Jump to content


  • Posts

  • Joined

  • Last visited

Everything posted by blakemw

  1. The quoted value for 90Sr beta emissions (which is higher energy than some other emitters) I found was 4 mm of aluminium which is much thicker than a spacesuit or the walls of something like the apollo lander. Of course the RTG thermocouples should also attenuate the radiation some. But basically no, a spacesuit or the walls of a light capsule won't block high energy beta radiation.
  2. I had to actually, but actually in Kerbalism it appears to be based on90Sr because it has a halflife of 28.8 y, hence it emits beta radiation, which certainly has the penetrating power to get through the spacesuit of a Kerbal who disrespects the square cube law and decides to hug the RTG.
  3. I was wondering how best to shield ships and bases from CMEs (especially bases, since they are harder to orientate to the sun), so did some testing. The first thing, is that it appears a raycast is done from the center of the habitat to the sun, and anything intersected counts as shielding: this means that a single surface mounted small battery can shield the habitat if it happens to be on that line between the sun and the center of the habitat. The next thing: is that it seems putting anything in between the sun and the habitat module reduces radiation by a lot: a flag reduced radiation by about 74%, a solar panel reduced radiation by about 91%, a structural panel reduced radiation by 95%, a Cargo Bay reduced radiation by 99.4%. Shielding layers stack well. Putting TWO flags between the habitat and the sun, reduced radiation by 86%, two solar panels reduced radiation by 99%. I tested "standoff", whether moving the shielding object further from the ship improves shielding. It does, but by very very little. For example in a particular solar storm, a flag on the surface of the habitat reduced radiation to 0.444 rad/h, the same flag at the maximum extension of a 1P4 telescoping cylinder, reduced radiation to 0.441 rad/h. So there is an effect, but it's absolutely negligible. The surface attached part can be inside the habitat, that is fine. However if the raycast starts INSIDE a part, that part does not contribute to shielding: embedding a habitat inside a larger part such as a fuel tank does not help. However if the larger part has internal walls such as a structural tube or a service bay then it does work, the principle is probably that the raycast has to hit a wall you can't see through: essentially if you can see the sun, the sun's radiation can see you. Also it doesn't appear to check other vehicles, a Kerbal on EVA can't shelter in the shadow of a ship. The upshot: higher density objects are more effective for radiation shielding, but they don't need to be very high density, solar panels work well, flags are a bit flimsy but significantly better than nothing. Stacking layers of shielding is probably far more effective than heavier shielding. Enclosing the habitat inside an object with inner walls like a cargo bay is useful. Standoff doesn't literally do nothing but is practically worthless. A habitat in a base can be shielded simply by blanketing it in surface attached parts or placing it inside a cargo bay or service bay. If you don't want to know exploits don't read this next part: This is the measured in-game behaviour. It might not be the intended behaviour. RTGs: They produce essentially no radiation. Technically they do, but a Kerbal has to snuggle up against one to get an increased dose. They do nothing like 1m away. Don't clip one to the center of a habitat, don't put one on the wall of a very small habitat. Otherwise they do nothing and no shielding strategy is required. I don't have a fission reactor mod installed atm to test them. LV-N "Nerv": They produce even less radiation than an RTG which is saying something, and also produce no additional radiation when firing. A Kerbal on EVA has to snuggle up against one to get even half the dose of an RTG. Don't attach one directly to the bottom of an Mk1 lander can or something like that and otherwise they are irrelevant. I don't have any nuclear engine mod installed atm.
  4. Hey I'm a globehead and my one complaint about AA's Fly-By-Wire mode is that it doesn't pitch the nose down to account for the curvature of a planet. Is there any way that AA can do this amoung it's multitude of bewildering settings? I'm aware that AA constant altitude can do this, but the context I want it for is controlled ascend/descent not merely maintaining a certain altitude.
  5. There's actually a marginally good reason to throttle back with certain engines in certain contexts. First is core-throttling on Falcon Heavy style rockets, that is you don't throttle everything, just the core. This give you more deltav and would even do so on a vacuum world because by saving fuel in the core for after the boosters are decoupled, that fuel goes only to accelerating the mass of the core+payload, and not boosters+core+payload. If you're using a rocket with boosters and aren't using crossfeed (which isn't even a possibility if the boosters are SRBs), then it is usually an improvement to keep the TWR at about 2.0, of course you could also bring more fuel but that costs more and you could technically probably use smaller engines, but there aren't a ton of engine choices available. (like if you could custom tailor every aspect of the engines you're using, this strategy probably wouldn't be employed. But it's like SpaceX, they want to make a Falcon Heavy using "off-the-shelf" Falcon 9 boosters, without designing entirely new boosters and core, the way to maximize performance is to heavily throttle the core soon after liftoff, until the boosters have detached) The second reason you might throttle is ISP: some launch engines, especially Swivel, get a lot more thrust per unit of fuel once out of the lower atmosphere (practically, 10000m is as good as a vacuum for ISP purposes), this often has a much more significant impact on efficiency than aerodynamic drag. Now, of course, you do need to climb and maintain a good enough TWR, most likely this will only be useful when used in conjunction with the first factor I mentioned: If you're using something like a Swivel core with Thumber SRBs then throttling back the swivel means not only are you saving fuel for when it no longer needs to accelerate the mass of the SRBs, you also get more thrust out of it thanks to a higher ISP due to altitude.
  6. I think the key is understanding lift. Lift is very potent at transforming vertical velocity ("falling") into horizontal velocity ("distance"). Even with the stock aerodynamics, a good spaceplane (one which needs wings to fly, not one with enough power to do it without wings) can glide about 1/4 of the way around Kerbin further than the ballistic prediction would indicate (and that's the simple suborbital prediction, not even the ballistic-with-drag prediction of a poorly configured Trajectories). The FAR aerodynamics are more "slippery" than stock and higher lift/drag ratios are possible, meaning gliding even further. When landing spaceplanes you need to aim to undershoot by A LOT because the plane is going to glide around the world - that is assuming you are pitching up a bit to slow descent and ease heating. Trajectories mod is so useful for modeling this but it's vital to tell it you'll be gaining lift (and nice thing is, you can adjust the angle of attack you want it to use for its prediction and see the projected trajectory change in response). If you've never seen it I super highly recommend this video about how the Apollo capsule used lift to narrow down, though mainly just for inspiration, the main take-home is that it's damn hard, you descend too fast and burn up/get destroyed, or you overshoot by half a world (and it's computer stuff, not "seat of the pants" stuff):
  7. Better yet install the Mod "Better Burn Time", fixes that issue and adds some other minor utility.
  8. I'm sorry but this advice is almost completely invalid when using FAR. The FAR atmosphere is a lot more "slippery" and lift is a lot more effective, if you try a "looney toons stop" then for one you'll retain more velocity because steering tends to be very effective in changing direction without losing velocity, for two you probably won't be able to because of the air restraining the plane's movement or you'll succeed but rip the wings off. Because lift is a lot more effective (i.e. lift/drag ratio is higher) by adjusting the angle of attack you can glide very long distances, much further than in stock. It's hard to bleed off velocity and overshooting is a very realistic possibility. Because of the slipperiness of the atmosphere heating can also be a serious constraint, you can get down to lower altitudes where it's hotter without slowing down enough. Even in stock, but moreso in FAR, with the right plane it's possible to glide further than the ballistic trajectory would indicate, with a high enough lift/drag ratio the vertical velocity can be converted into horizontal velocity effectively enough to more than compensate for drag - to the point where you can fly at least a 3rd of the way further around kerbin (than ballistic impact point) through the power of gliding. If you're re-entering a capsule then aim to overshoot, if a plane then aim to undershoot and glide to make the distance. (and if using trajectories it's essential to set a positive AOA so it knows you'll be gaining lift, the 0 AOA prediction is nearly meaningless for a plane)
  9. Use the Trajectories mod (it works with FAR). You need to tune it a little bit, tell it you'll be entering prograde and estimate your angle of attack at different regimes (or use 10%). Naturally the prediction made by trajectories is not perfect, it's probably not perfect even if you maintain the angle of attack you told it you would, but the prediction is much more useful than not having it, generally speaking if the prediction is undershooting the target you can pitch up a little, and if it's overshooting you can pitch down.
  10. Yup. For spaceplanes wings make by far the most mass-efficient liquid fuel storage - that is assuming you need the wings. This is because wings with fuel capacity are identical in terms of mass per lift/drag as wings without fuel capacity, the fuel capacity is a freebie in terms of mass.
  11. Hiding stuff inside fairings and cargo holds DOES work, though when things are attached to the walls the game might convince itself that the thing is really not inside the cargo hold. But if anything is offset deeply into the cargo hold it'll definitely be removed from the airstream. To clear up any confusion here is how it works: Node attachment: What a thing is attached to and relative node sizes matters*, offsetting and clipping doesn't matter in the slightest. Cargo holds/fairings: The things physical location relative to the cargo hold/fairing matters, the part it is attached to does not matter in the slightest. (yaaaay!) * There is a limited exception when dealing with truncated cone parts, for whatever reason you can stack parts like mk1 command pods on top of each other or place a part matching the bottom node on the top node (i.e. putting a 1.25m fuel tank on top of a mk1 command pod) and the game considers this to be properly streamlined.
  12. Some interesting alternative landing gear include decouplers / hard points and radiators. Radiators have an especially odd interaction with the terrain, they aren't exactly bouncy, but they will slide up on impact helping to absorb the shock. Service and Cargo Bay doors can also be used as landing gears particularly if you adjust how far they open. Doors are rigid, indestructible and deployable.
  13. Nose intakes are typically less draggy than nose cones, the Shockcone particularly has unnaturally low drag but all nose intakes are less draggy than comparable nose-cones. Intakes are typically heavier and (much) more expensive than nose cones and sometimes have less thermal tolerance. Surface attached intakes can be quite draggy because in general the drag model hates surface attachment, but there are only two of them and neither of them are that good. Certainly given the choice between a surface attached intake or inline/nose-intake you should not go with surface attached.
  14. The main reason to use a precooler is if you want an intake with reasonable performance and you can't use a shockcone on the nose of the ship because it will burn off (or other reasons, like wanting a shielded docking port). Shockcones have good thermal tolerance but bad other thermal properties, one factor which determines how easily something burns is "cross section exposed to heating" vs "skin area available to radiate heat", the Shock Cone being a short and stubby part can't radiate much heat. A reasonably pointy fairing makes the best nose cone for thermal tolerance and you can put stuff inside it like batteries, reaction wheels, probe core, antennas etc. A single shockcone can supply 6 RAPIERs with air with the right ascent profile, so is very much overkill for a spaceplane with only 1 or 2 engines - which a precooler will suffice for. Using a Precooler + Fairing Nose with light/cheap 0.625m parts inside it can be a nice optimization for smaller planes.
  15. It is possible but it's hard. You have to use some combination of leaving early in the transfer window to Duna and a fast transfer (low trip time), when using Transfer Window Planner it's arrival date you care about, you want it to be as soon as possible. This will result in you arriving late in the Duna->Kerbin transfer window, allowing you to (nearly) immediately transfer back. The deltaV requirements are not absurd, I can't remember exactly what, but budget maybe 2000m/s extra than for perfect Hohmann transfers.
  16. Ideally you want to leave and/or arrive at the ascending/descending node between Kerbin's orbit and the target planet's orbit around the sun, allowing the plane change to be performed at the same time as the ejection/insertion burn. Of course, this requires the planets to line up right, when using transfer window planner these will manifest as transfer windows with particularly low deltaV requirements and is most obvious with Dres, Moho and Eeloo. Of course you can also change the plane of the transfer orbit at time of ejection to give a ballistic trajectory, this might be more or less expensive than a mid course plane change. How you do this is you need to make the AN/DN relative to the sun, line up with the rendezvous with the target planet by adjusting the normal/anti-normal component of the ejection burn, this will often involve a steeply inclined ejection, the cost of this can be greatly reduced firstly by launching directly into that inclination (i.e. by doing the planning with a seperate craft, possible cheated), or by first raising your apoapsis around Kerbin to a 1day+ orbit, then doing the inclination change at apoapsis, then completing the ejection burn at periapsis, when the required inclination change is large this results in fuel savings.
  17. I like 600km. There is the concept of a gate orbit, that is the cheapest orbit to perform an ejection burn from to a certain destination. There are two factors: first is being further out of the gravity well which makes it cheaper, secondly is the reduced Oberth effect which makes it more expensive. Where the cost of the ejection burn is minimized is the gate orbit. 600km makes a pretty good orbit for a range of destinations (in general, "climbing out of the gravity well" is more valuable for near destinations like Duna and Eve resulting in a relatively high gate orbit, Oberth effect is more valuable for far destinations like Moho and Jool resulting in a relatively low gate orbit). But note: getting up to the gate orbit requires more fuel than ejecting directly from a lower orbit, but higher orbits can still be good places for fuel depots if that fuel is coming down from Minmus rather than up from Kerbin, or if you're lifting so much fuel at once you don't care.
  18. I don't use them much. Partly from playing hard career and more staging = more problems. Also the benefits are kind of marginal. Dirty secret is: most the payload to orbit benefit just comes from making the rocket bigger rather than cleverer. A rocket which is twice as big straightforwardly delivers twice as much payload to orbit. A rocket which is twice as clever (in terms of fancy staging) probably only launches 5% more compared with a simple two-stage rocket.
  19. 2.5m engine plates produce more drag than a tri-adapter and less than a quad-adapter assuming 3 or 4 1.25m engines are attached. The amount of drag they cause is actually dependent on the amount of surface that doesn't have an engine attached. So if you attach 4 engines to a plate, it'll produce less drag than if only 2 engines of the same size are attached, an engine plate smothered in engines causes basically no drag at all.
  20. Seems intentional. It's notable that the Convert-o-Tron 125 also had its max cooling modified, increased from 50kW to 75kW. It still can't help but overheat, but shouldn't overheat as badly now.
  21. Even nuclear engines still need reaction mass to throw out the back of the ship, so that is why hydrogen is needed. The reactors do the throwing, by heating that reaction mass up (either heating it directly, or indirectly via electricity).
  22. The inflatable heat shield is good only up to about 5500m/s (at Eve) because it lacks ablator. It can be used but you have to be careful to not come in too fast. The heat shields with ablator are good up to about 9000m/s.
  23. Feathering is consistently effective at Eve, a small scale example which I built for 9000m/s entry at Eve: It scales up well too. This rather exploits the fact that the game's drag model is too simplistic for hypersonic speeds, even stuff which should be in the wake of the capsule still generates full drag. So basically you can just add lots of these kind of flat pieces (spaced apart) to get as much drag as you like.
  24. This works for me: I didn't put a lot of effort into getting the thrust offset angle perfect but it's close enough the single reaction wheel can counteract the effect. Also, it's air-launched. Though I've launched similar things on a monowheel in the past.
  25. This is true for Ore to LF and/or Ox since both weigh 5kg/unit while ore weighs 10kg/unit. But monoprop weighs 4kg/unit so there is a 20% mass loss in the conversion from ore to monoprop.
  • Create New...