Eve One - Soliciting advice for an Eve Ascent Vehicle

[TranquilTempest]

48-7S should be used liberally for an Eve ascent craft: http://www.twitch.tv/soulwager/c/3872144

[KerikBalm]

Those last few thousand m/s of delta-v of an Eve ascent are like the upper atmosphere part of a Kerbin ascent. Drag is still quite significant, but as the terminal velocity starts rising rapidly, you need high thrust to keep up with it. You can keep accelerating and pushing your apoapsis higher even with low thrust, but what you really want to do is to keep the climb rate high enough while gaining horizontal speed, so that time to apoapsis keeps increasing.

Well, you don't actually want to keep up with terminal velocity the whole ascent. Its most obvious in the upper atmosphere where escape velocity is lower than terminal velocity.

Generally you want the dV losses of gravity drag to equal the dV loses of atmospheric drag. However, the dV loses of gravity drag correspond to the sin of your angle relative to the surface. When you thrust parallel to the ground, you have no gravity drag - of course in the case where you still have atmospheric drag, this is not good - they should be equal, so you should still be pointing up, rasing your apoapsis.

But assuming a proper ascent path, at the higher altitudes, well into your gravity turn, gravity drag becomes quite small due to your angle with respect to the surface, and as a result, you should also be travelling much lower than terminal velocity.

Also note that generally speaking, you need a TWR of 2:1 to maintain terminal velocity (whereas a TWR of 1.5 should get you 70% of the way there), as at that point the force of drag = the force of gravity. Once you're thrusting at 30 degrees from the horizon in your gravity turn, you'd only need half the thrust to maintain a given optimal velocity - however its true you don't need to just hold one speed at one instant, but you need to accelerate to reach the ever higher optimal ascent velocities.

So I'm not sure what the TWR you really need for the upper stages is... if you were already at a given optimal velocity, you need less thrust to maintain it than when in the vertical part of your ascent, but you also want to accelerate quite rapidly

[Red Iron Crown]

The new tweaks may have given me the delta-V I need, but I notice that it's now nearly impossible to steer. Common problem with the Aerospike, I know. Anybody have a suggestion on how to correct it?

A couple of reaction wheels on your middle ring stages should do the trick. You won't need them when you get down to your core stage, by then pod torque should be enough.

[Temstar]

It's absolutely critical that you make good use of mountain top locations to shave off delta-V requirement. A mountain top launch could cut something like 5000m/s off the requirement compared to a seal level launch. Imaging designing a rocket with 7500m/s of delta-V for your payload, then imagine designing another rocket that could lift that whole vehicle into LKO. How huge would that bottom rocket be? Instead of trying to come up with a solution for 5000m/s, you're better off designing a surface roaming stage for your 7500m/s rocket so that once it lands, it can relocate itself to a mountain on wheels.

See the thing is, it's easy to put downmass onto Eve's surface, it's hard to take stuff back. If you want to take sample of both surface and ocean you're better off dropping people down on a rover, drive to the ocean, take sample, drive back to the mountain top to meet with your return vehicle; rather than try to design a return vehicle with 13,000m/s of delta-V to launch from the coast line.

Here is an old manned Eve return mission in did using the rover concept:

http://forum.kerbalspaceprogram.com/threads/30909-Two-men-Eve-landing-and-return-%28image-heavy%29

[cerberusti]

My control items at liftoff were four winglets and two torque parts. I thought the winglets would work well on eve, and they did not disappoint.

If you check your log (F3) after liftoff you can see if it destroyed anything. Entering or exiting the water needs to be done slowly (I did it at a little under a meter a second, but a few meters a second will probably work.)

The mountains are fairly small, which makes landing there without mod assistance difficult (after looking at a terrain map I decided that my piloting skills were not up to the challenge, and I should design something which could return from anywhere.)

If you are planning to go for a drive, another option is to build a small ion powered plane with rover wheels. You can get better speeds this way, and the thick atmosphere means the speed required for liftoff is low (you can even take off with the rover wheels, and get a small climb from the ion engines.) This was what the spot in the middle of my rocket was intended to hold, but I got bored and launched it after my parachute test.

Edited March 18, 2014 by cerberusti

[Jouni]

So I'm not sure what the TWR you really need for the upper stages is... if you were already at a given optimal velocity, you need less thrust to maintain it than when in the vertical part of your ascent, but you also want to accelerate quite rapidly

If we assume a constant velocity, the total delta-v losses due to atmospheric drag are roughly proportional to the inverse of the climb rate. Letting the climb rate fall due to gravity favors a high initial climb rate even more.

A typical Eve ascent is pretty much vertical until about 30 km altitude, where the climb rate should be roughly 500 m/s. If we want to maintain that rate, vertical component of thrust and centrifugal force should offset climb rate losses due to gravity and atmospheric drag. At 60 km altitude, the gravity is roughly 85% of surface gravity. If the airspeed is 2000 m/s (about 50% of terminal velocity), the vertical component of thrust should correspond to surface TWR 0.94, assuming no centrifugal force. If the nose of the rocket points to 45 degrees, total (surface) TWR should be 1.33 at this point.

The reality is a bit more complicated, but I would say that we need TWR > 1, until we are close to orbital velocities.

[ScottyDoesKnow]

The number of landing legs you need to put a 500 ton vehicle down on eve is large, probably well over 100. Once you strut them and count other structure you will need, it starts to eat into the part budget. At that point you need girders and trusses (which also increases your part count by quite a bit.)

Landing legs aren't too bad actually. My eve lander uses 48 legs for 417t. Landing around 3m/s breaks a couple of the legs, which can be fixed by a kerbal. Landing around 1m/s doesn't break any. I don't even detach them since they're on outer stages (they're about 1% of the mass and are gone in the first 5% of the flight). Part count of the whole landing leg structure is 78 (8 legs on 3 short girders with 2 struts between adjacent girders, times 6).

Alright - tonight I tried a test water landing on Kerbin. Chute deployment went well and I had a nice, cushioned landing at 3 m/s. No parts fell off at splashdown. Chute ejection went as normal. On take off, however, the rocket promptly did a couple of loops and went right into the drink. It was almost like I had lost an engine somewhere along the line.

The new tweaks may have given me the delta-V I need, but I notice that it's now nearly impossible to steer. Common problem with the Aerospike, I know. Anybody have a suggestion on how to correct it?

You may have lost an engine when entering the water, you can check the debug log to see. For control, my Eve ascent vehicle has a dozen reaction wheels attached to outer stages, which are used to make sure it gets vertical at the start of the flight. Once it's vertical, a single LV-45 and the command pod torque are enough to keep it stable.

It's absolutely critical that you make good use of mountain top locations to shave off delta-V requirement. A mountain top launch could cut something like 5000m/s off the requirement compared to a seal level launch. Imaging designing a rocket with 7500m/s of delta-V for your payload, then imagine designing another rocket that could lift that whole vehicle into LKO. How huge would that bottom rocket be? Instead of trying to come up with a solution for 5000m/s, you're better off designing a surface roaming stage for your 7500m/s rocket so that once it lands, it can relocate itself to a mountain on wheels.

I wouldn't say absolutely critical. It's definitely useful to start higher up, but you'll either need to use MechJeb to get the precise landing or be able to do a very precise landing yourself. And some people just want the challenge of an Eve ascent vehicle that can launch from anywhere.

[cerberusti]

That assumes you burn constantly all the way up to orbital speed and height. This is very difficult to do manually, and a gravity turn which ends up too low is much worse than one which is too high (which makes the trajectory you want with a safety margin higher than the ideal.)

If you go end up with a wait time before you circularize, you can get away with a low TWR since all you really need to be able to do is circularize by the time you hit your desired ap.

Also another control note: reaction wheels work far better if placed close to the center of mass. Ideally you want them on it, but as close as you can reasonably get.

Edited March 18, 2014 by cerberusti

[capi3101]

Speaking of gravity turns, where would y'all recommend I begin it on Eve? I've heard 40,000 before; that was with Scott Manley's attempt IIRC, but what would y'all suggest? 30k? 35k? Where exactly? I realize I've got other problems to iron out at the moment but that particular piece of information would be nice to have when the time comes.

[cerberusti]

I started at 30k. Ideal is probably a little lower, but that gives some error margin (and yes, wastes a little bit of dv.)

Anyway, if you are using mechjeb just let it pick. The automatic turn probably calculates it for you. It will probably start the turn under 10k (but with more precision than you are likely to have if you fly it yourself.)

Edited March 18, 2014 by cerberusti

[Jouni]

That assumes you burn constantly all the way up to orbital speed and height. This is very difficult to do manually, and a gravity turn which ends up too low is much worse than one which is too high (which makes the trajectory you want with a safety margin higher than the ideal.)

If you go end up with a wait time before you circularize, you can get away with a low TWR since all you really need to be able to do is circularize by the time you hit your desired ap.

The thing is that you don't need that much delta-v to circularize. The orbital velocity is just 3150 m/s, so if you are using more than 1000 m/s to circularize, you are doing something wrong.

Besides, the longer you keep burning during the gravity turn, the lower TWR you need. As soon as you stop the burn, your climb rate starts falling, and your total delta-v losses from drag start increasing. In order to reach the apoapsis with reasonable horizontal velocity, your climb rate must be higher than if you continue burning with a lower TWR. This in turn means that you have spent more delta-v from high-TWR stages low in the atmosphere.

[cerberusti]

The stage in question goes from a kerbin TWR of 1.49 to 4.28 over two and a half minutes with a dv of 3616 on 3 48-7s, tuning that would have required more parts than I wanted to throw at it.

That entire stage is basically my safety margin for the vehicle. I probably could have gone without it, but Jeb insisted that I strap another rocket to my rocket before he would take the mission.

[capi3101]

Anyway, if you are using mechjeb just let it pick.

Not using Mechjeb yet; Mechjeb is the reward I'm giving myself if I manage to pull off a manned Eve ascent...

[cerberusti]

I did the same thing, adding mechjeb after I managed to return home from eve. I got really good at manual docking and intercepts on the eve trip, as I had half a dozen fuel flights in kerbin orbit (which motivated me to create a fuel vehicle which could bring enough to refill it in eve orbit in one trip.)

I estimated that a safe ascent which I could eyeball without mods should probably begin a gravity turn somewhere between 20k and 40k. I went with 30k, the less confident you are that you can fly a good path the higher I would begin it. If you fail to make orbit, you probably need to fly a better ascent to make it with your craft.

An ideal turn probably begins under 10k, but the savings are small and erring on the too soon or too shallow side is usually very bad compared to waiting too long. I think the real answer depends upon how close you are to being able to follow terminal velocity, but if you have some error margin a turn at 30k should do.

[capi3101]

Last night I tweaked the design again, this time working with a new lander leg configuration. I had forgotten there was a short, fat I-beam in and among the stock parts. I laid one sideways across a TT-38K decoupler, secured it to the decoupler with a strut and attached two of the big lander legs, and then set one of these leg pairs on each of the outboard and inboard boosters. Five parts total per leg set, Sixty parts in all. The reconfiguration didn't add any more legs to the design but it did redistribute the support across the entire width of the lander (as opposed to just the edges with the previous design). Parachute tests were successful; I was able to land the craft intact on land in a 3 m/s landing test. I might've taken off from that point except that I had fouled up the staging - the legs decoupled before the chutes. So that's how that test ended. I might try a full mock landing and takeoff again tonight.

Leaves me with one final problem to tackle - how to get Jeb out of the can and onto the ground and back. I'll have to post some revised screenies of the craft as it's standing right now, and I'll need to go ahead and make a version with the manned payload. My cheatsy method I'd intended hasn't panned out, on account of a lack of places to put ejectable lander cans near the bottom of the craft (I was going to Crew Manifest him from the center pod to the ejectable pods.

Meantime, other aspects of the overall mission plan are going well. I designed the return vehicle the day before last and successfully flew it intact to Kerbin orbit. I still need to do a mass study for the Barn Burner train (the refueling mission prior to Eve landing) and get a transfer stage designed for that as well. Overall things are progressing reasonably well and I hope to be able to make the mission attempt sometime in the next week or two.

[Red Iron Crown]

I reworked my ascent stage last night. Previously it was a 19 aerospike asparagus configuration with a small final stage, but I was unhappy with the TWR as it started at about 1.3 and fell to 1.1 as I shed stages.

Here's the rework, at first glance it looks like the previous design, with 19 aerospike boosters arranged asparagus style. The number of stages is wrong, though.

Strip away the outer ring of boosters and the difference appears. It's a small asparagus rocket with an LV909 as the core and 48-7Ss as the ring fed from 200 unit tanks, sitting on top of asparagused aerospike stages. The innovation (for me, anyway, I'm sure others have done it) is that as each aerospike drops away the 48-7S above it lights, helping to keep the TWR up. The upper asparagus is fed by fuel lines from the core stage below so the smaller rocket keeps its full fuel load. The result is TWR above 1.3 for all stages except the LV909 by itself in the final stage. Also visible is one of the two reaction wheels to give control authority with the non-gimbaled aerospikes. The gimbals on the 48-7Ss are locked as they spend a fair amount of time burning above the CoM.

A shot of the full readout. I would prefer a little more dV for a cushion, but the final LV909 stage will be working in almost vacuum so it will have almost 500m/s more. Now to strut it and add all the landing equipment.

TL;DR: Yo dawg, I heard you like asparagus, so I made an asparagus on top of an asparagus inside an asparagus.

[ScottyDoesKnow]

https://dl.dropboxusercontent.com/u/61004449/ksp/AdamCore.png

Strip away the outer ring of boosters and the difference appears. It's a small asparagus rocket with an LV909 as the core and 48-7Ss as the ring fed from 200 unit tanks, sitting on top of asparagused aerospike stages. The innovation (for me, anyway, I'm sure others have done it) is that as each aerospike drops away the 48-7S above it lights, helping to keep the TWR up. The upper asparagus is fed by fuel lines from the core stage below so the smaller rocket keeps its full fuel load. The result is TWR above 1.3 for all stages except the LV909 by itself in the final stage. Also visible is one of the two reaction wheels to give control authority with the non-gimbaled aerospikes. The gimbals on the 48-7Ss are locked as they spend a fair amount of time burning above the CoM.

Totally stole this design and reworked my eve lander. It looks better now and I was able to remove the 6 radial engines around my final column, as well as an entire layer of tanks (from 4 layers to 3, it's a fairly heavy payload). Thanks for the idea!

Edit: my signature has the old version if you want to see.

Edited March 22, 2014 by ScottyDoesKnow

[Red Iron Crown]

Totally stole this design and reworked my eve lander. It looks better now and I was able to remove the 6 radial engines around my final column, as well as an entire layer of tanks (from 4 layers to 3, it's a fairly heavy payload). Thanks for the idea!

Edit: my signature has the old version if you want to see.

Glad you liked it! I'm looking closely at your landing gear design on the craft in your sig. I just had some beams with a bunch of landing legs around the perimeter, but when I put tanks on top of the rocket to simulate the mass of the lander on Eve for testing they crumpled just sitting on the pad.

[Kasuha]

If you use seats rather than a pod, you can make the ascent stage even smaller. 19 FL-T800 with aerospikes with small circularizer on top can totally get a crew of two to orbit even from sea level.

[ScottyDoesKnow]

Glad you liked it! I'm looking closely at your landing gear design on the craft in your sig. I just had some beams with a bunch of landing legs around the perimeter, but when I put tanks on top of the rocket to simulate the mass of the lander on Eve for testing they crumpled just sitting on the pad.

Make sure you look at the actual landed pictures, since the shots are from different versions. In the end it's 3 girders with 8 legs total (all times 6) to land 417t. I had to add a strut from the girder to the tanks as you can see, and some will break if you land at more than ~1m/s (which you have to fix with a kerbal or it can screw with the staging on ascent). I use the same design for the new version (327t) which no longer needs the strut and can land safely at 3m/s without any breaking.

Edit: preview of the new version, which I think looks massively better than the old one (no more giant square of boosters). It's actually got more dV than that, since that's including the return craft and rover, and mechjeb screws up a bit with the eve lander staging.

Edited March 23, 2014 by ScottyDoesKnow

[Red Iron Crown]

If you use seats rather than a pod, you can make the ascent stage even smaller. 19 FL-T800 with aerospikes with small circularizer on top can totally get a crew of two to orbit even from sea level.

I know, but I don't use command seats as a self-imposed limitation for realism and aesthetic reasons. Though if I end up having to mount a rescue mission I might revisit that policy.

I am impressed with the way you set up the landing equipment on the pictured lander. Are you willing to share the craft file so I can have a look at how it is assembled in the VAB?

Make sure you look at the actual landed pictures, since the shots are from different versions. In the end it's 3 girders with 8 legs total (all times 6) to land 417t. I had to add a strut from the girder to the tanks as you can see, and some will break if you land at more than ~1m/s (which you have to fix with a kerbal or it can screw with the staging on ascent). I use the same design for the new version (327t) which no longer needs the strut and can land safely at 3m/s without any breaking.

I must be doing something wrong with the landing gear. I had 36 of the heavy lander legs on my ~150t lander but they weren't up to the task. Maybe I need to make them almost vertical when deployed so they're less likely to splay apart. In the pic below you can see that with the extra mass to simulate Eve conditions the gear settle until the aerospikes touch the surface and take most of the weight.

[Kasuha]

Are you willing to share the craft file so I can have a look at how it is assembled in the VAB?

Sure, the link is in that post.

I was just trying to reduce debris (so I don't have to kill too many things from tracking station) so it is all in two pieces. It makes the lander more bulky and heavier, though.

[Red Iron Crown]

Sure, the link is in that post.

I was just trying to reduce debris (so I don't have to kill too many things from tracking station) so it is all in two pieces. It makes the lander more bulky and heavier, though.

Wow, that lighter final stage pays big dividends. Almost 13km/s of atmospheric dV and TWR around 1.6 or higher for all but the last stages. Very nice!

Edit: Also, I have to admire someone that lands on Eve and says "This spot isn't low enough, let's get right down to sea level."

Edited March 23, 2014 by Red Iron Crown

[cerberusti]

This one also gets from sea level to eve orbit with the lifter at 68.8 tons. It does use a seat, and landing gear + parachutes take it up to nearly 100.

It has similar dv to my last one (a bit over 14k vaccum, 12.5k atomsphere.) It uses a mix of engines, which turns out to be fairly effective. 47-7s are the primary engines, aerospikes are too heavy to be efficient on upper stages. RT-10s are great for liftoff in order to quickly accelerate to terminal velocity, the Mark 55s are for acceleration and control immediately after launch during the solid booster stage.

In eve orbit (so you can see the eve relative TWR):

http://steamcommunity.com/sharedfiles/filedetails/?id=241562832

Top:

http://steamcommunity.com/sharedfiles/filedetails/?id=241562748

Side:

http://steamcommunity.com/sharedfiles/filedetails/?id=241562682

[ScottyDoesKnow]

I must be doing something wrong with the landing gear. I had 36 of the heavy lander legs on my ~150t lander but they weren't up to the task. Maybe I need to make them almost vertical when deployed so they're less likely to splay apart. In the pic below you can see that with the extra mass to simulate Eve conditions the gear settle until the aerospikes touch the surface and take most of the weight.

https://dl.dropboxusercontent.com/u/61004449/ksp/LandingGearTest.png

Think of it like a chain, it'll break at the weakest link. Mine has 2/3 legs on a girder attached directly to a tank, yours looks like it has 6 legs on girders attached to a radial decoupler. Radial decouplers are very weak, especially against shear forces. And no amount of struts can replace strong connections. (In my last image of my new ship, the top wobbled like crazy no matter how many struts I added. Then I added 6 docking ports on girders around the outside and no more problems.)