LG-2. Surface features transmit with 100% efficiency, but this meant that for surface samples, mystery goo, and so forth, I needed four copies of each experiment result, three of which would be carried back to Kerbin.

As of this writing, the run is incomplete, so not all of the vehicles have their details here yet. I've largely designed most of them, but want to make sure that everything works how I expect before explaining them here and posting them in the Steam workshop. As such, this section is incomplete.

Landers

Landers are the vehicles that are responsible for landing on a given planet or moon to recover any science data that requires that you be landed to run an experiment. In places where this is relevant, they are also responsible for collecting science while splashed down or in flight in the lower atmosphere. Those require considerable thrust to move from one biome to another. Experiments run in space or in an upper atmosphere can readily be recovered from orbit and are not the responsibility of landers.

Different planets or moons called for different landers in a variety of ways. Even so, on a majority of them, my light lander could start in orbit, land in several biomes, and return to orbit to dock with a space station again all on a single tank of fuel. When that was possible, I used the same light lander in many places. When it was not possible, I built a custom lander for that particular planet or moon to optimize it for local conditions. The light lander does not stage, but merely refills fuel for its single stage. The other landers all have multiple stages as a way to get sufficient delta-v to complete their mission.

Duna has light enough gravity that having multiple hops and then refueling is possible. The light lander won't work, however, because it has very bad aerodynamics. It also lacks a Atmospheric Fluid Spectro-Variometer. So I made a custom lander for Duna.

For Tylo and Eve, it was sufficient to make a single lander for use everywhere. On Tylo, a single lander is responsible for landing in multiple biomes before returning to orbit. On Eve, it is enough to land just once, but the lander had to be capable of either landing on a slope or splashing down in an ocean. For Kerbin, it was also have been possible to use a single lander, but I customized it some for particular launches, such as replacing wheels by lander legs when I knew that I wouldn't need to drive, or eliminating early rocket stages when I didn't need to go far.

On Laythe, I needed two separate lander vehicles for two different purposes. The simple lander can land in one place, pick up science, and return to orbit. But there was also a sinker vehicle that was designed to pick up science from the lower atmosphere on the way down, then land in the water and sink to the bottom. It could pick up the science for being splashed down as it sank, then for being landed in the same biome upon reaching the bottom of the ocean. This allowed picking up the science while splashed and landed in the same biome from a single vehicle. It was also necessary to have something that could sink in order to be able to recover the science while landed in certain biomes that are entirely covered in water.

You cannot land on Jool, but you can pick up science from its lower atmosphere. I decided to list my vehicles for doing so in this section.

Light lander

The purpose of my light lander is to land on a relatively small planet or moon and pick up all available science there. It is intended for use on all planets or moons that do not have an atmosphere, except for Tylo. It offers 4197 m/s of delta-v in a vacuum from a full load of fuel. On Gilly, Minmus, Pol, or Bop, it can easily land in all biomes and return to orbit on a single load of fuel. On Ike and Dres, it may be able to do that, but it will be tight. On Mun, Eeloo, Vall, or Moho, you're going to need to refuel.

The lander has a mass of 11.255 tons, and its single Terrier engine provides 60 kN of thrust, for about 5.33 m/s^2 of acceleration with full fuel. That is plenty enough for any location that lacks an atmosphere other than Tylo, as the highest such acceleration due to gravity is 2.7 m/s^2 at sea level on Moho, which gives the lander a thrust to weight ratio just under 2. Naturally, the mass goes down as you burn fuel, ending at 3.255 tons, for about 18.43 m/s^2 of acceleration.

With lander legs retracted, the lander has 4.5 m of height, 3.3 m width, and 6.5 m length. With legs extended, the height increases to 5.1 m and the width to 5.8 m. The four LT-2 Landing Struts thus provide a rectangular base of support when landed that is 5.8 m wide on the short side, and a little over 6 m on the longer side. With a center of mass that is only about 2 m above the bottom of the extended lander legs, the maximum slope on which the lander should be stable is about 55 to 60 degrees, depending on how it is rotated. That allows landing absolutely anywhere on the various planets or moons for which it is intended except for certain portions of Dres's canyons.

The design of the lander is that the central column has, from top to bottom, a Clamp-O-Tron Docking Port, an SC-9001 Science Jr., a [[Mk1 Lander Can], a Probodobodyne HECS2 probe core, and a terrier engine. As these are all of small radial size, they provide good structural stability when connecting to other components mounted on top and bottom as part of a stack to deliver the vehicle.

The fuel tanks are a pair of Rockomax X200-8 Fuel Tanks, attached to the HECS2 probe core on opposite sides. The four lander legs are radially attached to the fuel tanks as high as possible, in order to bring the rest of the lander as close to the ground as possible. The mystery goo is right outside of the lander can door for easy access to reset it. This allows a scientist in the lander can to go on EVA, reset both the science jr. and mystery goo, pick up a surface sample and EVA report, and then return to the lander can all without needing to move around outside at all.

The rest of the equipment is radially attached to the fuel tanks to bring the mass as low to the ground as possible. That includes the other relevant science equipment: a thermometer, barometer, seismometer, and gravmax. It also includes a single PB-NUK Radioisotope Thermoelectric Generator to ensure never running out of electricity. I include three Experiment Storage Units, to allow pickup up four copies of all science as I move around: three in the storage units and one on the HECS2. This allows returning three copies of all science data to Kerbin, while having a fourth available for a mobile science lab. Returning three copies rather than one to Kerbin allows getting over 98.4% of the maximum possible science from all experiments that the lander will do. For comparison, a single copy would gather less than 80% of available science for three of them.

For communications, I use a Communotron 16, which is the weakest and lightest of the dedicated antennas, but still strong enough to connect to an RA-2 Relay Antenna in orbit 1000 km away with a signal strength over 0.997. As it has 100 times the antenna rating of the weak antenna built into the probe cores, this gives you ten times the range to connect to relays.

In testing, the light lander is a tough little critter. A free-fall from over 1.3 km on Minmus that resulted in hitting the ground at over 35 m/s did not result in any damage at all, even though I had enough fuel on hand to have a little under 2000 m/s of delta-v still available. That's the benefit of having lander legs account for about 12% of the vehicle's dry weight. It also landed quite easily on a 47 degree slope on Minmus, which is the steepest that I've been able to find there.

Kerbin lander

The Kerbin lander is built for a peculiar purpose that likely isn't what you want. Nevertheless, it fills its goal admirably well, so I list it here. The basic goal is to pick up all science from Kerbin's surface and lower atmosphere, including landed at the special biomes such as the island airfield. From there, load all science from a given run onto a detachable rocket that can deliver it to a station in space. Furthermore, keep three copies of all science picked up (both from the lower atmosphere and on the ground) on the base piece that is left on the ground when the detachable copy takes the fourth copy of the science into space.

This is handled in three basic phases. The third and final phase is to build a rocket that can deliver the science data to the orbiting station. Building a rocket that will work at all is not at all difficult. The real challenge is to make it as small and light as possible, and with a shape that will not cause trouble for the earlier phases. I use a fairly simple design with three rocket stages and an initial mass of 2.34 tons. There is technically a fourth stage to discard the Aerodynamic Nose Cone once the rocket is at a high enough altitude to no longer need it. Considerable testing found that putting an aerodynamic nose cone on top in spite of its small radial size offered the best performance even though all of the in-line components in the rocket are of tiny radial size.

The first phase is also pretty simple. Regard the vehicle for the second phase as a payload and put enough thrust behind it to deliver it into a polar orbit about Kerbin with enough delta-v remaining to land safely. Even if this is done badly, the worst that will happen is a waste of funds. My design uses a skipper and four thumper boosters at launch, then discards the latter before the former as they run out of fuel. That's not enough for a stable orbit, of course, but I get additional thrust from the second phase.

The second phase is the real challenge, for several reasons. For starters, some of the places that it has to reach are very small. The vehicle needs to be able to splash down at tundra, grasslands, deserts, and badlands, all of which are naturally land biomes. Lakes in them have to be very small in order to not register as the shores or water biomes instead. The special biomes such as Woomerang and the Dessert Launch Site are much smaller yet, and nearly impossible to land on directly from orbit without a lot of thrust happening near the ground.

The solution here is to build a vehicle that can drive around after landing. That way, instead of failing because you missed the center of your target by 10 m, you can just land within a few km, then drive to the proper location. Yet this also creates additional complications. For starters, rover wheels explode at 1200 K, which means that you'll need a much larger retrograde burn to land without blowing up. It's also naturally harder to land on wheels rather than lander legs, and you need to be able to drive up whatever slopes you encounter on the way to your target.

And then there is the problem of tipping over. Rovers are naturally prone to tipping over. Vehicles that splash down in the water are even more prone to tipping over. And rovers that drive into lakes to recover data while splashed down are especially prone to tipping over. The normal solution to all of these is to make your vehicle short and wide. But short and wide doesn't get a rocket into orbit while using minimal mass, as is needed to deliver the science into space. So that detachable rocket on top makes the tipping over problems much, much worse. Avoiding that in the second stage is the real challenge of the design.

The design of the second phase is to have a central column that is, from top to bottom, a Mk1 Command Pod, a Service Bay (1.25m), an SC-9001 Science Jr., an RC-001S Remote Guidance Unit probe core, an Advanced Inline Stabilizer, another 1.25 m service bay, and a dart engine. Science equipment goes in the top service bay, while electrical equipment and experiment storage units go in the bottom service bay. With a scientist as the occupant of the command pod, this makes it easy to go on an EVA, reset the science equipment, and board again without needing ladders. Sometimes you have to move slightly to reach the science jr. unit and sometimes not.

But the real key to the design is that it has four FL-T800 Fuel Tanks connected to the central column by Modular Girder Segment XLs. That gets you the wide base. Each fuel tank has two TR-2L Ruggedized Vehicular Wheels at the bottom, which allows you to drive. Each fuel tank also has an [[Aerodynamic Nose Cone on top, as well as four Mk2-R Radial-Mount Parachutes and one Mk12-R Radial-Mount Drogue Chute near the top. That provides enough drag to splash down without breaking fuel tanks even when the fuel tanks are completely full. Having lightweight gear extended far out from the center keeps the vehicle from tipping over even while splashed down, and regardless of whether the fuel tanks are full, empty, or anywhere in between. Each fuel tank is also connected to the top of the detachable rocket via a strut for stability.

When the second phase fuel tanks are empty, the second phase has a dry mass of 11.725 tons. Putting that on top of eight TR-2L wheels means that you have under 1.5 tons per wheel, so ascending modest slopes is easily doable--such as up the ramp to the Woomerang or Dessert launch sites. It's hardly the nicest rover to drive around on, but it's not bad, either, especially given that it has to carry around a rocket on top able to dock in space above Kerbin.

The lander can also pick up the science around the Kerbal Space Center by removing the first phase and having the fuel tanks of the second phase start empty. While there, you might as well pick up the science landed and splashed on the shores biome, with the latter available by driving into the ocean. And then you still have the rocket on board available to carry that science into space.

Duna lander

all land science, enough delta-v to hop between a few nearby biomes before returning to space

Tylo lander

all land science in vacuum, stages to land repeatedly nearby and then return to space

Laythe lander

all land science, stages to land on Laythe using parachutes, then return to space

Laythe sinker

all land science, stages to land in Laythe's ocean, sink to bottom, recover both splashed and landed science, then return to space

Eve lander

Eve has a well-deserved reputation for being a hard place to land--and a harder place to leave alive, so we're going to have to empty the bag of tricks here. The lander needs to be able to land in any biome, or splash down. Regardless of whether it lands or splashes down, it needs to be able to return to orbit. Eve is relatively flat, so handling extreme slopes isn't necessary, but I didn't want for it to be necessary to find somewhere that is almost perfectly flat. Rather, my lander can handle slopes of a little over 20 degrees, which is enough that landing in a random spot will nearly always work fine. I haven't yet seen it fall over.

The full vehicle to go from low orbit of Eve to touching down and back to orbit is 126.855 tons. That's what you have to deliver to a low orbit of Eve before it can work. There are basically four phases of the lander:

1) Go from orbit down into Eve's lower atmosphere without burning up. 2) Go from Eve's lower atmosphere to down on the ground or splashed in the ocean and pick up all low flight and either landed or splashed science for that biome. 3) Go from on the ground to Eve's thin upper atmosphere while still moving up fast 4) Go from Eve's thin upper atmosphere to a stable orbit and then rendezvous with whichever vehicle is there to pick you up.

Each phase basically has to treat the subsequent ones as a payload, so let's tackle them in the reverse order. The final payload that we're trying to escape Eve with has a Mk1 Command Pod on top of a Service Bay (1.25m) with four Experiment Storage Units inside, and that's it. With the monopropellant removed from the command pod, that's a mere 1.1 tons for the final payload, to get four copies of the scientific data out. That's the lightest command pod to have decent aerodynamics. Being able to handle 2200 K rather than a mere 2000 K also helps when trying to escape Eve, as you absolutely can burn up in your ascent.

The final payload is also notable for what it doesn't contain: a probe core, electrical equipment, communications equipment, parachutes, or anything else not listed. The lack of a probe core means no need for electrical equipment to power it. The 50 electric charge in the command pod is enough to power the reaction wheels until the command pod can dock. The lack of a probe core for SAS also means that you need a pilot.

A circular orbit just outside Eve's atmosphere is 3218 m/s, and it's going to take a lot of delta-v go to from maybe 800 m/s at the start of this phase to a stable orbit. My fourth phase vehicle has four stages with a total mass at the start (including the command pod and so forth) of 10.94 tons for 4953 m/s of delta-v. The four stages each have a single fuel tank separated by a TD-12 Decoupler. The the top two tanks are an FL-T200 Fuel Tank, then an FL-T400 Fuel Tank, and finally a FL-T800 Fuel Tank on the bottom. The decouplers directly connect fuel tanks.

The decouplers also have fuel crossfeed enabled. For rockets, I use four cub engines, with two attached to the top fuel tank and two attached to the third from the top. The two pairs of rockets are each placed opposite each other for balance, and the lower pair is rotated 90 degrees from the upper so that the upper pair doesn't blast the lower. All four engines fire at the start of the phase, and continue firing after staging to drop the bottom tank. Thus, we have two stages where all four cub engines fire, and two stages where only the final two do. This keeps your thrust to weight ratio above 1 for the entire final three stages and part of the fourth, even in Eve's 16.7 m/s^s gravity.

There are two points to this approach. One is that using the same engines for multiple stages saves quite a bit on engine weight. I get through the final four stages with only 0.72 tons of engines. A single dart alone weighs more than that, and a terrier not that much less. Additionally, not having the engines in-line allows for a shorter rocket, and that allows it to be more stable. Excessively long rockets can become very wobbly on your ascent, and we're going to have to put some additional stages below this, too. Told you we were going to empty the bag of tricks, and we're just getting started.

This design actually confuses the game's delta-v calculator, and it isn't sure which order to use the fuel tanks. As such, I have to adjust the flow priority by adding -32 to the top tank, -21 to the next one down, and +31 to the bottom of the four tanks from the final phase. Otherwise, the game will try to empty the tanks in the wrong order, and then you can't stage whichever one is empty. It also means that the delta-v meter will show garbage, so you have to watch whichever tank is the lowest one left to see when it is empty, and stage it then.

It's worth it because it also means that you likely have about 1000 m/s of delta-v remaining when reaching orbit from sea level, or more from a higher landing point. That both allows some slop to cover it up if you're less efficient in your ascent, and also provides plenty to rendezvous with your vehicle that stayed in orbit.

The third phase is what leaves the ground from Eve. Including the fourth phase as a payload, it has a total mass of 71.46 tons and a total delta-v of 8704 m/s in a vacuum. Naturally, you're not starting out in a vacuum, so it's 3911 m/s of delta-v at sea level on Eve, with the cubs on top accounting for 1550 m/s of that. If you count the delta-v at the altitude where it gets used, this gives you somewhat over 3000 m/s of delta-v to reach an altitude around 50 km while traveling at perhaps 800 m/s or so.

The design of the rocket is that under the fourth phase, we stick another TD-12 decoupler, two more FL-T800 fuel tanks, and a vector engine. Meanwhile, we make a ring around it of six columns that each have an FL-T800 and an FL-T400 on top of it, with an Aerodynamic Nose Cone on top and a dart engine on the bottom. Attach them all to the central column with a TT-70 Radial Decoupler and add a strut and you have the basic design. You can either enable crossfeed on the decouplers or use FTX-2 External Fuel Ducts to get the proper asparagus staging, and I do some of each.

The idea is that as you leave the ground, you have the vector engine and all six darts going all out, which gets you a thrust to weight ratio of a meager 1.13 on Eve at sea level, but that goes up fast as your fuel level goes down. Once the fuel from a pair of opposite columns is gone, stage it to drop it and lose two darts. Do so three times until you just have the vector left, for four stages in phase 3. Decouplers that enable fuel crossfeed confuses the game's delta-v system, so I put a -31 flow priority on the fuel tanks in the central column to make them used last.

You don't want to keep all engines going all out all the way up, however, as a vector engine alone is too much thrust once you've burned a lot of fuel. Rather, you adjust the throttle to keep the thrust to weight ratio around 1.5 most of the way through the phase. Dart engines are more efficient than vector engines at any pressure, but especially so at high pressure. As such, we want to keep the dart engines going all out, while only scaling back the vector engine. That is done by adjusting the dart engines to ignore the main throttle, and instead, using an independent throttle meter that stays at 100% at all times. That will keep the engine going full blast from the moment it is enabled until it is out of fuel.

The second phase is to go from descending at several hundred m/s in Eve's lower atmosphere to safely on the ground and gathering the scientific data. This doesn't take any rockets, as gravity will do the work. It does take parachutes, lander legs, and science equipment, however. And we don't want to attach any of that to the fuel tanks that will be used for the ascent as we want to drop the weight.

The idea is to build some scaffolding around the ascent vehicle. Near the bottom of the fuel tanks in each of the six outer columns, attach a TT-70 radial decoupler, with a Modular Girder Segment XL attached horizontally to the decoupler. Put four LT-2 Landing Struts on the girder, with two near each end, and you've got your lander legs. On top of the far end of the girder, put another girder pointed vertically, and then a Service Bay (2.5m) on top of the girder. Stack two more girders on top of the service bay, with eight Mk2-R Radial-Mount Parachutes near the top of the top girder. That gets you your lander legs on the bottom and parachutes on top, so that you can land. Eve's thick atmosphere will slow you plenty enough that drogue parachutes don't matter.

The service bays in the outer scaffolding give you plenty of space to store your science equipment. I also attach a Communotron 16-S to an opposite pair of them to make sure that I can get a good connection to the nearby relays during my descent, as I don't like to lose KerbNet access to see where I'm going to land.

Remember that we're only bringing a pilot and not a scientist on this mission, so we can't reset a Mystery Goo™ Containment Unit or SC-9001 Science Jr. after it is used. That's readily handled by just bringing four of them. Those both have data vary by biome only when landed or splashed down, so they don't also need to get the low flight data. Just one vehicle can get the low flight data for those experiments, and I have the vehicle that drops my rover do that.

That brings is to the first phase, which raises the question of how to just drop into Eve without promptly burning up. My answer to that is to use a lot of Heat Shield (10m)s. That, of course, changes the question to how to use heat shields without flipping over and then still burning up. The answer is that some heat shields are there for the heat shielding and others are there for drag to keep you upright.

Underneath the vector engine, I put another TD-12 decoupler, an RC-001S Remote Guidance Unit (so that I can check KerbNet during my descent), a Rockomax Brand Adapter, a Z-4K Rechargeable Battery Bank, an Advanced Reaction Wheel Module, Large, a Rockomax X200-8 Fuel Tank, and a poodle engine. That allows some amount of maneuverability, and the large battery has far more power than I'll need even without any recharging capability. Around the central fuel tank, I radially attach three Modular Girder Segments, and put another Rockomax X200-8 fuel tank at the end of each. I then put an inflatable heat shield on the bottom of each of the fuel tanks. The additional fuel tanks mean that I now have 463 m/s of delta-v from the poodle engine to get into Eve's atmosphere. From a low orbit, I do about another 1/6 of a revolution about Eve from the time I start the poodle burn until I touch down.

The three columns with the heat shields are not lined up exactly with the six columns above. Rather, each of the three lower columns is halfway between two upper columns. That lines things up properly for each of the three lower heat shields to shelter two columns above it, as well as the central column.

Meanwhile, at the top of the six upper columns, just above the parachutes, I put more heat shields. I put a TD-06 Decoupler and an inflatable heat shield at the top of each column. The idea is that when all of the heat shields are inflated, much of the drag is above the center of mass. Furthermore, if the vehicle starts to tilt one way rather than pointing directly retrograde during the ascent, that exposes more of the heat shields on the side it tilts toward and less on the side it tilts away from. That allows drag to provide a stronger upward force on the side it is tilting toward and weaker on the side it is tilting away from. That combination means a strong net torque to force the vehicle back toward pointing retrograde. You'll want to get pointed retrograde while in orbit so that you don't burn up before the heat shields can rotate you properly, but once you're pointed retrograde while in Eve's atmosphere, you can turn off SAS and just let the heat shields keep you pointed retrograde until you're moving slowly enough to not need the heat shields anymore.

Because the heat shields inflate into the poodle engine, the game regards them as protecting the engine, so that it won't burn up. Meanwhile, the thrust from the engine comes out below the heat shields, so that you can continue to fire the engine and have it work properly. That's not at all how real physics works, as it can't simultaneously be above and below the heat shields. But I told you that we're using a lot of tricks here. Add some struts for stability and you've got the whole vehicle.

Phase 1 ends when you discard the heat shields, which we need to do. If not discarded, the outer columns of the ascent vehicle would hit the upper shields on the way up. You want to discard the upper heat shields before the lower ones. You need to wait until you're going slowly enough to not burn up, but you also want to discard the upper heat shields while you're still under heavy acceleration from drag, so that the air will rip them off and get them away fast. Because the upper heat shields inflate into each other, they will collide and explode when you stage to drop them. That will usually destroy some decouplers, often some parachutes, and sometimes some struts. None of those are still essential, but if it takes out your command pod, that's a mission failure. Discarding them while you're still dropping at 500 m/s seems to avoid that. Waiting until you're only going 200 m/s makes it far riskier. Regardless, save your game before staging the heat shields. You will need a lot of parachutes to land, but losing several of them is no big deal. If it's lopsided so that you're worried about tipping over, you can cut some parachutes opposite the ones that exploded for balance.

The lower heat shields are part of a single unit that gets staged by a single decoupler, so they don't necessarily explode upon being released. They do have a lot more drag than your vehicle, in spite of a lot less mass, however, so they'll try to go up relative to your vehicle. The way to handle this is to turn off SAS and wait until you get far into Eve's lower atmosphere, so that the atmosphere flips you over. Once the heat shields are above your vehicle, it's safer to stage them. You don't want to wait until you're too close to landing, however, as you don't want the heat shields to land on you. Still, as with the upper heat shields, save your game before staging.

Jool diver

parachutes to slow fall near end, transmit manned science, die hope to figure out how to return kerbal to orbit, but hard

Jool probe

probe to recover all unmanned atmospheric science, return to space

Rovers

The natural purpose of a rover is to scan surface features. Unlike some science experiments, surface features can be transmitted back at 100% efficiency. Thus, you can scan a feature once, transmit the data, and have all of the science you can get from it, at least apart from taking the results to a Mobile Processing Lab MPL-LG-2. Surface features are in very particular locations, though, making it difficult to land in just the right spot to scan one. Rather, it is much easier to land close, then drive over to one. And to drive, you want a rover.

I task my rovers with some additional work. Rather than just landing a rover and scanning surface features, I also use them to deliver the deployed science. The basic technique is that the rover will have a small lander connected to it via a stack separator. The lander will have equipment for all of the deployed science, a command pod with an engineer inside, and some rockets that can get the engineer back into orbit. The plan is that I can land the rover, remove the lander from the rover, and then let the rover go scan all of the surface features and transmit the science. It can then bring a copy of all surface features back to the lander, where the engineer picks them up and takes them into orbit, to deliver them to a mobile science lab. If the rover breaks a wheel, there is conveniently an engineer right there to fix it.

Some of the deployed science equipment needs to be deployed by an engineer, and some by a scientist. I usually have only the engineer land with the rover, and a scientist can land nearby using the normal lander for that world, then stop over to deploy the science experiments. The lander has to land in that biome sometime, after all.

I use two different models of rovers, as different planets and moons have different challenges for rovers. In heavy gravity, even a mediocre rover design probably won't flip over apart from doing something dumb. However, it can be difficult to be able to ascend the hills in your path. It's frustrating to see the surface feature you need right near you, but you can't get there because it's uphill. As such, I have a hill climbing rover that is designed to be able to climb steep hills.

In light gravity, it doesn't take nearly as much force to drive up a hill. Furthermore, even if a hill genuinely is too steep, you can just fly over it by firing up rocket engines. Rather, the problem in light gravity is that the gravity doesn't do such a good job of keeping your rover upright. Rather, it is likely to be very prone to flip over, and then parts break off and the mission fails. As such, I have an "unflippable" rover designed to be very resistant to flipping over for use in light gravity.

The unflippable rover has rockets of its own, so it can deliver itself and the deployed science to the surface. The hill climbing rover does not have rockets or fuel to save on weight (and thus be able to climb hills better), so it cannot deliver itself to a planet. Rather, it requires some additional infrastructure to land it that isn't part of the rover proper. That additional structure varies by planet, as does the escape vehicle to return the engineer to orbit. And, for that matter, on Kerbin, you just recover the engineer and return only an unmanned probe to orbit, while on Eve, you the vehicle to deliver the rover doesn't have the capability of returning to orbit, which requires yet another vehicle.

Unflippable rover

The unflippability of the unflippable rover is slightly exaggerated, but only slightly. There are four basic ways that rovers flip over: 1) be on too steep of a hill, 2) accelerate or brake too sharply, 3) turn too sharply at high speeds, or 4) hit a ramp at high speeds to get off the ground, then turn over while flying.

The first cause is pretty obvious, and also applies to landers. The second will probably flip over the first rover that most people build. The third can flip over just about any rover, though some can handle much sharper turns at much higher speeds than others. Absolutely any rover that is capable of moving at high speeds is vulnerable to the fourth, though it isn't really that big of a concern, as it requires intentionally flipping over. The larger concern is landing wrong and breaking something in the process.

The unflippable rover is basically immune to the first cause of flipping, at least apart from extreme situations such as the canyons of Dres or certain parts of Kerbin's mountains. It is also completely immune to the second cause of flipping at least on Minmus. I've tried going downhill on a 45 degree slope on Minmus at more than 30 m/s, and then slamming on the brakes and accelerating backwards to slow as quickly as possible--and with the advanced tweakables adjusted to turn off traction control and maximize braking power. I came to a dead stop without the wheels ever leaving the ground. On that basis, it is probably completely immune to the second cause of flipping on Minmus and Bop, and possibly also on Pol. Likely not on Gilly, though.

The unflippable rover can be flipped by the third cause, but it is far more forgiving than most. The very low center of mass allows turning more sharply at higher speeds than most other rovers. In the light gravity of Minmus, it can make its sharpest possible turns at 5.7 m/s without wheels leaving the ground. It can also drive in tight circles at over 3 m/s while on a slope steeper than 40 degrees. Meanwhile, the high moment of inertia means that if the rover does flip, it will flip slowly enough that you can likely react in time to prevent disaster. Even if you do leave the ground, you may be able to fire the rockets to get enough height to reorient yourself and land safely.

So how do you design a rover like that? The basic size specs give a lot of information: 1.4 m height, 10.2 m width, 13.1 m length. The places where the four wheels touch the ground gives a support base that is a rectangle over 10 m on a side. Meanwhile, the center of mass is less than 1 m off the ground. If the rover didn't slide down the hill, that should theoretically be stable on an 80 degree slope.

The rover has a total mass of 12.79 tons with full fuel. It is fueled by two Rockomax X200-8 Fuel Tanks, which feed four twitch engines. That provides a total of 2793 m/s of delta-v, which should be plenty enough to land from orbit on any of the planets or moons where the rocket is intended for use, and then still have plenty more to fly around further if needed. The engines offer a total of 64 kN of thrust, which offers more than 5 m/s^2 of acceleration on full fuel tanks, increasing to 13.36 m/s^2 as they exhaust their fuel.

The rover is designed to have a Probodobodyne HECS2 probe core that controls it at the center. Two Modular Girder Segment XLs protrude from the front and back of it, with the two fuel tanks at their other ends. Two more such XL girders connect from the sides of each fuel tank, each with a TR-2L Ruggedized Vehicular Wheel at the end, for four wheels in total. That keeps the base of the rover very low to the ground, as it is all just off the ground.

All other rover parts are radially attached to the fuel tanks. The four twitch engines are attached in a symmetric manner such that the center of thrust is exactly along the center of the HECS2 probe core at the center of the rover. The other parts are distributed such that exactly half of the remaining weight is attached to each fuel tank, to keep the center of mass almost exactly at the center of the probe core. One end has only an OP-E Large Scanning Arm. The other end has a Communotron 88-88, two PB-NUK Radioisotope Thermoelectric Generators, and two Z-400 Rechargeable Batterys.

With all wheels in use, you can drain the batteries at a net 12.55 electric charge per second. With a capacity of 1800 electric charge in total, this gives you over 143 seconds of electricity, not counting any reduced power usage from motors not going at 100%, nor any recharging while you coast. You can also park to recharge your batteries.

I use the very powerful antenna because the rover has to actually transmit science. Furthermore, as it is completely unmanned, losing your connection at a bad time is much worse than merely falling back to crew control. The antenna is able to reach all the way back to Kerbin from anywhere in the solar system, or able to get a very strong connection to any space station, even around a different planet entirely. That means that even if by chance you can't connect to any of the relays about the planet or moon where the rover is, you'll probably still be able to connect to something else, at least unless you're on the dark side of Eeloo.

The unflippable rover is able to survive hitting the ground at over 15 m/s without damage. That it attempts to be so low to the ground prevents it from surviving some of the more extreme falls that the light lander can take. It also tends to bounce if you land hard, as wheels can't stick nearly as well as lander legs, so you do want to let it down gently, as with any other rover.

Hill climbing rover

The hill climbing rover is for those who, upon reaching a steep slope in a mountainous area, wish to just drive right up it rather than having to stop and go around. This means a very high wheel to vehicle weight ratio, in order to get sufficient force to drive up the hill. As being able to ascend hills safely is the driving reason for this rover to exist, it is quite good at it. While the design is relatively low to the ground, it is much easier to flip over than my unflippable rover.

As best as I can tell, there are only three places in the game where there are hills that the unflippable rover cannot ascend. One is the canyons of Dres, where the rover has ample drive force, but would flip and fall on the nearly vertical slopes of the sides of the canyons. A second is the mountains of Kerbin, where the steepest slopes that the rover can climb are about 70 degrees. That's good enough to climb most of the slopes in Kerbin's mountains, but there are some steeper than that. The third is in Eve's mountainous areas, which do have some steep slopes, even though most of Eve is relatively flat. Even so, the hill climbing rover can go straight up most of the slopes in Eve's mountainous areas.

The hill climbing rover is a single stage with no rockets. It crams six TR-2L Ruggedized Vehicular Wheels into a mass of 1.82 tons, meaning that those wheels account for more than a third of the total mass of the rover. Its dimensions are 5.8 m length, 4.5 m width, and 2.8 m height. That may not seem particularly low to the ground, but merely removing the rover arm sticking up reduces the vehicle height to 1.5 m. Also removing the antenna makes it 1.2 m.

Because it isn't as low to the ground as the unflippable rover, the hill climbing rover can't handle as steep of slopes. While stationary, it can handle slopes of a little over 70 degrees, but it becomes awkward to maneuver without falling on slopes over about 60 degrees. I managed to create an unrecoverable save file of the rover stopped on Kerbin on a 69 degree slope, as it takes a little bit of time for the brakes to kick in when you load a saved game, and that was enough time to pick up enough speed that trying to brake and stop would cause the rover to flip.

The design of the rover is to build around a [[Z-4K Rechargeable Battery Bank] as its main body component. It has a wheel attached to either side, as well as Modular Girder Segments on the front and back. Each of the girders has a wheel on either side. The rover puts the OP-E Large Scanning Arm necessary for its mission on the front, then puts three PB-NUK Radioisotope Thermoelectric Generator and a Communotron 88-88 antenna on the back to balance the weight of the scanning arm. Add an RC-001S Remote Guidance Unit probe core on top of the center and you have the entire vehicle.

The hill climbing rover is intended for heavy gravity areas, where the gravity will greatly mitigate the risk of flipping due to accelerating or braking too fast. Lighter gravity can handle a lot more mass while keeping the wheels per weight ratio high enough to maneuver easily. As such, the hill climbing rover is intended for Eve, Kerbin, Laythe, and Tylo.

Unlike the unflippable rover, the hill climbing rover has no rockets, and hence, no way to deliver itself to a planet. Reaching Kerbin is easy because you start there. The others all require a custom delivery vehicle to put the rover on the surface. The details here vary considerably from target destination to the next.

On Kerbin, the rover was not responsible for delivering the deployed science equipment. Rather, that was a separate vehicle that started at the runway, got out to deploy the equipment, then recovered the main vehicle. With no need to deploy science equipment, there was no need for a manned rover. Rather, it loaded a rocket on top to deliver the scanning arm science to space, then drove from the Kerbal Space Center into the nearby mountains.

On Eve, the rover is delivered by a modified Eve lander vehicle. I put the deployed science and the rover under the ascent vehicle, then basically put the outer scaffolding on stilts. Put a Modular Girder Segment vertically under the outer edge of the scaffolding, and then four LT-2 Lander Struts on each of those lower girders. Upon touching the ground, I can stage to drop the rover to the ground (a short distance so it doesn't explode), then retract several lander legs to allow the rover to drive out. This design causes lander legs to bounce around and explode on Eve due to buggy physics, by that is fixable by reducing the spring strength to the minimum value allowed.

The deployed science requires both an engineer and a scientist to set it up properly, so I had to have two kerbals with me for the rover delivery, not one. Furthermore, they also needed to be able to reach the ground. The latter is easily managed by putting a Mk1 Lander Can low in the scaffolding, together with a Kelus-LV Bay Mobility Enhancer. Putting the science equipment that needs to be reset by the scientist near his command pod allowed him to reset it repeatedly to collect all of the science both for low flight and landed.

Needing to have two kerbals depart Eve, neither of which is a scientist, requires changing the upper payload from what the Eve lander used. From the top, I have an Aerodynamic Nose Cone, a TD-12 Decoupler (pointed downward to allow shedding the nose cone once far into Eve's upper atmosphere), a Mk1 Crew Cabin, a Service Bay (1.25m), and an RC-001S Remote Guidance Unit to pilot the vehicle. The service bay needs three Experiment Storage Units (with the probe core handling the fourth copy of science data) and a PB-NUK Radioisotope Thermoelectric Generator to keep the probe core powered. Kerbals can get from the lander cans to the crew cabin by using the transfer crew option. The nose cone blocks the crew hatch until it is removed, so staging it isn't just about dropping weight.

At this point, one might ask, how do you replace a 1.1 ton final payload by a 1.5 ton final payload and expect that it will just work? After all, the rover escape vehicle only has 7593 m/s of vacuum delta-v, as compared to 8704 m/s for the normal Eve lander. The answer to that is, land higher. The difference of landing at an altitude of about 6 km, as I did, roughly makes up for that difference. The normal lander has to splash down, so it has to escape from sea level. The rover just needs to land in one mountainous area and that's it. Aim for either the Olympus area or one of the other places with a large chunk of the peaks and foothills biomes and you'll be several km above sea level, which makes the ascent much easier.

(need to include some details for Laythe and Tylo)

Space stations

Space stations serve several purposes, so the basic design is to add whatever equipment is needed for each purpose until you have everything. First, a space station must serve as a very powerful relay. I attach 16 RA-100 Relay Antennas. That gets you eight times the antenna strength of just a single RA-100 relay. With such a station at each end, even with Eeloo at its farthest point from the Sun, you can connect back to Kerbin at a signal strength of over 0.93. The idea is that you can have many light relays with much weaker connections that pass a signal on to the big station that sends it to other planets.

Second, a space station should have a Mobile Processing Lab MPL-LG-2. Other than on Kerbin, you do get a little more science for doing your lab research while landed rather than in space. But you're not going to start using labs until you've unlocked everything, anyway, so I just do it in space. In addition to the science lab, that means electrical generators to be able to run the lab and a place to store the data.

Third, there are some scientific experiments that are only done in space. You don't want to weigh down a lander or rover with such equipment, so I keep it on the station. That includes an [[M700 Survey Scanner, a SENTINEL Infrared Telescope, and a Magnetometer Boom. Both the survey scanner and the desire to use the station as a relay mean putting it in a polar orbit. You can get into a low enough orbit to do the science in space near the planet and the survey, as well as being near so that refueling is easier, and then later move to a higher orbit to be more appropriate as the long-distance relay.

Fourth, after setting up the deployed science on the surface, you want to be able to have something crash into the surface to complete the seismometer science. The space station goes at the bottom of my stack of vehicles, and is directly connected to the rockets used to deliver the stack to its destination. That allows removing all other vehicles while still being able to fire rockets to reposition the station. Once everything is done, you can disconnect the rockets in order to have a large vehicle that is no longer otherwise needed, so it can crash into the surface at high speeds and instantly complete the grand slam passive seismometer science. Or at least, you can so long as you don't have a thick atmosphere to slow you down, as you do on Eve and Laythe.

There are actually three different models of space stations. The main difference is whether they need to be able to refuel the lander. On Gilly, Minmus, Pol, Bop, Ike, Dres, Mun, Eeloo, Vall, and Moho, you do need to be able to refuel the light lander, so I added a Rockomax Jumbo-64 Fuel Tank, which has enough fuel to completely refuel the lander four times. That is sufficient in each of those locations. I also added the necessary docking port to be able to connect to the lander.

Elsewhere, the landers are disposable, so you don't try to refuel them at a space station. As such, you don't need the fuel tank or docking port. It is still convenient to be able to attach to landers easily, however, so I added an [[Advanced Grabbing Unit. That is simpler to use than a docking port, and still good enough for transferring kerbals and data, even if it can't transfer fuel. The station without fuel is used on Duna, Laythe, Tylo, Eve, and Jool.

For Kerbin, since that is where you are connecting back to, I made an alternate version that added many more relays. Rather than 16 RA-100 relays, I put 162 on a station and put it in orbit about Kerbin. That allows for a theoretical maximum range for a connection between the Kerbin station and any other of over 1900 Gm, or about 15 times the greatest possible distance between Eeloo and Kerbin. Thus, at normal difficulty, the connection strength between the Kerbin station and any other station will always be greater than .987, at least if they can see each other.

Science return

The idea of a science return vehicle is that it can pick up science data from around whatever planet or moon the stack was sent to, then return that data to Kerbin. I want to be able to bring three copies of data home, so that is done by storing it in two experiment storage units and a probe core. You add the needed communications and electrical equipment, as well as parachutes to be able to land. I also have a docking port as the way to transfer data.

There are two quirks to the design. First, it needs to fit neatly in a stack, which means a relatively short central column consisting entirely of in-line parts of small radial size. That is readily done by, from top to bottom, a Clamp-O-Tron Docking Port, a Service Bay (1.25m), an RC-001S Remote Guidance Unit probe core, an FL-T200 Fuel Tank, and a terrier engine. Then you add however many more FL-T200 fuel tanks you need via asparagus staging.

There are actually two different models, with different amounts of staging. With two extra stages (for five fuel tanks in total), you get somewhat over 5000 m/s of delta-v, which is enough to get home from most places. That's not enough from Moho or a low orbit of Jool, however. Those add extra stages for the other model.

The other quirk is that I don't put the Communotron 88-88 antennas on the final stage, as you normally would. Rather, I put them on the next to last stage. It is anticipated that you'll be near Kerbin by the time you get to that last stage, so the weak antenna built into the probe core will suffice. Depending on where you're returning from, you could have a 2000 m/s retrograde burn to get into a low orbit of Kerbin, which is more delta-v than the final stage contains. Discarding the antennas a little sooner allows that final stage to provide more delta-v.

Light relay

The idea of the light relay is that you put several of them around every planet or moon that you orbit. That way, rather than other vehicles needing to connect to relays that are very far away or even all the way back to Kerbin, there are several that are at most a few thousand km away. That allows even a weak antenna to get a good connection.

You don't want to rely on a light relay to connect all the way back to Kerbin, however. The idea is that if each planet or moon has one very powerful space station and several light relays, then you just have to connect to some relay, and the relays can connect to each other to eventually get your signal to the powerful space station. That space station is responsible for the connection across interplanetary space.

The reason to make most of the relays very light is so that it is easy to get several of them to a planet. If you just have one huge space station, then when you're landed, more often than not, the planet you're landed on will block your access to the station. It will similarly block your access to any particular relay more often than not. But if you have eight light relays plus a station arranged randomly, the odds that you can connect to something at any given time while landed are about 99%--and much higher than that while in a higher orbit.

My own light relay design is very simple, with specs of 0.65 tons of mass, 2.3 m height, 1.0 m width, and 1.0m length. From top to bottom, I have in order, an RA-2 Relay Antenna, a Probodobodyne HECS probe core, a Z-200 Rechargeable Battery Bank, an Oscar-B Fuel Tank, and a spark engine. There are two OX-4L 1x6 Photovoltaic Panels connected to the probe core, opposite each other for balance. The engine offers 1154 m/s of delta-v in a vacuum, which is plenty to move the probe cores from one high orbit to another.

Transporting eight of those at once with the ability to detach them from a larger vessel thus adds less than 6 tons of weight to the vehicle. They each have their own rocket engine, which allows putting the eight small relays into eight very different orbits. That makes it cheap and easy to set up a good CommNet connection about any planet or moon.

Electricity is an issue, but the built-in batteries can last 140 minutes without recharging. That can be greatly extended if needed by putting the probe core into hibernation mode. Furthermore, the solar panels will tend to recharge the batteries much faster than they drain even without hibernation mode as far out as Jool. It gets dicier for Eeloo at its furthest point from the Sun, but there, you can put the probes in orbit, then the game will never check their electric charge so long as you don't come close. Eeloo is also completely fine when it is relatively near the Sun, as it likely will be when you approach.

I use eight light relays about every planet or moon except for Kerbin. They aren't needed on Kerbin unless you disable the extra groundstations. "Every planet or moon" includes Jool, where you may want to send a probe into Jool's atmosphere, but does not include the Sun.

Space science pickup

This is quite a simple vehicle whose design is largely dictated by its intended usage. Basically, in some places, it's easy to have your entire stack pick up all of the science available while in space. Get into a high orbit and wait until you cross every biome, running experiments as you go. Then get down to a low orbit and repeat. All that's left is landing. Even for a few small biomes that are hard to hit exactly, if you wanted to snap off a small, maneuverable vehicle to do it, the reusable lander that you've already brought is well-suited for the job. That's certainly the case on all of the planets and moons with relatively light gravity (Duna or smaller), and it's easy to hit all of Tylo's biomes from space.

But that leaves the problem of atmospheres. There, you need to be able to maneuver a bit, and putting an entire stack of vehicles into an atmosphere is awkward. On Laythe, you can snap off a sinker vehicle to do it before heading on its mission. But for Eve, the landers are huge, and for Jool, the other vehicles that are intended to go into its atmosphere are intended to reach the lower atmosphere, not just graze the upper atmosphere. Thus, for Eve and Jool, I needed another vehicle to handle picking up the upper atmosphere science. The Atmospheric Fluid Spectro-Variometer is the only experiment that varies by biome, but I needed to hit all biomes with it after running everything else once.

The light lander doesn't quite work for this, as it lacks that one piece of equipment for use in an atmosphere. So I made a similar vehicle with similar components, but arranged to be more compact and fit nicely into a stack. It only has 1/4 of the fuel of the light lander, for a total delta-v of 2337 m/s, but that's more than enough for the mission. While I've got a nice vehicle for grazing the upper atmosphere, I might as well use it to pick up the science in space, too. The space science pickup vehicle is used on Eve and Jool. I also did something similar for Kerbin, where it's easy to launch more vehicles, but that one needed parachutes to be able to land at the end of its mission, too, not just to dock.

Stacks

The idea of stacks is that, even if I want several vehicles on a given moon, that doesn't mean that I have to launch them all one at a time. Rather, I can attach them all to each other so that I only have to launch one vehicle from Kerbin, then split it up into components when the combined vehicle reaches its destination.

Stacking several unrelated vehicles is pretty much guaranteed to give a combined vehicle with extremely bad aerodynamics. It is also going to be prone to falling apart. The latter is fixable by adding a lot of struts. Even so, you don't want a bunch of unnecessary struts still attached to your real vehicles after disconnecting them. The solution is to have reaction wheels with I-beams attached in between vehicles. The struts to hold the different vehicles together all have their base on an I-beam, so that they are discarded as debris after detaching everything.

Light stack

The light stack has several vehicles. From bottom to top, it includes a station with fuel, an unflippable rover, a light lander, a science return vehicle, and eight light relays. I put the eight relays under a single AE-FF2 Airstream Protective Shell (2.5m).

The light stack is used on Mun, Minmus, Ike, Gilly, Dres, Vall, Bop, Pol, and Eeloo. Well, I didn't really use the full stack for Mun and Minmus. I did use the component vehicles there, but sent them in two separate chunks as part of completing the technology tree. That allowed me to recover science from vehicles I could build, then have the others join them later.

The light stack has about 5500 delta-v remaining after reaching a low orbit of Kerbin. That is enough to comfortably go from a low orbit about Kerbin to a low orbit about any other celestial body in the game, except for Moho, Jool, and the Sun. Moho is the only one of those three that matters, as you can't land on the other two, making the light stack wildly inappropriate for the job.

As the payload is only a little over 100 tons, that's easily managed, and I use rhino engines so that I can mostly have acceleration around 10 m/s in space, rather than waiting a long time for nerv engines to finish their very slow burns.

Moho stack

The Moho stack is nearly identical to the light stack. The only real difference is that it takes more delta-v to get to and from Moho. As such, I wanted more stages on the science return vehicle to ensure that it would have about 7000 m/s of delta-v and could actually get back to Kerbin. Everything else about the payload was the same, however.

The earlier stages to deliver the payload to Moho were substantially different, however. Rather than face the enormous cost of adding additional stages to an already heavy vehicle, I kept a single rhino engine as its last stage, but moved some previous stages to nerv engines. The much lower thrust of the nerv engines means much longer burns to get to Moho, but it allows having a little under 9000 m/s of delta-v to get there from a low orbit of Kerbin at a cost of about 500k extra funds.

Duna stack

4x Duna lander, hill climbing rover, station without fuel, science return, 8x light relay, 3000 delta-v after reaching Kerbin orbit

Tylo stack

4x Tylo lander, hill climbing rover, station without fuel, science return, 8x light relay, 5000 delta-v after reaching Kerbin orbit

Laythe stack

8x Laythe sinker, Laythe lander, hill climbing rover, station without fuel, science return, 8x light relay, 5000 delta-v after reaching Kerbin orbit

Eve stack

The Eve stack is actually pretty similar to the light stack. The landers and rover are delivered separately, so take those out of the stack. Then put in the space science pickup to replace them and you've got the Eve stack. From top to bottom, it has eight light landers, a space science pickup vehicle, a science return vehicle, a station with fuel, and the same delivery rockets as the light stack. The lighter vehicle means more delta-v from the same delivery mechanism, which is overkill, but I wasn't inclined to redesign it.

Jool stack

station without fuel, science return, all non-landed science equipment, 8x light relay, 7000 delta-v after reaching Kerbin orbit

Custom vehicles

In addition to the stacks of vehicles that I send to various planets and moons, there are also some custom vehicles that I launch independently for a variety of reasons. In the cases of Jool and Eve, it is because all of the vehicles that I want to send to a planet collectively have too much mass to be reasonable in a single launch, so they have to be broken up. For the Sun, because you can't land there or even enter its atmosphere, there is only ever a need for a single vehicle to pick up the science in space around it.

On Kerbin, there is also the issue that launching many separate vehicles is much easier. In some cases, the vehicles don't even have to reach a stable orbit to complete their mission. As such, I don't list vehicles that were launched separately while staying near Kerbin in this section.

Sun science pickup

The sun science pickup vehicle is simple enough for their to be no need for copious details. The basic vehicle just needs to have a single kerbal (as needed for a crew report and EVA report) in a command pod (needed for the crew report) together with all of the science equipment for experiments that can be done in space. Bring enough experiment storage units for however many copies you want to return to Kerbin, and choose a scientist kerbal to be able to reset the science jr. and mystery goo units. That plus electrical and communications equipment is readily doable in under three tons.

After that, you just need to put enough fuel tanks and rocket engines to have sufficient delta-v to get there. From a low orbit of Kerbin, it takes about 4000-4500 delta-v to get your periapsis about the Sun under 1 million km, which is the threshold for being in space near the sun rather than high in space above it. You can pick up the science high in space as soon as you leave Kerbin's sphere of influence. Then you wait until you get near periapsis to pick up the science near the sun.

To get back to Kerbin, you wait until you are back at apoapsis, then burn prograde to make it so that the sum of your periods before and after the burn totals one year. That means that your two orbits about the sun will match one Kerbin orbit, and the next time you're back near apoapsis, you'll have an encounter with Kerbin. The prograde burn takes about 3000-3500 delta-v. You can fine-tune your encounter from the periapsis of your second orbit. And then, of course, once you have your final encounter with Kerbin, you'll need some more delta-v to land safely. A total of 10000 delta-v remaining after reaching the initial low Kerbin orbit should be sufficient.

Jool diver

Jool probe

Eve landers

My Eve lander design is explained above. But each lander is single use, and I need to land on Eve 22 times to pick up all of the surface samples (13 biomes landed, 9 splashed). Trying to deliver 22 vehicles that are each over 120 tons to a low orbit about Eve all as part of a single stack isn't happening. And the vehicles are spatially awkward and unaerodynamic, for good measure. But I didn't want 22 separate launches, so I made a vehicle that could handily deliver three landers to Eve at a time.

Eve rover

The Eve rover ran into the same problem as the Eve landers. I probably could have put it into the main Eve stack, but that would have been hard enough that I didn't bother. The Eve rover gets its own separate vehicle.